A method and system for evaluating carbon dioxide sequestration capacity, and a computer-readable medium

Abstract

The embodiments of the present application provide a method and system for evaluating the carbon dioxide storage capacity, and a computer-readable medium, which can solve the technical problem of poor evaluation ability of the carbon dioxide storage capacity. The method includes constructing a structured geological grid model through multi-source data fusion, integrating target geological parameters, geochemical characteristics and engineering constraints, and generating a high-resolution three-dimensional attribute field. Relying on the dynamic numerical simulation of multiphase flow-chemistry-mechanics, combining the preset pressure safety threshold and the plume constraint range, eliminating the high-risk areas of overpressure and boundary crossing, generating a probabilistic leakage risk map, and realizing the safe conversion from the theoretical capacity to the engineering effective capacity. Taking the effective capacity and the risk heat map as inputs, solving the injection well position coordinates, the rate time schedule and the pressure monitoring threshold, to ensure the dynamic balance of maximizing the injection efficiency and minimizing the risk. Based on the real-time monitoring data, generating control instructions to form a closed loop of "monitoring-simulation-decision-making-feedback", and improving the evaluation ability.

Description

Technical Field

[0001] This application relates to the technical field of carbon dioxide storage, and in particular, to a method and system for evaluating carbon dioxide storage capacity, and a computer-readable medium. Background Art

[0002] Carbon Capture and Storage (CCS) captures carbon dioxide emitted from industries and, after safe transportation, stores the carbon dioxide underground in geological formations or deep sea for a long time. To ensure the safety and effectiveness of this technology, Carbon Storage Capacity Assessment (CSCA) is a core link. CSCA quantifies the carbon dioxide ( ) storage potential and risk parameters of a specific geological formation or technical solution, providing a scientific decision-making basis for large-scale emission reduction.

[0003] Currently, the maximum potential storage capacity (i.e., the theoretical value of the storage capacity) is calculated based on the volume and porosity of the geological formation. When determining the actual storage capacity, considering geological sealing, injection efficiency, and regulatory restrictions, the actual storage capacity is usually set at 10% - 40% of the theoretical value of the storage capacity.

[0004] However, in the evaluation of carbon dioxide storage capacity, the method of estimating the actual storage capacity based on the theoretical value at a fixed ratio (10% - 40%) oversimplifies the dynamic geological process and the coupling effect of multiple factors, resulting in a significant deviation risk in the evaluation result of the storage capacity. Summary of the Invention

[0005] Embodiments of this application provide a method and system for evaluating carbon dioxide storage capacity, and a computer-readable medium, which can solve the technical problem of poor carbon dioxide storage capacity evaluation ability.

[0006] To achieve the above object, the embodiments of this application adopt the following technical solutions:

[0007] In a first aspect, an embodiment of the present application provides a method for evaluating carbon dioxide sequestration capacity. The method for evaluating carbon dioxide sequestration capacity includes: obtaining a structured geological grid model based on training data; the training data includes target geological data, target geochemical data, and target engineering parameters; the structured geological grid model includes a three-dimensional grid structure, and its property fields include porosity, permeability, and saturation; inputting the structured geological grid model into a preset simulator, and obtaining a dynamic simulation result based on initial temperature and pressure conditions, carbon dioxide injection rate, and relative permeability curve; calculating an effective capacity and a risk probability map according to the dynamic simulation result, based on a preset pressure safety threshold and a preset plume range; obtaining well location coordinates, an injection rate schedule, and a pressure monitoring threshold through a preset well pattern value, a preset injection rate, and a preset pressure control threshold according to the effective capacity and the risk probability map; obtaining a real-time monitoring data set based on the well location coordinates; the real-time monitoring data set includes wellhead pressure, wellhead temperature, fiber optic time series data, and four-dimensional seismic difference volume; inputting the real-time monitoring data set into a machine learning model to generate an injection rate adjustment instruction and a leakage risk warning; updating dynamic numerical simulation parameters according to the injection rate adjustment instruction and the leakage risk warning, based on the well location coordinates, the injection rate schedule, and the pressure monitoring threshold, and calculating the carbon dioxide distribution and pressure evolution result.

[0008] Based on the above description of the method for evaluating carbon dioxide sequestration capacity provided by the embodiment of the present application, it can be seen that the method for evaluating carbon dioxide sequestration capacity includes constructing a structured geological grid model through multi-source data fusion, integrating target geological parameters (porosity, permeability), geochemical characteristics (mineral reaction activity), and engineering constraints (well pattern parameters), generating a high-resolution three-dimensional property field, breaking through the limitations of traditional homogenization models, accurately depicting reservoir heterogeneity, and laying a data foundation for multi-physics field coupling simulation at the millimeter-level grid scale. On this basis, relying on multi-phase flow-chemistry-mechanics dynamic numerical simulation, quantify The spatio-temporal expansion law of the plume, the propagation boundary of the pressure field, and the mineralization and storage potential are combined with the preset pressure safety threshold (0.9 times the fracture pressure) and the plume constraint range to eliminate the overpressure and high-risk areas beyond the boundary, generate a probabilistic leakage risk map, and achieve the safe conversion from the theoretical capacity to the engineering effective capacity. Further, through the well pattern optimization algorithm, with the effective capacity and the risk heat map as inputs, the optimal solutions of the injection well location coordinates, the rate schedule, and the pressure monitoring threshold are solved to ensure the dynamic balance of maximizing the injection efficiency and minimizing the risk. Finally, based on the real-time monitoring data (wellhead pressure, fiber optic strain, four-dimensional seismic), the model is used to identify abnormal modes and generate control instructions. By modifying the geological model parameters, a "monitoring - simulation - decision - feedback" closed loop is formed, enabling the system to have self-adaptability to the working conditions. Through the full-chain innovation of "data-driven modeling, dynamic safety assessment, and intelligent optimization and control", this method reduces the evaluation error of the carbon storage volume from ±50% to ±15%, improves the leakage warning response speed to the minute level, and increases the injection efficiency by 40%. Ultimately, it realizes the technological leap of carbon storage projects from rough estimation to precise control, providing high-reliability decision-making support for large-scale deployment.

[0009] In a feasible implementation manner of the first aspect, when performing the step of inputting the structured geological grid model into the preset simulator and obtaining the dynamic simulation results based on the initial temperature and pressure conditions, the carbon dioxide storage volume evaluation method further includes: obtaining the multiphase flow results based on the porosity and permeability according to the state equation and the chemical reaction model; the multiphase flow results include the phase behavior of carbon dioxide and liquid brine; and obtaining the dynamic simulation results according to the multiphase flow results.

[0010] In a feasible implementation manner of the first aspect, the carbon dioxide storage volume evaluation method further includes: calculating the theoretical capacity; confirming the target potential area according to the theoretical capacity; after performing the step of obtaining the well location coordinates, the injection rate schedule, and the pressure monitoring threshold through the preset well pattern value, the preset injection rate, and the preset pressure control threshold based on the effective capacity and the risk probability map, the carbon dioxide storage volume evaluation method further includes: confirming the overlapping coordinate value of the target potential area and the well location coordinates; and updating the well location coordinates based on the overlapping coordinate value.

[0011] Among them, the calculation formula of the theoretical capacity includes:

[0012] ;

[0013] Among them, A represents the reservoir area, with the unit of km²; h represents the effective thickness, with the unit of m; represents the average porosity, with the unit of %; represents the density of carbon dioxide; E represents the storage efficiency factor.

[0014] In a feasible implementation of the first aspect, before performing the step of training a three-dimensional geological model based on training data to obtain a structured geological grid model, the carbon dioxide sequestration capacity assessment method further includes: preprocessing the training data through Kriging interpolation and wavelet transform.

[0015] In a feasible implementation of the first aspect, when performing the step of inputting a real-time monitoring data set into a machine learning model to generate an injection rate adjustment instruction and a leakage risk warning, the carbon dioxide sequestration capacity assessment method further includes: inputting sensor data, fiber optic values, four-dimensional seismic data, and surface deformation values into the machine learning model to obtain a real-time plume map, a leakage risk warning result, and a policy dynamic adjustment result; generating an injection rate adjustment instruction and a leakage risk warning based on the real-time plume map, the leakage risk warning result, and the policy dynamic adjustment result.

[0016] In a feasible implementation of the first aspect, the calculation formula for the effective capacity includes:

[0017] ;

[0018] where V safe represents the safe sequestration volume that satisfies the pressure constraint p ≤ 0.9p frac ; represents the porosity; represents the carbon dioxide density; S g represents the carbon dioxide phase saturation; E dynamic represents the dynamic efficiency factor.

[0019] In a feasible implementation of the first aspect, the calculation formula for the risk probability map includes:

[0020] ;

[0021] where λ i represents the failure rate of the i-th type of risk event, in units of times / year; t represents the sequestration time, in units of years; N represents the total number of types of risk events.

[0022] In a feasible implementation of the first aspect, the calculation formula for the injection rate schedule includes:

[0023] ;

[0024] where Q j represents the injection rate of the j-th well, in units of kg / s; Q total represents the total injection amount constraint; P j represents the wellhead pressure of the j-th well; p frac represents the formation fracture pressure; ΔQ jis denoted as the rate adjustment step size; M is denoted as the total number of wells.

[0025] In a second aspect, an embodiment of the present application provides a carbon dioxide sequestration capacity evaluation system, which includes: at least one processor; a memory communicatively connected to the at least one processor; wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the method provided in the first aspect.

[0026] By executing the method provided in the first aspect, the carbon dioxide sequestration capacity evaluation system constructs a structured geological grid model through multi-source data fusion, integrates target geological parameters (porosity, permeability), geochemical characteristics (mineral reaction activity), and engineering constraints (well pattern parameters), generates a high-resolution three-dimensional attribute field, breaks through the limitations of traditional homogenization models, accurately depicts reservoir heterogeneity, and lays a data foundation for multi-physics field coupling simulation at the millimeter-scale grid. On this basis, relying on multi-phase flow-chemistry-mechanics dynamic numerical simulation, quantify the spatio-temporal expansion law of the plume, the propagation boundary of the pressure field, and the mineralization sequestration potential, combine the preset pressure safety threshold (0.9 times the fracture pressure) with the plume constraint range, eliminate overpressure and out-of-bounds high-risk areas, generate a probabilistic leakage risk map, and realize the safe conversion from theoretical capacity to engineering effective capacity. Further, through the well pattern optimization algorithm, with the effective capacity and the risk heat map as inputs, solve the multi-objective Pareto optimal solutions of the injection well location coordinates, rate schedule, and pressure monitoring threshold, ensuring the dynamic balance of maximizing injection efficiency and minimizing risk. Finally, based on real-time monitoring data (wellhead pressure, fiber optic strain, four-dimensional seismic), use the LSTM-CNN hybrid model to identify abnormal modes and generate control instructions, and dynamically correct the geological model parameters through ensemble Kalman filtering, forming a "monitoring-simulation-decision-feedback" closed loop, enabling the system to have working condition self-adaptability. Through the full-chain innovation of "data-driven modeling, dynamic safety assessment, and intelligent optimization control", this method reduces the sequestration capacity evaluation error from ±50% to ±15%, improves the leakage warning response speed to the minute level, increases the injection efficiency by 40%, and finally realizes the technical leap of carbon sequestration projects from rough estimation to precise control, providing high-reliability decision-making support for large-scale deployment.

[0027] In a third aspect, an embodiment of the present application provides a computer-readable medium, on which computer program instructions are stored, and the computer program instructions can be executed by a processor to implement the method provided in the first aspect.

[0028] Computer program instructions in a computer-readable medium provide a method by implementing the first aspect. A structured geological grid model is constructed through multi-source data fusion, integrating target geological parameters (porosity, permeability), geochemical characteristics (mineral reaction activity), and engineering constraints (well pattern parameters), generating a high-resolution three-dimensional attribute field, breaking through the limitations of traditional homogenization models, accurately depicting reservoir heterogeneity, and laying a data foundation for multi-physics field coupling simulation at the millimeter-scale grid. On this basis, relying on multi-phase flow-chemistry-mechanics dynamic numerical simulation, the spatio-temporal expansion law of the plume, the propagation boundary of the pressure field, and the mineralization and storage potential are quantified. Combining the preset pressure safety threshold (0.9 times the fracture pressure) and the plume constraint range, high-risk areas with overpressure and out-of-bounds are eliminated, and a probabilistic leakage risk map is generated to achieve a safe conversion from theoretical capacity to engineering effective capacity. Further, through a well pattern optimization algorithm, with the effective capacity and risk heat map as inputs, a multi-objective Pareto optimal solution for the injection well location coordinates, rate schedule, and pressure monitoring threshold is solved to ensure a dynamic balance between maximizing injection efficiency and minimizing risk. Finally, based on real-time monitoring data (wellhead pressure, fiber optic strain, four-dimensional seismic), an LSTM-CNN hybrid model is used to identify abnormal modes and generate control instructions, and the geological model parameters are dynamically corrected through ensemble Kalman filtering to form a "monitoring-simulation-decision-feedback" closed loop, enabling the system to have working condition self-adaptability. Through the full-chain innovation of "data-driven modeling, dynamic safety assessment, and intelligent optimization and control", this method reduces the sequestration volume assessment error from ±50% to ±15%, improves the leakage warning response speed to the minute level, and increases the injection efficiency by 40%, ultimately achieving a technological leap from extensive estimation to precise control in carbon sequestration projects and providing high-reliability decision-making support for large-scale deployment. The spatio-temporal expansion law of the plume, the propagation boundary of the pressure field, and the mineralization and storage potential are quantified. Combining the preset pressure safety threshold (0.9 times the fracture pressure) and the plume constraint range, high-risk areas with overpressure and out-of-bounds are eliminated, and a probabilistic leakage risk map is generated to achieve a safe conversion from theoretical capacity to engineering effective capacity. Further, through a well pattern optimization algorithm, with the effective capacity and risk heat map as inputs, a multi-objective Pareto optimal solution for the injection well location coordinates, rate schedule, and pressure monitoring threshold is solved to ensure a dynamic balance between maximizing injection efficiency and minimizing risk. Finally, based on real-time monitoring data (wellhead pressure, fiber optic strain, four-dimensional seismic), an LSTM-CNN hybrid model is used to identify abnormal modes and generate control instructions, and the geological model parameters are dynamically corrected through ensemble Kalman filtering to form a "monitoring-simulation-decision-feedback" closed loop, enabling the system to have working condition self-adaptability. Through the full-chain innovation of "data-driven modeling, dynamic safety assessment, and intelligent optimization and control", this method reduces the sequestration volume assessment error from ±50% to ±15%, improves the leakage warning response speed to the minute level, and increases the injection efficiency by 40%, ultimately achieving a technological leap from extensive estimation to precise control in carbon sequestration projects and providing high-reliability decision-making support for large-scale deployment. Brief Description of the Drawings

[0029] Figure 1 It is a schematic structural diagram of a carbon dioxide sequestration volume assessment system provided by an embodiment of the present application;

[0030] Figure 2 It is a schematic flow diagram of a carbon dioxide sequestration volume assessment method provided by an embodiment of the present application. Detailed Embodiments

[0031] Next, the technical solutions in the embodiments of the present invention will be described with reference to the accompanying drawings in the embodiments of the present invention. Among them, in the description of the embodiments of the present invention, unless otherwise specified, "a plurality" means two or more than two. "At least one (item)" or its similar expression refers to any combination of these items, including any combination of single item (item) or multiple items (items). For example, at least one (item) of a, b, or c can represent: a, b, c, a - b, a - c, b - c, or a - b - c, where a, b, c can be single or multiple.

[0032] In addition, for the convenience of clearly describing the technical solutions of the embodiments of the present invention, in the embodiments of the present invention, terms such as "first" and "second" are used to distinguish identical or similar items with basically the same functions and effects. Those skilled in the art can understand that terms such as "first" and "second" do not limit the quantity and execution order, and "first", "second", etc. do not necessarily mean different. At the same time, in the embodiments of the present invention, words such as "exemplary" or "for example" are used to represent examples, illustrations or explanations. Any embodiment or design solution described as "exemplary" or "for example" in the embodiments of the present invention should not be construed as being more preferred or having more advantages than other embodiments or design solutions. Exactly speaking, using words such as "exemplary" or "for example" aims to present relevant concepts in a specific way for easy understanding.

[0033] The principles and features of the present application are described below, and the examples given are only used to explain the present application and are not used to limit the scope of the present application.

[0034] In the assessment of carbon dioxide sequestration capacity, if there are significant deviations in the assessment results of the sequestration capacity, it may lead to derivative risks. Overestimating the sequestration capacity causes the calculation of the project investment return period to be distorted (for example, the preset sequestration period is 20 years, but it actually only takes 5 years to reach saturation). Underestimating the migration range may miss detecting leakage paths (for example, unrecognized microfaults become migration channels). All of these will make it difficult to achieve the carbon quota allocation or emission reduction targets formulated based on the incorrect sequestration capacity.

[0035] For example, in a certain carbon dioxide sequestration project, the theoretical sequestration capacity is 150 million tons (based on the volume and porosity of the saline aquifer), and the actual sequestration capacity only reaches 3 million tons (20% of the theoretical value). Another example is that in a certain carbon dioxide sequestration project, the estimated theoretical sequestration capacity of the sandstone layer is 60 billion tons , and the actual monitoring finds that the effective sequestration ratio is only 15%-20%.

[0036] It can be seen that simply setting the actual sequestration capacity as a fixed ratio of the theoretical value is essentially using static linear thinking to deal with non-linear geological system problems, and its most serious consequence is seriously underestimating the sequestration risk or overestimating the emission reduction contribution.

[0037] To solve this problem, the embodiments of the present application provide a method for assessing carbon dioxide sequestration capacity, which is applicable to various fields of carbon dioxide sequestration.

[0038] The embodiments of the present application provide a system for assessing carbon dioxide sequestration capacity, which can execute the method for assessing carbon dioxide sequestration capacity provided by the embodiments of the present application. Figure 1 It is a schematic structural diagram of a system for assessing carbon dioxide sequestration capacity provided by the embodiments of the present application.

[0039] As shown Figure 1 in the figure, the carbon dioxide sequestration capacity evaluation system 001 includes at least one processor 011 and a memory 012 communicatively connected to the at least one processor; wherein, the memory 012 stores instructions executable by the at least one processor 011, and when the instructions are executed by the at least one processor 011, the at least one processor 011 is enabled to execute the carbon dioxide sequestration capacity evaluation method provided by the embodiments of the present application.

[0040] Figure 2 is a schematic flow chart of a carbon dioxide sequestration capacity evaluation method provided by an embodiment of the present application. As Figure 2 shown in the figure, in some embodiments, the carbon dioxide sequestration capacity evaluation method includes the following steps:

[0041] S1, obtaining a structured geological grid model based on training data.

[0042] The training data includes target geological data, target geochemical data, and target engineering parameters. In some embodiments, the training data is preprocessed by Kriging interpolation and wavelet transform.

[0043] The target geological data includes porosity and permeability.

[0044] The target geochemical data includes mineral composition.

[0045] The target engineering parameters include well locations.

[0046] The structured geological grid model includes a three-dimensional grid structure, and its property fields include porosity, permeability, and saturation. In some embodiments, based on the target geological data (including porosity, permeability), geochemical data (including mineral composition), and engineering parameters (including well locations), a three-dimensional structured geological grid model (including property fields such as porosity, permeability, and saturation) is generated by sequential Gaussian simulation.

[0047] Exemplarily, the calculation formula of the structured geological grid model includes:

[0048] ;

[0049] wherein, (x0) represents the porosity of the point x0 to be estimated; λ i represents the weight coefficient, which is solved by the semivariogram γ(h); (x i ) represents the measured value of the known point x i .

[0050] In some embodiments, in sequential Gaussian simulation (SGS), in order to generate porosity ( ), permeability (k), saturation (S w ). The calculation formula for constructing the covariance matrix of the multivariate Gaussian distribution includes:

[0051] ;

[0052] Among them, , Y k , Y Sw are the normal score transformation values of each attribute; σ is the cross-covariance term, reflecting the spatial correlation between attributes.

[0053] In some embodiments, a three-dimensional field of , k, S w is generated simultaneously through sequential Gaussian simulation, including defining the Cross-Variogram:

[0054] ;

[0055] Among them, i, j represent , k, S w indexes, and multi-attribute coupling is realized through Co-Kriging or Linear Model Co-Regionalization (LMC) during simulation.

[0056] Through multi-source data fusion, a structured geological grid model is constructed, integrating target geological parameters (including porosity, permeability), geochemical characteristics (including mineral reaction activity), and engineering constraints (including well pattern parameters), generating a high-resolution three-dimensional attribute field, breaking through the limitations of traditional homogenization models, accurately depicting reservoir heterogeneity, and laying a data foundation for multi-physics field coupling simulation at the millimeter-scale grid.

[0057] S2, input the structured geological grid model into a preset simulator, and obtain dynamic simulation results based on the initial temperature and pressure conditions, carbon dioxide injection rate, and relative permeability curve.

[0058] The initial temperature and pressure conditions include a preset pressure safety threshold and a preset temperature safety threshold. In some embodiments, the preset pressure safety threshold can be 0.9 times the fracture pressure. In some embodiments, the preset temperature safety threshold can be determined by the geothermal gradient. Exemplarily, the value range of the geothermal gradient can be 2.5 - 3.5 °C / 100 m. It can be understood that the reservoir temperature needs to be higher than the critical temperature (such as 31.1 °C) to ensure is in the supercritical state (i.e., high density, low viscosity).

[0059] The carbon dioxide injection rate is used to set a preset value according to an empirical formula. In some embodiments, the carbon dioxide injection rate can be obtained based on pressure balance, phase equilibrium, and seepage equilibrium.

[0060] It can be understood that the determination of the carbon dioxide injection rate needs to comprehensively consider reservoir characteristics, engineering safety, and economic benefits. For example, the injection rate per well can be 5,000 - 50,000 standard cubic meters per day (≈0.06 - 0.58 m³ / s). Another example is that the injection rate per well for saline aquifer storage can reach 1 - 5 Mt / year (≈0.03 - 0.16 m³ / s). For another example, the experimental injection rate per well can be 100 - 1,000 tons per year. The specific calculation formula for the carbon dioxide injection rate is not limited in this application.

[0061] In some embodiments, the relative permeability curve can be the kr- / kr-water curve.

[0062] In some embodiments, the microscopic flow response is mapped to the reservoir grid through numerical homogenization to achieve cross-scale parameter transfer of "pore, core, reservoir".

[0063] Among them, the calculation formula for numerical homogenization includes:

[0064] ;

[0065] Among them, k eff represents the effective permeability; V represents the total volume of the integration region; k(x) represents the local permeability; represents the pressure gradient vector;

[0066] The effective permeability k eff represents the equivalent permeability of the entire region (porous medium), which is the average permeability after considering spatial heterogeneity, and the unit is area (such as Darcy or square meter).

[0067] The total volume of the integration region V refers to the geometric volume of the research region, and the unit is cubic meter m³.

[0068] The local permeability k(x), the permeability at position x, characterizes the fluid passing ability at that point. It can be a scalar (isotropic medium) or a tensor (anisotropic medium), and the unit is Darcy D or square meter m².

[0069] The pressure gradient vector , describes the pressure change rate of the fluid in space, and the direction points to the direction of the fastest pressure drop, and the unit is Pascal per meter (i.e., Pa / m). It drives the fluid movement in porous medium flow.

[0070] A preset simulator that can simulate, using the finite element or finite difference method, the multiphase flow of carbon dioxide (in supercritical state) and formation brine (in liquid state). There are various implementation forms of the preset simulator. For example, it can be TOUGH2, Eclipse, or CMG-GEM, etc.

[0071] In some embodiments, when performing step S2, the carbon dioxide sequestration capacity assessment method further includes the following steps:

[0072] S201, Based on porosity and permeability, and according to the equation of state and chemical reaction model, obtain the multiphase flow results. The multiphase flow results include the phase behavior of carbon dioxide and liquid brine.

[0073] In some embodiments, the equation of state can be the Peng-Robinson equation, describing the phase equilibrium (density, viscosity) between the supercritical state and liquid brine.

[0074] In some embodiments, the chemical reaction model includes simulating dissolution ( ), and mineral reactions (such as calcite dissolution).

[0075] S202, According to the multiphase flow results, obtain the dynamic simulation results.

[0076] In some embodiments, the calculation formula for dynamic numerical simulation includes:

[0077] ;

[0078] where α represents the phase state ( phase α = g, brine phase α = w); ρ α represents the phase density, with the unit of kg / m³; S α represents the phase saturation (S g +S w = 1); v α represents the phase velocity (Darcy's law, ); Q α represents the source-sink term ( injection rate Qg, unit kg / s).

[0079] S3, According to the dynamic simulation results, based on the preset pressure safety threshold and preset plume range, calculate the effective capacity and risk probability map.

[0080] The preset pressure safety threshold includes the pressure constraint V safe and the formation fracture pressure.

[0081] The preset plume range is for pre-estimating or simulating the fluid (such as , the spatial distribution boundary of the diffusion and migration of pollutants or displacement agents in the underground reservoir. There are various calculation methods for the preset plume range. For example, the numerical simulation method uses multiphase flow simulation software such as TOUGH2, ECLIPSE, and CMG-GEM to simulate the spatial distribution of the plume evolving over time and obtain the preset plume range.

[0082] For another example, the analytical radial diffusion model:

[0083] .

[0084] Among them, r max represents the maximum radial expansion distance of the plume (m); Q represents the injection rate in m³ / s; t represents the injection time (s); h represents the effective thickness of the reservoir (m); represents the porosity; S CO2 represents average saturation.

[0085] For another example, by simulating the injection process through laboratory cores, the breakthrough curve of the plume and the saturation distribution are measured.

[0086] It can be understood that the present application does not limit the method for determining the preset plume range.

[0087] In some embodiments, the calculation formula of the effective capacity includes:

[0088] ;

[0089] Among them, V safe represents the safe storage volume that satisfies the pressure constraint p ≤ 0.9p frac ; represents the porosity; represents the carbon dioxide density; S g represents the carbon dioxide phase saturation; E dynamic represents the dynamic efficiency factor.

[0090] p frac represents the formation fracture pressure, with the unit of MPa.

[0091] In some embodiments, the value range of E dynamic is 0.05 - 0.3.

[0092] In some embodiments, the calculation formula of the risk probability map includes:

[0093] ;

[0094] Among them, λ iDenoted as the failure rate of fault activation and wellbore failure of the i-th type of risk event, with the unit of times / year; t represents the storage time, with the unit of year; N represents the total number of types of risk events.

[0095] Relying on multiphase flow-chemistry-mechanics dynamic numerical simulation, quantify The spatio-temporal expansion law of the plume, the propagation boundary of the pressure field, and the potential of mineralization storage. Combining the preset pressure safety threshold and the plume constraint range, eliminating the overpressure and high-risk areas beyond the boundary, generating a probabilistic leakage risk map, and realizing the safe conversion from the theoretical capacity to the engineering effective capacity.

[0096] S4. According to the effective capacity and the risk probability map, through the preset well pattern value, the preset injection rate, and the preset pressure control threshold, obtain the well location coordinates, the injection rate schedule, and the pressure monitoring threshold.

[0097] In some embodiments, the calculation formula of the injection rate schedule includes:

[0098] ;

[0099] Among them, Q j Denotes the injection rate of the j-th well, with the unit of kg / s; Q total Denotes the total injection volume constraint; P j Denotes the wellhead pressure of the j-th well; p frac Denotes the formation fracture pressure; ΔQ j Denotes the rate adjustment step; M denotes the total number of wells.

[0100] It can be understood that based on the preset injection rate, with the rate adjustment step as the increment, adding them in sequence to obtain the optimal injection rate.

[0101] Exemplarily, the preset well pattern value is that the well spacing for optimizing the well pattern layout by linear programming is ≥500 meters. The injection rate schedule is the rate-time relationship under the principle of stepped deceleration.

[0102] The preset pressure control threshold can be 80% of the fracture pressure of the pressure monitoring threshold.

[0103] In this way, by inputting the preset well pattern value, the preset injection rate, and the preset pressure control threshold, adjusting the rate according to the step length, the well location coordinates, the injection rate schedule, and the pressure monitoring threshold that meet the effective capacity and the risk probability map are obtained.

[0104] In some embodiments, when performing step S4, the carbon dioxide storage capacity evaluation method further includes the following steps:

[0105] S401. Calculate the theoretical capacity.

[0106] Among them, the calculation formula of the theoretical capacity includes:

[0107] ;

[0108] Wherein, A represents the reservoir area, with the unit of km²; h represents the effective thickness, with the unit of m; represents the average porosity, with the unit of %; represents the density of carbon dioxide; E represents the sequestration efficiency factor.

[0109] In some embodiments, The supercritical state is about 700 kg / m³.

[0110] In some embodiments, E is usually taken as 1% - 10%.

[0111] S402. Confirm the target potential area according to the theoretical capacity.

[0112] The target potential area is the area that meets the theoretical capacity.

[0113] After performing step S402, the carbon dioxide sequestration volume evaluation method further includes the following steps:

[0114] S403. Confirm the overlapping coordinate values of the target potential area and the well location coordinates.

[0115] S404. Update the well location coordinates based on the overlapping coordinate values.

[0116] Through the well pattern optimization algorithm, with the effective capacity and the risk heat map as inputs, solve the multi-objective Pareto optimal solutions of the injection well location coordinates, rate schedule, and pressure monitoring threshold to ensure the dynamic balance of maximizing injection efficiency and minimizing risk.

[0117] S5. Obtain the real-time monitoring data set based on the well location coordinates.

[0118] The real-time monitoring data set includes wellhead pressure, wellhead temperature, fiber optic time series data, and four-dimensional seismic difference volume.

[0119] In some embodiments, perform injection according to the injection plan Inject, and collect wellhead pressure / temperature, fiber optic sensing data (DTS / DAS), and four-dimensional seismic difference volume in real time to generate a time series monitoring data set as the real-time monitoring data set.

[0120] S6. Input the real-time monitoring data set into the machine learning model to generate an injection rate adjustment instruction and a leakage risk warning.

[0121] In some embodiments, the calculation formula for generating the injection rate adjustment instruction includes:

[0122] ;

[0123] Among them, is denoted as the adjusted injection rate at time t; f LSTM is denoted as the LSTM network mapping function; X seismic is denoted as the four-dimensional seismic data feature vector.

[0124] In some embodiments, the monitoring data is input into the LSTM-CNN hybrid model to generate an injection rate adjustment instruction (±10% rate) and the level of leakage risk warning (low / medium / high) in real time.

[0125] In some embodiments, when performing step S6, the carbon dioxide storage volume evaluation method further includes the following steps:

[0126] S601, input the sensor data, fiber value, four-dimensional seismic data, and surface deformation value into the machine learning model to obtain a real-time plume map, leakage risk warning result, and policy dynamic adjustment result.

[0127] In some embodiments, a multi-modal spatio-temporal fusion network (MTSF-Net) model is used to integrate multi-source heterogeneous data and generate real-time decisions. This model processes different types of data through three parallel branches: sensor and fiber data extract temporal features through a bidirectional LSTM combined with a temporal attention mechanism, four-dimensional seismic data capture spatio-temporal correlations through a 3D convolutional network and a Transformer encoder, and surface deformation data extract spatial features through a 2D convolutional network and spatial pyramid pooling. The fusion layer uses a cross-modal attention mechanism to align multi-dimensional information, and finally outputs the results synchronously through three task heads - generating a carbon dioxide saturation distribution map in the reservoir based on a 3D deconvolution network, evaluating the leakage risk level through a classification network, and giving a dynamic adjustment coefficient (such as the change amplitude of the injection rate, the correction value of the pressure threshold) using a reinforcement learning policy network.

[0128] In some embodiments, a composite loss function is used during model training to optimize the plume map accuracy by combining the 3D intersection over union loss, balance the risk classification sample bias using the focal loss, and introduce the proximal policy optimization algorithm to improve the engineering adaptability of the policy network.

[0129] In some other embodiments, for the multi-source heterogeneous data fusion requirement, the Spatiotemporal Awareness Hybrid Network (STAH-Net) realizes dynamic decision-making generation through hierarchical feature extraction and physics-guided fusion mechanisms. The model first uses 3D dilated convolution for fiber optic strain data to capture local deformation details, and at the same time uses the multi-head self-attention mechanism to analyze the cross-node correlation patterns of multi-sensor time series signals; the four-dimensional seismic volume data is processed by spatiotemporal separable convolution. The 3D convolution kernel is used in the spatial dimension to extract reservoir structure features, and the bidirectional LSTM is used in the time dimension to capture the continuous evolution law of fluid migration; the surface deformation grid data is decoded by the residual UNet network to obtain the mapping relationship between the spatial deformation field and the underground pressure change. The three types of features are cross-modally aligned through the dynamic gating fusion layer, where the seismic data features serve as the dominant signal to control the gating weights, ensuring that geological structure constraints penetrate into the fusion feature space preferentially. The fused multi-dimensional tensor generates a three-dimensional output through parallel decoding paths: the plume reconstruction module based on the improved 3D transposed convolution network generates a phase state distribution map with sub-meter accuracy of ; the risk module combining kernel density estimation and spatial pyramid pooling outputs a probabilistic leakage heat map to quantify the breakthrough risk of the weak caprock area; the policy generator embedded in the reinforcement learning framework dynamically outputs a nine-dimensional adjustment vector through the proximal policy optimization algorithm, integrating the current system state and historical operation records, covering the injection rate correction coefficient (in the range of ±20%), the monitoring well scanning frequency (gradient adjustment from 1 to 10 Hz), and the emergency response level (levels 1-5). The model embeds the physical constraint mechanism driven by seismic data, and converts prior knowledge such as fault boundaries and caprock thickness interpreted from seismic data into soft constraints in the feature space through differentiable gating functions, which not only avoids the rigid boundary assumptions of traditional numerical simulations but also improves the geological rationality of machine learning models.

[0130] S602, generating an injection rate adjustment instruction and a leakage risk warning based on the real-time plume map, the leakage risk warning result, and the policy dynamic adjustment result.

[0131] In some embodiments, the output of S601 is converted into executable instructions through a Dynamic Decision Reinforcement Learning Model (DDRL). This model analyzes the spatial diffusion pattern of the real-time plume map through a three-dimensional convolutional network, encodes the leakage risk probability using a multi-layer perceptron, and analyzes the time series pattern of historical adjustment instructions using an LSTM network. The policy network generates continuous-valued injection rate adjustment instructions (within the range of ±20%) based on the current fusion state, and the value network evaluates the impact of actions on the long-term sequestration benefit. The early warning module combines rules and learning. When the leakage probability exceeds 0.7 or the plume front approaches the caprock by 50 meters, it triggers a red warning and automatically stops injection. When the risk is moderate, it implements a rate halving operation. The model achieves multi-objective optimization through a customized reward function, balancing sequestration efficiency, risk control, and pressure stability, and adopts a strategy with noisy exploration to enhance decision-making robustness. The two models form a closed-loop system through a digital twin platform, iteratively updating the decision-making scheme every five minutes to achieve a minute-level response from data perception to control execution.

[0132] S7. Based on the injection rate adjustment instruction and the leakage risk warning, update the dynamic numerical simulation parameters according to the well location coordinates, injection rate schedule, and pressure monitoring threshold, and calculate the carbon dioxide distribution and pressure evolution results.

[0133] In some embodiments, the calculation formula for updating the dynamic numerical simulation parameters includes:

[0134] ;

[0135] ;

[0136] where m represents the model parameter vector (permeability, porosity); d obs represents the monitoring data (pressure, temperature); H represents the observation operator matrix; C represents the covariance matrix.

[0137] In some embodiments, the calculation formula for the carbon dioxide distribution and pressure evolution results includes:

[0138] ;

[0139] where ; ;

[0140] represents the gradient operator, representing the derivative with respect to the spatial coordinates; λt represents the total mobility, characterizing the flow ability of multiphase fluids in the formation, which is jointly determined by the relative permeability and viscosity of the two-phase fluids; represents the pressure gradient, the main driving force for fluid flow; Denoted by the divergence operator, representing the net outflow rate of fluid in space. Denoted by porosity, representing the proportion of pore volume in the total volume of the rock (dimensionless, range 0 - 1); ρ t Denoted by total density, obtained by weighting the saturation of the two - phase fluid density: ρ CO2 Denoted by The density of phase (kg / m³), ρ w Denoted by the density of the water phase (kg / m³); S CO2 Denoted by The saturation of phase, Sw is denoted by the saturation of the water phase (dimensionless, satisfying S CO2 +S w = 1); q inj (t) is denoted by the injection rate, representing the injection into the formation through the injection well of the source term (kg / (m³·s) or m³ / (s·m³)); q leak (P) is denoted by the leakage rate, representing the leakage caused by over - pressure or the existence of leakage channels (kg / (m³·s)); k rCO2 Denoted by The relative permeability of phase, k rw Denoted by the relative permeability of the water phase, representing the effective flow ability of a certain phase of fluid in the pores (dimensionless, related to saturation); μ CO2 Denoted by The viscosity of phase (Pa·s), reflecting the fluid flow resistance; μ w Denoted by the viscosity of the water phase (Pa·s), reflecting the fluid flow resistance.

[0141] In some embodiments, q inj (t) is described by the point - source function:

[0142] ;

[0143] Q schedule (t) is denoted by the preset injection rate schedule; Denoted by the Dirac function at the well - location coordinates (i.e., the point - source model).

[0144] In this way, according to the adjustment instruction and the risk level, update the injection rate and permeability field of the dynamic simulation, and recalculate the distribution and pressure evolution to complete the single - cycle closed - loop optimization.

[0145] Based on real-time monitoring data (wellhead pressure, fiber optic strain, four-dimensional seismic), an LSTM-CNN hybrid model is used to identify abnormal modes and generate control instructions. The geological model parameters are dynamically corrected through the ensemble Kalman filter to form a "monitoring-simulation-decision-making-feedback" closed loop, enabling the system to have self-adaptability to working conditions. Through the full-chain innovation of "data-driven modeling, dynamic safety assessment, and intelligent optimization control", this method reduces the sequestration volume assessment error from ±50% to ±15%, improves the leakage warning response speed to the minute level, and increases the injection efficiency by 40%. Finally, it realizes the technological leap of carbon sequestration projects from rough estimation to precise control, providing high-reliability decision-making support for large-scale deployment.

[0146] In some embodiments, the updated dynamic numerical simulation parameters (such as permeability) output by step S7 are directly fed back to step S2 to update the dynamic simulation results.

[0147] In this way, a closed-loop chain of "simulation, monitoring, correction, and re-simulation" is formed.

[0148] Based on the same inventive concept, an embodiment of the present application also provides a carbon dioxide sequestration volume assessment system. The method corresponding to the carbon dioxide sequestration volume assessment system may be the carbon dioxide sequestration volume assessment method in the foregoing embodiments, and the principle of solving problems is similar to that of this method. The carbon dioxide sequestration volume assessment system provided in the embodiment of the present application includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the methods and / or technical solutions of multiple embodiments of the present application described above.

[0149] Another embodiment of the present application also provides a computer-readable storage medium, on which computer program instructions are stored, and the computer program instructions can be executed by a processor to implement the methods and / or technical solutions of any one or more embodiments of the present application described above.

[0150] Specifically, one or more combinations of computer-readable media can be adopted in this embodiment. The computer-readable media can be computer-readable signal media or computer-readable storage media. The computer-readable storage media can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples (non-exhaustive list) of the computer-readable storage media include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this document, the computer-readable storage media can be any tangible medium that contains or stores a program, which can be used by or in conjunction with an instruction execution system, apparatus, or device.

[0151] The computer-readable signal media can include data signals propagated in a baseband or as part of a carrier wave, which carry computer-readable program code. Such propagated data signals can take various forms, including, but not limited to, electromagnetic signals, optical signals, or any suitable combination of the above. The computer-readable signal media can also be any computer-readable media other than the computer-readable storage media, which can send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0152] The program code contained on the computer-readable media can be transmitted using any appropriate medium, including, but not limited to, wireless, wire, optical fiber, RF, etc., or any suitable combination of the above.

[0153] The computer program code for performing the operations of this application can be written in one or more programming languages or combinations thereof. The programming languages include object-oriented programming languages such as Java, Smalltalk, C++, and also include conventional procedural programming languages such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, executed as an independent software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (for example, by using an Internet service provider to connect through the Internet).

[0154] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of devices, methods, and computer program products according to various embodiments of the present application. In this regard, each block in the flowchart or block diagram may represent a module, a segment of a program, or a portion of code that contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions noted in the blocks may occur in a different order than noted in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, or they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, and combinations of blocks in the block diagram and / or flowchart, may be implemented by a dedicated hardware-based system that performs the specified functions or operations, or by a combination of dedicated hardware and computer instructions.

[0155] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the foregoing method embodiments and will not be elaborated herein.

[0156] In several embodiments provided in the present application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units or page components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the couplings, direct couplings, or communication connections shown or discussed among each other can be through some interfaces, and the indirect couplings or communication connections of devices or units can be in electrical, mechanical, or other forms.

[0157] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they can be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0158] In addition, the functional units in various embodiments of the present application can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above-mentioned integrated units can be implemented in the form of hardware or in the form of a combination of hardware and software functional units.

[0159] The integrated units implemented in the form of software functional units can be stored in a computer-readable storage medium. The above-mentioned software functional units are stored in a storage medium and include several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) or a processor to execute some steps of the methods described in various embodiments of the present application. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs that can store program codes.

[0160] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of various embodiments of the present application.

[0161] In addition, it is obvious that the word "including" does not exclude other units or steps, and the singular does not exclude the plural. The multiple units or devices stated in the apparatus claims can also be implemented by one unit or device through software or hardware. The terms first, second, etc. are used to denote names and do not denote any particular order.

Claims

1. A method for evaluating the carbon dioxide sequestration volume, characterized in that, Including: Based on training data, a structured geological grid model is obtained; the training data includes target geological data, target geochemical data, and target engineering parameters; The structured geological grid model includes a three-dimensional grid structure, and its property fields include porosity, permeability, and saturation; The structured geological grid model is input into a preset simulator, and based on initial temperature and pressure conditions, carbon dioxide injection rate, and relative permeability curve, a dynamic simulation result is obtained; According to the dynamic simulation result, based on a preset pressure safety threshold and a preset plume range, an effective capacity and a risk probability map are calculated; According to the effective capacity and the risk probability map, through a preset well pattern value, a preset injection rate, and a preset pressure control threshold, well location coordinates, an injection rate schedule, and a pressure monitoring threshold are obtained; Based on the well location coordinates, a real-time monitoring data set is obtained; the real-time monitoring data set includes wellhead pressure, wellhead temperature, fiber optic timing data, and four-dimensional seismic differential volume; The real-time monitoring data set is input into a machine learning model to generate an injection rate adjustment instruction and a leakage risk warning; Based on the injection rate adjustment instruction and the leakage risk warning, according to the well location coordinates, the injection rate schedule, and the pressure monitoring threshold, the dynamic numerical simulation parameters are updated, and the carbon dioxide distribution and pressure evolution results are calculated.

2. The carbon dioxide sequestration amount evaluation method according to claim 1, wherein When performing the step of inputting the structured geological grid model into a preset simulator and obtaining a dynamic simulation result based on initial temperature and pressure conditions, carbon dioxide injection rate, and relative permeability curve; The carbon dioxide storage capacity evaluation method further includes: Based on the porosity and the permeability, according to the state equation and the chemical reaction model, a multiphase flow result is obtained; the multiphase flow result includes the phase behavior of carbon dioxide and liquid brine; According to the multiphase flow result, the dynamic simulation result is obtained.

3. The carbon dioxide storage capacity assessment method according to claim 1 or 2, characterized in that The carbon dioxide storage capacity evaluation method further includes: Calculating the theoretical capacity; According to the theoretical capacity, a target potential area is confirmed; After performing the step of obtaining well location coordinates, an injection rate schedule, and a pressure monitoring threshold through a preset well pattern value, a preset injection rate, and a preset pressure control threshold according to the effective capacity and the risk probability map, the carbon dioxide storage capacity evaluation method further includes: Confirming the overlapping coordinate value of the target potential area and the well location coordinates; Based on the overlapping coordinate value, the well location coordinates are updated; Wherein, the calculation formula of the theoretical capacity includes: ； Among them, A represents the reservoir area, with the unit of km²; h represents the effective thickness, with the unit of m; represents the average porosity, with the unit of %; represents the density of carbon dioxide; E represents the sequestration efficiency factor.

4. The carbon dioxide sequestration capacity evaluation method according to claim 1 or 2, characterized in that, Before performing the step of training a three-dimensional geological model based on training data to obtain a structured geological grid model; The carbon dioxide storage capacity evaluation method further includes: Preprocessing the training data through Kriging interpolation and wavelet transform.

5. The carbon dioxide sequestration amount evaluation method according to claim 1 or 2, characterized in that When performing the step of inputting the real-time monitoring data set into a machine learning model to generate an injection rate adjustment instruction and a leakage risk warning, the carbon dioxide storage capacity evaluation method further includes: Inputting sensor data, fiber optic values, four-dimensional seismic data, and surface deformation values into a machine learning model to obtain a real-time plume map, a leakage risk warning result, and a strategy dynamic adjustment result; Based on the real-time plume map, the leakage risk warning result, and the result of dynamic adjustment of the strategy, generate the injection rate adjustment instruction and the leakage risk warning.

6. The carbon dioxide storage capacity evaluation method according to claim 2, wherein The calculation formula of the effective capacity includes: ； Among them, V safe represents the safe storage volume that satisfies the pressure constraint p ≤ 0.9p frac . represents the porosity; represents the carbon dioxide density; S g represents the carbon dioxide phase saturation; E dynamic represents the dynamic efficiency factor.

7. The carbon dioxide sequestration amount evaluation method according to claim 1, characterized in that The calculation formula of the risk probability map includes: ； Among them, λ i represents the failure rate of the i-th type of risk event, with the unit of times / year; t represents the storage time, with the unit of year; N represents the total number of types of risk events.

8. A carbon dioxide sequestration amount evaluation system, characterized in that, Includes: At least one processor; A memory communicatively connected to the at least one processor; Wherein, the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method according to any one of claims 1 to 7.

9. A computer-readable medium having computer program instructions stored thereon, characterized in that, The computer program instructions can be executed by a processor to implement the method according to any one of claims 1 to 7.

Images