Construction method, evaluation method and related device of CO2 oil displacement burying agent model

Abstract

The invention discloses a construction method, an evaluation method and a related device of a CO2 oil displacement burying agent model. The agent model construction method comprises the following steps: determining an input variable and an output variable; constructing an agent model containing parameters according to the input variables and the output variables; based on the obtained data sets corresponding to the input variables and the output variables, utilizing a machine learning strategy to train the agent model to obtain a CO2 oil displacement burying agent model; and based on the real data of the input variables and the real data of the output variables of the mining area site, carrying out iterative optimization on the CO2 oil displacement burying agent model. Compared with traditional prediction methods such as numerical analysis and experimental methods, the agent model and the evaluation method applying the agent model have the advantages of being short in calculation period, high in calculation precision, capable of saving equipment cost, high in risk resistance, convenient to apply and the like.

Description

Technical Field

[0001] This invention relates to the field of crude oil extraction technology, and in particular to a method for constructing, evaluating, and related apparatus for a CO2 flooding and storage proxy model. Background Technology

[0002] Collecting CO2 and injecting it into the reservoir serves two purposes: CO2 sequestration and enhanced oil recovery. Before undertaking a CO2 enhanced oil storage project, evaluating the effectiveness of CO2 enhancement in underground reservoirs is a crucial preliminary step. Current evaluation methods primarily include experimental and numerical methods. Experimental methods have limitations in the size of rock samples used, leading to significant discrepancies between the evaluation results and actual reservoir-scale results, and they also lack sufficient intelligence. Numerical methods suffer from drawbacks such as complex simulation processes, long calculation cycles, and high requirements for simulation equipment. Summary of the Invention

[0003] In view of the above problems, the purpose of this invention is to provide a method for constructing a CO2 flooding storage surrogate model, an evaluation method, and related apparatus.

[0004] In a first aspect, embodiments of the present invention provide a method for constructing a CO2 flooding surcharge surcharge model, comprising:

[0005] Determine the input and output variables used to build the proxy model;

[0006] Based on the input and output variables, a surrogate model containing parameters is constructed; the parameters include coefficients and exponents associated with the input variables;

[0007] Based on the datasets corresponding to the obtained input and output variables, the surrogate model is trained using a machine learning strategy to determine the coefficient values ​​and the exponent values ​​of the surrogate model, thereby obtaining a CO2 flooding storage surrogate model.

[0008] Based on real data of input variables and output variables from the mining area, the CO2 enhanced oil recovery and storage proxy model is iteratively optimized.

[0009] In one embodiment, the iterative optimization of the CO2 enhanced oil recovery and storage surrogate model based on real data of input variables and output variables from the mining area includes:

[0010] The surrogate model is used to fit the real data of the input variables and the real data of the output variables to determine the goodness-of-fit index.

[0011] If the fitting index meets the preset requirements, the CO2 flooding and storage surrogate model will not be optimized.

[0012] If the fitting degree index does not meet the preset requirements, the surrogate model is reconstructed, the reconstructed surrogate model is trained, and an updated CO2 flooding storage surrogate model is obtained, until the fitting degree index corresponding to the updated CO2 flooding storage surrogate model meets the preset requirements.

[0013] In one embodiment, constructing the proxy model containing parameters includes:

[0014] Based on the analysis of the input and output variables and the mechanistic relationships between the input variables, the combination methods between the input variables and the coefficients and exponents associated with the input variables are determined.

[0015] In one embodiment, the combination method includes: algebraic combination, power law combination, and algebraic-power law combination.

[0016] In one embodiment, the input variables include:

[0017] Porosity φ n Penetration rate K n Crude oil viscosity V is Permeability variation coefficient V k Oil layer thickness h, oil-bearing area S, initial oil saturation S oi Water drive recovery level R o degree of miscibility f p Well density D w and the injected gas volume Q ng (t).

[0018] In one embodiment, the output variable includes: oil and gas equivalent Q. o Or recovery rate E or .

[0019] In one embodiment, the proxy model containing parameters is as follows:

[0020]

[0021] in:

[0022]

[0023] φ n Porosity, %;

[0024] K n Penetration rate, 10 -3 μm 2 (mD);

[0025] V is Crude oil viscosity, mPa·s;

[0026] V k : Coefficient of variation of permeability, decimal;

[0027] h: Oil layer thickness, m;

[0028] S: Oil-bearing area, km² 2 ;

[0029] S oi Initial oil saturation, %;

[0030] R o : Water drive recovery rate, %;

[0031] f p : Degree of miscibility, decimal;

[0032] D w Well density, number of wells / km 2 ;

[0033] Q ng (t): Injected gas volume, m 3 ;

[0034] N n Cumulative oil and gas production equivalent, m 3 ;

[0035] Q o Oil and gas equivalent, m 3 ;

[0036] E or Recovery rate, %;

[0037] a,b,c,d,e,f,g,h,i,j,k,x1,x2,x3,x4,x5,x6,x7,x8,x9,x10: Parameters, including coefficients and exponents, are decimals;

[0038] a',b',c',d',e',f',g',h',i',j',k',y1,y2,y3,y4,y5,y6,y7,y8,y9,y10: Parameters, including coefficients and exponents, are decimals.

[0039] In one embodiment, the proxy model containing parameters is as follows:

[0040]

[0041]

[0042] in:

[0043] φ n Porosity, %;

[0044] K n Penetration rate, 10 -3 μm 2 (mD);

[0045] V is Crude oil viscosity, mPa·s;

[0046] V k : Coefficient of variation of permeability, decimal;

[0047] h: Oil layer thickness, m;

[0048] S: Oil-bearing area, km² 2 ;

[0049] S oi Initial oil saturation, %;

[0050] R o : Water drive recovery rate, %;

[0051] f p : Degree of miscibility, decimal;

[0052] D w Well density, number of wells / km 2 ;

[0053] Q ng (t): Injected gas volume, m 3 ;

[0054] Q o Oil and gas equivalent, m 3 ;

[0055] N n Cumulative oil and gas production equivalent, m 3 ;

[0056] E or Recovery rate, %;

[0057] x1,x2,x3,x4,x5,x6,x7,x8,x9,x10,x11,α,β: Parameters, including coefficients and exponents, are decimals;

[0058] y1,y2,y3,y4,y5,y6,y7,y8,y9,y10,y11: Parameters, including coefficients and exponents, are decimals.

[0059] Secondly, embodiments of the present invention provide an evaluation method for a CO2 flooding surviving surviving surviving model, characterized in that it includes:

[0060] Based on the input variable data obtained from the evaluation area and the pre-built CO2 flooding and storage surrogate model, the CO2 flooding and storage effect of the evaluation area is predicted.

[0061] Sensitivity analysis of input variables was performed on the CO2 flooding and storage surrogate model to identify the key input variables affecting the effectiveness of CO2 flooding and storage; based on the analysis of the key input variables, the CO2 flooding and storage measures in the evaluation area were optimized.

[0062] The CO2 flooding and storage surrogate model is obtained in advance using the surrogate model construction method described in the first aspect above.

[0063] Thirdly, embodiments of the present invention provide a related apparatus for a CO2 flooding and storage surrogate model, comprising:

[0064] The variable determination module is used to determine the input and output variables used to build the proxy model;

[0065] The model building module is used to construct a surrogate model containing parameters based on input and output variables; the parameters include coefficients and exponents associated with the input variables;

[0066] The model determination module is used to train the surrogate model based on the dataset corresponding to the acquired input and output variables using a machine learning strategy, and determine the coefficient value and the exponent value of the surrogate model to obtain the CO2 flooding storage surrogate model.

[0067] The iterative optimization module is used to iteratively optimize the CO2 enhanced oil recovery and storage proxy model based on the real data of the input variables and the real data of the output variables in the mining area.

[0068] Fourthly, embodiments of the present invention provide a computing device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the aforementioned method for constructing a CO2 flooding storage surrogate model or the method for evaluating a CO2 flooding storage surrogate model.

[0069] Fifthly, embodiments of the present invention provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the aforementioned method for constructing a CO2 flooding storage surrogate model or the method for evaluating a CO2 flooding storage surrogate model.

[0070] In a sixth aspect, embodiments of the present invention provide a computer program product, comprising: a computer program, wherein when the computer program is executed by a processor, it implements the aforementioned method for constructing a CO2 flooding storage surrogate model or the method for evaluating a CO2 flooding storage surrogate model.

[0071] The beneficial effects of the above-described technical solutions provided in the embodiments of the present invention include at least the following:

[0072] This invention provides a method for constructing a CO2 flooding and storage surrogate model. The constructed surrogate model is used to predict the CO2 flooding and storage effect in mining areas. Compared with traditional experimental or numerical simulation methods, this surrogate model has advantages such as reducing experimental costs, calculation cycles, and equipment costs. The surrogate model is clear, concise, easy to understand, and easy to promote and use.

[0073] Furthermore, the inventors pre-constructed input and output variables, determined by analyzing the impact of the mining area on CO2 enhanced oil recovery and storage, with each input and output variable having a clear physical meaning. The pre-defined combination method was determined based on the analysis of the input and output variables, as well as the mechanistic relationships between the input variables. Therefore, this proxy model is a powerful combination of mathematical thinking and objective practice, possessing high credibility and resilience.

[0074] Furthermore, when determining the parameter values ​​in the proxy model, a machine learning model is applied to train the parameterized proxy model based on a dataset consisting of a large amount of input and output variable data, and the parameter values ​​are fitted. The use of machine learning avoids interference from human factors and increases the accuracy of the proxy model.

[0075] Furthermore, during the iterative optimization of the surrogate model, the inventors started with the initially preset input and output variables, reconstructed the surrogate model, and then trained the reconstructed surrogate model until the updated CO2 enhanced oil recovery and storage surrogate model met the preset requirements. Starting from the very beginning of establishing the surrogate model ensures that it further improves the accuracy of the CO2 enhanced oil recovery and storage surrogate model based on a connection to and description of objective reality, thus guaranteeing the rigor and scientific nature of the final constructed surrogate model.

[0076] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0077] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0078] Figure 1 This is a flowchart illustrating the method for constructing a CO2 flooding storage proxy model in an embodiment of the present invention;

[0079] Figure 2 This is a flowchart of the evaluation method using the CO2 flooding and storage surrogate model in an embodiment of the present invention;

[0080] Figure 3This is a schematic diagram of the structure of the CO2 flooding storage proxy model related device in an embodiment of the present invention. Detailed Implementation

[0081] This invention provides a method for constructing and evaluating a CO2 flooding surcharge surcharge model, as well as related apparatus. While exemplary embodiments of this disclosure are shown in the accompanying drawings, it should be understood that this disclosure can be implemented in various forms and should not be limited to the embodiments set forth herein. Rather, these embodiments are provided to enable a more thorough understanding of this disclosure and to fully convey the scope of this disclosure to those skilled in the art.

[0082] This invention provides a method for constructing a CO2 enhanced oil recovery surcharge surcharge surcharge model, referring to... Figure 1 As shown, the method includes the following steps:

[0083] S11. Determine the input and output variables used to build the proxy model.

[0084] S12. Construct a proxy model containing parameters based on the input variables and output variables; the parameters include: coefficients and exponents associated with the input variables.

[0085] S13. Based on the datasets corresponding to the obtained input and output variables, the surrogate model is trained using a machine learning strategy to determine the coefficient values ​​and the exponent values ​​of the surrogate model, thereby obtaining the CO2 flooding storage surrogate model.

[0086] S14. Based on the real data of the input variables and the real data of the output variables in the mining area, the CO2 flooding and storage proxy model is iteratively optimized.

[0087] CO2 enhanced oil storage (CCUS-EOR) involves injecting CO2 into the reservoir to simultaneously drive away oil production and store CO2 underground. This method is significant because it can both increase oil production and reduce greenhouse gas emissions.

[0088] Step S11: Predetermine the input and output variables. If the pre-built surrogate model is considered as a function, then the output variable is the dependent variable of the function, and the input variable is the independent variable of the function. To improve the efficiency of building the surrogate model, it is necessary to predetermine the input and output variable systems. The input variable system includes the input variables, and the output variable system includes the input variables.

[0089] In one embodiment, the output variable includes: oil and gas equivalent Q. o Or recovery rate E or Input variables include: porosity φ n Penetration rate Kn Crude oil viscosity V is Permeability variation coefficient V k Oil layer thickness h, oil-bearing area S, initial oil saturation S oi Water drive recovery level R o degree of miscibility f p Well density D w and the injected gas volume Q ng (t).

[0090] The selection and determination of these input and output variables were derived by the inventors through analysis of various factors affecting the CO2 enhanced oil recovery and storage effect. Moreover, these input and output variables all have clear physical meanings. The determination of these input and output variables is crucial for the subsequent establishment of the CO2 enhanced oil recovery and storage surrogate model. The surrogate model pre-constructed by the inventors is not merely a theoretical model, but fully embodies the close connection between mathematical thinking and objective practice, reflecting the original intention of the pre-constructed surrogate model to solve practical problems in mining areas. The reasonable selection of input and output variables enables the subsequently established surrogate model to move beyond mere theoretical discussion and achieve a unified understanding and application.

[0091] Step S12 involves constructing a surrogate model containing parameters. In this embodiment, the surrogate model is a mathematical model, thus possessing advantages such as ease of calculation and convenient application. Furthermore, the physical meanings of the input and output variables are clear, and the mathematical modeling process is closely linked to actual working conditions; therefore, the surrogate model exhibits high reliability and accuracy. At this point, the values ​​of each parameter in the mathematical model are unknown. These parameters include coefficients and exponents associated with the input variables.

[0092] In one embodiment, a proxy model containing parameters is constructed based on input and output variables, including:

[0093] Based on the analysis of the mechanistic relationship between input and output variables, and the analysis of the mechanistic relationship between each input variable, the combination of each input variable is determined, and associated coefficients and exponents are added to each input variable.

[0094] In one embodiment, the combination methods include: algebraic combination, power law combination, and algebraic-power law combination, but the specific combination methods are not limited in the embodiments of the present invention.

[0095] Thus, a surrogate model containing undetermined parameter values ​​has been constructed. Firstly, the construction of this surrogate model fully considers various combinations of mathematical operations. Secondly, the combination method of this surrogate model is determined based on mechanistic analysis; this non-linear combination method can reflect the complex mechanistic relationship between input and output variables. This further demonstrates that the modeling process of this surrogate model fully considers practical situations and is not an arbitrarily constructed mathematical model, potentially giving the surrogate model a high degree of reliability.

[0096] In one embodiment, the proxy model that includes parameters is as follows:

[0097]

[0098] in:

[0099]

[0100] φ n Porosity, %;

[0101] K n Penetration rate, 10 -3 μm 2 (mD);

[0102] V is Crude oil viscosity, mPa·s;

[0103] V k : Coefficient of variation of permeability, decimal;

[0104] h: Oil layer thickness, m;

[0105] S: Oil-bearing area, km² 2 ;

[0106] S oi Initial oil saturation, %;

[0107] R o : Water drive recovery rate, %;

[0108] f p : Degree of miscibility, decimal;

[0109] D w Well density, number of wells / km 2 ;

[0110] Q ng (t): Injected gas volume, m 3 ;

[0111] N n Cumulative oil and gas production equivalent, m 3 ;

[0112] Qo Oil and gas equivalent, m 3 ;

[0113] E or Recovery rate, %;

[0114] a,b,c,d,e,f,g,h,i,j,k,x1,x2,x3,x4,x5,x6,x7,x8,x9,x10: Parameters, including coefficients and exponents, are decimals;

[0115] a',b',c',d',e',f',g',h',i',j',k',y1,y2,y3,y4,y5,y6,y7,y8,y9,y10: Parameters, including coefficients and exponents, are decimals.

[0116] In one embodiment, the proxy model that includes parameters is as follows:

[0117]

[0118] in:

[0119] φ n Porosity, %;

[0120] K n Penetration rate, 10 -3 μm 2 (mD);

[0121] V is Crude oil viscosity, mPa·s;

[0122] V k : Coefficient of variation of permeability, decimal;

[0123] h: Oil layer thickness, m;

[0124] S: Oil-bearing area, km² 2 ;

[0125] S oi Initial oil saturation, %;

[0126] R o : Water drive recovery rate, %;

[0127] f p : Degree of miscibility, decimal;

[0128] D w Well density, number of wells / km 2 ;

[0129] Q ng (t): Injected gas volume, m 3 ;

[0130] Q o Oil and gas equivalent, m 3 ;

[0131] N n Cumulative oil and gas production equivalent, m 3 ;

[0132] E or Recovery rate, %;

[0133] x1,x2,x3,x4,x5,x6,x7,x8,x9,x10,x11,α,β: Parameters, including coefficients and exponents, are decimals;

[0134] y1,y2,y3,y4,y5,y6,y7,y8,y9,y10,y11: Parameters, including coefficients and exponents, are decimals.

[0135] The above two embodiments represent two different proxy models constructed by the inventors, but the embodiments of the present invention are not limited to these two types of proxy models. In these two types of proxy models, the right side of the equal sign includes input variables, coefficients, and exponents, while the left side of the equal sign represents output variables.

[0136] In one embodiment, the input variable includes: porosity φ n Penetration rate K n Crude oil viscosity V is Permeability variation coefficient V k Oil layer thickness h, oil-bearing area S, initial oil saturation S oi Water drive recovery level R o degree of miscibility f p Well density D w and the injected gas volume Q ng (t).

[0137] In one embodiment, the output variable includes: oil and gas equivalent Q. o Or recovery rate E or .

[0138] Step S13: Determine the parameter values ​​in the proxy model, including coefficients and exponents.

[0139] In one embodiment, a dataset of input and output variables is obtained in advance. Methods for obtaining the dataset include survey methods, experimental methods, and numerical analysis methods. For example, surveys can be conducted on developed mining areas to obtain existing exploration, experimental, and other relevant variable data; sampling can be performed on a pre-defined area of ​​the mining area to conduct indoor mechanical experiments or in-situ tests to determine relevant variable data; or a geological model of the mining area can be established, and relevant variable data can be obtained through numerical analysis calculations.

[0140] In one embodiment, the experimental method for acquiring the dataset can employ laboratory testing, in-situ testing, or a combination of both. Because the finite element method (FEM) offers advantages such as ease of solution, convenient programming, and mature software, the numerical analysis method can utilize finite element simulation. The basic steps of the finite element method are: discretizing the continuous geological structure of the mining area into multiple finite elements; performing mechanical analysis on the discrete elements; establishing the relationship between nodal displacements and nodal forces of each element; and then combining all elements into a whole for comprehensive analysis. Through the finite element simulation process, multiple relevant variable data can be obtained, such as the underground CO2 reserves and oil and gas equivalent. This invention is not limited to survey methods, experimental methods, and numerical analysis methods, nor is it limited to different combinations of the above methods.

[0141] With the development of computer and network technologies, information and data sharing has gradually become a trend. In terms of acquiring datasets, large amounts of input and output variable data can be obtained quickly and conveniently through the network. However, since data obtained from the network often contains noise, redundant values, or missing values, preprocessing and feature extraction are necessary. Data noise refers to unreliable or inaccurate data, while redundant values ​​refer to repeated or redundant data. To ensure the accuracy of the subsequent surrogate model construction, it is necessary to preprocess and extract features from the acquired network data. Noise processing can employ methods such as mean filtering, median filtering, and Gaussian filtering. Feature extraction involves extracting useful and crucial information, such as using principal component analysis, linear discriminant analysis, and wavelet transform.

[0142] In one embodiment, based on the acquired dataset, a machine learning model is used to train the constructed surrogate model containing parameters, and the values ​​of each parameter in the surrogate model, including coefficient values ​​and exponential values, are determined. In the resulting CO2 flooding surrogate model, each parameter is uniquely determined.

[0143] To determine the values ​​of each parameter in the surrogate model, machine learning strategies can be applied. For example, a backpropagation (BP) neural network can be used to train the constructed surrogate model containing parameters, fitting and determining the coefficient and exponent values ​​in the surrogate model. Using the input and output variable data obtained through the above examples, the parameter-containing surrogate model is trained, enabling the constructed surrogate model to automatically learn the mapping relationship between the output variable data and each output variable. The training process of the BP neural network involves fitting the acquired dataset with the constructed surrogate model, ensuring that the surrogate model and the dataset achieve a preset degree of fit. A BP neural network is a multi-layer feedforward neural network trained using the backpropagation algorithm. Based on gradient search technology, it minimizes the mean square error between the actual and expected output values ​​of the neural network. Theoretically, a BP neural network model can fit any function.

[0144] In this embodiment of the invention, the constructed CO2 enhanced oil recovery surrogate model is a nonlinear mathematical model. The input to its BP neural network model is the input variable data from the dataset, and the model output is the predicted value of the output variable. Based on a preset loss function, the difference between the predicted output variable value and the actual output variable data in the dataset is determined. The parameters to be determined in the surrogate model are adjusted based on the magnitude of this difference. The BP neural network model is continuously trained until the value of the loss function reaches the convergence condition. The loss function can be the mean squared error, i.e., the mean of the squares of the predicted output variable value and the actual output variable data value, but this embodiment of the invention does not limit this.

[0145] To construct a more accurate surrogate model, the dataset used to fit the surrogate model should contain a sufficient amount of relevant data. To effectively utilize the large dataset and avoid unnecessary human interference in the modeling process, the inventors employed a machine learning strategy to fit the parameters in the determined surrogate model. The resulting model parameters exhibit higher reliability and accuracy compared to other fitting methods.

[0146] Once the parameters are determined, a CO2 flooding and storage surrogate model with unique parameters is obtained. The obtained CO2 flooding and storage surrogate model is presented as an intuitive mathematical model, which is easy to understand and promote.

[0147] Step S14: Iteratively optimize the obtained CO2 enhanced oil recovery surrogate model. First, determine or verify whether the surrogate model meets the preset requirements; then, based on the determination result, optimize the surrogate model as appropriate.

[0148] In one embodiment, the obtained CO2 enhanced oil recovery surrogate model is used to fit the real data of the input variables and the real data of the output variables in the mining area. Based on the fitting results, a goodness-of-fit index is determined. It is then determined whether the goodness-of-fit index meets the preset requirements. If it does, the obtained surrogate model has met the preset requirements and does not need to be optimized and can be used directly. If it does not meet the preset requirements, the surrogate model has not met the preset requirements and cannot be used directly. It needs to be optimized according to the preset strategy.

[0149] In one embodiment, if the CO2 flooding surviving surviving surviving model does not meet the preset requirements, it is optimized according to the following preset strategy: A new surviving model containing parameters is reconstructed. Based on the acquired dataset (including input and output variables), the surviving model is trained or fitted to determine the parameter values, thus obtaining an updated CO2 flooding surviving surviving surviving model. The updated surviving model is then reassessed to determine if it meets the preset requirements. If not, the above steps are repeated until the obtained CO2 flooding surviving surviving surviving model meets the preset requirements, at which point the process stops, ultimately yielding the CO2 flooding surviving surviving surviving model. After determining the parameters, the surviving model is iteratively optimized. The surviving model after iterative optimization has higher robustness and risk resistance.

[0150] In one embodiment, the coefficient of determination is used as the fit index, but this embodiment of the invention does not limit this. For example, when determining whether the surrogate model meets the preset requirements, if the coefficient of determination is greater than 85%, the surrogate is not optimized; if the coefficient of determination is less than or equal to 85%, the surrogate model is optimized.

[0151] The CO2 enhanced oil recovery surviving ...

[0152] When iteratively optimizing the surrogate model, the inventors reconstructed the surrogate model from the very beginning, starting with the initial mathematical model for any model that did not meet the preset requirements. Because the surrogate model constructed by the inventors comprehensively considers the real physical world (such as defined input and output variables) and mathematical algorithms (such as algebraic combinations, power-law combinations, and algebraic-power-law combinations), the inventors readjusted the corresponding input variables, output variables, and mathematical expressions from the very source of the surrogate model's creation, rather than starting optimization from a certain stage in the process of building the surrogate model. This ensures that the optimization process is rooted in the source and improves the accuracy of the CO2 enhanced oil recovery surrogate model by fully connecting it with the objective world.

[0153] This invention provides an evaluation method for a CO2 flooding surcharge surcharge model, referring to... Figure 2 As shown, it includes:

[0154] S21. Based on the input variable data obtained from the evaluation area and the pre-constructed CO2 flooding and storage surrogate model, predict the CO2 flooding and storage effect of the evaluation area.

[0155] S22. Perform sensitivity analysis on the input variables of the CO2 flooding and storage proxy model to determine the key input variables that affect the CO2 flooding and storage effect; based on the analysis of the key input variables, optimize the CO2 flooding and storage measures in the evaluation area.

[0156] Step S21: In one embodiment, input variable data is acquired in advance, and these input variables are substituted into the constructed CO2 flooding and storage surrogate model to calculate output variables such as oil and gas equivalent or recovery rate. Based on the magnitude of the output variable values, the CO2 flooding and storage effect in the evaluation area is predicted. The CO2 flooding and storage surrogate model is obtained in advance using the aforementioned method for constructing a CO2 flooding and storage surrogate model.

[0157] Step S22, in one embodiment, involves performing sensitivity analysis on the input variables in the CO2 flooding and storage surrogate model. For example, the value of one input variable is changed one by one while the values ​​of other input variables remain unchanged to obtain the change pattern of the corresponding output variable. Key input variables affecting the output variable, or the flooding and storage effect, can also be identified. Based on these results, a reasonable construction plan for CO2 flooding and storage in the evaluation area can be proposed. For example, for certain key input variables, the physical meaning of the input variables is analyzed, and construction measures such as fracturing and acidizing are proposed in conjunction with the actual geological conditions of the evaluation area to improve the values ​​of the corresponding key input variables.

[0158] As described above, this invention also provides a method for quantitatively predicting and evaluating the CO2 flooding effect in a new evaluation area using a constructed CO2 flooding surviv ...

[0159] In addition, the predictive analysis results of the evaluation area can help to know the economic and social benefits of the CO2 flooding and storage project in the evaluation area in advance, and provide direction and theoretical support for the subsequent project.

[0160] Based on the same inventive concept, this embodiment of the invention also provides a related device for a CO2 flooding and storage surrogate model. Since the principle of the problem solved by this device is similar to the aforementioned method for constructing a CO2 flooding and storage surrogate model, the implementation of this device can refer to the implementation of the aforementioned method, and the repeated parts will not be described again.

[0161] This invention provides a device for a CO2 flooding and storage proxy model, referring to... Figure 3 As shown, it includes:

[0162] Variable determination module 31 is used to determine the input and output variables used to construct the proxy model;

[0163] The model building module 32 is used to build a surrogate model containing parameters based on the input variables and the output variables; the parameters include: coefficients and exponents associated with the input variables;

[0164] The model determination module 33 is used to train the surrogate model based on the dataset corresponding to the acquired input variables and output variables using a machine learning strategy, and determine the coefficient value and the exponent value of the surrogate model to obtain the CO2 flooding storage surrogate model.

[0165] The iterative optimization module 34 is used to iteratively optimize the CO2 flooding and storage proxy model based on the real data of the input variables and the real data of the output variables in the mining area.

[0166] This invention provides a computing device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the aforementioned method for constructing a CO2 flooding storage surrogate model or the method for evaluating a CO2 flooding storage surrogate model.

[0167] This invention provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the aforementioned method for constructing a CO2 flooding surcharge surcharge surcharge model or the method for evaluating a CO2 flooding surcharge surcharge surcharge model.

[0168] This invention provides a computer program product, including: a computer program, which, when executed by a processor, implements the aforementioned method for constructing a CO2 flooding storage surrogate model or the method for evaluating a CO2 flooding storage surrogate model.

[0169] Obviously, those skilled in the art can make various modifications to this invention without departing from its spirit and scope. Therefore, if these modifications fall within the scope of the claims and their equivalents, this invention is also intended to include these modifications.

Claims

1. A method for constructing a CO2 enhanced oil recovery and storage surrogate model, characterized in that, include: Determine the input and output variables used to build the proxy model; Based on the input and output variables, a surrogate model containing parameters is constructed; the parameters include coefficients and exponents associated with the input variables; Based on the datasets corresponding to the obtained input and output variables, the surrogate model is trained using a machine learning strategy to determine the coefficient values ​​and the exponent values ​​of the surrogate model, thereby obtaining a CO2 flooding storage surrogate model. Based on real data of input variables and output variables from the mining area, the CO2 enhanced oil recovery and storage proxy model is iteratively optimized.

2. The method as described in claim 1, characterized in that, The CO2 enhanced oil recovery and storage surrogate model is iteratively optimized based on real data of input variables and output variables from the mining area, including: The surrogate model is used to fit the real data of the input variables and the real data of the output variables to determine the goodness-of-fit index. If the fitting index meets the preset requirements, the CO2 flooding and storage surrogate model will not be optimized. If the fitting degree index does not meet the preset requirements, the surrogate model is reconstructed, the reconstructed surrogate model is trained, and an updated CO2 flooding storage surrogate model is obtained, until the fitting degree index corresponding to the updated CO2 flooding storage surrogate model meets the preset requirements.

3. The method as described in claim 1, characterized in that, The construction of the proxy model containing parameters includes: Based on the analysis of the input and output variables and the mechanistic relationships between the input variables, the combination methods between the input variables and the coefficients and exponents associated with the input variables are determined.

4. The method as described in claim 3, characterized in that, The combination methods include: algebraic combination, power law combination, and algebraic-power law combination.

5. The method as described in claim 3, characterized in that, The input variables include: Porosity φ n Penetration rate K n Crude oil viscosity V is Permeability variation coefficient V k Oil layer thickness h, oil-bearing area S, initial oil saturation S oi Water drive recovery level R o degree of miscibility f p Well density D w and the injected gas volume Q ng (t).

6. The method as described in claim 3, characterized in that, The output variables include: Oil and gas equivalent Q o Or recovery rate E or .

7. The method as described in claim 3, characterized in that, The proxy model containing parameters is shown below: in: φ n Porosity, %; K n Penetration rate, 10 -3 μm 2 (mD); V is Crude oil viscosity, mPa·s; V k : Coefficient of variation of permeability, decimal; h: Oil layer thickness, m; S: Oil-bearing area, km² 2 ; S oi Initial oil saturation, %; R o : Water drive recovery rate, %; f p : Degree of multiphase, decimal; D w Well density, number of wells / km 2 ; Q ng (t): Injected gas volume, m 3 ; N n Cumulative oil and gas production equivalent, m 3 ; Q o Oil and gas equivalent, m 3 ; E or Recovery rate, %; a,b,c,d,e,f,g,h,i,j,k,x1,x2,x3,x4,x5,x6,x7,x8,x9,x10: Parameters, including coefficients and exponents, are decimals; a',b',c',d',e',f',g',h',i',j',k',y1,y2,y3,y4,y5,y6,y7,y8,y9,y10: Parameters, including coefficients and exponents, are decimals.

8. The method as described in claim 3, characterized in that, The proxy model containing parameters is shown below: in: φ n Porosity, %; K n Penetration rate, 10 -3 μm 2 (mD); V is Crude oil viscosity, mPa·s; V k : Coefficient of variation of permeability, decimal; h: Oil layer thickness, m; S: Oil-bearing area, km² 2 ; S oi Initial oil saturation, %; R o : Water drive recovery rate, %; f p : Degree of multiphase, decimal; D w Well density, number of wells / km 2 ; Q ng (t): Injected gas volume, m 3 ; Q o Oil and gas equivalent, m 3 ; N n Cumulative oil and gas production equivalent, m 3 ; E or Recovery rate, %; x1,x2,x3,x4,x5,x6,x7,x8,x9,x10,x11,α,β: Parameters, including coefficients and exponents, are decimals; y1,y2,y3,y4,y5,y6,y7,y8,y9,y10,y11: Parameters, including coefficients and exponents, are decimals.

9. An evaluation method for a CO2 flooding surviving surviving surviving model, characterized in that, include: Based on the input variable data obtained from the evaluation area and the pre-built CO2 flooding and storage surrogate model, the CO2 flooding and storage effect of the evaluation area is predicted. Sensitivity analysis of input variables was performed on the CO2 flooding and storage surrogate model to identify the key input variables affecting the effectiveness of CO2 flooding and storage; based on the analysis of the key input variables, the CO2 flooding and storage measures in the evaluation area were optimized. The CO2 flooding storage surrogate model is obtained in advance by the surrogate model construction method described in any one of claims 1-8.

10. A device related to a CO2 flooding and storage surrogate model, characterized in that, include: The variable determination module is used to determine the input and output variables used to build the proxy model; The model building module is used to construct a surrogate model containing parameters based on input and output variables; the parameters include coefficients and exponents associated with the input variables; The model determination module is used to train the surrogate model based on the dataset corresponding to the acquired input and output variables using a machine learning strategy, and determine the coefficient value and the exponent value of the surrogate model to obtain the CO2 flooding storage surrogate model. The iterative optimization module is used to iteratively optimize the CO2 enhanced oil recovery and storage proxy model based on the real data of the input variables and the real data of the output variables in the mining area.

11. A computing device, characterized in that, include: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the program, implements the method for constructing the CO2 flooding surcharge surcharge surcharge model according to any one of claims 1-8 or the method for evaluating the CO2 flooding surcharge surcharge surcharge model according to claim 9.

12. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the method for constructing the CO2 flooding storage surrogate model according to any one of claims 1-8 or the method for evaluating the CO2 flooding storage surrogate model according to claim 9.

13. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the method for constructing the CO2 flooding storage surrogate model according to any one of claims 1-8 or the method for evaluating the CO2 flooding storage surrogate model according to claim 9.

Images