Auxiliary Design System and Method for Repairing Leakage Channels in CO2 Sequestration Geological Bodies

Abstract

This application relates to the field of data design technology, and in particular to an auxiliary design system and method for repairing leakage channels in CO2 storage geological bodies. This application first accurately acquires multidimensional data of the channel, and then analyzes its morphology to pre-screen injectable fillers; then predicts the interfacial bonding strength between the filler and the channel and the overall sealing performance, and assesses long-term stability in conjunction with future environmental factors; finally, it automatically recommends the optimal filler based on the quantitative evaluation results. This solution elevates the repair design from relying on experience to being data-driven, ensuring the remediation solution's pertinence and long-term effectiveness.

Description

Technical Field

[0001] This application relates to the field of data design technology, and in particular to an auxiliary design system and method for repairing leakage channels in CO2-sealed geological bodies. Background Technology

[0002] In carbon dioxide geological storage (CCS) projects, the long-term integrity of the stored geological bodies (such as deep saline aquifers and depleted oil and gas reservoirs) is crucial. However, due to the complexity of geological structures, the heterogeneity of caprocks, or injection pressure, stored CO2 may leak through natural or induced channels such as fractures, faults, and abandoned wells. Once a leak occurs, it not only directly leads to storage failure, causing economic losses and failing to meet carbon reduction targets, but also poses potential risks to shallow groundwater, the surface ecosystem, and even atmospheric safety. Therefore, rapidly, accurately, and effectively repairing CO2 leakage channels is a core technical aspect of ensuring the long-term safety of CCS projects and enhancing public confidence.

[0003] Currently, the repair of CO2 leak channels mainly relies on traditional oil and gas well or underground engineering plugging techniques, with adaptive adjustments made to these. However, the repair of CO2 leak channels faces more complex and demanding operating conditions than conventional plugging, and existing technologies have significant shortcomings in addressing these challenges: the selection of repair materials (fillers) is based on a single approach and lacks systematic evaluation. Traditional plugging materials (such as cement slurry, gel, and resin) are often selected based on engineers' experience, material supplier recommendations, or simple rules for conventional water / gas leaks (e.g., using coarse particles for large leaks and fine particles for small leaks). For CO2 leak channels, the internal environment is characterized by high CO2 concentrations, acidity (carbonic acid environment), specific temperature-pressure fields, and complex flow phases (supercritical, gaseous, and dissolved phases). Existing methods struggle to systematically quantify the suitability of specific fillers in this complex chemical-physical coupling environment, including their injectability (whether they can successfully enter and fill the channel), reaction and curing behavior (whether curing time and strength development are affected by CO2), and long-term stability (resistance to carbonic acid corrosion and temperature / stress cycle resistance). This leads to a high degree of blind selection of materials, which easily results in problems such as the material not being able to be poured in, not being able to be fixed, or not being stable for long. Therefore, there is an urgent need to develop an auxiliary design technology for the repair of leakage channels in CO2-sealed geological bodies that can overcome the above limitations. Summary of the Invention

[0004] To overcome the shortcomings and deficiencies of existing technologies, this application first accurately acquires multidimensional channel data, then analyzes its morphology to pre-screen injectable fillers; subsequently, it predicts the interfacial bonding strength between the filler and the channel and the overall sealing performance, and assesses long-term stability in conjunction with future environmental factors; finally, it automatically recommends the optimal filler based on the quantitative evaluation results. This solution elevates the repair design from relying on experience to being data-driven, ensuring the targeted and long-term effectiveness of the repair solution.

[0005] To achieve the above objectives, this application adopts the following technical solution: Firstly, this application provides an auxiliary design system for repairing leakage channels in CO2 sequestration geological bodies, including the following modules: The system comprises a data acquisition module, a leakage channel analysis module, a packing connection prediction module, a connection performance prediction module, a reaction assessment module, and a packing selection module. The data acquisition module acquires information about the leak outlet, the gas conditions within the channel, and the corresponding environmental conditions. The leakage channel analysis module analyzes the filling and insertion of various packing materials based on the three-dimensional information of the leak outlet, the gas conditions within the channel, and the corresponding environmental conditions. The packing connection prediction module predicts the packing bonding performance based on the solidification of various packing materials in the given scenario. The connection performance prediction module predicts the connection performance based on the filling and insertion conditions of various packing materials and the predicted packing bonding performance. The reaction assessment module predicts future reaction impacts based on the characteristics of the packing materials, the gas conditions within the channel, and the corresponding environmental conditions, and assesses the future service life based on the predicted future reaction impacts and connection performance. The packing selection module selects repair packing materials based on the future service life assessment results.

[0006] In one implementation of this application, the leakage channel and leakage outlet conditions include obtaining images of the leakage outlet through acoustic scanning or downhole television, and reconstructing a three-dimensional image of the leakage channel and leakage outlet using three-dimensional reconstruction software to obtain the precise geometric shape of the leakage outlet; the gas conditions within the channel include dynamic parameters such as phase state, composition, temperature, pressure, flow rate, and pH value of the leaking fluid obtained through sensors; the environmental conditions at the corresponding location include determining the geostress field, geothermal temperature, and mineral composition around the leakage point and the channel.

[0007] This step utilizes various technologies such as acoustic scanning, downhole television, and sensors to accurately and comprehensively acquire information on the geometric morphology of the leak channel, internal fluid dynamic parameters, and surrounding geological environment. This provides a reliable and multi-dimensional data foundation for subsequent analysis, ensuring the targeted and accurate nature of the repair design.

[0008] In one implementation of this application, the analysis of the filling and insertion of various fillers includes the following specific steps: The first step is to obtain the geometry of the leak outlet channel, including the number of bends, bend angles, average inner diameter, and smoothness anomalies of the channel's inner wall. The smoothness anomaly is the average height of any protrusion or recess on the channel's inner surface divided by the safe height, which is the maximum height that does not affect the flow velocity of the corresponding packing material. This is obtained experimentally. The bend angle anomaly is obtained by dividing the channel bend angle by the safe bend angle. The overall bend anomaly is obtained by summing all the bend angle anomalies at the leak outlet. The safe bend angle is the maximum bend angle that does not affect the flow velocity of the corresponding packing material. The channel inner diameter anomaly is obtained by dividing the average inner diameter of the packing particles by the average inner diameter of the channel. The obstruction anomaly of the corresponding packing material filling the channel is obtained by weighted summing of the channel inner diameter anomaly, overall bend anomaly, and smoothness anomaly. The second step involves calling the reaction model from the repair packing database to simulate the hydration, gelation, carbonization, or cross-linking reaction process of the packing under corresponding CO2 concentrations, acidity, and specific temperature and pressure conditions. This allows for the acquisition of the maximum flow-curing distance of the packing during the injection process under specific flow rates and viscosities in the corresponding environment. This can be obtained by looking up historical experimental data. Simultaneously, multiple sets of experiments are conducted to observe how far the packing will flow in different environments before curing. The average value is then calculated to obtain the final curing distance, which is the maximum flow-curing distance. The obstruction effect is obtained by multiplying the obstruction anomaly in the corresponding packing filling channel by the obstruction influence coefficient. The actual flow-curing distance is obtained by subtracting the obstruction anomaly effect from the value 1 and then multiplying it by the maximum flow-curing distance. Finally, the filtration loss risk of the corresponding packing is obtained by dividing the internal depth of the leak channel by the actual flow-curing distance. The obstruction influence coefficient is obtained by averaging historical data. This step, by quantitatively analyzing the geometric characteristics of the channel and combining them with the physical properties of the packing, can scientifically assess the flow resistance and filtration risk during the injection of different packings, thereby pre-screening packings that are feasible to inject under specific channel morphologies. In one implementation of this application, the prediction of the filler bonding performance includes the following specific aspects: The first step is to call the reaction model in the repair filler database to simulate the hydration, gelation, carbonization or cross-linking reaction process of the filler under high CO2 concentration, acidity, and specific temperature and pressure conditions. At the same time, the stress field, ground temperature, number of bends, bend angle, average inner diameter of the channel, smoothness anomalies of the inner wall of the channel and mineral composition of the leak point and channel are called in to build a neural network prediction model to predict the bonding strength between the interface and the repair filler. The specific steps include: Construction of the repair filler database and invocation of the reaction model: Data source: Extracting the basic formulation, physical properties, reaction kinetic parameters, thermodynamic parameters, and constitutive model parameters of the target filler from the repair filler database; Reaction simulation invocation: Based on the above parameters, establishing a chemical-thermal-mechanical coupled finite element / finite volume model; Setting up in COMSOL Multiphysics, etc.: Chemical module: Defining hydration, carbonization, and crosslinking reaction equations, considering CO2 concentration and pH value as reactant concentration variables; Heat and mass transfer module: Simulating reaction heat, temperature field, pressure field, and diffusion-convection process of reactants in porous media / cracks; Solid mechanics module: Coupled with the volume change and stress development caused by the chemical reaction and the geostress field; Output: Through simulation, obtaining key field variables of the restoration body's evolution over time / space, such as: porosity distribution, hardness / modulus distribution, reaction degree, and internal microstress field; These will serve as filler-side features for subsequent prediction models.

[0009] Quantification of geological features of the leakage channel: Feature extraction: Transform geological parameters into standardized, computable numerical feature vectors: Stress field: Decomposed into maximum principal stress, minimum principal stress, and stress direction; Temperature field: Take the average temperature and temperature gradient along the leakage channel. Channel geometry and topology: number of bends; bend angle sequence; average inner diameter; inner wall smoothness; mineral composition: XRD or SEM-EDS analysis of the channel inner wall to obtain the mass percentage of major minerals; integrated output: a comprehensive geological feature vector; combining two vastly different types of input data—filler reaction evolution characteristics and geological channel static characteristics—a dual-flow feature fusion neural network is used; specific steps are as follows: Step 101: Analysis of Filler Evolution Feature Flow Based on Convolutional Neural Networks: Network Layer Construction and Parameter Settings: 3D Convolutional Layers: Function: Simultaneously extract local patterns in both spatial and temporal dimensions. For example, capturing the advancement of reaction fronts and the migration of stress concentration zones; Parameter Settings: Kernel Size: For example, (3,3,3); the first 3 indicates observing patterns over 3 consecutive time steps. Number of Filters: Start with a small number and gradually increase as the network deepens. This allows the network to learn spatiotemporal features from simple to complex. Stride and Padding: The stride is usually set to (1,1,1) to preserve information, and the same padding is used to maintain the feature map size; 3D Batch Normalization Layer: Purpose: Stabilizes the training of deep networks, accelerates convergence, and is especially important for multi-physics data with varying input scales; Parameter Settings: Default settings, immediately following each 3D convolutional layer; Activation Function: Choose ReLU or its variants; ReLU provides sparse activation, alleviates gradient vanishing, and is computationally efficient; 3D Pooling Layer: Purpose: Reduces the resolution of the feature map in the spatiotemporal dimension, increases the receptive field of subsequent layers, and controls overfitting; Type and Parameters: Usually 3D max pooling is used, with a pooling window such as (2,2,2) and a stride of 2; one is placed every 1-2 convolutional blocks; Spatiotemporal Global Average Pooling Layer: Purpose: At the end of the flow, the complex spatiotemporal feature map output by the last convolutional layer is averaged along the entire dimension, generating a scalar value for each channel. This transforms an input of arbitrary size into a fixed-length feature vector, which is well-suited for handling mesh sizes that may vary in simulations. The output is a one-dimensional vector that serves as a high-level abstract feature of the filler flow. Step 102: Analysis of Geological Static Feature Flow Based on Fully Connected Neural Networks: Network Layer Construction and Parameter Settings: Fully Connected Layers; Function: To learn complex nonlinear combinations and interactions between features. For example, to learn the impact of high stress and high clay content on the interface. Parameter Settings: Number of Layers and Neurons: Funnel-shaped design is adopted. The number of neurons in the first layer can be close to the number of input features, and the number of neurons in subsequent layers gradually decreases; 2-4 hidden layers are usually sufficient; Weight Initialization: He initialization is used, which works well with the ReLU activation function; Batch Normalization Layer: Function: Also used for stable training, especially when the dimensions of input features differ greatly. Parameter Settings: Placed between each fully connected layer and activation function; Activation Function Selection: ReLU is also used; Dropout Layer: Function: A key regularization method to prevent overfitting; Randomly shuts down a portion of neurons during training, forcing the network to learn more robust features; Parameter Settings: The dropout rate is usually set between 0.2 and 0.5; Step 103: The concatenation layer simply concatenates the first two outputs along the feature dimension to form a joint feature vector, which is the most direct and effective fusion method. The fully connected fusion layer's function is to learn the higher-order correlation between two different types of features and ultimately map it to the prediction target. Parameter settings: It consists of 1-3 fully connected layers; the number of neurons in the first layer is approximately half or one-third of the length of the concatenated vector, and subsequent layers continue to reduce this number; each layer is followed by ReLU, BN, and Dropout; the final output layer: a single adhesion strength is represented by one neuron; key training parameters: the loss function is the mean squared error loss; if multiple intensity values ​​are predicted, the MSE can be calculated for each output and then summed; the optimizer is Adam, the initial learning rate is set to 1e-4 to 1e-3, and a learning rate scheduler is used to reduce the learning rate when training stalls; regularization: in addition to Dropout, L2 weight decay can be used in the fully connected layers. The second step is to obtain the bond strength between the repair filler and the repair filler after all the repair fillers are predicted by the model and enter the leak channel.

[0010] This step utilizes a dual-flow feature fusion neural network to deeply integrate and model the spatiotemporal characteristics of the chemical reaction evolution of the filler under complex environments with the static geological characteristics of the channel, enabling high-precision prediction of the final bond strength between the filler and the channel interface.

[0011] In one implementation of this application, the connection performance prediction includes the following specific aspects: The filtration loss impact is obtained by multiplying the filtration loss risk of the corresponding packing by the filtration loss risk impact coefficient. The bond strength anomaly is obtained by dividing the safe bond strength by the bond strength of the leak channel and the repair packing. The performance prediction anomaly result of the corresponding packing is obtained by multiplying the sum of the filtration loss impact and the value of 1 by the bond strength anomaly. The filtration loss risk impact coefficient is obtained by fitting historical data to calculate the average value.

[0012] This step calculates the abnormal results of the packing performance prediction by comprehensively considering two core indicators: filtration loss risk and interfacial bonding strength. It realizes the quantitative evaluation of the comprehensive sealing performance of the packing in specific leakage scenarios, making the performance comparison between different packings intuitive and operable.

[0013] In one implementation of this application, the future usage status assessment includes the following specific steps: The acidity of the CO2 acidic fluid at the corresponding location and the fluctuation of the ambient temperature at the corresponding location are obtained. The acidity is represented by pH, and the temperature fluctuation is represented by the average of the maximum temperature difference within a single day. At the same time, the suitable pH range and suitable temperature fluctuation range of the corresponding packing are obtained. The standard deviation of the acidity and ambient temperature fluctuations from the corresponding suitable ranges are obtained, and then the weighted sum is used to obtain the future reaction impact anomaly. The reaction impact is obtained by multiplying the future reaction impact anomaly by the environmental impact coefficient. The future use status anomaly assessment result is obtained by subtracting the environmental reaction impact on the corresponding packing from the value of 1 and multiplying it by the performance prediction anomaly result of the corresponding packing. The future use status anomaly assessment results of various packings are sorted in ascending order, and the packing with the smallest future use status anomaly assessment result is selected as the remediation packing and sent to the client. The environmental impact coefficient is also obtained by fitting the average of historical data.

[0014] This step achieves a comprehensive assessment from short-term sealing effectiveness to long-term service stability by considering the potential impact of future environmental conditions on packing durability and combining future reaction shock anomalies with current performance predictions.

[0015] Secondly, this application also provides an auxiliary design method for repairing leakage channels in CO2 sequestration geological bodies, including the following specific steps: Obtain information on the leak point in the leak channel, the gas conditions inside the channel, and the environmental conditions at the corresponding location; The filling and entry of various packing materials are analyzed based on the three-dimensional situation of the leak, the gas situation in the channel, and the environmental situation at the corresponding location. Predict the bonding performance of fillers based on the solidification of various fillers in different scenarios; Connection performance is predicted based on the filling conditions of various fillers and the predicted results of filler bonding performance. Based on the characteristics of the packing, the gas conditions in the channel, and the environmental conditions at the corresponding location, future reaction impacts are predicted, and the future service status is assessed based on the prediction results of future reaction impacts and connection performance. The selection of repair fillers is based on the assessment results of future usage conditions.

[0016] Thirdly, this application provides an electronic device comprising: a processor and a memory, wherein the memory stores a computer program that can be called by the processor, and the processor executes an auxiliary design method for repairing leakage channels in CO2-sealed geological bodies by calling the computer program stored in the memory.

[0017] Fourthly, this application provides a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform an auxiliary design method for repairing leakage channels in CO2-sealed geological bodies.

[0018] Compared with the prior art, this application has the following advantages: First, accurate multidimensional data of the channel is obtained, and its morphology is analyzed to pre-screen injectable fillers. Then, the interfacial bonding strength between the filler and the channel and the overall sealing performance are predicted, and the long-term stability is evaluated in combination with future environmental factors. Finally, the optimal filler is automatically recommended based on the quantitative evaluation results. This solution improves the remediation design from relying on experience to being data-driven, ensuring the remediation solution is targeted and long-lasting. Attached Figure Description

[0019] Other features, objects, and advantages of this application will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings: Figure 1 This is a schematic diagram of the overall process structure of the method in this application; Figure 2 This is a schematic diagram of the system structure of this application. Detailed Implementation

[0020] The technical solution of this application will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments and specific features in the embodiments are detailed descriptions of the technical solution of this application, rather than limitations thereof. In the absence of conflict, the embodiments and technical features in the embodiments can be combined with each other.

[0021] Please see Figures 1 to 2 , Figure 1 This is a schematic diagram of the overall process of the auxiliary design method for repairing leakage channels in CO2 sequestration geological bodies provided in this application embodiment, which specifically includes the following steps: Obtain information on the leak point in the leak channel, the gas conditions inside the channel, and the environmental conditions at the corresponding location; In this embodiment, the leakage channel and leakage point conditions include images of the leakage point obtained through acoustic scanning or downhole television, and 3D reconstruction software is used to reconstruct the 3D image of the leakage channel and leakage point to obtain the precise geometric shape of the leakage point; the gas conditions within the channel include dynamic parameters such as the phase (supercritical CO2, gas phase, dissolved phase), composition (CO2 concentration, impurity gases, formation water chemical ions), temperature, pressure, flow rate, and pH value of the leaking fluid obtained through sensors; the environmental conditions at the corresponding location include determining the geostress field, geothermal temperature, and mineral composition around the leakage point and channel, all of which can be obtained through a data acquisition terminal; The filling and entry of various packing materials are analyzed based on the three-dimensional situation of the leak, the gas situation in the channel, and the environmental situation at the corresponding location. In this embodiment, the filling and insertion of various fillers are analyzed, including the following specific steps: The first step involves obtaining the geometry of the leak outlet channel, including the number of bends, bend angles, average inner diameter, and smoothness anomalies of the channel's inner wall. The smoothness anomaly is calculated by dividing the height of any protrusions or depressions on the channel's inner surface by the average of the safe height, which is the maximum height that does not affect the flow velocity of the corresponding packing material. This safe height is obtained experimentally. The bend angle anomaly is obtained by dividing the channel bend angle by the safe bend angle. The overall bend anomaly is obtained by summing all the bend angle anomalies at the leak outlet. The safe bend angle is the maximum bend angle that does not affect the flow velocity of the corresponding packing material. The channel inner diameter anomaly is obtained by dividing the average inner diameter of the packing particles by the average inner diameter of the channel. The obstruction anomaly of the corresponding packing material filling the channel is obtained by weighted summing of the channel inner diameter anomaly, overall bend anomaly, and smoothness anomaly. This transforms the complex, qualitative channel geometry (torsional, rough, narrow) into quantifiable and calculable numerical indicators (overall bend anomaly, smoothness anomaly, channel inner diameter anomaly). This provides a basis for objectively comparing the injectability of different packing materials in different channels, based on fluid mechanics and particle transport theory. The flow resistance of particles or slurries in cracks / pipes is directly related to their geometry. Bending increases local energy loss, roughness affects the boundary layer and particle entrapment, and inner diameter determines whether particles can pass through and the flow shear rate. Quantifying these factors as outliers provides a simplified yet effective mathematical representation of their hindering effect.

[0022] The second step involves calling the reaction model in the repair filler database (which stores the basic formulations and properties of cement, polymer gels, resins, etc.) to simulate the hydration, gelation, carbonization, or cross-linking reaction process of the filler under corresponding CO2 concentrations, acidity, specific temperatures, and pressures. This allows obtaining the maximum flow-curing distance of the filler during injection at specific flow rates and viscosities under corresponding environments. This can be obtained by looking up historical experimental data. Simultaneously, multiple sets of experiments are conducted to observe how far the filler flows before curing in different environments. The average value is then calculated to obtain the final curing distance, which is the maximum flow-curing distance. The obstruction effect is obtained by multiplying the obstruction anomaly in the corresponding filler-filled channel by the obstruction influence coefficient. The actual flow-curing distance is obtained by subtracting the obstruction anomaly effect from the numerical value and then multiplying it by the maximum flow-curing distance. Finally, the filtration risk of the corresponding filler is obtained by dividing the internal depth of the leak channel by the actual flow-curing distance. The obstruction influence coefficient is obtained by averaging historical data. The actual flow-curing distance combines the inherent properties of the filler (experimentally measured maximum flow-curing distance) with the obstruction conditions of a specific channel, predicting the farthest distance the filler can effectively migrate and cure in that real channel. This is crucial for assessing the adequacy of the sealing range; the filtration loss risk is directly provided as a risk indicator by comparing the channel depth with the actual flow solidification distance. A ratio >1 means that the packing material may solidify before reaching the leak source, leading to sealing failure (filtration loss); the safe height / safe bending angle values ​​must be obtained through targeted indoor core flow experiments or computational fluid dynamics (CFD) simulations. For each candidate packing material (or representative packing materials classified by particle size distribution and rheological properties), flow experiments are conducted in simulated channels (such as cracked plates with different roughness, bends with different angles). The critical geometric parameters at which the packing slurry flow velocity begins to decrease significantly (e.g., a 10% decrease) or significant blockage occurs are observed and defined as the safe value for that packing material; this step, by quantitatively analyzing the geometric characteristics of the channel (such as bends, inner diameter, and smoothness) and combining them with the physical properties of the packing material, can scientifically assess the flow resistance and filtration loss risk during the injection of different packing materials, thereby pre-screening packing materials that are feasible for injection under specific channel morphologies; Predict the bonding performance of fillers based on the solidification of various fillers in different scenarios; In this embodiment, the prediction of filler bonding performance includes the following specific aspects: The first step involves calling the reaction model in the repair filler database (which stores the basic formulas and properties of cement, polymer gels, resins, etc.) to simulate the hydration, gelation, carbonization, or cross-linking reaction process of the filler under high CO2 concentration, acidity, and specific temperature and pressure conditions. At the same time, the stress field, ground temperature, number of bends, bend angle, average inner diameter of the channel, smoothness anomalies of the channel inner wall, and mineral composition of the leakage point and channel are called up to build a neural network prediction model to predict the bonding strength between the interface and the repair filler. The specific steps include: Building a repair filler database and calling the reaction model: Data source: Extracting the basic formula (e.g., chemical composition, additive ratio), physical properties (e.g., density, particle size), reaction kinetic parameters (e.g., reaction rate constant, activation energy), thermodynamic parameters (e.g., heat of reaction, phase diagram information), and constitutive model parameters (e.g., elastic modulus, Poisson's ratio, strength development model) of the target filler (e.g., specific cement or gel) from the repair filler database; Reaction simulation calling: Based on the above parameters, establishing a chemical-thermal-mechanical coupled finite element / finite volume model; Setting up in COMSOL Multiphysics, etc. The system is structured as follows: Chemical Module: Defines hydration, carbonization, and cross-linking reaction equations, considering CO2 concentration and pH (e.g., acidity) as reactant concentration variables; Heat and Mass Transfer Module: Simulates reaction heat, temperature field (e.g., geothermal temperature, reaction exothermics), pressure field, and the diffusion-convection process of reactants (carbonic acid, etc.) in porous media / cracks; Solid Mechanics Module: Couples volume changes and stress development caused by chemical reactions with the geostress field; Output: Through simulation, key field variables of the restoration's evolution over time / space are obtained, such as porosity distribution, hardness / modulus distribution, degree of reaction, and internal microstress field; these will serve as filler-side characteristics for subsequent prediction models.

[0023] Quantification of geological features of the leakage channel: Feature extraction: Transforming geological parameters into standardized, computable numerical feature vectors: Geostress field: Decomposed into maximum principal stress, minimum principal stress, and stress direction (three scalars); Geothermal field: Taking the average temperature and temperature gradient along the leakage channel; Channel geometry and topology: number of bends; bend angle sequence; average inner diameter; inner wall smoothness; mineral composition: XRD or SEM-EDS analysis of the channel inner wall to obtain the mass percentage of major minerals (such as quartz, calcite, and clay minerals); integrated output: a comprehensive geological feature vector; combining two vastly different types of input data—filler reaction evolution characteristics and geological channel static characteristics—a dual-stream feature fusion neural network is used; specific steps are as follows: Step 101: Analysis of Filler Evolution Feature Flow Based on Convolutional Neural Networks: Network Layer Construction and Parameter Settings: 3D Convolutional Layers: Function: Simultaneously extract local patterns in both spatial and temporal dimensions. For example, capturing the advancement of reaction fronts and the migration of stress concentration zones; Parameter Settings: Convolutional Kernel Size: For example, (3,3,3) (time, height, width); the first 3 indicates observing patterns over 3 consecutive time steps. Number of Filters: Start with a smaller number (e.g., 32) and gradually increase as the network deepens (e.g., 64 to 128). This allows the network to learn spatiotemporal features from simple to complex. Stride and Padding: The stride is usually set to (1,1,1) to preserve information, and the same padding is used to maintain the feature map size; 3D Batch Normalization Layer: Purpose: Stabilizes the training of deep networks, accelerates convergence, and is especially important for multi-physics data with varying input scales; Parameter Settings: Default settings, immediately following each 3D convolutional layer; Activation Function: Choose ReLU or its variants (such as LeakyReLU); ReLU provides sparse activation, alleviates gradient vanishing, and is computationally efficient; 3D Pooling Layer: Purpose: Reduces the resolution of the feature map in the spatiotemporal dimension, increases the receptive field of subsequent layers, and controls overfitting; Type and Parameters: Usually 3D max pooling is used, with a pooling window of (2,2,2) and a stride of 2; one is placed every 1-2 convolutional blocks; Spatiotemporal Global Average Pooling Layer: Purpose: At the end of the flow, the complex spatiotemporal feature map output by the last convolutional layer is averaged along the entire dimension, generating a scalar value for each channel. This transforms an input of arbitrary size into a fixed-length feature vector (length equal to the final number of channels), which is well-suited for handling mesh sizes that may vary in simulations. The output is a one-dimensional vector that serves as a high-level abstract feature of the filler flow. Step 102: Analysis of Geological Static Feature Flow Based on Fully Connected Neural Networks: Network Layer Construction and Parameter Settings: Fully Connected Layers; Function: To learn complex nonlinear combinations and interactions between features. For example, to learn the impact of high stress and high clay content on the interface. Parameter Settings: Number of Layers and Neurons: Funnel-shaped design is adopted. The number of neurons in the first layer can be close to the number of input features (e.g., 128 or 256 neurons in the first layer if the input is 20-dimensional), and the number gradually decreases in subsequent layers (e.g., 64 to 32); 2-4 hidden layers are usually sufficient; Weight Initialization: He initialization is used, which works well with the ReLU activation function; Batch Normalization Layer: Function: Also used for stable training, especially when the dimensions of the input features are very different (stress is in the MPa range, percentage is 0-1). Parameter Settings: Placed between each fully connected layer and the activation function; Activation Function Selection: ReLU is also used; Dropout Layer: Function: A key regularization method to prevent overfitting; Randomly shuts down a portion of neurons during training, forcing the network to learn more robust features; Parameter Settings: The dropout rate is usually set between 0.2 and 0.5; Step 103: The splicing layer simply splices the first two outputs along the feature dimension to form a joint feature vector, which is the most direct and effective fusion method. The fully connected fusion layer's function is to learn the higher-order correlation between two different types of features and ultimately map it to the prediction target. Parameter settings: It consists of 1-3 fully connected layers; the number of neurons in the first layer is approximately half or one-third of the length of the spliced ​​vector, and subsequent layers continue to reduce this number; each layer is followed by ReLU, BN, and Dropout (before the last layer); the final output layer: a single adhesion strength is represented by one neuron; key training parameters: the loss function is the mean squared error loss; if multiple intensity values ​​are predicted, the MSE can be calculated for each output and then summed; the optimizer is Adam, the initial learning rate is set to 1e-4 to 1e-3, and a learning rate scheduler is used to reduce the learning rate when training stalls; regularization: in addition to Dropout, L2 weight decay can be used in the fully connected layers; this solves the problem of fusing spatiotemporal evolution data (high-dimensional, structured) and geological static data (low-dimensional, vector). CNN streams excel at extracting deep spatiotemporal features from images and field data, while FCN streams excel at handling nonlinear relationships in tabular features. Through deep learning, models can learn extremely complex nonlinear relationships, such as the carbonation enhancement effect of a certain cement on a quartzite surface at low temperatures under high CO2 concentrations, or how geostress competes with gel shrinkage to affect the cementation of clay mineral interfaces. This is something that traditional empirical formulas or simple physical models cannot achieve. The second step is to obtain the bond strength between the repair filler and the repair filler after all the repair fillers are predicted by the model and enter the leak channel.

[0024] This step utilizes a dual-flow feature fusion neural network to deeply integrate and model the spatiotemporal characteristics of the chemical reaction evolution of the filler under complex environments with the static geological characteristics of the channel, enabling high-precision prediction of the final bond strength between the filler and the channel interface. Connection performance is predicted based on the filling conditions of various fillers and the predicted results of filler bonding performance. In this embodiment, the connection performance prediction includes the following specific aspects: The filtration loss impact is obtained by multiplying the filtration loss risk of the corresponding packing by the filtration loss risk impact coefficient. The bond strength anomaly is obtained by dividing the safe bond strength by the bond strength of the leak channel and the repair packing. The performance prediction anomaly result of the corresponding packing is obtained by multiplying the sum of the filtration loss impact and the value of 1 by the bond strength anomaly. The filtration loss risk impact coefficient is obtained by fitting historical data to calculate the average value. Based on the characteristics of the packing, the gas conditions in the channel, and the environmental conditions at the corresponding location, future reaction impacts are predicted, and the future service status is assessed based on the prediction results of future reaction impacts and connection performance. In this embodiment, the future usage status assessment includes the following specific steps: The acidity of the CO2 acidic fluid at the corresponding location and the fluctuation of the ambient temperature at the corresponding location are obtained. The acidity is represented by pH, and the temperature fluctuation is represented by the average of the maximum temperature difference within a single day. At the same time, the suitable pH range and suitable temperature fluctuation range of the corresponding packing are obtained. The standard deviation of the acidity and ambient temperature fluctuations from the corresponding suitable ranges are obtained, and then the weighted sum is used to obtain the future reaction impact anomaly. The reaction impact is obtained by multiplying the future reaction impact anomaly by the environmental impact coefficient. The future use status anomaly assessment result is obtained by subtracting the environmental reaction impact on the corresponding packing from the value of 1 and multiplying it by the performance prediction anomaly result of the corresponding packing. The future use status anomaly assessment results of various packings are sorted in ascending order, and the packing with the smallest future use status anomaly assessment result is selected as the remediation packing and sent to the client. The environmental impact coefficient is also obtained by fitting the average of historical data.

[0025] This step considers the potential impact of future environmental conditions (such as pH and temperature fluctuations) on the durability of the packing material and combines future reaction shock anomalies with current performance predictions to achieve a comprehensive assessment from short-term sealing effect to long-term service stability. This is just an example to illustrate future environmental conditions. There are many possible future environmental conditions, and this is just an example of the most influential ones. For example, the impact of moisture content is not considered here. This does not mean that these impacts are not considered in practice. In practice, all these impacts are considered and then weighted and summed to obtain the impact of future environmental conditions. It is just that they are not listed one by one in this embodiment. The selection of repair fillers will be based on the assessment results of future usage conditions. The advantages of the above embodiments are: accurate acquisition of multi-dimensional channel data, followed by analysis of its morphology for pre-screening of injectable fillers; prediction of the interfacial bonding strength between the filler and the channel and the overall sealing performance, and assessment of long-term stability in conjunction with future environmental factors; and finally, automatic recommendation of the optimal filler based on the quantitative evaluation results. This solution elevates the remediation design from relying on experience to being data-driven, ensuring the remediation solution's pertinence and long-term effectiveness.

[0026] Here, it is necessary to explain the terminology used in this embodiment. Standardization means dividing the corresponding data by the standard value of the corresponding parameter; standard deviation means dividing the absolute value of the difference between the numerical value and the midpoint of the range by the range quantity. Please see Figure 2 , Figure 2 This is a schematic diagram of the auxiliary design system for repairing leakage channels in CO2 sequestration geological bodies provided in this application embodiment, including: The system comprises a data acquisition module, a leak channel analysis module, a packing connection prediction module, a connection performance prediction module, a reaction assessment module, and a packing selection module. The data acquisition module acquires information about the leak outlet, the gas conditions within the channel, and the corresponding environmental conditions. The leak channel analysis module analyzes the filling and insertion of various packings based on the 3D characteristics of the leak outlet, the gas conditions within the channel, and the corresponding environmental conditions. The packing connection prediction module predicts the bonding performance of various packings based on their solidification in the given scenario. The connection performance prediction module predicts the connection performance based on the filling and insertion conditions of various packings and the predicted bonding performance. The reaction assessment module predicts future reaction impacts based on the characteristics of the packings, the gas conditions within the channel, and the corresponding environmental conditions, and assesses the future service life based on the predicted reaction impacts and connection performance. The packing selection module selects repair packings based on the future service life assessment results.

[0027] The parameters and steps for implementing the corresponding functions of each unit module in the auxiliary design system for repairing leakage channels of CO2 storage geological bodies described above can be referred to the parameters and steps in the embodiments of the auxiliary design method for repairing leakage channels of CO2 storage geological bodies described above, and will not be repeated here.

[0028] Embodiments of this application also provide an electronic device, including a memory, a processor, and a communication bus; the memory and the processor are connected via the communication bus. The memory stores an auxiliary design method for repairing leakage channels in CO2-sealed geological bodies, which can be loaded by the processor and executed as provided in the above embodiments.

[0029] The memory can be used to store instructions, programs, code, code sets, or instruction sets. The memory may include a program storage area and a data storage area. The program storage area may store instructions for implementing an operating system, instructions for at least one function, and instructions for implementing the auxiliary design method for repairing leakage channels in CO2-sealed geological bodies provided in the above embodiments, etc. The data storage area may store data involved in the auxiliary design method for repairing leakage channels in CO2-sealed geological bodies provided in the above embodiments, etc.

[0030] A processor may include one or more processing cores. The processor executes instructions, programs, code sets, or instruction sets stored in memory, and calls data stored in memory to perform various functions and process data as described in this application. The processor may be at least one of a specific application-specific integrated circuit, a digital signal processor, a digital signal processing device, a programmable logic device, a field-programmable gate array, a central processing unit, a controller, a microcontroller, and a microprocessor. It is understood that, for different devices, the electronic devices used to implement the above-described processor functions may also be other types, and the embodiments of this application do not specifically limit this.

[0031] A communication bus may include a pathway for transmitting information between the aforementioned components. The communication bus can be a PCI bus or an EISA bus, etc. Communication buses can be categorized into address buses, data buses, control buses, etc.

[0032] This application provides a computer-readable storage medium storing a computer program that can be loaded by a processor and executed as described in the above embodiments, which is an auxiliary design method for repairing leakage channels in CO2-sealed geological bodies.

[0033] In this embodiment, a computer-readable storage medium can be a tangible device that holds and stores instructions used by an instruction execution device. The computer-readable storage medium can be, but is not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any combination thereof. Specifically, the computer-readable storage medium can be a portable computer disk, a hard disk, a USB flash drive, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), staging random access memory (SRAM), portable compact disc read-only memory (CD-ROM), digital multifunction disc (DVD), memory stick, floppy disk, optical disk, magnetic disk, mechanical encoding device, or any combination thereof.

[0034] The term includes, or any other variation thereof, is intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0035] The above description is merely a preferred embodiment of this application and an explanation of the technical principles employed. Those skilled in the art should understand that the scope of this application is not limited to technical solutions formed by specific combinations of the above-described technical features, but should also cover other technical solutions formed by arbitrary combinations of the above-described technical features or their equivalents without departing from the foregoing application concept. For example, technical solutions formed by substituting the above features with (but not limited to) technical features with similar functions claimed in this application.

Claims

1. An auxiliary design method for repairing leakage channels in CO2 sequestration geological bodies, characterized in that, The specific steps include the following: Obtain information on the leak point in the leak channel, the gas conditions inside the channel, and the environmental conditions at the corresponding location; The filling and entry of various packing materials are analyzed based on the three-dimensional situation of the leak, the gas situation in the channel, and the environmental situation at the corresponding location. Predict the bonding performance of fillers based on the solidification of various fillers in different scenarios; Connection performance is predicted based on the filling conditions of various fillers and the predicted results of filler bonding performance. Based on the characteristics of the packing, the gas conditions in the channel, and the environmental conditions at the corresponding location, future reaction impacts are predicted, and the future service status is assessed based on the prediction results of future reaction impacts and connection performance. The selection of repair fillers is based on the assessment results of future usage conditions.

2. The auxiliary design method for repairing leakage channels in CO2 sequestration geological bodies according to claim 1, characterized in that, The analysis of the filling and insertion of various fillers includes the following specific steps: Obtain the geometry of the leak outlet channel, including the number of bends, bend angles, average inner diameter, and smoothness anomalies of the channel's inner wall. The smoothness anomaly is the average height of any protrusion or recess on the channel's inner surface divided by the safe height, which is the maximum height that does not affect the flow velocity of the corresponding packing material. This is obtained experimentally. Divide the channel bend angle by the safe bend angle to obtain the bend angle anomaly. Sum all the bend angle anomalies at the leak outlet to obtain the overall bend anomaly. The safe bend angle is the maximum bend angle that does not affect the flow velocity of the corresponding packing material. Divide the average inner diameter of the packing particles by the average inner diameter of the channel to obtain the channel inner diameter anomaly. Weighted summation of the channel inner diameter anomaly, overall bend anomaly, and smoothness anomaly when the corresponding packing material is filled into the channel yields the obstruction anomaly when the corresponding packing material is filled into the channel. The reaction model in the repair packing database is called to simulate the hydration, gelation, carbonization, or cross-linking reaction process of the packing under corresponding CO2 concentration, acidity, and specific temperature and pressure conditions. The maximum flow-curing distance of the packing during the injection process under specific flow rate and viscosity in the corresponding environment is obtained. The obstruction abnormality effect is obtained by multiplying the obstruction abnormality of the corresponding packing into the channel by the obstruction effect coefficient. The actual flow-curing distance is obtained by subtracting the obstruction abnormality effect from the value 1 and then multiplying it by the maximum flow-curing distance. The filtration loss risk of the corresponding packing is obtained by dividing the internal depth of the leak channel by the actual flow-curing distance.

3. The auxiliary design method for repairing leakage channels in CO2 sequestration geological bodies according to claim 1, characterized in that, The prediction of the filler bonding performance includes the following specific contents: The reaction model in the repair filler database is called to simulate the hydration, gelation, carbonization or cross-linking reaction process of the filler under high CO2 concentration, acidity and specific temperature and pressure conditions. At the same time, the stress field, ground temperature, number of bends, bend angle, average inner diameter of the channel, smoothness anomaly of the inner wall of the channel and mineral composition of the leakage point and channel are called to construct a neural network prediction model to predict the bonding strength between the interface and the repair filler. Obtain the bond strength between the repair filler and the repair filler after all the repair filler materials predicted by the model enter the leak outlet channel.

4. The auxiliary design method for repairing leakage channels in CO2 sequestration geological bodies according to claim 1, characterized in that, The connectivity performance prediction includes the following specific details: The filtration loss effect is obtained by multiplying the filtration loss risk of the corresponding packing by the filtration loss risk influence coefficient. The bond strength anomaly is obtained by dividing the safe bond strength by the bond strength of the leak channel and the repair packing. The performance prediction anomaly result of the corresponding packing is obtained by multiplying the sum of the filtration loss effect and the value of 1 by the bond strength anomaly.

5. The auxiliary design method for repairing leakage channels in CO2 sequestration geological bodies according to claim 1, characterized in that, The future usage status assessment includes the following specific steps: The acidity of the CO2 acidic fluid at the corresponding location and the fluctuation of the ambient temperature at the corresponding location are obtained. The acidity is represented by pH, and the temperature fluctuation is represented by the average of the maximum temperature difference within a single day. At the same time, the suitable pH range and suitable temperature fluctuation range of the corresponding packing are obtained. The standard deviations of the acidity and ambient temperature fluctuations and the corresponding suitable ranges are obtained respectively. Then, the weighted sum is used to obtain the future reaction impact anomaly. The reaction impact is obtained by multiplying the future reaction impact anomaly by the environmental impact coefficient. The value of 1 is subtracted from the reaction impact of the environment on the corresponding packing and multiplied by the performance prediction anomaly result of the corresponding packing to obtain the future use status anomaly assessment result. The future use status anomaly assessment results of various packings are sorted in ascending order. The packing with the smallest future use status anomaly assessment result is selected as the repair packing and sent to the client.

6. The auxiliary design method for repairing leakage channels in CO2 sequestration geological bodies according to claim 1, characterized in that, The leakage channel and leakage point conditions include images of the leakage point obtained through acoustic scanning or downhole television, and 3D reconstruction software is used to reconstruct the 3D image of the leakage channel and leakage point to obtain the precise geometric shape of the leakage point; the gas conditions within the channel include dynamic parameters of the leakage fluid such as phase, composition, temperature, pressure, flow rate, and pH value obtained through sensors; the environmental conditions at the corresponding location include determining the geostress field, geothermal temperature, and mineral composition around the leakage point and channel.

7. A system for auxiliary design of repairing leakage channels in CO2-sealed geological bodies, used to implement the auxiliary design method for repairing leakage channels in CO2-sealed geological bodies as described in any one of claims 1-6, characterized in that, Includes the following modules: The system comprises a data acquisition module, a leakage channel analysis module, a packing connection prediction module, a connection performance prediction module, a reaction assessment module, and a packing selection module. The data acquisition module acquires information about the leak outlet, the gas conditions within the channel, and the corresponding environmental conditions. The leakage channel analysis module analyzes the filling and insertion of various packing materials based on the three-dimensional information of the leak outlet, the gas conditions within the channel, and the corresponding environmental conditions. The packing connection prediction module predicts the packing bonding performance based on the solidification of various packing materials in the given scenario. The connection performance prediction module predicts the connection performance based on the filling and insertion conditions of various packing materials and the predicted packing bonding performance. The reaction assessment module predicts future reaction impacts based on the characteristics of the packing materials, the gas conditions within the channel, and the corresponding environmental conditions, and assesses the future service life based on the predicted future reaction impacts and connection performance. The packing selection module selects repair packing materials based on the future service life assessment results.

8. An electronic device, comprising: A processor and a memory, wherein the memory stores a computer program that can be called by the processor; characterized in that the processor executes the auxiliary design method for repairing leakage channels of CO2 sequestration geological bodies as described in any one of claims 1-6 by calling the computer program stored in the memory.

Images