Method and device for calculating CO2 storage capacity of geologic body based on artificial intelligence

Abstract

The invention discloses a geologic body CO2 sequestration capacity calculation method and device based on artificial intelligence, and relates to the technical field of carbon sequestration of different types of geologic bodies, and the method comprises the steps: constructing three-dimensional geologic models of different types of oil reservoirs; according to the remaining oil distribution laws of the different types of oil reservoirs, the influence laws of the different types of oil reservoirs on CO2 burying under different reservoir conditions are analyzed, and then the change laws of various mechanism contribution rates under different conditions are determined; combining a numerical simulation technology and a change rule of a mechanism contribution rate, and through an artificial intelligence training model, forming an effective embedding coefficient plate under the influence of multiple factors; and establishing an effective carbon burying stock calculation method for different types of oil reservoirs to calculate the effective carbon burying stock and screen out favorable burying areas. According to the method for calculating the carbon burying stock of the different types of oil reservoirs, theoretical guidance is provided for backing up the potential of the oil field carbon reservoirs, the buried carbon amount of the different types of oil reservoirs can be rapidly and accurately calculated, and the carbon burying potential is defined.

Description

Technical Field

[0001] This application belongs to the field of carbon sequestration technology for different types of geological bodies, and in particular relates to a method and apparatus for calculating the CO2 sequestration capacity of geological bodies based on artificial intelligence. Background Technology

[0002] The rising concentration of greenhouse gases in the atmosphere is one of the main causes of frequent extreme weather events globally, with carbon dioxide being the most abundant greenhouse gas emitter. Under the "dual carbon" context, CO2 flooding for enhanced oil recovery and storage is currently the most economical and effective means to reduce greenhouse gas emissions and address global environmental issues. CCUS-EOR technology is applicable to various reservoirs, from low-permeability to medium-high permeability. Compared to other gases, CO2 not only dissolves easily in crude oil, increasing its volume factor and reducing its viscosity, but it also reaches a supercritical state at pressures greater than 7.38 MPa and temperatures above 31.2℃, increasing recovery by 10% to 30% compared to waterflooding. Furthermore, different types of reservoirs (light oil, heavy oil, and high-pour-point oil) have different characteristics in terms of reservoir structure, oil quality, temperature, pressure, saturation, trapping capacity, and CO2 miscibility. The water-rock reaction after CO2 injection has varying effects on reservoir space and properties, all of which influence the final carbon storage. Heavy oil reservoirs, due to their high viscosity and poor fluidity, often fail to achieve miscible flooding during CO2 injection. CO2 injection can significantly reduce crude oil viscosity and improve its flow characteristics. However, in practical applications, the low viscosity of CO2 can easily lead to channeling, and the effectiveness of CO2 injection in heavy oil reservoirs requires further research and evaluation. High-pour-point oils generally exhibit characteristics such as high wax content, high pour point, and high wax precipitation temperature. In the CO2-driven development of high-pour-point oil reservoirs, wax deposition is one of the key factors restricting the development effectiveness of such reservoirs.

[0003] The amount of CO2 stored and the establishment of calculation methods are closely related to the CO2 storage mechanism in oil reservoirs. After CO2 is injected into an oil reservoir, it is usually sealed in the form of mineralization, dissolution, structural and formation traps, and confined spaces. Most of the injected CO2 is sealed in the reservoir as free gas. The diffusion rate is determined by the properties of the fluid and the porosity and permeability of the reservoir, and is also affected by the boundary properties. The most important storage methods are structural space and confined space storage. Therefore, the key to calculating the theoretical CO2 storage in an oil reservoir is to determine the geometric space in the reservoir that can supply CO2 storage. However, with the extension of storage time, the dissolution of CO2 in crude oil and water cannot be ignored. This issue should be fully considered in the calculation. Furthermore, previous studies have suggested that the volume of stored CO2 depends on the pressure and temperature of the reservoir. Previous definitions of theoretical storage capacity refer to the maximum capacity that can be stored in an oilfield. Effective storage capacity is calculated by considering factors such as buoyancy, gravity overburden, mobility ratio, heterogeneity, and water saturation. Actual storage capacity considers technical, legal, governmental, and economic factors related to CO2 storage, and it varies with changes in technology, policies, data, and economic conditions. Matched storage capacity is evaluated based on the volume of the storage body, injection volume, and supply volume; this capacity represents the minimum. Therefore, calculating the effective CO2 storage capacity by comprehensively considering all influencing factors is crucial for solving practical field problems.

[0004] In recent years, artificial intelligence (AI) technology has been increasingly widely applied in CCUS-EOR (Chemical Coal-to-Oil Retention). Various machine learning and deep learning methods can analyze potential carbon burial patterns from complex historical data. However, it is still necessary to select appropriate methods for different geological conditions and application scenarios to maximize the utilization of computational resources. Traditional methods mainly compare schemes through numerical simulation to select the scheme with the largest burial volume as the recommended scheme. This method is not true optimization, but merely the selection of the best among the available schemes. Furthermore, traditional optimization methods only perform single static optimizations of the entire CCUS-EOR process, using unchanging well control parameters throughout, which does not meet the requirements of oilfield operations. Therefore, a statistical or machine learning method is needed to replace numerical simulation. By training this method with thousands, tens of thousands, or even more sets of numerical simulation data, and inputting specific parameters (such as permeability, temperature, and pressure) into the surrogate model, results can be obtained rapidly within seconds. Utilizing AI technology to solve practical problems in oilfields is an essential path for the rapid development of oilfields. Summary of the Invention

[0005] The embodiments of this application provide a method and apparatus for calculating the CO2 sequestration capacity of geological bodies based on artificial intelligence, which solves three technical problems in the prior art.

[0006] Other features and advantages of this application will become apparent from the following detailed description, or may be learned in part from practice of this application.

[0007] According to a first aspect of the embodiments of this application, such as Figure 1 As shown, an artificial intelligence-based method for calculating the CO2 sequestration capacity of geological bodies is provided, including:

[0008] Three-dimensional geological models of different types of reservoirs were constructed to determine the distribution patterns of remaining oil in different types of reservoirs;

[0009] Based on the distribution patterns of remaining oil in different types of reservoirs, the influence of different reservoir conditions on CO2 burial is analyzed, and the variation patterns of the contribution rates of various mechanisms under different conditions are determined. These different reservoir conditions include permeability, reservoir pressure, and temperature.

[0010] Combining numerical simulation technology with the variation law of mechanism contribution rate, a method for calculating CO2 effective storage coefficient was developed, and an effective storage coefficient chart under the influence of multiple factors was generated by training the model with artificial intelligence.

[0011] Based on the effective carbon storage coefficient chart, a method for calculating the effective carbon storage of different types of reservoirs is established to calculate the effective carbon storage and screen out favorable storage areas.

[0012] According to a second aspect of the embodiments of this application, such as Figure 2 As shown, an artificial intelligence-based CO2 sequestration capacity calculation device for geological bodies is provided, comprising:

[0013] Module: Used to build three-dimensional geological models of different types of reservoirs to determine the distribution patterns of remaining oil in different types of reservoirs;

[0014] Determine the module: Based on the distribution pattern of remaining oil in different types of reservoirs, analyze the influence of different reservoir conditions on CO2 burial in different types of reservoirs, and then determine the variation of the contribution rate of various mechanisms under different conditions, including permeability, reservoir pressure and temperature;

[0015] Training module: This module combines numerical simulation techniques with the variation law of mechanism contribution rate to develop a method for calculating the effective CO2 storage coefficient, and uses artificial intelligence to train the model to generate an effective storage coefficient chart under the influence of multiple factors.

[0016] Calculation module: Used for effective carbon storage coefficient charts, establishing calculation methods for effective carbon storage in different types of reservoirs, in order to calculate effective carbon storage and screen out favorable storage areas.

[0017] According to a third aspect of the embodiments of this application, an artificial intelligence-based geological body CO2 sequestration capacity calculation device is provided, including a processor and a memory, wherein the memory stores computer program instructions that can be executed by the processor, and when the processor executes the computer program instructions, it implements the steps of the method described in any of the first aspects above.

[0018] According to a fourth aspect of the embodiments of this application, a computer-readable storage medium is provided, wherein computer program instructions are stored therein, and when executed by a processor, the computer program instructions cause the processor to perform the steps of the method as described in any of the first aspects above.

[0019] This invention addresses a technical bottleneck that existing technologies struggle to overcome: for various types of oil reservoirs, it identifies key factors influencing CO2 storage and establishes methods for calculating carbon storage in different types of oil reservoirs, providing theoretical guidance for assessing the potential of oilfield carbon reserves.

[0020] This invention proposes an artificial intelligence-based method for calculating the effective CO2 sequestration capacity of different types of geological bodies, which can quickly and accurately calculate the carbon sequestration potential of different types of reservoirs and clarify the carbon sequestration potential.

[0021] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0022] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application. It is obvious that the drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort. In the drawings:

[0023] Figure 1 This is a schematic diagram of the L-block proxy model verification for thin oil.

[0024] Figure 2 This is a schematic diagram of the proxy model verification for heavy oil D-block;

[0025] Figure 3 This is a schematic diagram of the proxy model verification for the S block of high-pour-point oil;

[0026] Figure 4 This is a schematic diagram of model integration;

[0027] Figure 5 This is a graph showing the calculation results of the effective burial coefficient of the L block of light oil (28MPa).

[0028] Figure 6This is a graph showing the calculation results of the effective burial coefficient of the L block of light oil (29MPa).

[0029] Figure 7 This is a graph showing the calculation results of the effective burial coefficient of the L block of light oil (30MPa).

[0030] Figure 8 This is a graph showing the calculation results of the effective burial coefficient of the L block of light oil (31MPa).

[0031] Figure 9 This is a chart showing the calculation results of the effective burial coefficient of heavy oil block D (12MPa);

[0032] Figure 10 This is a graph showing the calculation results of the effective burial coefficient of heavy oil block D (13MPa);

[0033] Figure 11 This is a chart showing the calculation results of the effective burial coefficient of heavy oil block D (14MPa);

[0034] Figure 12 This is a chart showing the calculation results of the effective burial coefficient of heavy oil block D (15MPa).

[0035] Figure 13 This is a chart showing the calculation results of the effective burial coefficient of the high-pour-point oil S block (33MPa);

[0036] Figure 14 This is a chart showing the calculation results of the effective burial coefficient of the high-pour-point oil S block (34MPa);

[0037] Figure 15 This is a chart showing the calculation results of the effective burial coefficient of high-pour-point oil in Block S (35MPa);

[0038] Figure 16 This is a chart showing the calculation results of the effective burial coefficient of the high-pour-point oil S block (336MPa);

[0039] Figure 17 A schematic diagram of a method for calculating the CO2 sequestration capacity of a geological body based on artificial intelligence is shown in one embodiment.

[0040] Figure 18 A schematic diagram of an artificial intelligence-based geological body CO2 sequestration capacity calculation device is shown in one embodiment.

[0041] Figure 19 A schematic diagram of an artificial intelligence-based geological body CO2 sequestration capacity calculation device is shown in one embodiment. Detailed Implementation

[0042] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0043] Furthermore, the described features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. Numerous specific details are provided in the following description to give a thorough understanding of embodiments of this application. However, those skilled in the art will recognize that the technical solutions of this application can be practiced without one or more of the specific details, or other methods, components, apparatuses, steps, etc., can be employed. In other instances, well-known methods, apparatuses, implementations, or operations are not shown or described in detail to avoid obscuring various aspects of this application.

[0044] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0045] The flowcharts shown in the accompanying drawings are merely illustrative and do not necessarily include all content and operations / steps, nor do they necessarily have to be performed in the described order. For example, some operations / steps can be broken down, while others can be combined or partially combined; therefore, the actual execution order may change depending on the specific circumstances.

[0046] Currently, there is a lack of theoretical and technical research on the impact mechanism, key factors, and calculation methods of different types of reservoir conditions on CO2 storage. For various types of reservoirs, it is necessary to identify the key factors affecting CO2 storage, and to quickly and accurately calculate the carbon storage capacity of reservoirs by establishing calculation methods for different types of reservoirs, thereby clarifying the carbon storage potential and providing theoretical guidance for assessing the carbon pool potential of oilfields.

[0047] Specifically, Figure 17 This document illustrates a flowchart of an AI-based method for calculating the CO2 sequestration capacity of geological bodies, as shown in one embodiment. For various types of oil reservoirs, it is necessary to identify the key factors affecting CO2 sequestration. By establishing calculation methods for different types of oil reservoirs, the potential for CO2 sequestration can be quickly and accurately calculated, clarifying the carbon sequestration potential and providing theoretical guidance for assessing the carbon pool potential of oilfields. Figure 1 As shown, an artificial intelligence-based method for calculating the CO2 sequestration capacity of geological bodies is provided, which may include the following steps 100 to 400.

[0048] S100. Construct three-dimensional geological models of different types of reservoirs to determine the distribution patterns of remaining oil in different types of reservoirs.

[0049] It should be noted that step S100 includes S101 and S102, wherein:

[0050] S101. Using geological modeling software and geological history research data, construct three-dimensional structural and sedimentary facies models, establish attribute models, combine well logging interpretation results and oil, gas and water distribution data to form saturation and net-to-gross ratio models, and calculate original geological reserves.

[0051] S102. Numerical simulation software is used to fit the PVT phase parameters of the reservoir. Based on the reservoir geological model, phase simulation data and state equations are imported, and the distribution pattern of the remaining oil in the reservoir is obtained through historical fitting calculations.

[0052] Understandably, based on the results of detailed geological research, three-dimensional structural and sedimentary facies models are constructed using geological modeling software. A facies-controlled attribute model is established, and saturation and net-to-gross ratio models are established by combining well logging interpretation results and oil, gas, and water distribution data to calculate the original geological reserves. Numerical simulation facies simulation software is used to fit the reservoir's PVT facies parameters. Then, based on the reservoir geological modeling, the geological model is imported into the numerical simulation software, and the state equations generated by the facies simulation, which represent the actual fluid properties of the reservoir, are imported into the model to calculate the mechanisms reflecting the dissolution, expansion, and viscosity reduction of gas in oil and water. Historical data fitting is performed to study the distribution patterns of remaining oil in the reservoir.

[0053] S200. Based on the distribution patterns of remaining oil in different types of reservoirs, analyze the influence of different reservoir conditions on CO2 burial in different types of reservoirs, and then determine the variation patterns of the contribution rates of various mechanisms under different conditions, including permeability, reservoir pressure and temperature.

[0054] It should be noted that the influence of CO2 on the oil displacement effect of different types of reservoirs under different reservoir conditions (permeability, pressure and temperature) was studied. By analyzing the recovery results obtained from numerical simulation, the contribution rate of various mechanisms of CO2 burial under different reservoir conditions was obtained.

[0055] Therefore, the first step in this process is to collect geological, production, and fluid property data for different types of reservoirs. This data provides the foundation for analyzing the distribution patterns of remaining oil. Through geostatistical methods and numerical simulations, a three-dimensional distribution model of the remaining oil can be established. Reservoir conditions, such as permeability, reservoir pressure, and temperature, have a significant impact on fluid flow and CO2 burial processes within the reservoir. For example, high permeability may increase the CO2 injection rate, but it may also accelerate CO2 escape; changes in reservoir pressure affect the solubility and phase of CO2; temperature directly affects the viscosity and density of the fluid, thus influencing CO2 diffusion and reservoir fluidity. After determining the reservoir conditions, the contribution of different mechanisms to CO2 burial effects is further analyzed. These mechanisms may include dissolution, capillary action, gravitational differentiation, and diffusion. For example, dissolution can increase crude oil fluidity and improve recovery; capillary action can affect the distribution of CO2 in the reservoir. Through experimental data and numerical simulations, the changing trends of the contribution rates of each mechanism under different reservoir conditions can be observed. For example, capillary forces may dominate in low-permeability reservoirs, while dissolution may be more important in high-permeability reservoirs.

[0056] S300, combining numerical simulation technology and the variation law of mechanism contribution rate, developed a method for obtaining CO2 effective storage coefficient, and formed an effective storage coefficient chart under the influence of multiple factors by training the model with artificial intelligence.

[0057] In this step, the effective storage coefficient is determined as follows: Numerical simulation is the most effective method for calculating recovery rate and effective storage coefficient. CO2 is affected by factors such as fluid viscosity differences, fluid density differences, formation heterogeneity, water saturation, and strong water bodies. Compared with empirical methods for obtaining correlation coefficients, using "numerical simulation technology + experimental measurements" to calculate key parameters is more reliable. This method calculates the actual CO2 storage volume considering the influence of multiple factors, and then calculates the effective storage coefficient. A method for determining the effective storage coefficient of CO2 is developed, and the calculation formula for determining the effective storage coefficient of CO2 is as follows:

[0058] Effective storage coefficient = Actual CO2 storage volume / Theoretical CO2 storage volume

[0059] In the formula, the actual CO2 storage amount = the amount of CO2 injected - the amount of CO2 produced, and the theoretical CO2 storage amount = the amount of CO2 stored in the pore space + the amount of CO2 dissolved in water + the amount of CO2 dissolved in oil.

[0060] Acquisition of the surrogate model: Data sets of different types of reservoirs are obtained using numerical simulation methods. The models are then trained and continuously optimized to generate surrogate models that meet the accuracy requirements.

[0061] It is understood that step S300 includes S301, S302, S303, and S304, wherein:

[0062] S301. Obtain numerical simulation datasets for different types of oil reservoirs. Based on the mean and standard deviation of the original data, perform z-score standardization to obtain standardized datasets. The specific method is as follows:

[0063]

[0064] In the formula: 'a' represents the dataset; s represents the data mean; s represents the standard deviation; n represents the number of data points. To eliminate the influence of the indicator's dimensions, the indicator values ​​are standardized using the z-score method based on the original data mean and standard deviation.

[0065] S302. Based on the standardized dataset, after training the machine learning model, R-squared, MSE and RMSE are determined as the evaluation indicators to obtain the preliminary optimized machine learning model.

[0066] It should be noted that the coefficient of certainty represents the numerical characteristic of the relationship between a random variable and multiple random variables. It is used to reflect the regression model and is a statistical indicator of the reliability of changes in the dependent variable.

[0067]

[0068] Mean squared error is a measure of the difference between the estimator and the estimated quantity.

[0069]

[0070] The root mean square error (RMSE) is the square root of the ratio of the square of the deviation between the predicted and actual values ​​to the number of observations (n). It is used to measure the deviation between the observed and actual values.

[0071]

[0072] In the formula: wi represents the weight; n represents the number of data points; Indicates the prediction result; Indicates the actual result; This represents the average of the actual results.

[0073] S303. Based on the preliminary optimized machine learning model, after parameter adjustment and kernel function selection, the six types of machine learning algorithms are further subdivided to obtain 24 subdivided models. The six types of machine learning algorithms include support vector regression, Gaussian process regression, tree ensemble, neural network, linear regression, and regression tree.

[0074] It should be noted that 24 commonly used machine learning methods across 6 major categories were used to build parameter optimization models, including major algorithms such as support vector regression, Gaussian process regression, tree ensemble, neural networks, linear regression, and regression trees. These algorithms can be further subdivided into 24 subcategories based on parameters such as kernel function and model depth. Different kernel functions and model depths have varying degrees of adaptability. Parameter optimization and fitting were performed on all 24 models, as detailed in the table below.

[0075]

[0076]

[0077] S304. Based on 24 subdivided models, the models are trained by dividing them into training, testing, and validation sets to obtain training results. Based on the training results, the final prediction results of reservoir numerical simulation are obtained through model integration and weighted averaging.

[0078] The models trained previously with R-squared values ​​greater than a certain threshold are integrated into a single combined model. The weighted average of the results calculated by each model is used as the final predicted extraction level. The prediction results from the above models are weighted and averaged according to MSE, calculated as follows:

[0079]

[0080] In the formula: C represents the effective embedding coefficient of the final prediction; Ci represents the prediction result of the i-th model; MSEi represents the mean square error of the i-th model.

[0081] It should be noted that step S304 includes S3041 and S3042, wherein:

[0082] S3041. Through parameter optimization and fitting calculation, the dataset is divided into training set, test set and validation set in a ratio of 7:1.5:1.5, and the model is trained to obtain the training results of each model.

[0083] The processed dataset was divided into training, testing, and validation datasets in a 7:1.5:1.5 ratio to begin training the model. R-squared was used as the learning criterion to obtain the reservoir model training results.

[0084] Understandably, splitting the dataset into training, testing, and validation sets effectively evaluates the model's performance on unseen data, improves its generalization ability, and uses R-squared as the evaluation metric to select the best-performing model, thereby improving prediction accuracy. In practical applications, multiple machine learning models need to be trained and validated to determine the model best suited for specific reservoir data.

[0085] S3042. Based on the training results of each model, select the model with R-square greater than a certain value, and after weighted integration, obtain the final recovery degree prediction value. Then, according to the MSE, perform a weighted average of the prediction results of each model to obtain the final prediction result of reservoir numerical simulation and machine learning optimization.

[0086] Understandably, by selecting models with high R-squared values ​​and performing weighted ensemble, the accuracy of predictions can be improved. Using MSE for weighted averaging can balance the prediction errors of different models and further reduce the uncertainty of the overall prediction. This method allows for the comprehensive consideration of the prediction results of multiple models, thereby obtaining a more comprehensive and accurate reservoir dynamic prediction.

[0087] S400. Based on the effective carbon storage coefficient chart, establish a method for calculating the effective carbon storage of different types of reservoirs, so as to calculate the effective carbon storage and screen out favorable storage areas.

[0088] It should be noted that step S400 includes the formation of the effective storage coefficient chart under the influence of multiple factors, wherein the calculation formula for the effective storage amount of CO2 is as follows:

[0089] Effective CO2 storage capacity = Effective storage coefficient × Theoretical CO2 storage capacity

[0090]

[0091] Where: M t ρ represents the theoretical CO2 reserves in the reservoir. r The value represents the density of CO2 under reservoir conditions; ER represents the oil recovery rate; A represents the reservoir area; and h represents the reservoir thickness. Indicates reservoir porosity; S wi V represents the saturation of bound water in the reservoir. iw V represents the amount of water injected into the reservoir; pw Indicates the amount of water produced from the reservoir; C ws C represents the solubility coefficient of CO2 in water. os E represents the solubility coefficient of CO2 in oil. Rb E represents the oil recovery rate before CO2 levels were exceeded. Rh This indicates the oil recovery rate when a certain volume of CO2 is injected.

[0092] It is understood that in the steps of this embodiment, an effective carbon storage coefficient chart is obtained. This chart integrates data on CO2 storage efficiency under the influence of multiple factors. According to the characteristics of different types of reservoirs, the corresponding effective carbon storage coefficient is selected. The selected coefficient is applied and combined with the specific geological and production data of the reservoir to establish a calculation model for effective carbon storage. According to the calculation model, the effective carbon storage is quantitatively calculated for different reservoir areas, and favorable storage areas with high CO2 storage potential are evaluated and screened.

[0093] In other words, by using effective burial coefficient charts, the CO2 burial potential of different reservoir areas can be assessed more accurately, and quantitative calculation results help identify the most favorable burial areas, thereby improving the efficiency and economic benefits of CO2 burial. In practical applications, it is necessary to integrate geological, production, and geochemical data of the reservoir to construct accurate calculation models. When selecting favorable burial areas, geological risks, operational risks, and environmental risks also need to be considered. This method not only improves the scientific rigor and target orientation of CO2 burial but also provides a powerful decision support tool for reservoir management and CCS (carbon capture and storage) projects.

[0094] To provide a clearer understanding of the technical features, objectives, and effects of the present invention, specific embodiments are now described with reference to the accompanying drawings. In some embodiments, the following steps are described in detail:

[0095] Step 1: Three-dimensional geological models of three types of reservoirs, including light oil, heavy oil, and high-pour-point oil, were constructed respectively (the typical light oil reservoir was selected as block L, the heavy oil reservoir as block D, and the high-pour-point oil reservoir as block S; all abbreviations are used throughout the text). The distribution pattern of remaining oil in the three types of reservoirs was determined, providing a basis for subsequent exploration of CO2 flooding storage potential and optimization of effects.

[0096] Step 1.1: Based on the results of detailed geological research, a three-dimensional structural model and sedimentary facies model are constructed using geological modeling software. A facies-controlled attribute model is established. A saturation model and a net-to-gross ratio model are established in combination with well logging interpretation results and oil, gas and water distribution. The original geological reserves are then calculated.

[0097] Step 1.2: The reservoir's PVT phase parameters are fitted using numerical simulation software. Based on the reservoir geological model, the geological model is imported into the numerical simulation software. The state equations generated by the phase simulation, representing the actual fluid properties of the reservoir, are imported into the model to calculate mechanisms reflecting gas dissolution, expansion, and viscosity reduction in oil and water. Historical data fitting is then performed to study the distribution patterns of remaining oil in the reservoir.

[0098] Step 2: The influence of CO2 on oil recovery in three reservoirs (L, D, and S) under different reservoir conditions (permeability, pressure, and temperature) was studied. Analysis of the recovery rate results obtained from numerical simulations revealed the variation patterns of the contribution rates of various CO2 burial mechanisms under different reservoir conditions.

[0099] In the L-block of light oil, simulations of CO2 flooding under different permeability, temperature, and pressure conditions yielded the following results: 1. With increasing permeability, the contribution rate of dissolved burial in oil and water decreases, while the contribution rate of tectonic-bound burial increases, accounting for 50%–60% of the total burial volume; temperature has little effect on the contribution rate of different burial mechanisms; with increasing initial formation pressure, the contribution rate of dissolved burial in oil and water increases, while the contribution rate of tectonic-bound burial decreases. 2. The effective burial coefficient increases with increasing permeability and temperature; with increasing initial formation pressure, the effective burial coefficient first increases and then decreases, such as… Figure 1 As shown.

[0100] Heavy oil Block D: Simulations of CO2 flooding under different permeability, temperature, and pressure conditions yielded the following results: 1. With increasing permeability, the contribution rate of dissolved burial in the oil decreases, while the contribution rate of dissolved burial in water increases, and the contribution rate of tectonic burial initially increases and then decreases; with increasing temperature, the contribution rate of dissolved burial in the oil decreases, while the contributions of dissolved burial in water and tectonic-bound burial increase; the original formation pressure has little effect on the contribution rate of different burial mechanisms; 2. With increasing permeability, the effective burial coefficient decreases; with increasing temperature, the effective burial coefficient of CO2 increases significantly; when the original formation pressure exceeds a certain value, the effective burial coefficient decreases, such as... Figure 2 As shown.

[0101] High-pour-point oil block S: Simulations of CO2 flooding under different permeability, temperature, and pressure conditions yielded the following results: I. With increasing permeability, the contribution rates of dissolved burial in both oil and water decrease, while the contribution rate of tectonic burial increases. With increasing temperature, the contribution rate of dissolved burial in oil decreases, dissolved burial in water remains essentially unchanged, and the contribution rate of tectonic confinement burial increases. The original formation pressure has little effect on the contribution rates of different burial mechanisms. II. With increasing permeability, the effective burial coefficient increases. With increasing temperature, the effective burial coefficient of CO2 increases to some extent. When the original formation pressure exceeds a certain value, the effective burial coefficient decreases slightly, such as... Figure 3 As shown.

[0102] Step 3: Based on numerical simulation technology, establish a method for obtaining the effective storage coefficient of CO2. Combined with an artificial intelligence training model, generate an effective storage coefficient chart under multiple factors (permeability, reservoir pressure, temperature), which can be used for calculating the effective storage volume and screening favorable storage areas.

[0103] Step 3.1: Use the 21 sets of data obtained from numerical simulation of the L, D, and S blocks to train and continuously optimize the model, generating a surrogate model that meets the accuracy requirements. Figure 1 , 2 Figures 3 and 4 represent the proxy model verification graphs for the three blocks. The horizontal axis represents the actual value, and the vertical axis represents the predicted value. Points falling on the 45° line indicate complete accuracy.

[0104] The specific process is as follows:

[0105] (1) Model input parameter preprocessing

[0106]

[0107] (2) Evaluation indicators of model prediction performance

[0108]

[0109] (3) Machine learning model for single extraction degree

[0110] We employed 24 commonly used machine learning methods across 6 major categories to build parameter-optimized models, including major algorithms such as support vector regression, Gaussian process regression, tree ensemble, neural networks, linear regression, and regression trees. These algorithms can be further subdivided into 24 subcategories based on parameters such as kernel function and model depth. Different kernel functions and model depths exhibit varying degrees of adaptability. Parameter optimization and fitting were performed on all 24 models. The processed dataset was then split into training, testing, and validation datasets in a 7:1.5:1.5 ratio, and model training began using R-squared as the learning metric. The training results for the three blocks are shown in the table below.

[0111] The Gaussian process regression with a quadratic rational kernel and the Gaussian process regression with an exponential kernel have the best model fit for the L-block of light oil. Models with a fit greater than 0.8 include quadratic rational GPR, exponential GPR, and quadratic exponential GPR. It can be seen that Gaussian process regression has a stronger adaptability to the L-block of light oil.

[0112] Training results of a single model in the L-block of thin oil

[0113]

[0114]

[0115] The model with the highest good fit for the D block of heavy oil is the support vector machine with a cubic polynomial kernel function. Models with a good fit greater than 0.8 include: interaction effect linear regression, stepwise linear regression, quadratic SVM, cubic SVM, quadratic exponential GPR, Matern 5 / 2 GPR, and three-layer neural network. It can be seen that Gaussian process regression and support vector regression have stronger adaptability to the D block of heavy oil.

[0116] Training results of a single model for heavy oil block D

[0117]

[0118]

[0119]

[0120] The Gaussian process regression with the squared exponential kernel function had the highest model fit in the high-pour-point oil S block. Models with a fit greater than 0.8 included: quadratic rational GPR, squared exponential GPR, exponential GPR, narrow neural network, medium-sized neural network, wide neural network, and two-layer neural network. It can be seen that Gaussian process regression and neural network have stronger adaptability to the high-pour-point oil S block.

[0121] Training results of a single model in the S block of high-pour-point oil

[0122]

[0123]

[0124] (4) Multiple extraction degree machine learning models

[0125] Integrate the previously trained models with R-squared greater than 0.8 into a combined model, such as... Figure 4 As shown, the weighted average calculated by each model is used as the final predicted value of extraction degree.

[0126] The prediction results of the above model are weighted and averaged according to MSE, and the calculation method is as follows:

[0127]

[0128] The results show that the R-square of the three-block combination model is 0.95 (L block), 0.92 (D block), and 0.93 (S block), calculated as follows. This is a significant improvement over the highest value of the single model, with the final model errors being 5% (L block), 8% (D block), and 7% (S block).

[0129]

[0130] Step 3.2: Establish effective burial coefficient charts for blocks L, D, and S, as follows: Figure 5-16As shown: 1. Calculation results of effective burial coefficient of thin oil block L: The effective burial coefficient of thin oil block L under different temperatures, pressures and permeabilities was obtained. The effective burial volume of the model has favorable burial areas under different pressure conditions, which can provide a basis for the selection of burial areas. As the pressure increases, the effective burial coefficient is greatly improved.

[0131] 2. Calculation results of effective burial coefficient of heavy oil block D: The effective burial coefficient of heavy oil block D under different temperatures, pressures and permeabilities was obtained. The effective burial volume of the model has favorable burial areas under different pressure conditions, which can provide a basis for the selection of burial areas. The effective burial coefficient of heavy oil block D is very sensitive to temperature. The higher the temperature, the larger the effective burial coefficient. As the pressure increases, the effective burial coefficient decreases.

[0132] 3. Calculation results of effective burial coefficient of high pour point oil S block: The effective burial coefficient of high pour point oil S block under different temperatures, pressures and permeabilities was obtained. The effective burial volume of the model has favorable burial areas under different pressure conditions, which can provide a basis for the selection of burial areas. The effective burial coefficient of high pour point oil S block increases with the increase of pressure.

[0133] Step 4: Establishment of calculation methods for effective CO2 storage capacity of different types of oil reservoirs.

[0134] Effective CO2 storage capacity = Effective storage coefficient × Theoretical CO2 storage capacity

[0135] Theoretical CO2 reserves:

[0136]

[0137] The calculation method for the L-block of thin oil is as follows:

[0138] Relevant parameters of the L block of thin oil reservoir

[0139] Original formation pressure of the oil reservoir 30.325MPa average permeability of oil reservoir 26.53mD reservoir temperature 83.1℃ Pore ​​volume <![CDATA[6108620m 3 ]]> Bound water saturation 0.3 <![CDATA[Recovery factor before CO2 breakthrough]]> 13.63％ <![CDATA[Recovery factor after CO2 breakthrough]]> 33.4％ Buried pressure limit To the original formation pressure

[0140] Theoretical CO2 reserves = 709.88 × [(0.4 × 13.63% + 0.6 × 33.4%) × 6108620 × 0.7 + 0.09031 × 6108620 × 0.3 + 0.6794 × (1 - 0.4 × 13.63% - 0.6 × 33.4%) × 6108620 × (1 - 0.7)] / 10,000,000 = 1,550,000 tons.

[0141] According to the map, the effective storage coefficient is 0.472. Therefore, the final CO2 storage volume is 155 × 0.62 = 961,000 tons, which is 2.4% different from the numerical simulation result of 984,700 tons.

[0142] The calculation method for heavy oil block D is as follows:

[0143] Heavy oil block D reservoir related parameters

[0144] Original formation pressure of the oil reservoir 13.5MPa average permeability of oil reservoir 889mD reservoir temperature 69℃ Pore ​​volume <![CDATA[2311201m 3 ]]> Bound water saturation 0.3 <![CDATA[Recovery factor before CO2 breakthrough]]> 13.54％ <![CDATA[Recovery factor after CO2 breakthrough]]> 23.13％ Buried pressure 17MPa

[0145] Theoretical CO2 reserves = 554.40 × [(0.4 × 13.53% + 0.6 × 23.13%) × 2311201 × 0.7 + 0.1455 × 2311201 × 0.3 + 0.368525 × (1 - 0.4 × 13.53% - 0.6 × 23.13%) × 2311201 × (1 - 0.7)] / 10000000 = 343,280 tons.

[0146] According to the map, the effective burial coefficient is 0.484. Therefore, the final CO2 burial volume is 34328 × 0.49 = 165,000 tons, which is 2.9% different from the numerical simulation result of 170,000 tons.

[0147] The calculation method for the high-pour-point oil S block is as follows:

[0148] Relevant parameters of the high-pour-point-oil S block reservoir

[0149]

[0150]

[0151] Theoretical CO2 reserves = 647.04 × [(0.4 × 12.4% + 0.6 × 15%) × 17,625,970 × 0.7 + 0.08457 × 17,625,970 × 0.3 + 0.339726 × (1 - 0.4 × 12.4% - 0.6 × 15%) × 17,625,970 × (1 - 0.7)] / 10,000,000 = 1,535,000 tons

[0152] According to the map, the effective burial coefficient is 0.472. Therefore, the final burial amount of CO2 is 153.5 × 0.49 = 724,500 tons, which is 4.6% different from the numerical simulation result of 691,500 tons.

[0153] In summary, this invention is used to conduct differentiated design research on CO2 storage methods for different types of oil reservoirs, clarifying the various storage mechanisms and dynamic changes during CO2 flooding storage in different types of oil reservoirs; it is used for numerical simulation of CO2 carbon flooding storage under different reservoir conditions (permeability, reservoir pressure, temperature) for different types of oil reservoirs, establishing a method for calculating the effective CO2 storage coefficient, and combining it with an artificial intelligence training model to form an effective storage coefficient chart under multiple factors (permeability, reservoir pressure, temperature), quantifying the final storage capacity; this invention is used to identify the key factors affecting CO2 storage for various types of oil reservoirs, establish storage calculation methods for different types of oil reservoirs, and can be used for effective storage calculation and selection of favorable storage areas, providing theoretical guidance for verifying the carbon pool potential of oilfields.

[0154] This embodiment provides an artificial intelligence-based device for calculating the CO2 sequestration capacity of geological bodies, such as... Figure 18 As shown, the device includes:

[0155] Module: Used to build three-dimensional geological models of different types of reservoirs to determine the distribution patterns of remaining oil in different types of reservoirs;

[0156] Determine the module: Based on the distribution pattern of remaining oil in different types of reservoirs, analyze the influence of different reservoir conditions on CO2 burial in different types of reservoirs, and then determine the variation of the contribution rate of various mechanisms under different conditions, including permeability, reservoir pressure and temperature;

[0157] Training module: This module combines numerical simulation techniques with the variation law of mechanism contribution rate to develop a method for calculating the effective CO2 storage coefficient, and uses artificial intelligence to train the model to generate an effective storage coefficient chart under the influence of multiple factors.

[0158] Calculation module: Used for effective carbon storage coefficient charts, establishing calculation methods for effective carbon storage in different types of reservoirs, in order to calculate effective carbon storage and screen out favorable storage areas.

[0159] Furthermore, the building module includes:

[0160] Establishment Unit: Used to construct three-dimensional structural and sedimentary facies models using geological modeling software and geological history research data, and to establish attribute models. Combined with well logging interpretation results and oil, gas and water distribution data, saturation and net-to-gross ratio models are formed, and original geological reserves are calculated.

[0161] Fitting Unit: Used to fit the PVT phase parameters of the reservoir using numerical simulation software, and based on the reservoir geological model, import phase simulation data and state equations, and obtain the distribution pattern of the remaining oil in the reservoir through historical fitting calculations.

[0162] Furthermore, the training module includes:

[0163] Processing unit: Used to obtain numerical simulation datasets for different types of reservoirs. Based on the mean and standard deviation of the original data, it is standardized by z-score to obtain a standardized dataset.

[0164] Determining the unit: This is used to determine the preliminary optimized machine learning model based on a standardized dataset and after training with a machine learning model, using R-squared, MSE, and RMSE as evaluation metrics.

[0165] Subdivision Unit: Based on the initially optimized machine learning model, after parameter adjustment and kernel function selection, it is used to subdivide six types of machine learning algorithms into 24 subdivision models;

[0166] Unit partitioning: Based on 24 subdivision models, the model is trained by dividing it into training, test and validation sets, and the training results are obtained. Based on the training results, the final prediction results of reservoir numerical simulation are obtained through model integration and weighted averaging.

[0167] Furthermore, the subdivision unit includes:

[0168] Optimization unit: Used to divide the dataset into training, test and validation sets in a ratio of 7:1.5:1.5 through parameter optimization and fitting calculation, and to train the model to obtain the training results of each model;

[0169] Prediction Unit: Based on the training results of each model, selects the model with an R-square greater than a certain value, performs weighted integration to obtain the final recovery degree prediction value, and then performs a weighted average of the prediction results of each model according to the MSE to obtain the final prediction result of reservoir numerical simulation and machine learning optimization.

[0170] It should be noted that the specific manner in which each module performs its operation in the apparatus described in the above embodiments has been described in detail in the embodiments of the method, and will not be elaborated here.

[0171] Corresponding to the above method embodiments, this embodiment also provides an artificial intelligence-based geological body CO2 sequestration capacity calculation device. The artificial intelligence-based geological body CO2 sequestration capacity calculation device described below and the artificial intelligence-based geological body CO2 sequestration capacity calculation method described above can be referred to in correspondence.

[0172] Figure 19 This is a block diagram illustrating an artificial intelligence-based CO2 sequestration capacity calculation device 800 for geological bodies, according to an exemplary embodiment. Figure 19As shown, the AI-based geological body CO2 sequestration capacity calculation device 800 includes a processor 801 and a memory 802. The AI-based geological body CO2 sequestration capacity calculation device 800 also includes one or more of a multimedia component 803, an I / O interface 804, and a communication component 805.

[0173] The processor 801 controls the overall operation of the AI-based geological body CO2 sequestration capacity calculation device 800 to complete all or part of the steps in the aforementioned AI-based geological body CO2 sequestration capacity calculation method. The memory 802 stores various types of data to support the operation of the AI-based geological body CO2 sequestration capacity calculation device 800. This data may include, for example, instructions for any application or method operating on the AI-based geological body CO2 sequestration capacity calculation device 800, as well as application-related data such as contact data, sent and received messages, images, audio, video, etc. The memory 802 can be implemented using any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The multimedia component 803 may include a screen and an audio component. The screen may be, for example, a touchscreen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signals may be further stored in the memory 802 or transmitted via the communication component 805. The audio component also includes at least one speaker for outputting audio signals. I / O interface 804 provides an interface between processor 801 and other interface modules, such as a keyboard, mouse, or buttons. These buttons can be virtual or physical. Communication component 805 is used for wired or wireless communication between the AI-based geological CO2 storage capacity computing device 800 and other devices. Wireless communication includes, for example, Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G, or 4G, or a combination thereof. Therefore, the corresponding communication component 805 may include a Wi-Fi module, a Bluetooth module, or an NFC module.

[0174] In an exemplary embodiment, the AI-based geological body CO2 sequestration capacity calculation device 800 may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to execute the aforementioned AI-based geological body CO2 sequestration capacity calculation method.

[0175] In another exemplary embodiment, a computer-readable storage medium including program instructions is also provided. When executed by a processor, these program instructions implement the steps of the above-described AI-based method for calculating the CO2 sequestration capacity of geological bodies. For example, the computer-readable storage medium may be the memory 802 including the program instructions, which may be executed by the processor 801 of the AI-based geological body CO2 sequestration capacity calculation device 800 to complete the above-described AI-based method for calculating the CO2 sequestration capacity of geological bodies.

[0176] Corresponding to the above method embodiments, this embodiment also provides a readable storage medium. The readable storage medium described below can be referred to in conjunction with the above-described method for calculating the CO2 storage capacity of geological bodies based on artificial intelligence.

[0177] A computer program is stored on a readable storage medium, and when the computer program is executed by a processor, it implements the steps of the artificial intelligence-based method for calculating the CO2 sequestration capacity of geological bodies in the above-described method embodiments.

[0178] In summary, this invention, during the logging process of tight sandstone gas reservoirs, comprehensively applies conventional gas logging techniques combined with new technologies such as rock and mineral scanning and elemental analysis to form a "five-property" evaluation method for tight sandstone gas reservoirs. This method enables the selection of superior reservoirs from heterogeneous ones, providing support for high and stable production in horizontal wells. High-precision electron microscopy allows for observation of minerals and pore fractures at the micro- and nano-scale, improving geological understanding and solving the problem of distorted or missing logging data caused by complex conditions. Furthermore, it enables analysis while drilling, saving logging time after completion, improving construction efficiency, and reducing engineering risks. This achievement has been applied in over 20 wells, with 18 wells achieving unobstructed flow rates exceeding 50,000 cubic meters per second.

[0179] The readable storage medium can specifically be a USB flash drive, external hard drive, read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk, or any other readable storage medium capable of storing program code.

[0180] The above description is merely an embodiment of this application and is not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of the claims of this application.

Claims

1. A method for calculating the CO2 sequestration capacity of geological bodies based on artificial intelligence, characterized in that, include: Three-dimensional geological models of different types of reservoirs were constructed to determine the distribution patterns of remaining oil in different types of reservoirs; Based on the distribution patterns of remaining oil in different types of reservoirs, the influence of different reservoir conditions on CO2 burial is analyzed, and the variation patterns of the contribution rates of various mechanisms under different conditions are determined. The different reservoir conditions include permeability, reservoir pressure and temperature. Combining numerical simulation technology with the variation law of mechanism contribution rate, a method for calculating CO2 effective storage coefficient was developed, and an effective storage coefficient chart under the influence of multiple factors was generated by training the model with artificial intelligence. Based on the effective carbon storage coefficient chart, a method for calculating the effective carbon storage of different types of reservoirs is established to calculate the effective carbon storage and screen out favorable storage areas.

2. The method for calculating the CO2 sequestration capacity of geological bodies based on artificial intelligence according to claim 1, characterized in that, The construction of three-dimensional geological models for different types of oil reservoirs is intended to determine the distribution patterns of remaining oil in these reservoirs, including: Using geological modeling software and geological history research data, a three-dimensional structural and sedimentary facies model was constructed, and an attribute model was established. Combined with well logging interpretation results and oil, gas and water distribution data, a saturation and net-to-gross ratio model was formed, and the original geological reserves were calculated. Numerical simulation software was used to fit the PVT phase parameters of the reservoir. Based on the reservoir geological model, phase simulation data and state equations were imported, and the distribution pattern of the remaining oil in the reservoir was obtained through historical fitting calculations.

3. The method for calculating the CO2 sequestration capacity of geological bodies based on artificial intelligence according to claim 1, characterized in that, The method for determining the effective CO2 storage coefficient is described above, and the calculation formula for the effective CO2 storage coefficient is as follows: Effective storage coefficient = Actual CO2 storage volume / Theoretical CO2 storage volume In the formula, the actual CO2 storage amount = the amount of CO2 injected - the amount of CO2 produced, and the theoretical CO2 storage amount = the amount of CO2 stored in the pore space + the amount of CO2 dissolved in water + the amount of CO2 dissolved in oil.

4. The method for calculating the CO2 sequestration capacity of geological bodies based on artificial intelligence according to claim 1, characterized in that, The model trained using artificial intelligence includes: Numerical simulation datasets of different types of reservoirs were obtained, and standardized datasets were obtained by z-score standardization based on the mean and standard deviation of the original data. Based on a standardized dataset, a machine learning model was trained, and R-squared, MSE, and RMSE were determined as evaluation metrics to obtain a preliminary optimized machine learning model. Based on the preliminary optimized machine learning model, after parameter adjustment and kernel function selection, the six types of machine learning algorithms were further subdivided to obtain 24 subdivided models; Based on 24 subdivided models, the models were trained by dividing them into training, testing, and validation sets. The training results were obtained, and the final prediction results of reservoir numerical simulation were obtained by integrating the models and weighting the average based on the training results.

5. The method for calculating the CO2 sequestration capacity of geological bodies based on artificial intelligence according to claim 4, characterized in that, The six types of machine learning algorithms include support vector regression, Gaussian process regression, tree ensemble, neural networks, linear regression, and regression trees.

6. The method for calculating the CO2 sequestration capacity of geological bodies based on artificial intelligence according to claim 4, characterized in that, The model is trained by dividing the dataset into training, testing, and validation sets to obtain training results. Based on these results, the final prediction results of the reservoir numerical simulation are obtained through model ensemble and weighted averaging, including: Through parameter optimization and fitting calculation, the dataset was divided into training set, test set and validation set in a ratio of 7:1.5:1.5, and the model was trained to obtain the training results of each model. Based on the training results of each model, the model with an R-square greater than a certain value is selected, and after weighted integration, the final recovery degree prediction value is obtained. Then, the prediction results of each model are weighted and averaged according to MSE to obtain the final prediction result of reservoir numerical simulation and machine learning optimization.

7. The method for calculating the CO2 sequestration capacity of geological bodies based on artificial intelligence according to claim 1, characterized in that, The effective storage coefficient chart formed under the influence of multiple factors includes the following formula for calculating the effective storage amount of CO2: Effective CO2 storage capacity = Effective storage coefficient × Theoretical CO2 storage capacity Where: M t ρ represents the theoretical CO2 reserves in the reservoir. r The value represents the density of CO2 under reservoir conditions; ER represents the oil recovery rate; A represents the reservoir area; and h represents the reservoir thickness. S represents reservoir porosity; wi V represents the saturation of bound water in the reservoir. iw V represents the amount of water injected into the reservoir; pw Indicates the amount of water produced from the reservoir; C ws C represents the solubility coefficient of CO2 in water. os E represents the solubility coefficient of CO2 in oil. Rb E represents the oil recovery rate before CO2 levels were exceeded. Rh This indicates the oil recovery rate when a certain volume of CO2 is injected.

8. A calculation device for the CO2 sequestration capacity of geological bodies based on artificial intelligence, characterized in that, include: Module: Used to build three-dimensional geological models of different types of reservoirs to determine the distribution patterns of remaining oil in different types of reservoirs; The determination module is used to analyze the influence of different reservoir conditions on CO2 burial in different types of reservoirs based on the distribution patterns of remaining oil in different types of reservoirs, and then determine the variation patterns of the contribution rates of various mechanisms under different conditions, including permeability, reservoir pressure and temperature. Training module: This module combines numerical simulation techniques with the variation law of mechanism contribution rate to develop a method for calculating the effective CO2 storage coefficient, and uses artificial intelligence to train the model to generate an effective storage coefficient chart under the influence of multiple factors. Calculation module: Used for effective carbon storage coefficient charts, establishing calculation methods for effective carbon storage in different types of reservoirs, in order to calculate effective carbon storage and screen out favorable storage areas.

9. The AI-based geological body CO2 sequestration capacity calculation device according to claim 8, characterized in that, The building module includes: Establishment Unit: Used to construct three-dimensional structural and sedimentary facies models using geological modeling software and geological history research data, and to establish attribute models. Combined with well logging interpretation results and oil, gas and water distribution data, saturation and net-to-gross ratio models are formed, and original geological reserves are calculated. Fitting Unit: Used to fit the PVT phase parameters of the reservoir using numerical simulation software, and based on the reservoir geological model, import phase simulation data and state equations, and obtain the distribution pattern of the remaining oil in the reservoir through historical fitting calculations.

10. The AI-based geological body CO2 sequestration capacity calculation device according to claim 8, characterized in that, The training module includes: Processing unit: Used to obtain numerical simulation datasets for different types of reservoirs. Based on the mean and standard deviation of the original data, it is standardized by z-score to obtain a standardized dataset. Determining the unit: This is used to determine the preliminary optimized machine learning model based on a standardized dataset and after training with a machine learning model, using R-squared, MSE, and RMSE as evaluation metrics. Subdivision Unit: Based on the initially optimized machine learning model, after parameter adjustment and kernel function selection, it is used to subdivide six types of machine learning algorithms into 24 subdivision models; Unit partitioning: Based on 24 subdivision models, the model is trained by dividing it into training, test and validation sets, and the training results are obtained. Based on the training results, the final prediction results of reservoir numerical simulation are obtained through model integration and weighted averaging.

Images