Carbon dioxide sequestration prediction method and device based on mechanism and data fusion driving

Abstract

The invention discloses a carbon dioxide sequestration prediction method and device based on mechanism and data fusion driving. The method comprises the steps that reservoir parameters, well parameters and fluid parameters of a shale reservoir are obtained; establishing a multi-scale seepage control model according to the reservoir parameters, the well parameters and the fluid parameters; embedding the multi-scale seepage control model into a pre-trained carbon dioxide sequestration deep learning model to obtain a shale reservoir carbon dioxide sequestration prediction model; and carrying out sealing prediction on the reservoir parameters, the well parameters and the fluid parameters through the shale reservoir carbon dioxide sealing prediction model to obtain a prediction result.

Description

Technical Field

[0001] The present invention relates to the technical field of carbon capture and storage (CCUS), and particularly to a method and device for predicting carbon dioxide storage driven by the fusion of mechanism and data. Background Art

[0002] With the development of industrialization, the large-scale emission of carbon dioxide (CO2) has triggered the problem of global warming. Carbon capture and storage (CCUS) technology is regarded as an important means to address climate change. Among them, geological storage technology can not only slow down the increase in the concentration of carbon dioxide in the atmosphere but also improve the oil and gas recovery rate by injecting CO2 into deep underground reservoirs (such as shale formations). The widely developed micro-nano scale pores inside the shale reservoir provide good geological conditions for CO2 storage. However, the internal structure of the shale reservoir after volume fracturing is more complex, including a large number of nano-scale organic pores and micro-nano scale clay mineral pores, as well as natural fractures at the micron to millimeter scale and artificial fractures at a larger scale, forming a strongly heterogeneous and cross-scale multi-scale storage and seepage system. The complex pore and fracture structure makes the CO2 show significant multi-scale seepage characteristics during the storage process, seriously restricting the reliability of CO2 storage dynamic prediction.

[0003] The existing technology mainly relies on numerical simulation methods for CO2 storage dynamic prediction. Numerical simulation can accurately describe the seepage process by constructing a reservoir mathematical model and performing discrete solutions. However, the current numerical simulators for CO2 storage in shale reservoirs have systematic deficiencies in integrating micro and macro multi-scale transport mechanisms (such as surface diffusion, slip flow, adsorption and desorption, and Darcy flow, etc.), lacking effective cross-scale physical mechanism coupling, resulting in a lack of precise connection between the micro process and the macro flow, and thus causing large deviations in key prediction parameters. At the same time, the traditional numerical simulation method is limited by the grid division and discretization efficiency. When facing a complex reservoir structure from nano-pores to kilometer-scale fractures, the calculation amount is huge, and the single simulation cycle is long, making it difficult to meet the requirements of rapid response and real-time decision-making for CO2 storage dynamic management.

[0004] In recent years, a large number of training samples have been generated using reservoir numerical simulators, and a non-linear mapping relationship between the input conditions and the storage dynamic results has been established through machine learning methods, achieving a relatively high prediction efficiency. Although this proxy modeling method takes into account the prediction efficiency and accuracy to a certain extent, due to the lack of physical constraints during the training process, the model shows typical black-box characteristics, making it difficult to clearly infer the internal learning mechanism and reducing the reliability and credibility in engineering applications. Therefore, there is an urgent need to develop a new intelligent prediction method that integrates multi-scale seepage physical mechanisms while maintaining high computational efficiency, so as to significantly reduce the generalization error of CO2 storage dynamic prediction and provide a technical solution with high efficiency, accuracy, and physical interpretability for actual storage projects.

[0005] This section aims to provide background or context for the embodiments of the present invention described in the claims. The description herein is not admitted to be prior art merely by virtue of its inclusion in this section. Summary of the Invention

[0006] Embodiments of the present invention provide a carbon dioxide sequestration prediction method driven by the fusion of mechanism and data, which is used to realize the fusion modeling of multi-scale seepage mechanism and deep learning model of carbon dioxide sequestration, and improve the prediction accuracy and calculation efficiency of the carbon dioxide sequestration process.

[0007] The carbon dioxide sequestration prediction method driven by the fusion of mechanism and data includes:

[0008] Obtain reservoir parameters, well parameters and fluid parameters of the shale reservoir;

[0009] According to the reservoir parameters, the well parameters and the fluid parameters, establish a multi-scale seepage control model;

[0010] Embed the multi-scale seepage control model into a pre-trained deep learning model of carbon dioxide sequestration to obtain a carbon dioxide sequestration prediction model for the shale reservoir;

[0011] Perform sequestration prediction on the reservoir parameters, the well parameters and the fluid parameters through the carbon dioxide sequestration prediction model for the shale reservoir to obtain a prediction result.

[0012] In one embodiment, the establishing a multi-scale seepage control model according to the reservoir parameters, the well parameters and the fluid parameters includes:

[0013] Establish a horizontal well injection model according to the reservoir parameters, the well parameters and the fluid parameters;

[0014] According to the reservoir parameters and the fluid parameters, establish a multi-scale seepage control model corresponding to the horizontal well injection model.

[0015] In one embodiment, the establishing a horizontal well injection model according to the reservoir parameters, the well parameters and the fluid parameters includes:

[0016] Determine injection engineering parameters according to the reservoir parameters, the well parameters and the fluid parameters;

[0017] Determine the system composition of the horizontal well injection model and its corresponding fluid seepage mechanism according to the reservoir parameters and the fluid parameters

[0018] Determine the initial conditions and boundary conditions of carbon dioxide sequestration according to the injection engineering parameters and the system composition of the horizontal well injection model;

[0019] The horizontal well injection model is established based on the system composition of the horizontal well injection model, its corresponding fluid seepage mechanism, the injection engineering parameters, the initial conditions, and the boundary conditions.

[0020] In one embodiment, the training steps of the carbon dioxide sequestration deep learning model include:

[0021] Based on the reservoir parameters, the fluid parameters, and the injection engineering parameters, a multi-modal training dataset is established;

[0022] The high-dimensional data in the multi-modal training dataset is input into the carbon dioxide sequestration deep learning model for encoding processing to obtain a low-dimensional feature space;

[0023] Transformer coupling is performed in the low-dimensional feature space to obtain a coupled relationship feature vector;

[0024] The coupled relationship feature vector is decoded to obtain high-dimensional prediction data.

[0025] In one embodiment, the performing Transformer coupling in the low-dimensional feature space to obtain a coupled relationship feature vector includes:

[0026] A linear transformation operation is performed on the low-dimensional feature space to respectively generate a query vector, a key vector, and a value vector;

[0027] According to the query vector, the key vector, the value vector, and the softmax function, the fused coupled relationship feature vector is determined.

[0028] In one embodiment, embedding the multi-scale seepage control model into a pre-trained carbon dioxide sequestration deep learning model to obtain a shale reservoir carbon dioxide sequestration prediction model includes:

[0029] The multi-scale seepage control model is discretized to obtain a multi-scale seepage control model in differential form;

[0030] The high-dimensional prediction data is input into the multi-scale seepage control model in differential form to obtain a physical residual loss;

[0031] According to the high-dimensional prediction data, a data fitting loss, a boundary condition residual, and an initial condition residual are respectively determined;

[0032] According to the physical residual loss, the data fitting loss, the boundary condition residual, the initial condition residual, and their corresponding weight factors, a total loss function is determined;

[0033] Update the model parameters through the Adam optimizer and the total loss function, and repeat the above steps for iterative training until the total loss function converges to obtain the shale reservoir carbon dioxide sequestration prediction model.

[0034] In one embodiment, based on the reservoir parameters, the fluid parameters, and the injection engineering parameters, a multi-modal training dataset is established, including:

[0035] Construct a multi-field coupling numerical model through a multi-physics simulation platform;

[0036] Construct a multi-modal training dataset according to the reservoir parameters, the fluid parameters, the injection engineering parameters, and the multi-field coupling numerical model.

[0037] In one embodiment, the establishment of the multi-modal training dataset according to the reservoir parameters, the fluid parameters, the injection engineering parameters, and the multi-field coupling numerical model includes:

[0038] Perform Latin hypercube sampling on the reservoir parameters, the fluid parameters, and the injection engineering parameters to generate multiple groups of input samples;

[0039] Input each group of the input samples into the multi-field coupling numerical model for simulation to obtain corresponding output samples;

[0040] Perform standardization processing on each group of the input samples;

[0041] Construct input-output training sample pairs according to the standardized input samples and their corresponding output samples to obtain the multi-modal training dataset.

[0042] In one embodiment, the performance of Latin hypercube sampling on the reservoir parameters, the fluid parameters, and the injection engineering parameters to generate multiple groups of input samples includes:

[0043] Determine the value ranges of the reservoir parameters, the fluid parameters, and the injection engineering parameters;

[0044] Equally and probabilistically divide the value range corresponding to each input parameter according to a preset number of samples to generate multiple sample intervals;

[0045] Perform random sampling within each of the sample intervals to obtain the sample values corresponding to each of the sample intervals;

[0046] Randomly combine the sample values sampled for each input parameter to obtain multiple groups of input samples.

[0047] In one embodiment, the reservoir parameters include: a first pore structure parameter; the fluid parameters include: a first fluid parameter; the horizontal well injection model includes an artificial fracture system; establishing the multi-scale seepage control model corresponding to the horizontal well injection model according to the reservoir parameters and the fluid parameters includes:

[0048] Determine the multi-scale seepage control model corresponding to the artificial fracture system according to the first pore structure parameter and the first fluid parameter; wherein, the first pore structure parameter includes: the permeability of the artificial fracture system, the artificial fracture width and the reservoir thickness; the first fluid parameter includes: the gas viscosity of the artificial fracture system, the gas volume factor.

[0049] In one embodiment, the reservoir parameters further include: a second pore structure parameter and a third pore structure parameter; the fluid parameters further include: a second fluid parameter and a third fluid parameter; the horizontal well injection model further includes a natural fracture system; establishing the multi-scale seepage control model corresponding to the horizontal well injection model according to the reservoir parameters and the fluid parameters includes:

[0050] Determine the multi-scale seepage control model corresponding to the natural fracture system according to the second pore structure parameter, the third pore structure parameter, the second fluid parameter and the third fluid parameter;

[0051] Wherein, the second pore structure parameter includes: the radius of the natural fracture system, the permeability of the natural fracture system, the natural fracture stress sensitivity modulus, the reservoir pressure of the natural fracture system, the interface radius between the fracture system and the inorganic matrix system, the porosity of the natural fracture system and the reservoir temperature of the natural fracture system; the third pore structure parameter includes: the shape factor of the inorganic matrix system, the permeability of the inorganic matrix system, the reservoir pressure of the inorganic matrix system and the radius of the inorganic matrix system; the second fluid parameter includes: the gas density in the natural fracture system, the gas viscosity of the natural fracture system, the gas compressibility factor, the compressibility factor and the ideal gas constant; the third fluid parameter includes: the gas density of the inorganic matrix system.

[0052] In one embodiment, the reservoir parameters further include: a fourth pore structure parameter; the fluid parameters further include: a fourth fluid parameter; the horizontal well injection model further includes an inorganic matrix system; establishing the multi-scale seepage control model corresponding to the horizontal well injection model according to the reservoir parameters and the fluid parameters includes:

[0053] Determine the multi-scale seepage control model corresponding to the inorganic matrix system according to the third pore structure parameter, the fourth pore structure parameter, the third fluid parameter, the fourth fluid parameter and the preset first adsorption model parameter;

[0054] Among them, the third pore structure parameter further includes: clay mineral content, the interfacial radius between the kerogen system and the inorganic matrix system, and the porosity of the inorganic matrix system; the fourth pore structure parameter includes: kerogen content, the shape factor of the kerogen system, the permeability of the kerogen system, the reservoir pressure of the kerogen system, and the radius of the kerogen system; the third fluid parameter further includes: the gas slip coefficient in the inorganic matrix system and the gas compressibility of the inorganic matrix system; the fourth fluid parameter includes: the gas density in the kerogen system; the first adsorption model parameter includes: the isothermal adsorption volume of the inorganic matrix system and the isothermal adsorption pressure of the inorganic matrix system.

[0055] In one embodiment, the reservoir parameter further includes: the fifth pore structure parameter; the fluid parameter further includes: the fifth fluid parameter; the horizontal well injection model further includes a porous kerogen system; establishing the multi-scale seepage control model corresponding to the horizontal well injection model according to the reservoir parameter and the fluid parameter includes:

[0056] Determining the multi-scale seepage control model corresponding to the porous kerogen system according to the fourth pore structure parameter, the fifth pore structure parameter, the fourth fluid parameter, the fifth fluid parameter, and the preset second adsorption model parameter;

[0057] Among them, the fourth pore structure parameter further includes: the porosity of the kerogen system; the fifth pore structure parameter includes: the shape factor of the organic matter system and the particle radius of the organic matter system; the fourth fluid parameter further includes: the gas slip coefficient of the kerogen system and the gas compressibility of the kerogen system; the fifth fluid parameter includes: the gas density, the gas effective diffusion coefficient in the organic matter system, and the gas concentration in the organic matter system; the second adsorption model parameter includes: the isothermal adsorption volume of the kerogen system and the isothermal adsorption pressure of the kerogen system.

[0058] In one embodiment, the horizontal well injection model further includes: the organic matter system, and establishing the multi-scale seepage control model corresponding to the horizontal well injection model according to the reservoir parameter and the fluid parameter includes:

[0059] Determining the multi-scale seepage control model corresponding to the organic matter system according to the fifth pore structure parameter and the fifth fluid parameter.

[0060] The embodiment of the present invention further provides a carbon dioxide sequestration prediction device driven by the fusion of mechanism and data, which is used to realize the fusion modeling of the multi-scale seepage mechanism and the carbon dioxide sequestration deep learning model, and improves the prediction accuracy and calculation efficiency of the carbon dioxide sequestration process.

[0061] The carbon dioxide sequestration prediction device driven by the fusion of mechanism and data includes:

[0062] A parameter acquisition module for acquiring reservoir parameters, well parameters, and fluid parameters of a shale reservoir;

[0063] A model construction module for establishing a multi-scale seepage control model based on the reservoir parameters, the well parameters, and the fluid parameters;

[0064] A model fusion module for embedding the multi-scale seepage control model into a pre-trained deep learning model for carbon dioxide sequestration to obtain a prediction model for carbon dioxide sequestration in a shale reservoir;

[0065] A prediction output module for performing sequestration prediction on the reservoir parameters, the well parameters, and the fluid parameters through the prediction model for carbon dioxide sequestration in a shale reservoir to obtain a prediction result.

[0066] An embodiment of the present invention further provides a computer device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the above-mentioned carbon dioxide sequestration prediction method based on mechanism and data fusion driving is implemented.

[0067] An embodiment of the present invention further provides a computer-readable storage medium. The computer-readable storage medium stores a computer program. When the computer program is executed by a processor, the above-mentioned carbon dioxide sequestration prediction method based on mechanism and data fusion driving is implemented.

[0068] An embodiment of the present invention further provides a computer program product. The computer program product includes a computer program. When the computer program is executed by a processor, the above-mentioned carbon dioxide sequestration prediction method based on mechanism and data fusion driving is implemented.

[0069] The carbon dioxide sequestration prediction method and device driven by the integration of mechanism and data provided by the embodiments of the present invention, in the process of establishing the multi-scale seepage control model, in view of the characteristics of the complex heterogeneous structure of shale reservoirs, the reservoir system is divided into multiple scale levels such as artificial fracture system, natural fracture system, inorganic matrix system, kerogen system and organic matter system, and the corresponding differential form seepage control equations are constructed respectively. This multi-scale seepage control model fully considers the multi-mechanism coupling migration behaviors such as Darcy flow, slip flow, dissolution diffusion and multi-component competitive adsorption, comprehensively reflects the migration and sequestration characteristics of carbon dioxide in different scale sub-domains, and effectively improves the physical description ability and adaptability of the multi-scale seepage control model for the dynamic process of carbon dioxide sequestration. Further, the above multi-scale seepage control model is embedded into the U-Net neural network structure to construct a deep learning model for carbon dioxide sequestration that integrates physical constraints. By introducing the physical residual term, boundary condition consistency term and initial condition constraint term of the differential control equation, a hybrid loss function is constructed, which not only improves the fitting accuracy of the model, but also ensures that the prediction results strictly conform to the physical field constraints, enhancing the interpretability and prediction stability of the model. Through the carbon dioxide sequestration prediction model for shale reservoirs, under the condition of inputting any combination of reservoir parameters, well parameters and fluid parameters, the efficient prediction of the carbon dioxide sequestration process can be realized, and the injection dynamics, pressure distribution and phase transformation results at multiple time scales can be quickly output. Compared with the traditional numerical simulation method, the present invention significantly improves the prediction accuracy and calculation efficiency, and has good engineering practicability. BRIEF DESCRIPTION OF THE DRAWINGS

[0070] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained according to these drawings. In the drawings:

[0071] Figure 1 is a flow chart of the carbon dioxide sequestration prediction method driven by the integration of mechanism and data in an embodiment of the present invention;

[0072] Figure 2 is a flow chart of the carbon dioxide sequestration prediction method driven by the integration of mechanism and data in another embodiment of the present invention;

[0073] Figure 3 is a flow chart of the carbon dioxide sequestration prediction method driven by the integration of mechanism and data in another embodiment of the present invention;

[0074] Figure 4Flowchart of the carbon dioxide sequestration prediction method driven by the fusion of mechanism and data in another embodiment of the present invention;

[0075] Figure 5 Flowchart of the carbon dioxide sequestration prediction method driven by the fusion of mechanism and data in another embodiment of the present invention;

[0076] Figure 6 Flowchart of the carbon dioxide sequestration prediction method driven by the fusion of mechanism and data in another embodiment of the present invention;

[0077] Figure 7 Flowchart of the carbon dioxide sequestration prediction method driven by the fusion of mechanism and data in another embodiment of the present invention;

[0078] Figure 8 Flowchart of the carbon dioxide sequestration prediction method driven by the fusion of mechanism and data in another embodiment of the present invention;

[0079] Figure 9 Flowchart of the carbon dioxide sequestration prediction method driven by the fusion of mechanism and data in another embodiment of the present invention;

[0080] Figure 10 Schematic structural diagram of the horizontal well injection model in an embodiment of the present invention;

[0081] Figure 11 Schematic structural diagram of the carbon dioxide sequestration prediction model for shale reservoirs in an embodiment of the present invention;

[0082] Figure 12 Flowchart of the carbon dioxide sequestration prediction method driven by the fusion of mechanism and data in another embodiment of the present invention;

[0083] Figure 13 Flowchart of the carbon dioxide sequestration prediction method driven by the fusion of mechanism and data in another embodiment of the present invention;

[0084] Figure 14 Flowchart of the carbon dioxide sequestration prediction method driven by the fusion of mechanism and data in another embodiment of the present invention;

[0085] Figure 15 Effect diagram of pressure difference prediction of the carbon dioxide sequestration prediction model for shale reservoirs under the numerical simulation data set in an embodiment of the present invention;

[0086] Figure 16 Effect diagram of pressure difference prediction of the carbon dioxide sequestration prediction model for shale reservoirs under the actual production data of shale reservoirs in an embodiment of the present invention;

[0087] Figure 17 Effect diagram of pressure difference prediction during the carbon dioxide injection process in an embodiment of the present invention;

[0088] Figures 18a to 18c It is the parameter inversion and dynamic prediction effect diagram of the carbon dioxide sequestration process under different data noise conditions in the embodiment of the present invention;

[0089] Figure 19 It is the dynamic prediction effect diagram under different sequestration mechanisms in the whole process of carbon dioxide sequestration in shale reservoirs in the embodiment of the present invention;

[0090] Figure 20 It is the flowchart of the carbon dioxide sequestration prediction method based on mechanism and data fusion drive of the carbon dioxide sequestration prediction device based on mechanism and data fusion drive in the embodiment of the present invention;

[0091] Figure 21 It is the schematic diagram of the entity structure of the electronic device provided by the embodiment of the present invention. Detailed implementation manners

[0092] To make the objectives, technical solutions and advantages of the embodiments of the present invention clearer and more understandable, the following further describes the embodiments of the present invention in detail with reference to the accompanying drawings. Herein, the illustrative embodiments of the present invention and their descriptions are used to explain the present invention, but not to limit the present invention.

[0093] The information collected in the technical solutions of this application is information and data authorized by the user or fully authorized by all parties. Moreover, the processing of relevant data, such as collection, storage, use, processing, transmission, provision, disclosure and application, all comply with the relevant laws, regulations and standards of relevant countries and regions, take necessary confidentiality measures, do not violate public order and good customs, and provide corresponding operation entrances for users to choose to authorize or refuse.

[0094] Provide corresponding operation entrances for users to choose to agree or refuse the results of automated decision-making; if the user chooses to refuse, the expert decision-making process will be entered.

[0095] To solve the technical problems existing in the existing carbon dioxide sequestration prediction method based on mechanism and data fusion drive, such as high numerical simulation calculation cost, large parameter uncertainty and lack of physical consistency, the present invention provides a carbon dioxide sequestration prediction method based on mechanism and data fusion drive. Based on a multi-physical field simulation platform (such as COMSOL), this method constructs a multi-field coupling numerical model that couples the seepage field, chemical field, temperature field and mechanical field, combines the Latin hypercube sampling and SOBOL global sensitivity analysis methods to generate a high-quality training data set, and based on the U-Net neural network, introduces the discrete residuals of the multi-scale seepage equation to construct a physical consistency loss function, realizing the rapid and accurate prediction of carbon dioxide sequestration in shale reservoirs. The present invention effectively improves the calculation efficiency and physical rationality of sequestration dynamic prediction, and enhances the adaptability and generalization ability of the model to complex reservoir conditions.

[0096] As Figure 1 shown, the carbon dioxide sequestration prediction method driven by mechanism and data fusion includes steps 101 to 104.

[0097] Step 101: Obtain reservoir parameters, well parameters, and fluid parameters of the shale reservoir.

[0098] Step 102: Establish a multi-scale seepage control model according to the reservoir parameters, well parameters, and fluid parameters.

[0099] Step 103: Embed the multi-scale seepage control model into a pre-trained deep learning model for carbon dioxide sequestration to obtain a prediction model for carbon dioxide sequestration in the shale reservoir.

[0100] Step 104: Perform sequestration prediction on the reservoir parameters, well parameters, and fluid parameters through the prediction model for carbon dioxide sequestration in the shale reservoir to obtain a prediction result.

[0101] In the embodiment of the present invention, during the establishment of the multi-scale seepage control model, in view of the characteristics of the complex heterogeneous structure of the shale reservoir, the reservoir system is divided into multiple scale levels such as an artificial fracture system, a natural fracture system, an inorganic matrix system, a kerogen system, and an organic matter system, and the corresponding differential form seepage control equations are constructed respectively. The multi-scale seepage control model fully considers the multi-mechanism coupling migration behaviors such as Darcy flow, slip flow, dissolution diffusion, and multi-component competitive adsorption, comprehensively reflects the migration and sequestration characteristics of carbon dioxide in different scale sub-domains, and effectively improves the physical characterization ability and adaptability of the multi-scale seepage control model for the dynamic process of carbon dioxide sequestration. Further, the above multi-scale seepage control model is embedded into the U-Net neural network structure to construct a deep learning model for carbon dioxide sequestration with physical constraints. By introducing the physical residual term, boundary condition consistency term, and initial condition constraint term of the differential control equation, a hybrid loss function is constructed, which not only improves the fitting accuracy of the model but also ensures that the prediction result strictly conforms to the physical field constraints, enhancing the interpretability and prediction stability of the model. Through the prediction model for carbon dioxide sequestration in the shale reservoir, under the condition of inputting any combination of reservoir parameters, well parameters, and fluid parameters, the efficient prediction of the carbon dioxide sequestration process can be realized, and the injection dynamics, pressure distribution, and phase transformation results at multiple time scales can be quickly output. Compared with the traditional numerical simulation method, the present invention significantly improves the prediction accuracy and calculation efficiency and has good engineering practicability.

[0102] As Figure 1 shown, the carbon dioxide sequestration prediction method driven by mechanism and data fusion includes steps 101 to 104.

[0103] Step 101: Obtain reservoir parameters, well parameters, and fluid parameters of the shale reservoir.

[0104] Specifically, reservoir parameters such as reservoir temperature, reservoir pressure, porosity, permeability, kerogen content, clay mineral content, natural fracture permeability, and reservoir compressibility of the shale reservoir can be obtained through core analysis experiments, well logging data interpretation, well testing, and geological modeling, etc. The present invention is not limited thereto.

[0105] In addition, well parameters related to the fracture structure such as artificial fracture width, fracture reconstruction radius, and number of artificial fractures can be obtained according to the fracturing design document, fracturing construction record, and microseismic monitoring results. Combining with wellbore design data and wellbore trajectory measurement data, well parameters related to the wellbore structure such as wellbore trajectory, wellbore radius, and wellbore storage coefficient can be obtained. The present invention is not limited thereto.

[0106] Furthermore, fluid parameters such as CO2 diffusion coefficient, CO2 solubility coefficient, gas density, gas viscosity, gas compressibility, adsorption volume, adsorption pressure, and slip coefficient can be obtained through methods such as PVT experiments and EOS (Equation of State) modeling. The present invention is not limited thereto.

[0107] The above reservoir parameters, well parameters, and fluid parameters are shown in Table 1 below.

[0108] Table 1

[0109]

[0110] Step 102: Establish a multi-scale seepage control model based on the reservoir parameters, well parameters, and fluid parameters.

[0111] In one embodiment, as Figure 2 shown, Step 102 includes Step 201 to Step 202.

[0112] Step 201: Establish a horizontal well injection model (i.e., a physical model of a multi-scale system for carbon dioxide sequestration in shale reservoirs) based on the reservoir parameters, well parameters, and fluid parameters.

[0113] Specifically, for the target shale reservoir, a dense artificial fracture network structure can be formed through volume fracturing transformation to enhance the seepage capacity of the reservoir. Using a horizontal well as the injection well type and implementing carbon dioxide injection operations can construct a complete injection and sequestration scenario, and then carry out simulation modeling and dynamic prediction of the carbon dioxide sequestration behavior inside the shale reservoir.

[0114] As Figure 10As shown, the established multi-scale physical model of carbon dioxide sequestration in shale reservoirs takes into account the heterogeneity of reservoir media and the multi-scale characteristics of pore structures. This multi-scale system includes: an artificial fracture system, a natural fracture system, an inorganic matrix system, a kerogen system, and an organic matter system. The above systems correspond to different fluid seepage mechanisms and sequestration behaviors respectively.

[0115] In one embodiment, as Figure 3 shown, step 201 includes steps 301 to 304.

[0116] Step 301: Determine the injection engineering parameters based on reservoir parameters, well parameters, and fluid parameters.

[0117] Specifically, combining reservoir parameters such as reservoir porosity, permeability, and natural fracture distribution characteristics, well parameters such as wellbore structure, artificial fracture geometry, and their conductivity, and fluid parameters such as carbon dioxide density, carbon dioxide viscosity, compression factor, carbon dioxide solubility, and diffusivity, comprehensively analyze the fluid seepage mechanism and sequestration behavior during the carbon dioxide injection process, and determine the injection engineering parameters (Injection Engineering Parameters) and engineering design conditions (Design Configuration Parameters) applicable to the carbon dioxide sequestration process in shale reservoirs.

[0118] Among them, the engineering design conditions include: injection stage setting, well type selection, and reservoir stimulation methods, etc. The injection engineering parameters include: injection medium type, injection rate, injection time, injection pressure, and total injection volume, etc.

[0119] Furthermore, the well type adopts a horizontal well structure arranged along the shale bedding direction to increase the contact area between the wellbore and the reservoir. The reservoir stimulation method adopts volume fracturing technology to form a high-density artificial fracture system in the reservoir by hydraulic action, which can improve the overall permeability of the shale reservoir. The injection medium is carbon dioxide in a supercritical state, which has high solubility and fluidity. The injection rate should be controlled within a safe range that does not cause abnormal fracture propagation or reservoir stress sensitivity effects. The injection pressure should be higher than the reservoir pore pressure to ensure that the injected fluid smoothly enters the reservoir. At the same time, the injection pressure should be controlled within the allowable range of the wellbore pressure-bearing capacity to ensure the safety of the injection process. The injection time and total injection volume can be optimized according to the sequestration target, reservoir accommodation capacity, and injection safety.

[0120] Step 302: Determine the system composition of the horizontal well injection model and its corresponding fluid seepage mechanism based on reservoir parameters and fluid parameters. Among them, this horizontal well injection model is the multi-scale physical model of carbon dioxide sequestration in shale reservoirs.

[0121] Specifically, in the artificial fracture system, carbon dioxide mainly migrates along the fracture channels in the form of limited conductivity. Its fluid seepage mechanism adopts a modified Darcy flow equation to characterize the fluid transport characteristics under the condition of limited fracture conductivity.

[0122] In the natural fracture system, the seepage mechanism of carbon dioxide follows the traditional Darcy's law, and the permeability can be corrected according to the stress-sensitive characteristics, that is, introducing a stress-sensitive permeability function to reflect the influence of the fracture conductivity change with the formation stress.

[0123] In the inorganic matrix system, this system is mainly composed of an inorganic pore structure formed by clay minerals and other non-organic minerals and micro-nano scale inorganic pores distributed in the organic matter. Carbon dioxide undergoes slip flow in such micro-nano pores, and a slip-corrected Darcy flow equation needs to be used to depict the discontinuous flow behavior under the scale effect. At the same time, the influence of pore size on the Knudsen number needs to be considered to correct the traditional seepage equation for scale. In addition, in the surface area of clay minerals and organic matter, carbon dioxide molecules will undergo reversible physical adsorption, and its adsorption behavior obeys the Langmuir isothermal adsorption model to characterize the non-linear relationship between the adsorption capacity and the partial pressure.

[0124] In the kerogen system, carbon dioxide mainly undergoes slip flow in nano-scale pores, accompanied by competitive adsorption between the primary gas (such as methane) and carbon dioxide. At the same time, there is a certain proportion of dissolution and diffusion behavior of carbon dioxide molecules. To accurately characterize the above non-convective migration process, a dissolution and diffusion term needs to be introduced into the model to describe the diffusion-adsorption process of carbon dioxide molecules into the organic matrix and realize the simulation of the molecular-scale sequestration effect.

[0125] In the organic matter system, the migration of carbon dioxide still shows the characteristics of slip flow, and there is a competitive adsorption process with the primary gas at the same time, and the adsorption behavior conforms to the multi-component Langmuir isothermal adsorption model.

[0126] The seepage mechanisms of the above various systems together constitute the physical flow model of the multi-scale system for carbon dioxide injection and sequestration in shale reservoirs.

[0127] Step 303: Determine the initial conditions and boundary conditions for carbon dioxide sequestration according to the injection engineering parameters and the system composition of the horizontal well injection model.

[0128] Specifically, the initial conditions mainly include the physical state information of the shale reservoir before carbon dioxide injection, including the reservoir pressure field, temperature field, porosity and permeability distributions of systems at various scales, and the concentration field of the original gas (such as methane), etc. The above initial conditions can be comprehensively determined based on historical production data, well logging data interpretation results, well test analysis, and geological modeling data, so as to accurately depict the original physical field state of the reservoir before the start of carbon dioxide sequestration.

[0129] The boundary conditions are set according to the constructed sequestration working condition scenarios, usually including types such as the wellbore boundary of the injection well, far-field boundary, and vertical boundaries (caprock and bottom plate). The injection well boundary can be set as a constant injection rate boundary or a constant injection pressure boundary according to the engineering design parameters to control the injection driving conditions. The far-field boundary can be set as a constant pressure boundary or a no-flow boundary (Neumann boundary) to simulate the constraints or responses of the peripheral area of the sequestration system to the seepage behavior. The vertical boundaries are usually set as adiabatic or impermeable boundaries, which can ensure the vertical closure of the system and conform to the physical constraint conditions during the shale sequestration process.

[0130] In addition, to accurately simulate the seepage and migration behavior of carbon dioxide in the multi-scale system, appropriate continuity conditions, including pressure continuity conditions, mass flux continuity conditions, etc., need to be applied at the interfaces of different sub-domains in the horizontal well injection model to ensure the stability and physical consistency of the horizontal well injection model during the numerical solution process.

[0131] Step 304: Establish a horizontal well injection model based on the system composition of the horizontal well injection model, its corresponding fluid seepage mechanism, injection engineering parameters, initial conditions, and boundary conditions.

[0132] Specifically, according to the spatial structures and distribution characteristics of the artificial fracture system, natural fracture system, inorganic matrix system, kerogen system, and organic matter system, and their corresponding fluid seepage mechanisms, combined with the injection engineering parameters, initial conditions, and boundary conditions, they are uniformly integrated into a coupled physical modeling framework to construct a multi-scale system physical model for the whole process of carbon dioxide injection - migration - sequestration in the shale reservoir.

[0133] According to the spatial structures and distribution characteristics of the artificial fracture system, natural fracture system, inorganic matrix system, kerogen system, and organic matter system, as well as the corresponding seepage mechanisms and migration behaviors of each system, and combined with injection engineering parameters such as the type of injection medium, injection rate, and injection pressure, and initial conditions and boundary constraint conditions such as the pressure field, temperature field, and component distribution before injection, they are uniformly integrated into a multi-scale coupled physical modeling framework, thereby constructing a multi-scale system physical model that describes the entire process of carbon dioxide injection, migration, and storage in shale reservoirs. This multi-scale system physical model can accurately reflect the complex flow characteristics and multi-mechanism storage behaviors inside porous media, providing a theoretical basis and boundary input for subsequent multi-physical field numerical simulations and U-Net neural network training.

[0134] Step 202: Establish a multi-scale seepage control model corresponding to the horizontal well injection model according to reservoir parameters and fluid parameters.

[0135] Specifically, based on the constructed multi-scale system physical model for carbon dioxide storage in shale reservoirs, combined with the spatial structure characteristics and distribution laws of the artificial fracture system, natural fracture system, inorganic matrix system, kerogen system, and organic matter system, seepage control equations for the migration behavior of carbon dioxide in each scale system are established respectively (i.e., the multi-scale seepage control model for carbon dioxide injection and storage in shale reservoirs). This multi-scale seepage control model fully considers the differences in pore scale, seepage mechanism, adsorption behavior, and non-convective migration mechanism between different medium types, and can accurately characterize the multi-scale flow laws of carbon dioxide during the injection-migration-storage process in shale reservoirs.

[0136] In one embodiment, the above reservoir parameters include first pore structure parameters, the above fluid parameters include first fluid parameters, and the above horizontal well injection model includes an artificial fracture system.

[0137] Step 202 specifically includes determining the multi-scale seepage control model (i.e., the multi-scale seepage control equation) corresponding to the artificial fracture system according to the first pore structure parameters and the first fluid parameters. Among them, the first pore structure parameters include the permeability K of the artificial fracture system hf , the width ω of the artificial fracture hf and the vertical thickness h of the fracture section penetrating the reservoir. The first fluid parameters include the viscosity μ of the injected gas and the gas volume coefficient B.

[0138] Specifically, after carbon dioxide is compressed on the ground and enters the supercritical state, it is injected into the shale reservoir through the horizontal wellbore and first enters the fracture system jointly composed of the artificial fracture system and the natural fracture system.

[0139] In the artificial fracture system, carbon dioxide migrates in a form of limited diversion along the fracture channels. Its seepage behavior is affected by fracture geometry constraints and stress sensitivity effects. Therefore, a modified Darcy seepage model is used for description, and finite diversion and stress sensitivity correction terms are introduced to characterize the actual fluid transport characteristics under the condition of limited fracture diversion capacity.

[0140] The multi-scale seepage control equation of the artificial fracture system is as follows:

[0141]

[0142] Among them, K hf is the permeability of the artificial fracture system, ω hf is the width of the artificial fracture, h is the vertical thickness of the fracture segment penetrating the reservoir, μ is the viscosity of the injected gas, B is the gas volume coefficient, p hf is the pressure distribution function along the fracture direction in the artificial fracture segment, q schf(x,t) is the gas flow rate function per unit length of the fracture segment at position x and time t, x is the spatial coordinate of the artificial fracture segment, and t is the time variable.

[0143] In one embodiment, the above reservoir parameters further include: the second pore structure parameter and the third pore structure parameter. The above fluid parameters further include: the second fluid parameter and the third fluid parameter. The above horizontal well injection model further includes a natural fracture system.

[0144] Step 202 specifically includes determining the multi-scale seepage control model corresponding to the natural fracture system according to the second pore structure parameter, the third pore structure parameter, the second fluid parameter, and the third fluid parameter. Among them, the second pore structure parameter includes: the radius r of the natural fracture system, the permeability K lf of the natural fracture system, the natural fracture stress sensitivity modulus α, the reservoir pressure of the natural fracture system, the interface radius R m between the fracture system and the inorganic matrix system, the porosity φ fir of the natural fracture system, and the reservoir temperature T. The third pore structure parameter includes: the shape factor β m of the inorganic matrix system, the permeability K m of the inorganic matrix system, the reservoir pressure p m of the inorganic matrix system, and the radius r m of the inorganic matrix system. The second fluid parameter includes: the gas density ρ ir in the natural fracture system, the gas viscosity μ, the gas molar mass M g , the gas compressibility factor C gir , the compressibility factor Z, and the ideal gas constant R. The third fluid parameter includes: the gas density ρ m of the inorganic matrix system.

[0145] Specifically, in the natural fracture system, the migration process of carbon dioxide follows the traditional Darcy seepage law, and its permeability can be corrected by the stress-sensitive permeability function to characterize the variation characteristics of fracture conductivity with the change of formation stress.

[0146] The multi-scale seepage control equation of the natural fracture system is as follows:

[0147]

[0148] Among them, r is the radius of the natural fracture system, ρ ir is the gas density in the natural fracture system, K lf is the permeability of the natural fracture system, α is the stress-sensitive modulus of the natural fracture, p i is the initial reservoir pressure of the natural fracture system, p ir is the current reservoir pressure of the natural fracture system, μ is the gas viscosity, R m is the interface radius between the fracture system and the inorganic matrix system, ρ m is the gas density in the inorganic matrix system, β m is the shape factor of the inorganic matrix system, K m is the permeability of the inorganic matrix system, p m is the reservoir pressure of the inorganic matrix system, r m is the radius of the inorganic matrix system, M g is the molar mass of the gas, C gir is the gas compressibility factor of the natural fracture system, φ fir is the porosity of the natural fracture system, Z is the compressibility factor, R is the ideal gas constant, T is the reservoir temperature, and t is the time variable.

[0149] In one embodiment, the above reservoir parameters further include: the fourth pore structure parameter. The above fluid parameters further include: the fourth fluid parameter. The above horizontal well injection model further includes an inorganic matrix system.

[0150] Step 202 specifically includes: determining the multi-scale seepage control model corresponding to the inorganic matrix system according to the third pore structure parameter, the fourth pore structure parameter, the third fluid parameter, the fourth fluid parameter, and the preset first adsorption model parameter.

[0151] Among them, the third pore structure parameter further includes: the clay mineral content f c , the interface radius R k between the kerogen system and the inorganic matrix system, and the porosity φ m of the inorganic matrix system. The fourth pore structure parameter includes: the kerogen content f k , the shape factor β k of the kerogen system, the permeability K k, the reservoir pressure p of the kerogen system k and the radius r of the kerogen system k . The third fluid parameter also includes: the gas slip coefficient α in the inorganic matrix system m and the gas compressibility C of the inorganic matrix system gm . The fourth fluid parameter includes: the gas density ρ in the kerogen system k . The first adsorption model parameter includes: the isothermal adsorption volume V of the inorganic matrix system Lc and the isothermal adsorption pressure p of the inorganic matrix system Lc .

[0152] Specifically, carbon dioxide migrates from the fracture system to the inorganic matrix system in the shale matrix. The inorganic matrix system consists of a micro-nano scale inorganic pore structure with a small pore size, where the mean free path of gas molecules is comparable to the pore size, significantly affecting the seepage behavior of carbon dioxide. In this context, the flow of carbon dioxide in the inorganic matrix system exhibits obvious slip flow characteristics, and its overall seepage mechanism is described by a slip-corrected Darcy flow model to reflect the physical properties of non-continuous flow in micro-nano scale pores.

[0153] Furthermore, carbon dioxide molecules can compete with the native gas molecules (such as methane) on the surface of clay particles in the inorganic pore structure to form a reversible sequestration mechanism mainly based on physical adsorption. This physical adsorption process follows the Langmuir isothermal adsorption model, which can be used to characterize the non-linear relationship between the adsorption capacity and the local partial pressure, and further reflect the static sequestration capacity of carbon dioxide in the shale matrix system.

[0154] The multi-scale seepage control equation of the inorganic matrix system is as follows:

[0155]

[0156] where r m is the radius of the inorganic matrix system, ρ m is the gas density in the inorganic matrix system, β m is the shape factor of the inorganic matrix system, α m is the gas slip coefficient in the inorganic matrix system, K m is the permeability of the inorganic matrix system, μ is the gas viscosity, p m is the reservoir pressure of the inorganic matrix system, f c is the clay mineral content, V Lc is the Langmuir isothermal adsorption volume of the clay mineral, p Lc is the Langmuir isothermal adsorption pressure of the clay mineral, R k the interfacial radius between the kerogen system and the inorganic matrix system, ρ kis the gas density in the kerogen system, f k is the kerogen content, β k is the shape factor of the kerogen system, K k The permeability of the kerogen system, p k is the reservoir pressure of the kerogen system, r k is the radius of the kerogen system, M g is the molar mass of the gas, C gm is the gas compressibility factor of the inorganic matrix system, φ m is the porosity of the inorganic matrix system, Z is the compressibility factor, R is the ideal gas constant, T is the reservoir temperature, and t is the time variable.

[0157] In one embodiment, the reservoir parameters further include: the fifth pore structure parameter. The fluid parameters further include: the fifth fluid parameter. The horizontal well injection model further includes a porous kerogen system.

[0158] Step 202 specifically includes: determining the multi-scale seepage control model corresponding to the porous kerogen system according to the fourth pore structure parameter, the fifth pore structure parameter, the fourth fluid parameter, the fifth fluid parameter, and the preset second adsorption model parameters.

[0159] Among them, the fourth pore structure parameter further includes: φ k is the porosity of the kerogen system. The fifth pore structure parameter includes: the shape factor β of the organic matter system o and the particle radius r of the organic matter system o . The fourth fluid parameter further includes: the gas slip coefficient α of the kerogen system k and the gas compressibility factor C of the kerogen system gk . The fifth fluid parameter includes: the gas density ρ in the organic matter system o , the effective gas diffusion coefficient D E and the gas concentration C in the organic matter system o . The second adsorption model parameters include: the isothermal adsorption volume V of the kerogen system Lk and the isothermal adsorption pressure p of the kerogen system Lk .

[0160] Specifically, carbon dioxide further migrates from the inorganic matrix system to the organic matter system dominated by kerogen. Since the pore scale of the kerogen system is usually in the nanometer range and the pore diameter is small, carbon dioxide mainly exists in the form of slip flow in the kerogen system and is significantly affected by the coupling effect of the pore scale and the gas mean free path. At the same time, carbon dioxide molecules can compete for adsorption with the native gas (such as methane) on the pore inner wall surface, and the adsorption process is reversible. Its adsorption mechanism can be described by the multi-component Langmuir isothermal adsorption model, so as to accurately characterize the adsorption distribution of different gas components on the kerogen pore surface.

[0161] To accurately simulate the complex migration mechanism of carbon dioxide in the kerogen system, a coupling mechanism of slip flow, competitive adsorption, and dissolution diffusion is introduced into the multi-scale seepage control model of the kerogen system. This coupling mechanism can not only reflect the convective migration characteristics of carbon dioxide in the pores of kerogen, but also simulate the non-convective migration behavior of carbon dioxide penetrating into the organic matrix inside the pore wall, thus realizing the quantitative characterization of the whole process of carbon dioxide sequestration in the organic matter system.

[0162] The multi-scale seepage control equation of the porous kerogen system is as follows:

[0163]

[0164] Among them, r k is the radius of the kerogen system, ρ k is the gas density in the kerogen system, β k is the shape factor of the kerogen system, α k is the gas slip coefficient of the kerogen system, K k is the permeability of the kerogen system, μ is the gas viscosity, p k is the reservoir pressure of the kerogen system, V Lk is the Langmuir isothermal adsorption volume of the kerogen system, p Lk is the Langmuir isothermal adsorption pressure of the kerogen system, φ k is the porosity of the kerogen system, f k is the kerogen content, R o is the initial particle radius of the organic matter system, ρ o is the gas density in the organic matter system, β o is the shape factor of the organic matter system, D E is the effective diffusion coefficient of the gas, C o is the gas concentration in the organic matter system, r o is the particle radius of the organic matter system, M g is the molar mass of the gas, C gk is the gas compressibility factor of the kerogen system, Z is the compressibility factor, R is the ideal gas constant, T is the reservoir temperature, and t is the time variable.

[0165] In one embodiment, the horizontal well injection model further includes an organic matter system. Step 202 specifically includes: determining the multi-scale seepage control model corresponding to the organic matter system according to the fifth pore structure parameter and the fifth fluid parameter.

[0166] Specifically, carbon dioxide diffuses from the inner wall surface of the nano-scale kerogen pores into the interior of the organic matter particles along the concentration gradient and dissolves in the organic matter system in molecular form, thus forming a solid-phase sequestration form mainly in the dissolved state. To accurately depict this non-convection-dominated migration behavior, a coupling mechanism of slip flow, competitive adsorption, multi-component Langmuir isothermal adsorption, and dissolution diffusion is introduced into the multi-scale seepage control model of the organic matter system, and then the complex migration and sequestration process of carbon dioxide in the organic matter structure at the ultra-micro scale is simulated.

[0167] The multi-scale seepage control equation of the organic matter system is as follows:

[0168]

[0169] Among them, r o is the particle radius of the organic matter system, ρ o is the gas density in the organic matter system, β o is the shape factor of the organic matter system, D E is the effective diffusion coefficient of the gas, C o is the gas concentration in the organic matter system, and t is the time variable.

[0170] In one embodiment, as Figure 4 shown, the training steps of the carbon dioxide sequestration deep learning model include step 401 to step 404.

[0171] Step 401: Based on reservoir parameters, fluid parameters, and injection engineering parameters, establish a multi-modal training data set.

[0172] In one embodiment, as Figure 5 shown, step 401 includes step 501 to step 502.

[0173] Step 501: Construct a multi-field coupling numerical model through a multi-physics simulation platform.

[0174] Specifically, based on the COMSOL multi-physics simulation platform (COMSOL Multiphysics), a numerical modeling method combining the finite element method and the volume averaging method is adopted. The multi-scale seepage characteristics of the shale reservoir and the multi-physics process coupling mechanism of carbon dioxide - water - shale are comprehensively considered. According to the above-established physical model of the multi-scale system for carbon dioxide sequestration in the shale reservoir and the corresponding multi-scale seepage control model, a multi-field coupling numerical model coupling the seepage field, chemical field, temperature field, and mechanical field is established.

[0175] As Figure 12 shown, step 501 includes step 1201 to step 1206.

[0176] Step 1201: Construct a geometric model of the multi-scale system in the COMSOL multi-physics simulation platform.

[0177] Exemplarily, in the COMSOL multi-physics simulation platform, create a new 3D component (such as Component1 (3D)), and select a basic physical interface suitable for complex fluid structures (such as Porous Media Flow or Transport of Diluted Species, etc.).

[0178] As Figure 13 shown, Step 120 includes Step 1301 to Step 1304.

[0179] Step 1301: Construct a Representative Elementary Volume (REV) shell geometry.

[0180] Exemplarily, draw a cube or cuboid structure (such as a size of 1m × 1m × 1m) under the Geometry node as the computational domain of the multi-scale shale reservoir.

[0181] Step 1302: Embed a Discrete Fracture Network (DFN) inside the REV to establish an artificial fracture system and a natural fracture system.

[0182] Exemplarily, inside the REV geometry, create multiple planar fractures through the Work Plane node. The fractures can be modeled as ultra-thin cuboids or film volumes, with the fracture thickness set to the millimeter level, and set the spatial position, distribution direction, and geometric dimensions of the fractures to truly reflect the structural characteristics of artificial fractures and natural fractures.

[0183] Step 1303: Construct an inorganic matrix system and a kerogen system inside the REV.

[0184] Exemplarily, generate point cloud or volume mesh data of the microscopic pore network with the help of third-party software (such as Avizo, PoreSpy, or OpenPNM), and save the data in STL or OBJ format. Then import the above microscopic pore structure in the Import Geometry node of the COMSOL multi-physics simulation platform, and specify the imported pore-throat network structure as a "hollow domain", and set the corresponding porosity and slip-corrected flow characteristic parameters to establish a three-dimensional geometric model of the inorganic matrix system and the kerogen system.

[0185] Step 1304: After completing the modeling of the artificial fracture system, natural fracture system, inorganic matrix system, and kerogen system, physically connect different systems to obtain a multi-scale geometric model.

[0186] Exemplarily, the Identity Pair node or Continuity Boundary Conditions node in the COMSOL multiphysics simulation platform is used to set physical coupling for the boundaries between systems, ensuring the continuity of flow variables (such as pressure and concentration, etc.) at the boundaries of each system, which can meet the physical closure of the coupling model.

[0187] In addition, during the above-mentioned modeling process, different porous medium parameters can be set according to the characteristics of each scale system. For example: the fracture system area is set to have high permeability, high porosity, and low adsorption capacity; the inorganic matrix area is set to have low permeability, medium porosity, and a slip flow correction term is added; the kerogen system area is set to have extremely low permeability and high adsorption capacity to reflect the main controlling role of kerogen in carbon dioxide sequestration.

[0188] Step 1202: Construct multiple physical field control modules in the COMSOL multiphysics simulation platform. Among them, the physical fields include a seepage field, a chemical field, a temperature field, and a mechanical field.

[0189] Exemplarily, in the Component node of the COMSOL engineering file, the following modules are added in sequence: the Darcy’s Law module (for the seepage field), the Transport of Diluted Species module (for the chemical field), the Heat Transfer in Porous Media module (for the temperature field), and the Solid Mechanics module (for the mechanical field).

[0190] Step 1203: Set the key physical parameters of each physical field control module.

[0191] Exemplarily, in the Darcy’s Law module, set the porosity, permeability, and physical properties related to fluid flow such as gas density and gas viscosity in the porous medium area. In the Transport of Diluted Species module, set the concentration and diffusion coefficient of the injected medium (such as CO2 and related mineral ions), and add reaction equations describing the CO2 dissolution reaction and mineral precipitation reaction. In the Heat Transfer in Porous Media module, set the heat transfer parameters such as thermal conductivity, specific heat capacity, and initial temperature field. In the Solid Mechanics module, set the mechanical parameters such as the elastic modulus, Poisson's ratio, and Biot coefficient of the rock material to describe the pore-elastic coupling behavior, but the present invention is not limited thereto.

[0192] Step 1204: Establish the coupling relationships between various physical fields in the Multiphysics node of the COMSOL project file.

[0193] Exemplarily, the Poroelasticity Coupling interface can be selected to establish the coupling relationship between the Darcy’s Law module and the Solid Mechanics module, so as to achieve the two-way coupling between the seepage field and the mechanical field. Further, the Non-Isothermal Flow Coupling interface is selected to achieve the coupling between the Darcy’s Law module and the Heat Transfer in Porous Media module. During the non-isothermal flow coupling process, the fluid velocity field is set to drive the heat convection term, and the dependence of temperature on fluid parameters (such as gas density and gas viscosity, etc.) is configured to characterize the feedback regulation effect of temperature change on the seepage behavior. In addition, in the Transport of Diluted Species module, the mass flux term caused by the seepage field is added through the Source Term node to construct the coupling mechanism between the seepage field and the chemical field, and the coupled modeling of the carbon dioxide migration and the mineral reaction process is realized.

[0194] Step 1205: Set the custom partial differential equations in the COMSOL multiphysics simulation platform. The partial differential equations include: the mineral dissolution rate equation, the porosity evolution equation, and the permeability dynamic change equation, which are not limited to this in the present invention.

[0195] Exemplarily, in the Add Physics node of the COMSOL project file, the Weak Form PDE module is added, and the custom PDE expression is input.

[0196] For example:

[0197] The mineral dissolution rate equation is as follows:

[0198]

[0199] Where C is the mineral concentration or the mass of undissolved minerals, t is the time variable, k diss is the mineral dissolution rate constant, and C co2 is the CO2 concentration in the reaction environment.

[0200] The porosity evolution equation is as follows:

[0201]

[0202] Where φ is the porosity at the current moment, φ0 is the initial porosity, and △φ reaction represents the change in porosity caused by the reaction between minerals and fluids.

[0203] The dynamic change equation of permeability is as follows:

[0204]

[0205] where K is the permeability at the current moment, K o is the initial permeability, φ is the porosity at the current moment, φ0 is the initial porosity, and n is an empirical exponent, usually taking values from 2 to 5.

[0206] Step 1206: Based on the above multi-physical field control module, set the injection parameters, boundary conditions, and initial field, and perform simulations to obtain a multi-field coupling numerical model.

[0207] Exemplarily, in the Darcy’s Law module, for the boundary corresponding to the injection wellbore, apply boundary conditions, and a constant injection rate boundary or a constant injection pressure boundary can be selected.

[0208] In the Transport of Diluted Species module, apply the Inflow condition at the same injection boundary and set the concentration value of the injection medium (such as carbon dioxide). In the Heat Transfer in Porous Media module, set the temperature condition at the injection boundary and the temperature value of the injection medium.

[0209] After completing the physical field assignment of the initial state of the model through the above parameter settings and boundary configurations, the simulation calculation can be started, and then a multi-field coupling numerical model including the seepage field, chemical field, temperature field, and mechanical field can be obtained, providing basic data support for the subsequent simulation of the carbon dioxide dynamic storage process and the training of the U-Net neural network.

[0210] Step 502: Construct a multi-modal training dataset according to the reservoir parameters, fluid parameters, injection engineering parameters, and multi-field coupling numerical model.

[0211] In one embodiment, as Figure 6 shown, step 502 includes: steps 601 to 604.

[0212] Step 601: Perform Latin hypercube sampling on the reservoir parameters, fluid parameters, and injection engineering parameters to generate multiple groups of input samples.

[0213] Specifically, use the Latin hypercube sampling method to generate samples for the input parameters, and obtain multiple groups of input samples. Among them, the input parameters include reservoir parameters, fluid parameters, and injection engineering parameters.

[0214] In one embodiment, as Figure 7 shown, step 601 includes: steps 701 to 704.

[0215] Step 701: Determine the value ranges of reservoir parameters, fluid parameters, and injection engineering parameters.

[0216] Exemplarily, the value ranges of the input parameters are shown in Table 2 below, and the present invention is not limited thereto.

[0217] Table 2

[0218]

[0219] Step 702: Divide the value range corresponding to each input parameter into equal-probability segments according to a preset number of samples to generate multiple sample intervals.

[0220] Exemplarily, the number of samples is usually set to 100 groups, and the present invention is not limited thereto.

[0221] Divide the porosity [0.03, 0.05], matrix permeability [10 -18 , 10 -15 ], fracture permeability [10 -16 , 10 -13 ], fracture density [0.01, 0.1], injection pressure [20, 40], and injection temperature [310, 340] into 100 equal-probability sample intervals respectively.

[0222] Step 703: Conduct random sampling within each sample interval to obtain the sample value corresponding to each sample interval.

[0223] Specifically, randomly select a sample value from each sample interval corresponding to each input parameter.

[0224] Exemplarily, in the 23rd sample interval [0.042, 0.0432] of the porosity, randomly select a sample value, such as 0.0427, as the sample value corresponding to this sample interval.

[0225] Step 704: Randomly combine all the sample values sampled from each input parameter to obtain multiple groups of input samples.

[0226] Specifically, each group of input samples is composed of a sample value vector, and the input sample set is X = {x1, x2, …, x k , …, x n}, where x k is the kth group of input samples, and n is the total number of samples.

[0227] For example:

[0228] x k={Porosity, Permeability, Fracture density, Injection pressure, …, Injection temperature} = {0.035, 1.5×10 -16 , 0.05, 28.7, …, 321}。

[0229] Step 602: Input each group of input samples into the multi-field coupling numerical model for simulation to obtain the corresponding output samples.

[0230] Specifically, input the above input samples into the constructed multi-field coupling numerical model in sequence for simulation to obtain the dynamic data of carbon dioxide sequestration in the shale reservoir corresponding to different combinations of input parameters, that is, the output samples.

[0231] The output sample set is Y = {y1, y2, …, y k , …, y n}.

[0232] Among them, y k is the k-th group of output samples, n is the total number of groups of output samples, and y k contains variables such as injection dynamics, pressure distribution, and concentration field during the carbon dioxide sequestration process.

[0233] Step 603: Perform standardization processing on each group of input samples.

[0234] Specifically, use the following standardization formula to normalize each component x k of the input sample vector x ki .

[0235]

[0236] Among them, x* ki is the i-th input component after standardization, x ki is the original value of the i-th input component before standardization, μ(x k ) is the sample mean, and σ(x k ) is the sample standard deviation.

[0237] Step 604: Construct input-output training sample pairs based on the standardized input samples and their corresponding output samples to obtain a multi-modal training data set.

[0238] Specifically, pair the standardized input sample set X* = {x1*, x2*, …, x k *, …, x n *} with the corresponding output sample set Y = {y1, y2, …, y k , …, y n} to form an input-output training sample pair set (X*, Y), thereby forming a multi-modal training data set.

[0239] Preferably, in one embodiment, as Figure 14 shown, the key control parameters in the process of carbon dioxide sequestration in shale reservoirs are screened out by the Sobol global sensitivity analysis method, including steps 1401 to 1408.

[0240] Step 1401: Determine the value range of input parameters and establish the correlation constraints between input parameters. Among them, the input parameters include: reservoir parameters, fluid parameters, well parameters, injection engineering parameters, etc.

[0241] Exemplarily, the value range of porosity can be set between 0.03 and 0.15, and the value range of matrix permeability can be set between 10 -18 and 10 -15 m 2 , and the correlation coefficient between porosity and matrix permeability is specified as 0.8, indicating a highly positive correlation.

[0242] Step 1402: Perform Saltelli extended sampling on each input parameter to generate an independent Sobol sequence sample matrix for each input parameter.

[0243] Exemplarily, perform Saltelli extended sampling on porosity and permeability respectively to obtain the corresponding sample matrices A and B1.

[0244] Step 1403: Perform Copula transformation on the generated independent Sobol sequence sample matrix to construct a joint sample matrix that satisfies the correlation constraints.

[0245] Exemplarily, perform Copula transformation on sample matrices A and B1 to obtain a joint sample matrix C with a correlation coefficient of 0.8.

[0246] Step 1404: Perform inverse transformation on the joint sample matrix to obtain multiple Sobol input samples.

[0247] Specifically, map the joint sample matrix after Copula transformation from the standard normal space back to the original physical quantity space to obtain Sobol input samples that satisfy the set correlation coefficient.

[0248] Step 1405: Input the Sobol input samples into a multi-field coupling numerical model for simulation to obtain the corresponding output samples, and construct a set of input-output sample pairs.

[0249] Step 1406: Perform Bootstrap resampling on the set of input-output sample pairs, and calculate the Sobol sensitivity index after each round of sampling. Among them, the Sobol sensitivity index includes: the first-order effect index S i 、total effect index S Tiand the second-order effect index S ij etc., the present invention is not limited thereto.

[0250] Specifically, the input-output sample pair set is resampled with replacement by the Bootstrap method. After each round of resampling, the first-order effect index S i , the total effect index S Ti and the second-order effect index S ij are calculated respectively. Repeat the above process until the preset number of sampling times (such as 1000 times) is reached to construct the sample set of each sensitivity index.

[0251] Exemplarily, calculate the first-order effect index S i of porosity and permeability, and the first-order effect index of porosity is 0.32, and the first-order effect index of permeability is 0.06.

[0252] Step 1407: Perform statistical analysis on the sample set of each sensitivity index to obtain the confidence interval of the sensitivity index.

[0253] Specifically, sort each group of sensitivity index sample sets from smallest to largest in value, and extract the 2.5th percentile and the 97.5th percentile respectively to obtain the corresponding 95% confidence interval.

[0254] For example: the first-order effect index of porosity is 0.32, and its confidence interval CI is [0.28, 0.35]. The first-order effect index of permeability is 0.06, and its confidence interval CI of permeability is [0.02, 0.09].

[0255] Step 1408: Screen the sensitivity index according to the preset threshold to obtain the key control parameter set.

[0256] Exemplarily, the screening criterion can be set as: if the first-order effect index S i of an input parameter > 0.05 or the second-order effect index S ij with other parameters > 0.01, then the input parameter belongs to the key control parameter.

[0257] For example, the first-order effect index of porosity is 0.32 > 0.05, and the first-order effect index of permeability is 0.06 > 0.05, then both can be determined as key control parameters.

[0258] Step 402: Input the high-dimensional data in the multi-modal training dataset into the carbon dioxide sequestration deep learning model for encoding processing to obtain a low-dimensional feature space. Among them, the carbon dioxide sequestration deep learning model is a neural network constructed based on the U-Net structure, including an encoder (Encoder), a decoder (Decoder), and a Transformer layer based on the attention mechanism.

[0259] Specifically, the high-dimensional data φ at a certain moment is input into the U-Net neural network model. Through the encoder of the U-Net neural network model, multi-layer convolution and pooling operations are performed to map the input high-dimensional data φ t to a low-dimensional high-order feature space, obtaining low-dimensional compressed features. Among them, the high-dimensional data φ t belongs to the multi-modal training data set. t belongs to the multi-modal training data set.

[0260] Step 403: Perform Transformer coupling in the low-dimensional feature space to obtain a coupled relationship feature vector.

[0261] Specifically, in the low-dimensional feature space, through the Transformer layer in the U-Net neural network, self-attention mechanism fusion operations between input data are performed to obtain a coupled relationship feature vector Attention(q, k, v).

[0262] In one embodiment, as Figure 8 shown, Step 403 includes Step 801 to Step 802.

[0263] Step 801: Perform a linear transformation operation on the low-dimensional feature space to generate a query vector q, a key vector k, and a value vector v respectively.

[0264] Specifically, through the fully connected layer in the U-Net neural network, a linear transformation operation is performed on the low-dimensional compressed features output by the encoder to obtain a query vector q, a key vector k, and a value vector v.

[0265] Step 802: Determine the fused coupled relationship feature vector Attention(q, k, v) according to the query vector q, the key vector k, the value vector v, and the Softmax function.

[0266] Specifically, according to the formula of the attention mechanism, calculate the coupled relationship feature vector Attention(q, k, v).

[0267]

[0268] where q is the query vector, k is the key vector, v is the value vector, and d k is the vector dimension constant.

[0269] First, by calculating the similarity between the query vector q and the key vector k, obtain the attention weight coefficient qk T . Secondly, after normalization through the Softmax function and the vector dimension constant d k , obtain the attention weights between each feature point Finally, the value vectors v are weighted and summed based on the attention weights to obtain the fused coupled relationship feature vector Attention(q, k, v).

[0270] Step 404: Decode the coupled relationship feature vector Attention(q, k, v) to obtain high-dimensional prediction data.

[0271] Specifically, the decoder in the U-Net neural network performs multi-layer upsampling and convolution operations on the coupled relationship feature vector Attention(q, k, v), gradually restoring the spatial resolution, and finally reconstructing the multi-modal high-dimensional data φ at time t + δt. t+δt It can realize the time-series prediction of the dynamic carbon dioxide sequestration in shale reservoirs.

[0272] Step 103: Embed the multi-scale seepage control model into a pre-trained deep learning model for carbon dioxide sequestration to obtain a prediction model for carbon dioxide sequestration in shale reservoirs.

[0273] In one embodiment, as Figure 9 and Figure 11 shown, Step 103 includes Step 901 to Step 905.

[0274] Step 901: Discretize the multi-scale seepage control model to obtain a multi-scale seepage control model in differential form.

[0275] Specifically, the multi-scale seepage control model includes seepage control equations for artificial fracture systems, natural fracture systems, inorganic matrix systems, kerogen systems, and organic matter systems. The above seepage control equations are discretized using the finite difference method or the finite volume method to obtain differential expressions.

[0276] Step 902: Input the high-dimensional prediction data into the multi-scale seepage control model in differential form to obtain a physical residual loss.

[0277] Specifically, the high-dimensional prediction data φ t+δt output by the U-Net neural network model at time t + δt is substituted into the differential expressions of the multi-scale seepage control equations, and the discrete equation residuals of each physical system are calculated item by item to obtain physical residual values.

[0278] Furthermore, based on the discrete equation residuals of each physical system, the physical residual loss L of the U-Net neural network model is calculated according to the following formula (11) f .

[0279]

[0280] where Residual irepresents the residual value corresponding to the i-th predicted sample in the difference expression, and N represents the total number of predicted samples.

[0281] Step 903: Determine the data fitting loss L, boundary condition residual L s , and initial condition residual L b respectively according to the high-dimensional prediction data. i .

[0282] Specifically, based on the high-dimensional prediction result φ t+δt output by the U-Net neural network model at time t+δt, construct the data fitting loss L s , boundary condition residual L b , and initial condition residual L i and other loss functions.

[0283] Among them, the data fitting loss L s is used to measure the error between the predicted output of the U-Net neural network model (i.e., the high-dimensional prediction result φ t+δt ) and the high-quality numerical simulation result (i.e., the real observed data), and can be calculated by the mean square error (MSE) or the mean absolute error (MAE).

[0284] The boundary condition residual L b is used to evaluate whether the prediction result (i.e., the high-dimensional prediction result φ t+δt ) satisfies the consistency constraint of the preset physical boundary conditions (such as fixed pressure boundary, zero flow boundary or injection boundary) on the spatial boundary. It is obtained by substituting the high-dimensional prediction result φ t+δt into the control equation at the boundary position and calculating its residual.

[0285] The initial condition residual loss L i is used to evaluate whether the prediction result of the U-Net neural network model (i.e., the high-dimensional prediction result φ t+δt ) is consistent with the initial state condition, ensuring that the starting state of the predicted carbon dioxide sequestration process is consistent with the initial configuration of the physical model. It is obtained by comparing the error between the prediction result at t = 0 and the initial field configuration.

[0286] Step 904: Determine the total loss function L according to the physical residual loss L f , data fitting loss L s , boundary condition residual L b , initial condition residual L i and their corresponding weight factors.

[0287] Specifically, calculate the total loss function L according to formula (12).

[0288]

[0289] Among them, L s is the data fitting loss, L f is the physical residual loss, L b is the boundary condition residual, L i is the initial condition residual, λ s 、λ f 、λ b and λ i are weight factors.

[0290] Step 905: Update the model parameters through the Adam optimizer and the total loss function, and repeat the above steps for iterative training until the total loss function converges to obtain a prediction model for carbon dioxide sequestration in shale reservoirs.

[0291] Specifically, the Adam optimizer is used to update the parameters of the U-Net neural network model to minimize the above total loss function L, and steps 902 to 905 are iteratively executed until the loss function converges or reaches the set number of training epochs to complete the training of the prediction model for carbon dioxide sequestration in shale reservoirs.

[0292] In the embodiment of the present invention, by discretizing the multi-scale seepage control equation by difference and embedding it into the U-Net neural network structure, a prediction model for carbon dioxide sequestration in shale reservoirs integrating physical constraints is constructed, significantly reducing the computational cost of simulating the sequestration process and improving the overall computational efficiency. Further, by introducing physical residual terms, boundary condition consistency terms and initial condition constraint terms, a hybrid loss function is constructed, which improves the physical interpretability and convergence stability of the model results while ensuring the prediction accuracy. In addition, combined with a multi-field coupling model constructed based on the COMSOL multi-physics simulation platform, which covers the coupling relationships of the seepage field, temperature field, mechanical field and chemical field, it can accurately depict the migration mechanism of carbon dioxide in the multi-scale system of shale reservoirs (including artificial fracture systems, natural fracture systems, inorganic matrix systems, kerogen systems and nano-organic matter systems), and comprehensively reflect the physical and chemical behaviors such as competitive adsorption, slip flow, dissolution diffusion and mineralization reaction between supercritical carbon dioxide and native gas in actual engineering. Further, by generating a representative sample set through Latin hypercube sampling and screening key influencing parameters based on the SOBOL global sensitivity analysis method, the diversity and representativeness of the training data set are improved, and the generalization ability and stability of the prediction model are enhanced. The prediction model for carbon dioxide sequestration in shale reservoirs provided by the present invention can be adapted to numerical simulation data and on-site actual monitoring data, has good engineering applicability, and is suitable for dynamic prediction and safety assessment of carbon dioxide sequestration under complex geological conditions.

[0293] Step 104: Use the carbon dioxide sequestration prediction model for shale reservoirs to perform sequestration prediction on the reservoir parameters, well parameters, and fluid parameters, and obtain the prediction results.

[0294] Specifically, based on the trained carbon dioxide sequestration prediction model for shale reservoirs, input the reservoir parameters, well parameters, and fluid parameters of the target shale reservoir, and perform rapid prediction and quantitative evaluation on the dynamic changes during the carbon dioxide injection process.

[0295] As Figure 15 shown is the pressure difference prediction effect diagram of the carbon dioxide sequestration prediction model for shale reservoirs under the numerical simulation data set. Among them, the horizontal axis is the prediction result Δp of the carbon dioxide sequestration prediction model for shale reservoirs 预测 (unit: MPa), and the vertical axis is the simulation result Δp output by the COMSOL multi-physics simulation platform 模拟 (unit: MPa).

[0296] Use the Pearson correlation coefficient ρ and the goodness of fit R 2 to evaluate the prediction accuracy of the carbon dioxide sequestration prediction model for shale reservoirs. As Figure 15 shown, it is calculated that ρ = 0.993, R 2 = 0.986, indicating that the carbon dioxide sequestration prediction model for shale reservoirs has extremely high prediction accuracy and consistency on the multi-modal training data set.

[0297] As Figure 16 shown is the pressure difference prediction result of the carbon dioxide sequestration prediction model for shale reservoirs after inputting the actual production data of the shale reservoir. Among them, the horizontal axis is the prediction result Δp of the carbon dioxide sequestration prediction model for shale reservoirs 预测 (unit: MPa), and the vertical axis is the actual result Δp measured on site 实测 (unit: MPa).

[0298] Use the Pearson correlation coefficient ρ and the goodness of fit R 2 to evaluate the prediction accuracy of the carbon dioxide sequestration prediction model for shale reservoirs. As Figure 16 shown, it is calculated that ρ = 0.990, R 2 = 0.980, further verifying that the carbon dioxide sequestration prediction model for shale reservoirs also has high generalization ability in dealing with actual engineering data.

[0299] As Figure 17 shown are the carbon dioxide injection dynamic prediction curves of two typical samples in the test data set of the carbon dioxide sequestration prediction model for shale reservoirs. The blue curve represents the simulated data, and the red curve represents the prediction result of the carbon dioxide sequestration prediction model for shale reservoirs. It can be seen that the model prediction curve is highly consistent with the simulated curve at different time scales.

[0300] As Figures 18a to 18c shown, the dynamic prediction accuracy of the carbon dioxide sequestration prediction model for shale reservoirs is evaluated at different levels of observation noise (such as 0%, 10%, 20%). Figure 18a For the prediction result of the carbon dioxide sequestration prediction model for shale reservoirs without noise interference, the pressure change curve constructed by the carbon dioxide sequestration prediction model for shale reservoirs almost completely coincides with the real data. Figure 18b and Figure 18c are the prediction results of the carbon dioxide sequestration prediction model for shale reservoirs under 10% and 20% noise perturbation conditions respectively. Although there are slight fluctuations in the prediction curve, the overall trend still shows good consistency with the real curve, indicating that the carbon dioxide sequestration prediction model for shale reservoirs has good robustness and parameter inversion ability for observation errors.

[0301] Figure 19 are the dynamic prediction effect diagrams under different sequestration mechanisms during the whole process of carbon dioxide sequestration in shale reservoirs. As the sequestration years progress, carbon dioxide gradually migrates from the supercritical state to the dissolved state, adsorbed state, ionic state and mineralized state. After 300 years of prediction, the sequestration amount in the supercritical state reaches 5.12×10 5 tons, the sequestration amount in the adsorbed state is 3.08×10 5 tons, and the dissolved CO2 is 9.58×10 4 tons. The total sequestration amount based on the physical sequestration mechanism is 9.158×10 5 tons, and the total sequestration amount of the mineralization and ionic state conversion part is 1.082×10 5 tons. Figure 19 Demonstrates the quantitative ability of the carbon dioxide sequestration prediction model for shale reservoirs in predicting long-term transformation paths.

[0302] In the embodiments of the present invention, the carbon dioxide sequestration prediction model for shale reservoirs can not only accurately predict the injection dynamics and pressure changes, but also has good adaptability and robustness, and can effectively reflect the multi-field coupling mechanism and the time evolution process of the sequestration path in shale reservoirs, and can be widely applied to engineering scenarios such as carbon dioxide sequestration assessment and optimized injection design in shale reservoirs.

[0303] In the embodiments of the present invention, a carbon dioxide sequestration prediction device driven by the fusion of mechanism and data is also provided, as described in the following embodiments. Since the principle of solving problems by this device is similar to that of the carbon dioxide sequestration prediction method driven by the fusion of mechanism and data, the implementation of this device can refer to the implementation of the carbon dioxide sequestration prediction method driven by the fusion of mechanism and data, and the repeated parts will not be elaborated.

[0304] As Figure 20As shown in the figure, the carbon dioxide sequestration prediction device 2000 driven by the fusion of mechanism and data includes: a parameter acquisition module 2001, a model construction module 2002, a model fusion module 2003, and a prediction output module 2004.

[0305] The parameter acquisition module is used to acquire reservoir parameters, well parameters, and fluid parameters of the shale reservoir;

[0306] The model construction module is used to establish a multi-scale seepage control model according to the reservoir parameters, well parameters, and fluid parameters;

[0307] The model fusion module is used to embed the multi-scale seepage control model into a pre-trained deep learning model for carbon dioxide sequestration to obtain a carbon dioxide sequestration prediction model for the shale reservoir;

[0308] The prediction output module is used to perform sequestration prediction on the reservoir parameters, well parameters, and fluid parameters through the carbon dioxide sequestration prediction model for the shale reservoir to obtain a prediction result.

[0309] In one embodiment, the model construction module includes: a horizontal well injection model construction unit and a seepage mechanism modeling unit.

[0310] The horizontal well injection model construction unit is used to establish a horizontal well injection model according to the reservoir parameters, well parameters, and fluid parameters;

[0311] The seepage mechanism modeling unit is used to establish a multi-scale seepage control model corresponding to the horizontal well injection model according to the reservoir parameters and fluid parameters.

[0312] In one embodiment, the horizontal well injection model construction unit includes: a parameter determination unit, a system identification unit, a boundary initial value setting unit, and a model construction execution unit.

[0313] The parameter determination unit is used to determine injection engineering parameters according to the reservoir parameters, well parameters, and fluid parameters;

[0314] The system identification unit is used to determine the system composition of the horizontal well injection model and its corresponding fluid seepage mechanism according to the reservoir parameters and fluid parameters;

[0315] The boundary initial value setting unit is used to determine the initial conditions and boundary conditions for carbon dioxide sequestration according to the injection engineering parameters and the system composition of the horizontal well injection model;

[0316] The model construction execution unit is used to establish the horizontal well injection model according to the system composition, seepage mechanism, injection parameters, initial conditions, and boundary conditions.

[0317] In one embodiment, the training module of the carbon dioxide sequestration deep learning model includes a dataset construction module, an encoding module, a feature coupling module, and a decoding module.

[0318] The dataset construction module is used to establish a multi-modal training dataset based on the reservoir parameters, the fluid parameters, and the injection engineering parameters;

[0319] The encoding module is used to input the high-dimensional data in the multi-modal training dataset into the carbon dioxide sequestration deep learning model for encoding processing to obtain a low-dimensional feature space;

[0320] The feature coupling module is used to perform Transformer coupling in the low-dimensional feature space to obtain a coupled relationship feature vector;

[0321] The decoding module is used to perform decoding processing on the coupled relationship feature vector to obtain high-dimensional prediction data.

[0322] In one embodiment, the feature coupling module includes: a vector generation unit and a coupling fusion unit.

[0323] The vector generation unit is used to perform a linear transformation operation on the low-dimensional feature space to respectively generate a query vector, a key vector, and a value vector;

[0324] The coupling fusion unit is used to determine the fused coupled relationship feature vector according to the query vector, the key vector, the value vector, and the softmax function.

[0325] In one embodiment, the model fusion module includes: a physical model discretization unit, a physical loss acquisition unit, a boundary loss calculation unit, a total loss function construction unit, and an optimization training unit.

[0326] The physical model discretization unit is used to perform discretization processing on the multi-scale seepage control model to obtain a multi-scale seepage control model in differential form;

[0327] The physical loss acquisition unit is used to input the high-dimensional prediction data into the multi-scale seepage control model in differential form to obtain a physical residual loss;

[0328] The boundary loss calculation unit is used to respectively determine a data fitting loss, a boundary condition residual, and an initial condition residual according to the high-dimensional prediction data;

[0329] The total loss function construction unit is used to determine a total loss function according to the physical residual loss, the data fitting loss, the boundary condition residual, the initial condition residual, and their corresponding weight factors;

[0330] The optimization training unit is used to update the model parameters through the Adam optimizer and the total loss function, and repeat the above steps for iterative training until the total loss function converges, obtaining the shale reservoir carbon dioxide sequestration prediction model.

[0331] In one embodiment, the dataset construction module includes: a simulation model construction unit and a sample data generation unit.

[0332] The simulation model construction unit is used to construct a multi-field coupling numerical model through a multi-physics simulation platform;

[0333] The sample data generation unit is used to construct a multi-modal training dataset according to the reservoir parameters, the fluid parameters, the injection engineering parameters, and the multi-field coupling numerical model.

[0334] In one embodiment, the sample data generation unit includes: a sampling unit, a simulation unit, a sample normalization unit, and a training sample construction unit.

[0335] The sampling unit is used to perform Latin hypercube sampling on the reservoir parameters, the fluid parameters, and the injection engineering parameters to generate multiple groups of input samples;

[0336] The simulation unit is used to input each group of the input samples into the multi-field coupling numerical model for simulation to obtain corresponding output samples;

[0337] The sample normalization unit is used to perform normalization processing on each group of the input samples;

[0338] The training sample construction unit is used to construct input-output training sample pairs according to the input samples after normalization processing and their corresponding output samples, obtaining the multi-modal training dataset.

[0339] In one embodiment, the sampling module includes: a parameter determination unit, an interval generation unit, a random sampling unit, and a random combination unit.

[0340] The parameter determination unit is used to determine the value ranges of the reservoir parameters, the fluid parameters, and the injection engineering parameters;

[0341] The interval generation unit is used to equally and probabilistically divide the value range corresponding to each input parameter according to a preset number of samples to generate multiple sample intervals;

[0342] The random sampling unit is used to perform random sampling within each of the sample intervals to obtain the sample values corresponding to each of the sample intervals;

[0343] The random combination unit is used to randomly combine the sample values sampled for each input parameter to obtain multiple groups of input samples.

[0344] In one embodiment, the model construction module further includes an artificial fracture modeling unit.

[0345] The artificial fracture modeling unit is used to determine a multi-scale seepage control model corresponding to the artificial fracture system according to the first pore structure parameters and the first fluid parameters;

[0346] Among them, the first pore structure parameters include: the permeability of the artificial fracture system, the artificial fracture width, and the reservoir thickness; the first fluid parameters include: the gas viscosity of the artificial fracture system and the gas volume factor.

[0347] In one embodiment, the model construction module further includes a natural fracture modeling unit.

[0348] The natural fracture modeling unit is used to determine a multi-scale seepage control model corresponding to the natural fracture system according to the second pore structure parameters, the third pore structure parameters, the second fluid parameters, and the third fluid parameters;

[0349] Among them, the second pore structure parameters include: the radius of the natural fracture system, the permeability of the natural fracture system, the natural fracture stress sensitivity modulus, the reservoir pressure of the natural fracture system, the interface radius between the fracture system and the inorganic matrix system, the porosity of the natural fracture system, and the reservoir temperature of the natural fracture system; the third pore structure parameters include: the shape factor of the inorganic matrix system, the permeability of the inorganic matrix system, the reservoir pressure of the inorganic matrix system, and the radius of the inorganic matrix system; the second fluid parameters include: the gas density in the natural fracture system, the gas viscosity of the natural fracture system, the gas compressibility factor, the compressibility factor, and the ideal gas constant; the third fluid parameters include: the gas density of the inorganic matrix system.

[0350] In one embodiment, the model construction module further includes an inorganic matrix modeling unit.

[0351] The inorganic matrix modeling unit is used to determine a multi-scale seepage control model corresponding to the inorganic matrix system according to the third pore structure parameters, the fourth pore structure parameters, the third fluid parameters, the fourth fluid parameters, and the preset first adsorption model parameters;

[0352] Among them, the third pore structure parameter further includes: clay mineral content, the interfacial radius between the kerogen system and the inorganic matrix system, and the porosity of the inorganic matrix system; the fourth pore structure parameter includes: kerogen content, the shape factor of the kerogen system, the permeability of the kerogen system, the reservoir pressure of the kerogen system, and the radius of the kerogen system; the third fluid parameter further includes: the gas slip coefficient in the inorganic matrix system and the gas compressibility of the inorganic matrix system; the fourth fluid parameter includes: the gas density in the kerogen system; the first adsorption model parameter includes: the isothermal adsorption volume of the inorganic matrix system and the isothermal adsorption pressure of the inorganic matrix system.

[0353] In one embodiment, the model construction module further includes a porous kerogen modeling unit.

[0354] The porous kerogen modeling unit is used to determine the multi-scale seepage control model corresponding to the porous kerogen system according to the fourth pore structure parameter, the fifth pore structure parameter, the fourth fluid parameter, the fifth fluid parameter, and the preset second adsorption model parameter.

[0355] Among them, the fourth pore structure parameter further includes: the porosity of the kerogen system; the fifth pore structure parameter includes: the shape factor of the organic matter system and the particle radius of the organic matter system; the fourth fluid parameter further includes: the gas slip coefficient and the gas compressibility of the kerogen system; the fifth fluid parameter includes: the gas density, the effective gas diffusion coefficient, and the gas concentration in the organic matter system in the organic matter system; the second adsorption model parameter includes: the isothermal adsorption volume of the kerogen system and the isothermal adsorption pressure of the kerogen system.

[0356] In one embodiment, the model construction module further includes an organic matter system modeling unit.

[0357] The organic matter system modeling unit is used to determine the multi-scale seepage control model corresponding to the organic matter system according to the fifth pore structure parameter and the fifth fluid parameter.

[0358] Figure 21 The schematic physical structure diagram of the electronic device provided by the embodiment of the present invention is as Figure 5 shown, the electronic device includes: a processor 2101, a memory 2102, and a bus 2103.

[0359] Among them, the processor 2101 and the memory 2102 complete communication with each other through the bus 2103.

[0360] The processor 2101 is used to call the program instructions in the memory 2102 to execute the methods provided by the above method embodiments.

[0361] An embodiment of the present invention further provides a computer-readable storage medium storing a computer program, which when executed by a processor implements the above-mentioned carbon dioxide sequestration prediction method driven by mechanism and data fusion.

[0362] An embodiment of the present invention further provides a computer program product including a computer program, which when executed by a processor implements the above-mentioned carbon dioxide sequestration prediction method driven by mechanism and data fusion.

[0363] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk memory, CD-ROM, optical memory, etc.) containing computer-usable program code.

[0364] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, and the combination of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in Figure 1 one or more flows and / or blocks Figure 1 one or more blocks.

[0365] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including instruction means that implement the functions specified in Figure 1 one or more flows and / or blocks Figure 1 one or more blocks.

[0366] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process, so that the instructions executed on the computer or other programmable device provide means for implementing the functions specified in Figure 1 one or more flows and / or blocksFigure 1 Steps of functions specified in one or more boxes.

[0367] In the specific embodiments described above, the purpose, technical solutions and beneficial effects of the present invention have been further described in detail. It should be understood that the above are only specific embodiments of the present invention and are not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A carbon dioxide sequestration prediction method driven by the fusion of mechanism and data, characterized in that, Including: Obtaining reservoir parameters, well parameters, and fluid parameters of a shale reservoir; Establishing a multi-scale seepage control model based on the reservoir parameters, the well parameters, and the fluid parameters; Embedding the multi-scale seepage control model into a pre-trained carbon dioxide sequestration deep learning model to obtain a shale reservoir carbon dioxide sequestration prediction model; Performing sequestration prediction on the reservoir parameters, the well parameters, and the fluid parameters through the shale reservoir carbon dioxide sequestration prediction model to obtain a prediction result.

2. The method according to claim 1, characterized in that, The establishing of the multi-scale seepage control model according to the reservoir parameters, the well parameters, and the fluid parameters includes: Establishing a horizontal well injection model based on the reservoir parameters, the well parameters, and the fluid parameters; Establishing a multi-scale seepage control model corresponding to the horizontal well injection model according to the reservoir parameters and the fluid parameters.

3. The method according to claim 2, wherein The establishing of the horizontal well injection model according to the reservoir parameters, the well parameters, and the fluid parameters includes: Determining injection engineering parameters according to the reservoir parameters, the well parameters, and the fluid parameters; Determining the system composition of the horizontal well injection model and its corresponding fluid seepage mechanism according to the reservoir parameters and the fluid parameters; Determining the initial conditions and boundary conditions of carbon dioxide sequestration according to the injection engineering parameters and the system composition of the horizontal well injection model; Establishing the horizontal well injection model according to the system composition of the horizontal well injection model and its corresponding fluid seepage mechanism, the injection engineering parameters, the initial conditions, and the boundary conditions.

4. The method according to claim 3, characterized in that The training steps of the carbon dioxide sequestration deep learning model include: Establishing a multi-modal training data set based on the reservoir parameters, the fluid parameters, and the injection engineering parameters; Inputting the high-dimensional data in the multi-modal training data set into the carbon dioxide sequestration deep learning model for encoding processing to obtain a low-dimensional feature space; Performing Transformer coupling in the low-dimensional feature space to obtain a coupled relationship feature vector; Performing decoding processing on the coupled relationship feature vector to obtain high-dimensional prediction data.

5. The method according to claim 4, characterized in that The performing of Transformer coupling in the low-dimensional feature space to obtain a coupled relationship feature vector includes: Performing a linear transformation operation on the low-dimensional feature space to respectively generate a query vector, a key vector, and a value vector; Determining the fused coupled relationship feature vector according to the query vector, the key vector, the value vector, and the softmax function.

6. The method according to claim 5, characterized in that The embedding of the multi-scale seepage control model into a pre-trained carbon dioxide sequestration deep learning model to obtain a shale reservoir carbon dioxide sequestration prediction model includes: Performing discretization processing on the multi-scale seepage control model to obtain a multi-scale seepage control model in differential form; Inputting the high-dimensional prediction data into the multi-scale seepage control model in differential form to obtain a physical residual loss; Respectively determining a data fitting loss, a boundary condition residual, and an initial condition residual according to the high-dimensional prediction data; Determine the total loss function according to the physical residual loss, the data fitting loss, the boundary condition residual, the initial condition residual, and their corresponding weight factors; Update the model parameters through the Adam optimizer and the total loss function, and repeat the above steps for iterative training until the total loss function converges to obtain the shale reservoir carbon dioxide sequestration prediction model.

7. The method according to claim 4, characterized in that, Based on the reservoir parameters, the fluid parameters, and the injection engineering parameters, establish a multi-modal training dataset, including: Construct a multi-field coupling numerical model through a multi-physics simulation platform; Construct a multi-modal training dataset according to the reservoir parameters, the fluid parameters, the injection engineering parameters, and the multi-field coupling numerical model.

8. The method according to claim 7, wherein The establishment of a multi-modal training dataset according to the reservoir parameters, the fluid parameters, the injection engineering parameters, and the multi-field coupling numerical model includes: Perform Latin hypercube sampling on the reservoir parameters, the fluid parameters, and the injection engineering parameters to generate multiple groups of input samples; Input each group of the input samples into the multi-field coupling numerical model for simulation to obtain the corresponding output samples; Perform normalization processing on each group of the input samples; Construct input-output training sample pairs according to the normalized input samples and their corresponding output samples to obtain the multi-modal training dataset.

9. The method according to claim 8, characterized in that The performing Latin hypercube sampling on the reservoir parameters, the fluid parameters, and the injection engineering parameters to generate multiple groups of input samples includes: Determine the value ranges of the reservoir parameters, the fluid parameters, and the injection engineering parameters; Equally and probabilistically divide the value range corresponding to each input parameter according to a preset number of samples to generate multiple sample intervals; Perform random sampling within each of the sample intervals to obtain the sample values corresponding to each of the sample intervals; Randomly combine the sample values sampled for each input parameter to obtain multiple groups of input samples.

10. The method according to claim 2, characterized in that The reservoir parameters include: the first pore structure parameter; the fluid parameters include: the first fluid parameter; the horizontal well injection model includes an artificial fracture system; the establishment of the multi-scale seepage control model corresponding to the horizontal well injection model according to the reservoir parameters and the fluid parameters includes: Determine the multi-scale seepage control model corresponding to the artificial fracture system according to the first pore structure parameter and the first fluid parameter; wherein, the first pore structure parameter includes: the permeability of the artificial fracture system, the artificial fracture width, and the reservoir thickness; the first fluid parameter includes: the gas viscosity and the gas volume coefficient of the artificial fracture system.

11. The method according to claim 2, characterized in that, The reservoir parameters further include: the second pore structure parameter and the third pore structure parameter; the fluid parameters further include: the second fluid parameter and the third fluid parameter; the horizontal well injection model further includes a natural fracture system; the establishment of the multi-scale seepage control model corresponding to the horizontal well injection model according to the reservoir parameters and the fluid parameters includes: Determine the multi-scale seepage control model corresponding to the natural fracture system according to the second pore structure parameter, the third pore structure parameter, the second fluid parameter, and the third fluid parameter; Among them, the second pore structure parameters include: the radius of the natural fracture system, the permeability of the natural fracture system, the stress sensitivity modulus of the natural fracture, the reservoir pressure of the natural fracture system, the interface radius between the fracture system and the inorganic matrix system, the porosity of the natural fracture system, and the reservoir temperature of the natural fracture system; the third pore structure parameters include: the shape factor of the inorganic matrix system, the permeability of the inorganic matrix system, the reservoir pressure of the inorganic matrix system, and the radius of the inorganic matrix system; the second fluid parameters include: the gas density in the natural fracture system, the gas viscosity of the natural fracture system, the gas compressibility factor, the compressibility factor, and the ideal gas constant; the third fluid parameters include: the gas density of the inorganic matrix system.

12. The method according to claim 11, wherein, The reservoir parameters further include: the fourth pore structure parameters; the fluid parameters further include: the fourth fluid parameters; the horizontal well injection model further includes the inorganic matrix system; establishing the multi-scale seepage control model corresponding to the horizontal well injection model according to the reservoir parameters and the fluid parameters includes: Determining the multi-scale seepage control model corresponding to the inorganic matrix system according to the third pore structure parameters, the fourth pore structure parameters, the third fluid parameters, the fourth fluid parameters, and the preset first adsorption model parameters; Among them, the third pore structure parameters further include: the clay mineral content, the interface radius between the kerogen system and the inorganic matrix system, and the porosity of the inorganic matrix system; the fourth pore structure parameters include: the kerogen content, the shape factor of the kerogen system, the permeability of the kerogen system, the reservoir pressure of the kerogen system, and the radius of the kerogen system; the third fluid parameters further include: the gas slip coefficient in the inorganic matrix system and the gas compressibility factor of the inorganic matrix system; the fourth fluid parameters include: the gas density in the kerogen system; the first adsorption model parameters include: the isothermal adsorption volume of the inorganic matrix system and the isothermal adsorption pressure of the inorganic matrix system.

13. The method according to claim 12, wherein The reservoir parameters further include: the fifth pore structure parameters; the fluid parameters further include: the fifth fluid parameters; the horizontal well injection model further includes the porous kerogen system; establishing the multi-scale seepage control model corresponding to the horizontal well injection model according to the reservoir parameters and the fluid parameters includes: Determining the multi-scale seepage control model corresponding to the porous kerogen system according to the fourth pore structure parameters, the fifth pore structure parameters, the fourth fluid parameters, the fifth fluid parameters, and the preset second adsorption model parameters; Among them, the fourth pore structure parameters further include: the porosity of the kerogen system; the fifth pore structure parameters include: the shape factor of the organic matter system and the particle radius of the organic matter system; the fourth fluid parameters further include: the gas slip coefficient of the kerogen system and the gas compressibility factor of the kerogen system; the fifth fluid parameters include: the gas density in the organic matter system, the effective gas diffusion coefficient, and the gas concentration in the organic matter system; the second adsorption model parameters include: the isothermal adsorption volume of the kerogen system and the isothermal adsorption pressure of the kerogen system.

14. The method according to claim 13, wherein The horizontal well injection model further includes: the organic matter system, and establishing a multi-scale seepage control model corresponding to the horizontal well injection model according to the reservoir parameters and the fluid parameters, including: Determining a multi-scale seepage control model corresponding to the organic matter system according to the fifth pore structure parameter and the fifth fluid parameter.

15. A carbon dioxide sequestration prediction device driven by the fusion of mechanism and data, characterized in that, Including: A parameter acquisition module for acquiring reservoir parameters, well parameters, and fluid parameters of a shale reservoir; A model construction module for establishing a multi-scale seepage control model according to the reservoir parameters, the well parameters, and the fluid parameters; A model fusion module for embedding the multi-scale seepage control model into a pre-trained carbon dioxide sequestration deep learning model to obtain a carbon dioxide sequestration prediction model for a shale reservoir; A prediction output module for performing sequestration prediction on the reservoir parameters, the well parameters, and the fluid parameters through the carbon dioxide sequestration prediction model for a shale reservoir to obtain a prediction result.

16. A computer device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the computer program, the method according to any one of claims 1 to 14 is implemented.

17. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 14 is implemented.

18. A computer program product, characterized in that, The computer program product includes a computer program, and when the computer program is executed by a processor, the method according to any one of claims 1 to 14 is implemented.

Images