Prediction method for characteristic parameters of stratum based on carbon dioxide geological sequestration

The prediction model of stratigraphic characteristic parameters is constructed through the autoregressive Fourier neural operator model, which solves the problem of low prediction efficiency of stratigraphic characteristic parameters in carbon dioxide geological storage, and achieves fast and accurate prediction, reducing calculation costs.

CN119067255APending Publication Date: 2024-12-03CHINA UNIV OF PETROLEUM (BEIJING)
View PDF 0 Cites 5 Cited by

Patent Information

Application Number
CN202411120087.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-15
Publication Date
2024-12-03

AI Technical Summary

Technical Problem

The prediction efficiency of the formation characteristic parameters in the existing carbon dioxide geological storage process is low, especially in the numerical simulation of the long-term mineralization storage mechanism, which leads to slow prediction speed.

Method used

Using the prediction method based on the autoregressive Fourier neural operator model, the first prediction model for predicting formation pressure, carbon dioxide saturation and fluid density and the prediction of mineral volume fraction and formation by obtaining the geological parameters of the target formation are constructed. The second prediction model of permeability is performed, coupled training is performed to improve prediction efficiency.

Benefits of technology

The prediction time of formation characteristic parameters is shortened, the prediction speed and efficiency are improved, the prediction accuracy is ensured, and the calculation cost is reduced.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119067255A_ABST
    Figure CN119067255A_ABST
Patent Text Reader

Abstract

The invention discloses a method for predicting characteristic parameters of a stratum based on carbon dioxide geological sequestration, and belongs to the technical field of oil-gas exploration and development. The method comprises the following steps: acquiring a first geological parameter and a second geological parameter of a target stratum; based on a first target prediction model used for predicting first prediction stratum characteristic parameters, the first prediction stratum characteristic parameters of the target stratum are obtained according to the first geological parameters, and the first prediction stratum characteristic parameters comprise a stratum pressure prediction value, a carbon dioxide saturation prediction value and a stratum fluid density prediction value; based on a second target prediction model used for predicting second prediction stratum characteristic parameters, the second prediction stratum characteristic parameters of the target stratum are obtained according to the first prediction stratum characteristic parameters and the second geological parameters, and the second prediction stratum characteristic parameters comprise a mineral volume fraction prediction value and a stratum permeability prediction value. The prediction efficiency of the stratum characteristic parameters in the carbon dioxide geological sequestration process can be improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the technical field of oil and gas exploration and development, and in particular to a method for predicting characteristic parameters of a formation based on geological storage of carbon dioxide. Background Art

[0002] In the process of geological storage of carbon dioxide, due to the complexity of the physical and chemical mechanisms of this process, when the characteristic parameters of the formation in the geological storage process of carbon dioxide are predicted by traditional numerical simulation methods, the requirements for numerical simulation (especially for the numerical simulation of long-term mineralization storage mechanism) are high, resulting in a slow prediction speed for predicting the characteristic parameters of the formation through numerical simulation and a low prediction efficiency. Summary of the invention

[0003] The purpose of the embodiment of the present application is to provide a method for predicting characteristic parameters of a formation based on geological storage of carbon dioxide, so as to solve the problem of low efficiency in predicting characteristic parameters of a formation based on geological storage of carbon dioxide.

[0004] In order to achieve the above-mentioned purpose, the first aspect of the embodiment of the present application provides a method for predicting characteristic parameters of a formation based on geological storage of carbon dioxide, comprising:

[0005] Acquire a first geological parameter and a second geological parameter of a target formation, wherein the first geological parameter includes a target geophysical parameter of the target formation, and the second geological parameter includes a formation temperature and a target geochemical parameter of the target formation;

[0006] Based on a pre-constructed first target prediction model for predicting first predicted formation characteristic parameters, first predicted formation characteristic parameters of the target formation are obtained according to the first geological parameters, wherein the first predicted formation characteristic parameters include a predicted value of formation pressure, a predicted value of carbon dioxide saturation, and a predicted value of formation fluid density;

[0007] Based on a pre-constructed second target prediction model for predicting second predicted formation characteristic parameters, second predicted formation characteristic parameters of the target formation are obtained according to the first predicted formation characteristic parameters and the second geological parameters, wherein the second predicted formation characteristic parameters include a predicted value of mineral volume fraction and a predicted value of formation permeability.

[0008] In an embodiment of the present application, the construction of the first target prediction model and the construction of the second target prediction model include: obtaining a first initial prediction model and a second initial prediction model, wherein the first initial prediction model and the second initial prediction model are both autoregressive Fourier neural operator models; according to a pre-generated first prediction model training data set and a second prediction model training data set, coupling training is performed on the first initial prediction model and the second initial prediction model to construct the first target prediction model and the second target prediction model.

[0009] In an embodiment of the present application, the first initial prediction model and the second initial prediction model are coupled trained according to the pre-generated first prediction model training data set and the second prediction model training data set to construct a first target prediction model and a second target prediction model, including: training the first initial prediction model according to the first prediction model training data set to obtain the formation pressure output value, the carbon dioxide saturation output value and the formation fluid density output value output by the first initial prediction model; training the second initial prediction model according to the formation pressure output value, the carbon dioxide saturation output value, the formation fluid density output value and the second prediction model training data set to obtain the formation permeability output value output by the second initial prediction model; updating the first prediction model training data set according to the formation permeability output value, and re-updating the first initial prediction model according to the updated first prediction model training data set until the first initial prediction model and the second initial prediction model both meet the preset training termination conditions to obtain the first target prediction model and the second target prediction model.

[0010] In an embodiment of the present application, the prediction method also includes: respectively determining the total loss function, hyperparameters and model evaluation strategy corresponding to the first initial prediction model and the second initial prediction model; based on the total loss function, hyperparameters and model evaluation strategy corresponding to the first initial prediction model and the second initial prediction model, according to the first prediction model training data set and the second prediction model training data set, coupling training is performed on the first initial prediction model and the second initial prediction model.

[0011] In an embodiment of the present application, the prediction method also includes: verifying the first target prediction model based on a pre-generated first prediction model verification data set, and verifying the second target prediction model based on a pre-generated second prediction model verification data set.

[0012] In an embodiment of the present application, the prediction method also includes: obtaining a set of numerical simulation input parameters; based on a preset numerical simulation model, obtaining a set of numerical simulation output parameters according to the numerical simulation input parameter set, the numerical simulation output parameter set including a first prediction model data set and a second prediction model data set; dividing the first prediction model data set according to a first preset ratio to obtain a first prediction model training data set and a first prediction model verification data set; and dividing the second prediction model data set according to a second preset ratio to obtain a second prediction model training data set and a second prediction model verification data set.

[0013] In the embodiment of the present application, the target geophysical parameters include formation temperature, formation pressure, carbon dioxide saturation, permeability in all directions of the formation, porosity, rock compression coefficient, capillary force, relative permeability, injection well location and injection rate, and the target geochemical parameters include mineral volume fraction and pH value.

[0014] In an embodiment of the present application, the first target prediction model includes a formation pressure prediction model, a carbon dioxide saturation prediction model, and a fluid density prediction model, and the second target prediction model includes a mineral volume fraction prediction model and a formation permeability prediction model.

[0015] A second aspect of an embodiment of the present application provides a processor configured to execute the above-mentioned method for predicting characteristic parameters of a formation based on geological storage of carbon dioxide.

[0016] A third aspect of an embodiment of the present application provides a machine-readable storage medium having instructions stored thereon, the instructions being used to enable a machine to execute the above-mentioned method for predicting characteristic parameters of a formation based on geological storage of carbon dioxide.

[0017] The above technical solution obtains the first geological parameter and the second geological parameter of the target formation, and then obtains the first predicted formation characteristic parameter of the target formation according to the first geological parameter based on the pre-constructed first target prediction model for predicting the first predicted formation characteristic parameter, and further obtains the second predicted formation characteristic parameter of the target formation according to the first predicted formation characteristic parameter and the second geological parameter based on the pre-constructed second target prediction model for predicting the second predicted formation characteristic parameter. The present application directly predicts the first predicted formation characteristic parameter through the first target prediction model, and directly predicts the second predicted formation characteristic parameter through the second target prediction model and the first predicted formation characteristic parameter, thereby shortening the prediction time of the first predicted formation characteristic parameter and the second predicted formation characteristic parameter, and improving the prediction speed of the first predicted formation characteristic parameter and the second predicted formation characteristic parameter, thereby being able to improve the prediction efficiency of the characteristic parameters of the formation in the process of geological storage of carbon dioxide while ensuring the accuracy of the prediction.

[0018] Other features and advantages of the embodiments of the present application will be described in detail in the subsequent specific implementation section. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] The accompanying drawings are used to provide a further understanding of the embodiments of the present application and constitute a part of the specification. Together with the following specific implementations, they are used to explain the embodiments of the present application, but do not constitute a limitation on the embodiments of the present application. In the accompanying drawings:

[0020] Figure 1The flowchart of a method for predicting characteristic parameters of a formation based on geological storage of carbon dioxide according to an embodiment of the present application is schematically shown. DETAILED DESCRIPTION

[0021] In order to make the purpose, technical scheme and advantages of the embodiments of the present application clearer, the technical scheme in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. It should be understood that the specific implementation methods described herein are only used to illustrate and explain the embodiments of the present application, and are not used to limit the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of this application.

[0022] It should be noted that if the embodiments of the present application involve directional indications (such as up, down, left, right, front, back...), the directional indications are only used to explain the relative position relationship, movement status, etc. between the components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0023] In addition, if there are descriptions involving "first", "second", etc. in the embodiments of the present application, the descriptions of "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or suggesting their relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but they must be based on the ability of ordinary technicians in the field to implement them. When the combination of technical solutions is contradictory or cannot be implemented, it should be deemed that such combination of technical solutions does not exist and is not within the scope of protection required by this application.

[0024] Figure 1 The following schematically shows a flow chart of a method for predicting characteristic parameters of a formation based on geological storage of carbon dioxide according to an embodiment of the present application. Figure 1 As shown, the embodiment of the present application provides a method for predicting characteristic parameters of a formation based on geological storage of carbon dioxide. The method is described by taking the application of the method to a processor as an example. The method may include the following steps:

[0025] Step S101: acquiring first geological parameters and second geological parameters of a target formation, wherein the first geological parameters include target geophysical parameters of the target formation, and the second geological parameters include formation temperature and target geochemical parameters of the target formation.

[0026] Step S102: Based on a pre-constructed first target prediction model for predicting first predicted formation characteristic parameters, first predicted formation characteristic parameters of the target formation are obtained according to the first geological parameters, wherein the first predicted formation characteristic parameters include a predicted value of formation pressure, a predicted value of carbon dioxide saturation, and a predicted value of formation fluid density.

[0027] Step S103: Based on a pre-constructed second target prediction model for predicting second predicted formation characteristic parameters, second predicted formation characteristic parameters of the target formation are obtained according to the first predicted formation characteristic parameters and the second geological parameters, wherein the second predicted formation characteristic parameters include a predicted value of mineral volume fraction and a predicted value of formation permeability.

[0028] In the process of geological storage of carbon dioxide, its physical and chemical mechanism is complex, and the numerical simulation of geological storage requires a lot of computing cost. The deep learning method is an effective means to replace the traditional numerical simulation method because of its lower computing cost and more reliability. To this end, the present application can quickly and accurately simulate the multiphase flow of carbon dioxide in a short time in three-dimensional space during the underground storage process, the carbon dioxide dissolution process, and the carbon dioxide mineralization process in a long time through a deep learning method based on Fourier neural operators under the time and space scale, and predict the formation characteristic parameters of carbon dioxide when it is geologically sealed in the target formation.

[0029] To achieve the above-mentioned purpose, in an embodiment of the present application, the processor can obtain the first geological parameter and the second geological parameter of the target formation. Wherein, the target formation refers to the reservoir for storing carbon dioxide when simulating the geological storage process of carbon dioxide. The first geological parameter refers to the target geophysical parameter of the target formation, and the second geological parameter includes the formation temperature and the target geochemical parameter of the target formation. Since in the process of carbon dioxide storage, the phase state of the fluid includes the liquid phase and the supercritical state, wherein carbon dioxide exists in the form of dissolved state, free state in the supercritical state, and carbonate in the mineral precipitation. Then, it can be known from the control equations of multiphase flow and solute migration that the formation temperature, formation pressure, carbon dioxide saturation, permeability in all directions of the formation, porosity, rock compression coefficient, capillary force, relative permeability, injection well position and injection rate and other geophysical parameters have an influence on the formation pressure and carbon dioxide saturation, so it can be determined that the target geophysical parameters include formation temperature, formation pressure, carbon dioxide saturation, permeability in all directions of the formation, porosity, rock compression coefficient, capillary force, relative permeability, injection well position and injection rate and other geophysical parameters. At the same time, from the equilibrium calculation formula of reaction kinetics minerals, it can be seen that the reaction kinetics of non-local equilibrium minerals is characterized by the fact that the reaction rate changes with the changes in mineral content, pH value, temperature and pressure. Therefore, the target geochemical parameters including mineral volume fraction, ion concentration and pH value can be determined.

[0030] In this way, the processor can input the first geological parameter and the second geological parameter into the pre-constructed first target prediction model, so as to obtain the first predicted formation characteristic parameter of the target formation output by the first target prediction model. Subsequently, the processor can input the first predicted formation characteristic parameter and the second geological parameter into the pre-constructed second target prediction model, so as to obtain the second predicted formation characteristic parameter of the target formation output by the second target prediction model. Among them, the first target prediction model is a pre-trained model for predicting the first predicted formation characteristic parameter, and the second target prediction model is a pre-trained model for predicting the second predicted formation characteristic parameter. The first predicted formation characteristic parameter includes a predicted value of formation pressure, a predicted value of carbon dioxide saturation, and a predicted value of formation fluid density, and the second predicted formation characteristic parameter includes a predicted value of mineral volume fraction and a predicted value of formation permeability. Among them, the predicted value of mineral volume fraction varies due to the difference in the lithology of the rock of the target formation. In one example, if the lithology of the rock of the target formation is gravel-bearing fine sandstone and gravel-bearing mudstone, the predicted value of the mineral volume fraction refers to the predicted volume fraction of major minerals such as quartz, potassium feldspar, calcium feldspar, calcite, ankerite and kaolin.

[0031] It can be understood that in the process of predicting the first predicted formation characteristic parameter and the second predicted formation characteristic parameter, the processor can update the first geological parameter according to the prediction result obtained by the second target prediction model, that is, the predicted value of the formation permeability in the second predicted formation characteristic parameter, so as to predict the first predicted formation characteristic parameter at the next moment, and so on, until the prediction of the first predicted formation characteristic parameter and the second predicted formation characteristic parameter at all moments is completed. In this way, the coupled prediction of the first target prediction model and the second target prediction model can be realized, thereby completing the prediction of the characteristic parameters of the formation during the geological storage of carbon dioxide.

[0032] The above technical solution obtains the first geological parameter and the second geological parameter of the target formation, and then obtains the first predicted formation characteristic parameter of the target formation according to the first geological parameter based on the pre-constructed first target prediction model for predicting the first predicted formation characteristic parameter, and further obtains the second predicted formation characteristic parameter of the target formation according to the first predicted formation characteristic parameter and the second geological parameter based on the pre-constructed second target prediction model for predicting the second predicted formation characteristic parameter. The present application directly predicts the first predicted formation characteristic parameter through the first target prediction model, and directly predicts the second predicted formation characteristic parameter through the second target prediction model and the first predicted formation characteristic parameter, thereby shortening the prediction time of the first predicted formation characteristic parameter and the second predicted formation characteristic parameter, and improving the prediction speed of the first predicted formation characteristic parameter and the second predicted formation characteristic parameter, thereby being able to improve the prediction efficiency of the characteristic parameters of the formation in the process of geological storage of carbon dioxide while ensuring the accuracy of the prediction.

[0033] In an embodiment of the present application, the first target prediction model includes a formation pressure prediction model, a carbon dioxide saturation prediction model, and a fluid density prediction model, and the second target prediction model includes a mineral volume fraction prediction model and a formation permeability prediction model.

[0034] Specifically, the first target prediction model includes a formation pressure prediction model, a carbon dioxide saturation prediction model, and a fluid density prediction model. By inputting the first geological parameter into the formation pressure prediction model, the formation pressure prediction value output by the formation pressure prediction model can be obtained. By inputting the first geological parameter into the carbon dioxide saturation prediction model, the carbon dioxide saturation prediction value output by the carbon dioxide saturation prediction model can be obtained. By inputting the first geological parameter into the fluid density prediction model, the formation fluid density prediction value output by the fluid density prediction model can be obtained. And, by inputting the second geological parameter into the mineral volume fraction prediction model, the mineral volume fraction prediction value output by the mineral volume fraction prediction model can be obtained. By inputting the second geological parameter into the formation permeability prediction model, the formation permeability prediction value output by the formation permeability prediction model can be obtained.

[0035] When predicting the characteristic parameters of the formation through the first target prediction model and the first target prediction model, the first geological parameter and the second geological parameter (both the first geological parameter and the second geological parameter are four-dimensional spatiotemporal data) can be decomposed into three-dimensional spatial data in the time dimension t, and a data pair is constructed according to the field data at the time of input t=i and output t=i+1 (i=0,1,2…n-1) to obtain the predicted values ​​through the first target prediction model and the first target prediction model respectively. After recoupling the predicted values ​​to the time dimension t, four-dimensional spatiotemporal data, namely the first predicted formation characteristic parameter and the second predicted formation characteristic parameter, can be obtained. In this way, based on the formation pressure prediction model, the carbon dioxide saturation prediction model, the fluid density prediction model, the mineral volume fraction prediction model and the formation permeability prediction model, the formation pressure prediction value, the carbon dioxide saturation prediction value, the formation fluid density prediction value, the mineral volume fraction prediction value and the formation permeability prediction value can be predicted.

[0036] In an embodiment of the present application, the prediction method also includes: obtaining a set of numerical simulation input parameters; based on a preset numerical simulation model, obtaining a set of numerical simulation output parameters according to the numerical simulation input parameter set, the numerical simulation output parameter set including a first prediction model data set and a second prediction model data set; dividing the first prediction model data set according to a first preset ratio to obtain a first prediction model training data set and a first prediction model verification data set; and dividing the second prediction model data set according to a second preset ratio to obtain a second prediction model training data set and a second prediction model verification data set.

[0037] The processor can pre-generate a first prediction model training data set, a first prediction model verification data set, a second prediction model training data set, and a second prediction model verification data set. Specifically, the processor can pre-build a preset numerical simulation model. In one example, a model of a target formation with no flow boundaries at the top and bottom can be established. The model outer boundary of the target formation is closed and far away from the injection well, and can be regarded as an infinite oil reservoir. Supercritical carbon dioxide is injected into the target formation at a constant rate of 500 to 2000 cubic meters per day through a vertical injection well with a radius of 0.1m. The vertical injection well can be perforated throughout the reservoir thickness or within a selected depth range. At the same time, the number of grids of the preset numerical simulation model can be set to 35*35*10, and the grid size can capture the migration and pressure accumulation of carbon dioxide plumes while maintaining computational tractability. In this way, the injection of supercritical carbon dioxide for 30 years can be simulated, and the migration of carbon dioxide plumes and the mineralization reaction of carbon dioxide with formation minerals for the next 170 years can be monitored. During the injection period, the time step of the numerical simulation results output for the first 20 years is 0.5 years, the time step of the numerical simulation results output from the 20th to the 30th year is 1 year, and the time step of the numerical simulation results output during the monitoring period (i.e. the next 170 years) is 10 years.

[0038] Subsequently, the processor can obtain a set of numerical simulation input parameters. The set of numerical simulation input parameters includes input parameters such as the initial concentration range of ions, the range of mineral volume fractions, and kinetic parameters. Among them, the range of mineral volume fractions is affected by the dissolution and precipitation reactions of minerals during the injection of carbon dioxide. Therefore, it can be known that the range of mineral volume fractions and kinetic parameters are as follows:

[0039]

[0040] In addition, the total mineralization of the formation water is 60,000 to 75,000, which is a calcium chloride water type, which mainly contains seven kinds of anions and cations, including hydrogen ions, calcium ions, magnesium ions, sodium ions, potassium ions, iron ions, silicon dioxide, bicarbonate ions, sulfate ions, chloride ions and aluminate ions. The initial concentration range of these ions is as follows:

[0041]

[0042] In addition, for flow problems, the input parameters in the numerical simulation input parameter set also include the total number of reservoir grids, the grid size in the first direction, the grid size in the second direction, the grid size in the third direction, the reservoir area, the reservoir thickness, the reservoir top depth, the reservoir initial pressure range, the reservoir initial temperature range, the porosity range, the residual gas saturation, the residual water saturation, the permeability range in the first direction, the permeability range in the second direction, the permeability range in the third direction, the rock compressibility, the water density, the well diameter, and the injection rate range. The parameter values ​​are as follows:

[0043]

[0044]

[0045] The injection rate of the grid at the injection well location is set to a positive number, and that at other locations is set to 0, thereby converting the injection rate into a vector.

[0046] It should be noted that the aforementioned specific values ​​are values ​​that can be adjusted according to actual conditions and are not fixed values.

[0047] Further, the processor may input a numerical simulation input parameter set into a preset numerical simulation model, thereby obtaining a numerical simulation output parameter set output by the preset numerical simulation model. The numerical simulation output parameter set includes a first prediction model data set and a second prediction model data set. The first prediction model data set is a data set used to construct a first target prediction model, and the second prediction model data set is a data set used to construct a second target prediction model. Then, the processor may divide the first prediction model data set according to a first preset ratio to obtain a first prediction model training data set and a first prediction model verification data set, and at the same time divide the second prediction model data set according to a second preset ratio to obtain a second prediction model training data set and a second prediction model verification data set. Both the first preset ratio and the second preset ratio may be adjusted according to actual conditions, and the two may be the same or different. For example, both the first preset ratio and the second preset ratio may be 9:1. In this way, a first prediction model training data set, a first prediction model verification data set, a second prediction model training data set, and a second prediction model verification data set may be generated so as to subsequently construct the first target prediction model and the second target prediction model.

[0048] It can be understood that the first target prediction model can be trained based on the first prediction model training data set, and the first prediction model verification data set can be used to verify the reliability of the first target prediction model. The first prediction model training data set and the first prediction model verification data set both include first input data and first output data. Among them, the first input data includes formation temperature, formation pressure, carbon dioxide saturation, permeability in all directions of the formation, porosity, rock compression coefficient, capillary force, relative permeability, injection well location and injection rate. The first output data includes formation pressure, carbon dioxide saturation and formation fluid density. The second target prediction model can be trained based on the second prediction model training data set, and the second prediction model verification data set can be used to verify the reliability of the second target prediction model. The second prediction model training data set and the second prediction model verification data set both include second input data and second output data. The second input data includes mineral volume fraction, pH value, and formation temperature. The second output data includes formation permeability and mineral volume fraction.

[0049] In addition, according to actual needs, the processor can also adjust the first preset ratio and the second preset ratio, so that the first prediction model data set can be divided into a first prediction model training data set, a first prediction model verification data set and a first prediction model test data set, so that the second prediction model data set can be divided into a second prediction model training data set, a second prediction model verification data set and a second prediction model test data set.

[0050] In an embodiment of the present application, the construction of the first target prediction model and the construction of the second target prediction model include: obtaining a first initial prediction model and a second initial prediction model, wherein the first initial prediction model and the second initial prediction model are both autoregressive Fourier neural operator models; according to a pre-generated first prediction model training data set and a second prediction model training data set, coupling training is performed on the first initial prediction model and the second initial prediction model to construct the first target prediction model and the second target prediction model.

[0051] For time series problems, autoregressive models are usually used for prediction. Traditional autoregressive models use linear combinations of random variables at previous moments to describe random variables at one or several future times, thereby achieving predictions of future variables. Unlike traditional autoregressive models, the embodiments of the present application use an autoregressive Fourier neural operator model. The autoregressive Fourier neural operator model combines the idea of ​​finite differences and simulates multiphase flow, carbon dioxide dissolution, mass transfer and mineralization processes of four-dimensional space-time data (x, y, z, t) based on the current Fourier neural operator model and domain decomposition method.

[0052] Specifically, the processor can obtain the first initial prediction model and the second initial prediction model. The first initial prediction model and the second initial prediction model are both autoregressive Fourier neural operator models. The first initial prediction model and the second initial prediction model both include multiple columns, each column has the same structure and the same weight, and are complete Fourier neural operator models, which can output prediction results at a single moment. In addition, the mth column can process data at time t, and the m+1th column can process data at time t+1 until the data is processed. The output results of each column are used not only as output at the current moment, but also as input at the next moment. In this way, the processor can couple the first initial prediction model and the second initial prediction model according to the pre-generated first prediction model training data set and the second prediction model training data set to construct a first target prediction model and a second target prediction model. In this way, through the first target prediction model and the second target prediction model, a joint prediction of the characteristic parameters of the formation during geological storage of carbon dioxide can be performed.

[0053] In an embodiment of the present application, the first initial prediction model and the second initial prediction model are coupled trained according to the pre-generated first prediction model training data set and the second prediction model training data set to construct a first target prediction model and a second target prediction model, including: training the first initial prediction model according to the first prediction model training data set to obtain the formation pressure output value, the carbon dioxide saturation output value and the formation fluid density output value output by the first initial prediction model; training the second initial prediction model according to the formation pressure output value, the carbon dioxide saturation output value, the formation fluid density output value and the second prediction model training data set to obtain the formation permeability output value output by the second initial prediction model; updating the first prediction model training data set according to the formation permeability output value, and re-updating the first initial prediction model according to the updated first prediction model training data set until the first initial prediction model and the second initial prediction model both meet the preset training termination conditions to obtain the first target prediction model and the second target prediction model.

[0054] The processor can couple the first initial prediction model and the second initial prediction model according to the pre-generated first prediction model training data set and the second prediction model training data set. Specifically, the processor can input the first input data (including formation temperature, formation pressure, carbon dioxide saturation, formation permeability in all directions, porosity, rock compressibility, capillary force, relative permeability, injection well position and injection rate) in the first prediction model training data set into the first initial prediction model to train the first initial prediction model and obtain the formation pressure output value, carbon dioxide saturation output value and formation fluid density output value output by the first initial prediction model. After processing the formation pressure output value, carbon dioxide saturation output value and formation fluid density output value, the formation pressure output value, carbon dioxide saturation output value, formation fluid density output value and the second input data (including mineral volume fraction, pH value, formation temperature) in the second prediction model training data set are input into the second initial prediction model to train the second initial prediction model, thereby obtaining the mineral volume fraction output value and formation permeability output value output by the second initial prediction model.

[0055] Further, the processor can input the formation permeability output value and the carbon dioxide saturation output value into a pre-constructed relative permeability update model (for example, a van Genuchten-Mualem model) to determine the updated relative permeability, and update the first prediction model training data set according to the updated relative permeability, and then re-update the first initial prediction model according to the updated first prediction model training data set, until the first initial prediction model and the second initial prediction model both meet the preset training termination condition, so as to obtain the first target prediction model and the second target prediction model. The preset training termination condition can be that the number of iterations of the model reaches a preset number, that the training time of the model reaches a preset duration, or that the error between the output result of the first initial prediction model and the first output data is less than a preset error and the error between the output result of the second initial prediction model and the second output data is less than a preset error. The specific training termination condition can be determined according to actual needs. In this way, the construction of the first target prediction model and the second target prediction model can be completed.

[0056] In an embodiment of the present application, the prediction method also includes: respectively determining the total loss function, hyperparameters and model evaluation strategy corresponding to the first initial prediction model and the second initial prediction model; based on the total loss function, hyperparameters and model evaluation strategy corresponding to the first initial prediction model and the second initial prediction model, according to the first prediction model training data set and the second prediction model training data set, coupling training is performed on the first initial prediction model and the second initial prediction model.

[0057] Before coupling training the first initial prediction model and the second initial prediction model, the processor may respectively determine the total loss function, hyperparameters, and model evaluation strategies corresponding to the first initial prediction model and the second initial prediction model. The total loss function, hyperparameters, and model evaluation strategies corresponding to the first initial prediction model and the second initial prediction model may be the same or different, and are not limited here. Specifically, the total loss function includes three parts, namely, the L2 loss function, the boundary condition loss function, and the initial condition loss function. The function value of the total loss function satisfies the following formula:

[0058]

[0059] in, is the function value of the total loss function, w1 is the weight corresponding to the function value of the L2 loss function, L normal is the function value of the L2 loss function, w2 is the weight corresponding to the function value of the boundary condition loss function, L OBC is the function value of the boundary condition loss function, w3 is the weight corresponding to the function value of the initial condition loss function, L IBC is the function value of the initial condition loss function.

[0060] The function value of the L2 loss function satisfies the following formula:

[0061]

[0062] Among them, L normal is the function value of the L2 loss function, u is the true value, is the predicted output value, d u is the change in the true value, is the change in the predicted output value, ||||2 represents the L2 norm, and β is a constant.

[0063] In addition, since the outer boundary of the target formation model is closed, the formation pressure and carbon dioxide saturation cannot be transmitted outward, so the derivative in the direction of the outer normal is limited to zero. Therefore, the function value of the boundary condition loss function satisfies the following formula:

[0064]

[0065] Since the vertical injection well has injection rate control, the function value of the initial condition loss function satisfies the following formula:

[0066]

[0067] Among them, L OBC is the function value of the boundary condition loss function, L IBC is the function value of the initial condition loss function, N xbcis the grid of the first direction boundary, i is the i-th grid in the first direction, N ybc is the grid at the boundary of the second direction, j is the jth grid in the second direction, k is the kth grid in the third direction, x, y, z represent the first direction, the second direction and the third direction respectively, N zbc is the grid of the third boundary, N inj is the grid where the vertical injection well is located, is the predicted output value, n is the external normal direction, r is the radius of the vertical injection well, and const is the injection rate of the vertical injection well.

[0068] In addition, hyperparameters include learning rate, weight decay, and activation function. Assume that the learning rate is 10 -3 Up to 10 -5 The weight decay is between 10 -2 Up to 10 -4 , and the activation function is any one of relu, leaky-relu, selu and silu. After testing, the hyper parameters of the embodiment of the present application are determined as follows: learning rate 10 -4 , weight decay 10 -4 , the activation function is relu. In addition, the learning function can use the Adam optimizer, the batch size is selected as 4, and the number of epochs of the Fourier neural operator model is set to 63. The embodiment of the present application also adopts a stopping criterion. If the verification loss does not decrease further after a certain number of epochs, continue training. Otherwise, stop training.

[0069] For the model evaluation strategy, normalized residual, mean absolute percentage error, absolute coefficient and structural similarity index matrix are used to evaluate the accuracy and reliability of the model. The mean absolute percentage error can satisfy the following formula:

[0070]

[0071] The normalized residual can satisfy the following formula:

[0072]

[0073] The absolute coefficient can satisfy the following formula:

[0074]

[0075] The structural similarity index matrix can satisfy the following formula:

[0076]

[0077] Among them, MAPE is the mean absolute percentage error, NRE is the normalized residual, R 2is the absolute coefficient, SSIM is the structural similarity index matrix, y itrue is the output value of each grid obtained by numerical simulation, y true is the output value obtained by numerical simulation, is the average value of the output value obtained by numerical simulation, y true_max is the maximum value of the output value obtained by numerical simulation, y true_min is the minimum output value obtained by numerical simulation, y ipredicted is the predicted value of each grid obtained by the first target prediction model or the second target prediction model, y predicted is the predicted value obtained by the first target prediction model or the second target prediction model, is the average value of the prediction values ​​obtained by the first target prediction model or the second target prediction model, n samples is the total number of samples; u and v are the true value and predicted value (three-dimensional matrix), respectively. The structural similarity index matrix SSIM is calculated in m local windows of the u and v images; μ represents the mean, σ represents the standard deviation; K1 and K2 are both decimals to avoid zero values ​​in the denominator. The minimum value of the structural similarity index matrix SSIM is 0 (completely different) and the maximum value is 1 (completely the same).

[0078] Then, based on the total loss function, hyperparameters and model evaluation strategy corresponding to the first initial prediction model and the second initial prediction model, the processor can perform coupled training on the first initial prediction model and the second initial prediction model according to the first prediction model training data set and the second prediction model training data set.

[0079] In an embodiment of the present application, the prediction method also includes: verifying the first target prediction model based on a pre-generated first prediction model verification data set, and verifying the second target prediction model based on a pre-generated second prediction model verification data set.

[0080] The first prediction model verification data set is obtained by dividing the first prediction model data set generated by the preset numerical simulation model according to the first preset ratio, and the second prediction model verification data set is obtained by dividing the second prediction model data set generated by the preset numerical simulation model according to the second preset ratio. After the training is completed to obtain the first target prediction model and the second target prediction model, the processor can verify the first target prediction model according to the pre-generated first prediction model verification data set, and verify the second target prediction model according to the pre-generated second prediction model verification data set, so as to verify the prediction accuracy of the first target prediction model and the second target prediction model. When the verification fails, the model parameters of the first target prediction model and the second target prediction model can be adjusted based on the verification results, or the first prediction model training data set and the second prediction model training data set can be adjusted based on the verification results.

[0081] In a specific embodiment of the present application, a single injection well in a heterogeneous low-porosity, low-permeability gravel-containing sandstone reservoir is used as the target formation to simulate and predict the geological storage process of carbon dioxide. The target formation generally exhibits a positive cycle characteristic of thinning thickness and finer grain size upward, with a porosity ranging from 2.3% to 15%, an average of 10.2%, and a permeability ranging from 0.03mD to 50mD, an average of 23.73mD. The reservoir parameters and simulation information of the target formation are as follows:

[0082]

[0083] The main ions affecting mineralization in the target formation include bicarbonate ions, calcium ions and magnesium ions, among which the ion concentration of bicarbonate ions is in the range of 0.0791 to 1.4450 mol / kg H2O, and the initial average ion concentration is 0.0801 mol / kg H2O. The ion concentration of calcium ions is in the range of 0.0107-0.0404 mol / kg H2O, and the initial average ion concentration is 0.0108 mol / kg H2O. The ion concentration of magnesium ions is in the range of 0.0010-0.0040 mol / kg H2O, and the initial average ion concentration is 0.0011 mol / kg H2O.

[0084] The main minerals involved in the mineralization reaction in the target formation include potassium feldspar, calcium feldspar, calcite, iron dolomite and kaolin. Specifically, the volume fraction of potassium feldspar ranges from 0.0809 to 0.0904, with an average volume fraction of 0.0855. The volume fraction of calcium feldspar ranges from 0.2438 to 0.2724, with an average volume fraction of 0.2576. The volume fraction of calcite ranges from 0.1135 to 0.1267, with an average volume fraction of 0.1199. The volume fraction of kaolin ranges from 0.0463 to 0.0519, with an average volume fraction of 0.0490. The volume fraction of iron dolomite ranges from 0.1132 to 0.1268, with an average volume fraction of 0.1198.

[0085] Under the above parameters, supercritical carbon dioxide is injected into the saline layer within a set time period and monitored after the injection is completed. The dissolved supercritical carbon dioxide gas is converted into carbonic acid, which reacts with the formation to cause mineral dissolution and precipitation. During the injection process, the formation pressure, carbon dioxide saturation and fluid density continue to increase, and as the injection stops and the mineralization reaction proceeds, the formation pressure, carbon dioxide saturation and fluid density gradually decrease. During the injection and monitoring process, the volume fraction of calcite changes dynamically. At the front of the carbon dioxide sweep, calcite dissolves, and then calcite begins to precipitate due to the chemical reaction between carbon dioxide and water. At the same time, as the mineralization reaction proceeds, calcium feldspar and potassium feldspar continue to dissolve, and kaolinite and ankerite precipitate. For potassium feldspar, calcium feldspar, calcite, ankerite and kaolinite, the reaction rate is relatively slow at the beginning, but as more carbon dioxide dissolves in the saline layer, the reaction rate gradually increases.

[0086] The embodiment of the present application also provides a processor configured to execute the above-mentioned method for predicting characteristic parameters of a formation based on geological storage of carbon dioxide.

[0087] Specifically, in an embodiment of the present application, the processor can be configured to: obtain a first geological parameter and a second geological parameter of the target formation, the first geological parameter including a target geophysical parameter of the target formation, and the second geological parameter including a formation temperature and a target geochemical parameter of the target formation; based on a pre-constructed first target prediction model for predicting the first predicted formation characteristic parameter, obtain the first predicted formation characteristic parameter of the target formation according to the first geological parameter, wherein the first predicted formation characteristic parameter includes a predicted value of formation pressure, a predicted value of carbon dioxide saturation, and a predicted value of formation fluid density; based on a pre-constructed second target prediction model for predicting the second predicted formation characteristic parameter, obtain the second predicted formation characteristic parameter of the target formation according to the first predicted formation characteristic parameter and the second geological parameter, wherein the second predicted formation characteristic parameter includes a predicted value of mineral volume fraction and a predicted value of formation permeability.

[0088] In one embodiment, the processor is further configured to: obtain a first initial prediction model and a second initial prediction model, wherein the first initial prediction model and the second initial prediction model are both autoregressive Fourier neural operator models; and perform coupling training on the first initial prediction model and the second initial prediction model according to a pre-generated first prediction model training data set and a second prediction model training data set to construct a first target prediction model and a second target prediction model.

[0089] In one embodiment, the processor is further configured to: train the first initial prediction model according to the first prediction model training data set to obtain the formation pressure output value, carbon dioxide saturation output value and formation fluid density output value output by the first initial prediction model; train the second initial prediction model according to the formation pressure output value, carbon dioxide saturation output value, formation fluid density output value and the second prediction model training data set to obtain the formation permeability output value output by the second initial prediction model; update the first prediction model training data set according to the formation permeability output value, and re-update the first initial prediction model according to the updated first prediction model training data set until the first initial prediction model and the second initial prediction model both meet the preset training termination conditions to obtain the first target prediction model and the second target prediction model.

[0090] In one embodiment, the processor is further configured to: respectively determine the total loss function, hyperparameters and model evaluation strategy corresponding to the first initial prediction model and the second initial prediction model; based on the total loss function, hyperparameters and model evaluation strategy corresponding to the first initial prediction model and the second initial prediction model, perform coupling training on the first initial prediction model and the second initial prediction model according to the first prediction model training data set and the second prediction model training data set.

[0091] In one embodiment, the processor is further configured to: verify the first target prediction model based on a pre-generated first prediction model verification data set, and verify the second target prediction model based on a pre-generated second prediction model verification data set.

[0092] In one embodiment, the processor is further configured to: obtain a set of numerical simulation input parameters; based on a preset numerical simulation model, obtain a set of numerical simulation output parameters according to the numerical simulation input parameter set, the numerical simulation output parameter set including a first prediction model data set and a second prediction model data set; divide the first prediction model data set according to a first preset ratio to obtain a first prediction model training data set and a first prediction model verification data set; and divide the second prediction model data set according to a second preset ratio to obtain a second prediction model training data set and a second prediction model verification data set.

[0093] In one embodiment, the processor is further configured such that: the target geophysical parameters include formation temperature, formation pressure, carbon dioxide saturation, permeability in all directions of the formation, porosity, rock compressibility, capillary force, relative permeability, injection well location and injection rate; and the target geochemical parameters include mineral volume fraction and pH value.

[0094] In one embodiment, the processor is further configured such that: the first target prediction model includes a formation pressure prediction model, a carbon dioxide saturation prediction model, and a fluid density prediction model; and the second target prediction model includes a mineral volume fraction prediction model and a formation permeability prediction model.

[0095] The above technical solution obtains the first geological parameter and the second geological parameter of the target formation, and then obtains the first predicted formation characteristic parameter of the target formation according to the first geological parameter based on the pre-constructed first target prediction model for predicting the first predicted formation characteristic parameter, and further obtains the second predicted formation characteristic parameter of the target formation according to the first predicted formation characteristic parameter and the second geological parameter based on the pre-constructed second target prediction model for predicting the second predicted formation characteristic parameter. The present application directly predicts the first predicted formation characteristic parameter through the first target prediction model, and directly predicts the second predicted formation characteristic parameter through the second target prediction model and the first predicted formation characteristic parameter, thereby shortening the prediction time of the first predicted formation characteristic parameter and the second predicted formation characteristic parameter, and improving the prediction speed of the first predicted formation characteristic parameter and the second predicted formation characteristic parameter, thereby being able to improve the prediction efficiency of the characteristic parameters of the formation in the process of geological storage of carbon dioxide while ensuring the accuracy of the prediction.

[0096] An embodiment of the present application further provides a machine-readable storage medium having instructions stored thereon, the instructions being used to enable a machine to execute the above-mentioned method for predicting characteristic parameters of a formation based on geological storage of carbon dioxide.

[0097] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0098] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0099] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0100] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0101] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0102] The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.

[0103] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.

[0104] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.

[0105] The above are only embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application should be included within the scope of the claims of the present application.

Claims

1. A method for predicting characteristic parameters of a formation based on geological storage of carbon dioxide, It is characterized in that include: Acquire a first geological parameter and a second geological parameter of a target formation, wherein the first geological parameter includes a target geophysical parameter of the target formation, and the second geological parameter includes a formation temperature and a target geochemical parameter of the target formation; Based on a pre-constructed first target prediction model for predicting first predicted formation characteristic parameters, obtaining first predicted formation characteristic parameters of the target formation according to the first geological parameters, wherein the first predicted formation characteristic parameters include a predicted value of formation pressure, a predicted value of carbon dioxide saturation, and a predicted value of formation fluid density; Based on a pre-constructed second target prediction model for predicting second predicted formation characteristic parameters, second predicted formation characteristic parameters of the target formation are obtained according to the first predicted formation characteristic parameters and the second geological parameters, wherein the second predicted formation characteristic parameters include a predicted value of mineral volume fraction and a predicted value of formation permeability.

2. The prediction method according to claim 1, It is characterized in that The construction of the first target prediction model and the construction of the second target prediction model include: Acquire a first initial prediction model and a second initial prediction model, wherein both the first initial prediction model and the second initial prediction model are autoregressive Fourier neural operator models; According to the pre-generated first prediction model training data set and the second prediction model training data set, the first initial prediction model and the second initial prediction model are coupled trained to construct the first target prediction model and the second target prediction model.

3. The prediction method according to claim 2, It is characterized in that The coupling training of the first initial prediction model and the second initial prediction model according to the pre-generated first prediction model training data set and the second prediction model training data set to construct the first target prediction model and the second target prediction model includes: The first initial prediction model is trained according to the first prediction model training data set to obtain a formation pressure output value, a carbon dioxide saturation output value, and a formation fluid density output value output by the first initial prediction model; Training the second initial prediction model according to the formation pressure output value, the carbon dioxide saturation output value, the formation fluid density output value and a second prediction model training data set to obtain a formation permeability output value output by the second initial prediction model; The first prediction model training data set is updated according to the formation permeability output value, and the first initial prediction model is re-updated according to the updated first prediction model training data set until the first initial prediction model and the second initial prediction model both meet the preset training termination conditions to obtain the first target prediction model and the second target prediction model.

4. The prediction method according to claim 2, It is characterized in that The prediction method further comprises: Determine the total loss function, hyperparameters, and model evaluation strategy corresponding to the first initial prediction model and the second initial prediction model respectively; Based on the total loss function, the hyperparameters and the model evaluation strategy corresponding to the first initial prediction model and the second initial prediction model, the first initial prediction model and the second initial prediction model are coupled trained according to the first prediction model training data set and the second prediction model training data set.

5. The prediction method according to claim 2, It is characterized in that The prediction method further comprises: The first target prediction model is verified according to a pre-generated first prediction model verification data set, and the second target prediction model is verified according to a pre-generated second prediction model verification data set.

6. The prediction method according to claim 5, It is characterized in that The prediction method further comprises: Get the numerical simulation input parameter set; Based on a preset numerical simulation model, a numerical simulation output parameter set is obtained according to the numerical simulation input parameter set, wherein the numerical simulation output parameter set includes a first prediction model data set and a second prediction model data set; Dividing the first prediction model data set according to a first preset ratio to obtain the first prediction model training data set and the first prediction model verification data set; and The second prediction model data set is divided according to a second preset ratio to obtain the second prediction model training data set and the second prediction model verification data set.

7. The prediction method according to claim 1, It is characterized in that The target geophysical parameters include formation temperature, formation pressure, carbon dioxide saturation, formation permeability in all directions, porosity, rock compression coefficient, capillary force, relative permeability, injection well location and injection rate; the target geochemical parameters include mineral volume fraction and pH value.

8. The prediction method according to claim 1, It is characterized in that The first target prediction model includes a formation pressure prediction model, a carbon dioxide saturation prediction model and a fluid density prediction model, and the second target prediction model includes a mineral volume fraction prediction model and a formation permeability prediction model.

9. A processor, It is characterized in that The method is configured to execute the method for predicting characteristic parameters of a formation based on geological storage of carbon dioxide according to any one of claims 1 to 8.

10. A machine-readable storage medium, It is characterized in that The machine-readable storage medium stores instructions for causing a machine to execute a method for predicting characteristic parameters of a formation based on geological storage of carbon dioxide according to any one of claims 1 to 8.

Citation Information

Cited By

  • Formation pressure prediction method and device for high water injection rate working condition

    CN119940084A

  • A method and device for formation pressure prediction for high water injection rate conditions

    CN119940084B

  • Gas permeability monitoring method and system based on carbon dioxide geological sequestration

    CN120294836A

  • Gas permeability monitoring method and system based on carbon dioxide geological storage

    CN120294836B

  • Method for simulating and predicting saturation field and pressure field in underground carbon dioxide injection process based on deep learning

    CN121936309A