A simulation method and system for CO2 storage filling and sealing

Abstract

The application discloses a CO2 storage filling and sealing simulation method and a simulation system, and belongs to the technical field of carbon sealing and mine filling. The method comprises the following steps: establishing an engineering and configuring engineering parameters; performing a main process evaluation of a goaf, wherein the main process evaluation comprises design checking, process checking, injection reaction sealing simulation and risk evaluation; wherein the design checking determines the injection pressure boundary according to the chain relationship of span ratio, allowable span, plastic zone radius, bearing width, equivalent bearing pressure, support safety factor and maximum allowable injection pressure; the process checking determines whether the injection sealing is allowed to start based on the time sequence dependent relationship of construction, filling, curing, sealing detection, injection sealing and pumping pressure regulation; when the support safety factor is insufficient, the sealing detection is unqualified, the peak pressure is out of limit or the risk level is too high, the system automatically triggers parameter rollback adjustment and / or enhanced analysis, and the results are fed back to the main process for reiteration; the application can integrate the goaf structure stability, process sealing, injection reaction sealing and risk control into a closed-loop workflow, and improve the efficiency and reliability of the early-stage demonstration, scheme comparison, trial operation early warning and on-site decision support of the CO2 sealing scheme of the goaf.

Description

Technical Field

[0001] This invention relates to the fields of carbon capture, utilization and storage (CCUS) and functional backfilling technology in mines, specifically to a simulation method and system for CO2 storage backfilling. Background Technology

[0002] Coal mine goafs, as underground space resources, have always been a research hotspot in mine ecological restoration and resource utilization. Transforming goafs into CO2 sequestration storage facilities can not only achieve large-scale solid waste disposal, but also promote the green and low-carbon transformation of the coal industry, with significant environmental and economic benefits. At present, research on CO2 sequestration in goafs mainly focuses on numerical simulation of single physical fields (such as seepage field and stress field) or laboratory-scale reaction kinetic tests, and there is a lack of systematic simulation tools from the perspective of rapid engineering evaluation.

[0003] In existing technologies, some studies use general-purpose finite element or finite difference software (such as FLAC and COMSOL) to simulate the stability of goaf areas or CO2 migration. However, such software usually requires professional modeling, a large number of input parameters, and a long calculation cycle, making it difficult to quickly complete the comparison and preliminary demonstration of multiple schemes in the early stages of engineering. In addition, the construction of goaf reservoirs involves multiple aspects such as process flow (construction, filling, injection and sealing), design verification (roof span, coal pillar safety factor), CO2 injection, reaction and sealing coupling process, and leakage risk assessment. Existing software often only covers one part of these aspects and lacks a unified evaluation platform that integrates design, process, reaction, risk, local multiphysics analysis and monitoring data playback. Moreover, existing technologies generally lack quantitative processing of geological parameter uncertainties, lack cross-scale multiphysics adaptive solution strategies, and lack global multi-objective optimization based on artificial intelligence and real-time error compensation mechanisms based on physical information neural networks, making it difficult to cope with the complex working conditions in actual mines and provide highly reliable dynamic decision support.

[0004] Therefore, how to provide a simulation method that can perform closed-loop coupled assessment of the unique characteristics of the roof-filling body-floor joint bearing, airtightness detection, injection pressure constraint and risk feedback adjustment in coal mine goaf areas, and automatically trigger local multiphysics field verification and digital twin prediction and deduction when risks or curve deviations occur, has become a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the prior art and provide a CO2 storage filling and sealing simulation method and simulation system.

[0006] To achieve the above objectives, the present invention provides the following technical solution: This application provides a method for simulating the filling and sealing of CO2 storage facilities, including: Establish a project and configure project parameters, which include at least the geometric parameters of the goaf, the mechanical parameters of the roof, the bearing capacity parameters of the floor, the parameters of the filling body, the airtightness testing parameters, the injection and drainage process parameters, and the reaction and storage parameters. The main process assessment of the goaf is performed, which includes design verification, process validation, injection reaction sealing simulation, and risk assessment. The design verification determines the injection pressure boundary in a chain-like sequence, and the process validation determines whether injection sealing is allowed to start based on a time-series dependency graph. The maximum permissible injection pressure determined by the design verification and the process validation results are used as the injection reaction sealing criteria. The simulation constraints are used to calculate pressure, gaseous CO2 mass and mineralization, and the risk level is determined through risk assessment. When the safety factor of the support is insufficient, the airtightness test fails, the peak pressure exceeds the limit, or the risk level is too high, the parameter rollback adjustment and / or enhancement analysis are automatically triggered, and the adjustment results are fed back to the main process for re-iteration. Export the evaluation results.

[0007] Furthermore, this invention introduces an uncertainty quantification method based on Monte Carlo simulation and geostatistics, which sets key parameters such as burial depth and tensile strength as random variables, and outputs failure probability index, reliability index and comprehensive risk probability index through multiple sets of parallel calculations, thereby replacing the traditional single deterministic safety coefficient and risk level evaluation.

[0008] Furthermore, in the simulation of injection reaction sealing, this invention adopts an adaptive time step dynamic control strategy based on rate of change feedback, which monitors various rates of change in real time and dynamically adjusts the calculation time step. At the same time, it constructs a macro-micro cross-scale coupled model and automatically switches to a micro discrete model in key regions, thus achieving an intelligent balance between computational accuracy and efficiency.

[0009] Furthermore, this invention adds a real-time rapid deduction and multi-objective optimization unit based on an artificial intelligence agent model to the closed-loop coupled control module. It uses a deep neural network to train a lightweight agent model and takes maximizing the safety factor, minimizing pressure fluctuations, and maximizing the cumulative mineralization as multi-objective functions. It uses a particle swarm optimization algorithm to output the globally optimal real-time control strategy, enabling the system to have the ability to actively seek optimization and self-evolve.

[0010] Furthermore, this invention triggers an adaptive dynamic plugging and pressure oscillation synergistic strategy when local fracture leakage occurs. It dynamically calculates the gel plugging agent parameters and regulates the wellhead pressure to generate periodic pressure oscillation pulses, thereby fully utilizing the mineralization reaction space. In the digital twin analysis stage, a real-time error compensation mechanism based on a Physical Information Neural Network (PINN) is introduced to construct a dynamic error compensation field to generate a twin-corrected solution and output a probabilistic risk warning containing a confidence interval, thus significantly improving the reliability of early warnings for abnormal operating conditions in the field.

[0011] Through the above-mentioned technical means, this invention upgrades traditional numerical simulation into an intelligent closed-loop decision-making system with the capabilities of perception, optimization, self-correction, and probabilistic early warning.

[0012] Preferably, the design verification includes calculating the span ratio, theoretical allowable span of the plate, theoretical allowable span of the beam, recommended span, plastic zone radius, bearing width, equivalent bearing pressure, support safety factor, corrected support safety factor, and maximum allowable injection pressure in sequence according to a chain decision-making order based on the following formulas:

[0013]

[0014]

[0015]

[0016]

[0017]

[0018]

[0019]

[0020]

[0021]

[0022] Where a represents the actual span of the goaf reservoir, L represents the reservoir length, λ represents the span ratio, h represents the roof thickness, ν represents Poisson's ratio, and σ t α represents the tensile strength of the top plate, q represents the uniformly distributed load, and α represents the tensile strength of the top plate. s a represents the theoretical parameters of the beam related to the strength and thickness of the top slab. plate The plate theory allows for a span, a beam This indicates that beam theory allows for a certain span, a rec Indicates the recommended span, β a r represents the span safety reduction factor. pH represents the radius of the plastic region. c Indicates the height of the coal pillar or backfill support, α p φ represents the plastic zone coefficient, K represents the internal friction angle, K represents the lateral pressure coefficient, γ represents the bulk density, H represents the burial depth, θ represents the stress reduction angle, c represents the cohesion, b represents the bearing width, and σ represents the internal pressure coefficient. c P represents the compressive strength of the support. eq FS represents the equivalent bearing pressure, and FS represents the safety factor of the support. ′ χ represents the corrected safety factor of the support structure. s P represents the local stress concentration factor. inj,max Indicates the maximum permissible injection pressure, η p P represents the injection pressure reduction factor. ref Indicates the reference allowable pressure, FS th The safety factor threshold is represented by the span ratio, allowable span, plastic zone radius, bearing width, equivalent bearing pressure, support safety factor and maximum allowable injection pressure being sequentially related so that the design verification results can constrain the pressure boundary conditions of the injection reaction sealing simulation.

[0023] Preferably, the injection reaction sealing simulation includes calculating the pressure and mineralization amount based on the following formulas, and using the maximum allowable injection pressure obtained from the design verification as the pressure boundary condition:

[0024]

[0025]

[0026] Where P represents the storage pressure, n represents the total amount of gas in the storage, R represents the ideal gas constant, T represents the temperature, and V represents the volume of gas. eff Δn represents the effective volume of the storage facility. min k represents the change in the amount of CO2 caused by the mineralization reaction. r f represents the reaction rate constant. h Represents the humidity factor, n g C represents the amount of gaseous CO2. rem C0 represents the initial reaction capacity, Δt represents the calculation time step, and P represents the remaining reaction capacity. inj,max It indicates the maximum allowable injection pressure for design verification feedback; and outputs the pressure time history, gas phase CO2 mass time history, cumulative mineralization time history, porosity and pH change trends.

[0027] Preferably, the process verification includes: constructing a time-series dependency diagram comprising the construction stage, filling stage, curing stage, airtightness testing stage, injection and storage stage, and extraction and pressure regulation stage; verifying, based on the time-series dependency diagram, whether the preconditions, ventilation status, airtight door status, curing strength of the filling body, airtightness testing results, injection / extraction switches, and flow parameters of each stage meet the process logic of the goaf CO2 storage facility; when there are process parameters that violate the time-series dependency, the system locates conflict nodes and conflict constraint edges, and generates a parameter correction scheme based on the injection reaction and storage simulation results.

[0028] Preferably, the risk assessment includes: identifying risks of structural instability in the goaf, sealing failure, local CO2 accumulation, fault fracture leakage, pipeline leakage, and ecological impact; mapping the probability and severity of each risk type to a risk matrix to determine the risk level; when any risk level is high or extremely high, the system feeds back the corresponding risk item to the closed-loop coupled control process to trigger local multiphysics analysis, injection flow adjustment, drainage and pressure regulation, local sealing, or design parameter rollback adjustment.

[0029] Preferably, the local multiphysics analysis includes: constructing a two-dimensional mesh model for a local region automatically determined by high-risk terms, setting boundary conditions and geological features, and performing thermal, hydrological, and mechanical coupling; The system performs analysis or a combination of thermal, hydrological, mechanical, and chemical analyses, and outputs at least one of the following: pressure field, component concentration field, reaction progress field, porosity field, permeability field, instability index field, and leakage flux. The pressure field is used to determine the pressure concentration factor, the instability index field is used to correct the support safety factor, the leakage flux is used to correct the risk level, and the permeability field is used to correct the equivalent permeability parameters in the injection reaction sealing simulation.

[0030] Preferably, the digital twin analysis includes: when the deviation of the pressure curve, gas phase CO2 mass curve, or cumulative mineralization curve from the preset theoretical trend exceeds a set threshold, importing sensor observation data, performing quality control and time-series alignment on the observation data, mapping the sensor data to model observation variables, using ensemble Kalman filtering or least squares methods for parameter assimilation, and performing prediction and inference based on the assimilated model to output the pressure change trend, gas phase CO2 enrichment trend, leakage risk change trend, and the handling effect under different emergency measures within a preset time period.

[0031] Preferably, the exported evaluation results include: generating an evaluation report that includes at least an engineering overview, design verification results, scheme comparison table, process verification events, airtightness test results, response trends, risk conclusions, closed-loop iterative process, and key control parameters, and / or exporting one or more of the following according to the user: a one-page decision summary, detailed technical report, field process operation card, diagnostic data package, grid snapshot file, time series data file, and twin result file.

[0032] Preferably, it further includes: performing formula verification and parameter sensitivity analysis steps, comparing the calculation results of the preset reference case with the expected value, outputting the verification pass rate and error information, and calculating the sensitivity of each input parameter to the support safety factor, peak pressure and final mineralization amount, and outputting the sensitivity ranking and key control parameters; the key control parameters are used to determine the priority adjustment order of automatic adjustment and scheme comparison selection.

[0033] This application also provides a CO2 storage filling and sealing simulation system, including: The project management module is used to create or open projects and configure project parameters; The main process evaluation module includes a design verification unit, a process validation unit, an injection reaction sealing simulation unit, and a risk assessment unit. The enhanced analysis module includes a local multiphysics analysis unit and a digital twin analysis unit; The closed-loop coupling control module is used to transmit feedback parameters between the design verification unit, process verification unit, injection reaction sealing simulation unit, risk assessment unit, and enhancement analysis module, and automatically adjust design parameters, injection boundary conditions, or process control parameters based on the feedback results. At the same time, the closed-loop coupling control module is equipped with a real-time rapid deduction and multi-objective optimization unit based on an artificial intelligence agent model, which is used to perform digital twin self-correction and error feedback closed-loop correction. The results export module is used to output evaluation reports, scheme comparison tables, process operation cards, and data files. The system is used to perform a CO2 storage filling and sealing simulation method.

[0034] Compared with the prior art, this application has the following beneficial effects: This invention proposes a CO2 storage filling and sealing simulation method and system. After establishing a project and configuring parameters, a main process assessment is executed, including design verification, process validation, injection reaction sealing simulation, and risk assessment. The system also integrates the collaborative bearing capacity verification of the goaf roof-fill body-floor plate, airtightness testing, maximum allowable injection pressure constraints, and risk feedback adjustments into the same closed-loop workflow. When the safety factor is insufficient, peak pressure exceeds limits, airtightness testing fails, or the risk level is too high, the system can automatically generate [response data]. The system generates design adjustment schemes, flow adjustment schemes, drainage and pressure regulation schemes, or local sealing schemes, and re-executes subsequent evaluations until the risk is reduced to an acceptable range or the number of iterations is exhausted. Furthermore, local multiphysics field analysis and digital twin analysis are automatically triggered by the main process results, and their results are used to reverse correct the equivalent bearing pressure, safety factor, equivalent permeability parameters, maximum allowable injection pressure, and risk level. This allows the local heterogeneous characteristics of the goaf, monitoring data deviations, and the effectiveness of emergency measures to all be included in the main process evaluation, improving the reliability of the goaf CO2 storage pre-construction demonstration, scheme comparison, trial operation early warning, and on-site decision support.

[0035] Furthermore, this invention resolves the contradiction between computational efficiency and accuracy by introducing an adaptive time step based on rate of change feedback and a macro-micro cross-scale coupling model; it addresses the engineering risks caused by the unmeasurability of geological parameters by introducing geological parameter uncertainty and probabilistic risk indicators based on Monte Carlo simulation; and it enables the system to possess the capabilities of proactive optimization, dynamic state perception, and self-evolution by introducing multi-objective real-time optimization based on surrogate models and digital twin technology based on PINN error compensation, thus achieving a leap from traditional numerical simulation to intelligent perception, intelligent control decision-making, and intelligent evolution. Attached Figure Description

[0036] Figure 1 This is a schematic diagram of the simulation system and its closed-loop coupling control relationship provided in the embodiments of this application. Detailed Implementation

[0037] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0038] Furthermore, in this invention, an element referred to as fixed to or disposed on another element may be directly disposed on the other element, or there may be an intermediate element. When an element is considered to be connected to another element, it may be directly connected to the other element, or there may be an intermediate element present simultaneously. The terms vertical, horizontal, left, right, and similar expressions used herein are for illustrative purposes only and do not represent the only possible implementation.

[0039] See Figure 1 This application provides a CO2 storage filling and sealing simulation method, including: steps S1-S5: Step S1. Create a project and configure project parameters; Step S2. Perform the main process assessment of the goaf, which includes design verification, process validation, injection reaction sealing simulation, and risk assessment. Step S3. Using the maximum allowable injection pressure determined by the design verification and the process verification results as constraints for the injection reaction sealing simulation, calculate the pressure, gaseous CO2 mass and mineralization amount, and determine the risk level through risk assessment; Step S4. When the safety factor of the support is lower than the preset safety threshold, the airtightness test fails, the peak pressure exceeds the maximum allowable injection pressure, or any risk item reaches the preset high risk level, the system automatically triggers parameter rollback adjustment and / or enhancement analysis, and feeds the adjustment results back to the main process for re-iteration; Step S5. Export the evaluation results.

[0040] In practical applications, users first create a new project or open an existing project through the graphical interface and configure the project parameters.

[0041] Engineering parameters include geological parameters such as burial depth, top plate thickness, Poisson's ratio, uniformly distributed load, and volumetric weight; structural parameters such as span, length, column height, and plastic zone parameters; material parameters such as compressive strength of the support, reaction components, reaction rate constant, and humidity factor; and process parameters such as three-stage duration, gate status, injection switch, extraction switch, and injection mass flow rate. After the parameters are configured, the software executes the main process evaluation, which sequentially runs design verification, process validation, injection reaction sealing simulation, and risk assessment.

[0042] Design verification is performed in a chain-like sequence based on span ratio, allowable span, recommended span, plastic zone radius, bearing width, equivalent bearing pressure, support safety factor, and maximum allowable injection pressure. Further, a coordinated bearing capacity verification of the top plate, filling body, and bottom plate is conducted. Process validation constructs a time-series dependency diagram encompassing construction, filling, curing, airtightness testing, injection and sealing, and pressure regulation. It automatically identifies conflict nodes and recommends corrective measures when pre-constraints are violated. Injection reaction and sealing simulation uses the maximum allowable injection pressure as the pressure boundary. Risk assessment determines the risk level. When the risk level is too high or the pressure and mineralization curves deviate beyond limits, the system automatically triggers local multiphysics analysis or digital twin analysis, feeding back the pressure concentration factor, instability index, leakage flux, permeability correction factor, and assimilation prediction results to the main process. The assessment is re-executed until the risk is acceptable or the number of iterations is exhausted.

[0043] In this embodiment, the design verification and risk assessment adopts an uncertainty quantification method based on Monte Carlo simulation and geostatistics. The key input parameters are set as random variables under probability distribution. Through a large number of parallel calculations, the failure probability index and comprehensive risk probability index are output to solve the problem that the single safety factor in conventional methods cannot truly reflect the random fluctuations of geological parameters.

[0044] Meanwhile, the injection reaction sealing simulation in this embodiment introduces an adaptive time step dynamic control strategy based on rate of change feedback, which monitors the rate of change of pressure, concentration and mineralization rate in real time, dynamically adjusts the time step, and uses a macro-micro cross-scale coupling model to automatically switch to a micro discrete model in fault and injection point regions, effectively balancing the overall simulation accuracy and computational resource consumption.

[0045] In one specific implementation, the design verification includes calculating the span ratio, theoretical allowable span of the plate, theoretical allowable span of the beam, recommended span, plastic zone radius, bearing width, equivalent bearing pressure, support safety factor, modified support safety factor, and maximum allowable injection pressure in sequence according to a chain decision-making order based on the following formulas:

[0046]

[0047]

[0048]

[0049]

[0050]

[0051]

[0052]

[0053]

[0054]

[0055] Where a represents the actual span of the goaf reservoir, L represents the reservoir length, λ represents the span ratio, h represents the roof thickness, ν represents Poisson's ratio, and σ t α represents the tensile strength of the top plate, q represents the uniformly distributed load, and α represents the tensile strength of the top plate. s a represents the theoretical parameters of the beam related to the strength and thickness of the top slab. plate The plate theory allows for a span, a beam This indicates that beam theory allows for a certain span, a rec Indicates the recommended span, β ar represents the span safety reduction factor. p H represents the radius of the plastic region. c Indicates the height of the coal pillar or backfill support, α p φ represents the plastic zone coefficient, K represents the internal friction angle, K represents the lateral pressure coefficient, γ represents the bulk density, H represents the burial depth, θ represents the stress reduction angle, c represents the cohesion, b represents the bearing width, and σ represents the internal pressure coefficient. c P represents the compressive strength of the support. eq FS represents the equivalent bearing pressure, and FS represents the safety factor of the support. ′ χ represents the corrected safety factor of the support structure. s P represents the local stress concentration factor. inj,max Indicates the maximum permissible injection pressure, η p P represents the injection pressure reduction factor. ref Indicates the reference allowable pressure, FS th The safety factor threshold is represented by the span ratio, allowable span, plastic zone radius, bearing width, equivalent bearing pressure, support safety factor and maximum allowable injection pressure being sequentially related so that the design verification results can constrain the pressure boundary conditions of the injection reaction sealing simulation.

[0056] In this embodiment, the system reads the reservoir span, reservoir length, roof thickness, Poisson's ratio, roof tensile strength, uniformly distributed load, plastic zone parameters, coal pillar or backfill support parameters, backfill compressive strength, backfill elastic modulus, floor bearing parameters, and injection pressure control parameters. Subsequently, the system performs design verification in a chain sequence of span ratio—allowable span—recommended span—plastic zone radius—bearing width—equivalent bearing pressure—safety factor—maximum allowable injection pressure.

[0057] First, the system calculates the span ratio λ and substitutes it into both the slab-theoretical allowable span model and the beam-theoretical allowable span model to obtain the slab-theoretical allowable span and the beam-theoretical allowable span, respectively. Then, the smaller of the two values ​​is taken and multiplied by a safety reduction factor to obtain the recommended span. The recommended span is compared with the actual span input by the user. When the actual span is greater than the recommended span, the system determines that the current span scheme has a risk of roof instability and considers reducing the span as a candidate for automatic adjustment.

[0058] Secondly, the system calculates the radius of the plastic zone around the coal pillar or backfill support, and uses the actual span and the radius of the plastic zone together to calculate the bearing width; the system calculates the equivalent bearing pressure based on the bearing width and the overburden load, and then calculates the safety factor of the support by the ratio of the compressive strength of the support to the equivalent bearing pressure; when the local multiphysics analysis outputs the local pressure concentration factor or stress concentration factor, the system further uses the corrected safety factor for judgment.

[0059] Next, the system determines the maximum allowable injection pressure based on the corrected safety factor and feeds this maximum allowable injection pressure back to the injection reaction sealing simulation. The lower the safety factor, the lower the maximum allowable injection pressure; thus, the structural stability results of the goaf can directly constrain the subsequent injection boundary conditions.

[0060] When the safety factor of the support is lower than the preset safety threshold, the system automatically generates multiple sets of candidate adjustment schemes, including reducing the span of the storage tank, increasing the width of the coal pillar or backfill support, increasing the compressive strength of the backfill, extending the solidification time of the backfill, and reducing the maximum allowable injection pressure. For each set of candidate schemes, the system recalculates the safety factor, effective volume of the storage tank, volume loss rate, and maximum allowable injection pressure, and generates a scheme comparison table.

[0061] Furthermore, the system performs a coordinated load-bearing verification of the top plate, filling body, and bottom plate. It estimates the deflection of the top plate based on the equivalent bearing pressure, estimates the compression of the filling body based on the elastic modulus and height of the filling body, estimates the additional deformation of the bottom plate based on the bearing parameters of the bottom plate, and determines whether the deflection of the top plate can be coordinated with the compression of the filling body and the additional deformation of the bottom plate. When the coordinated deformation condition is not met, the system upgrades the structural instability risk level and generates adjustment schemes to increase the strength of the filling body, increase the support width, or reduce the injection pressure.

[0062] In one specific embodiment, the injection reaction sequestration simulation includes calculating the pressure and mineralization based on the following formula, and using the maximum allowable injection pressure obtained from the design verification as the pressure boundary condition:

[0063]

[0064]

[0065] Where P represents the storage pressure, n represents the total amount of gas in the storage, R represents the ideal gas constant, T represents the temperature, and V represents the volume of gas. eff Δn represents the effective volume of the storage facility. min k represents the change in the amount of CO2 caused by the mineralization reaction. r f represents the reaction rate constant. h Represents the humidity factor, n g C represents the amount of gaseous CO2. rem C0 represents the initial reaction capacity, Δt represents the calculation time step, and P represents the remaining reaction capacity. inj,max It indicates the maximum allowable injection pressure for design verification feedback; and outputs the pressure time history, gas phase CO2 mass time history, cumulative mineralization time history, porosity and pH change trends.

[0066] Specifically, the system treats the goaf storage as a lumped volume, assumes the gas is an ideal gas, reads the temperature, effective storage volume, total amount of material, and maximum allowable injection pressure from the design verification feedback, calculates the pressure according to the ideal gas law, and determines whether the pressure meets the maximum allowable injection pressure constraint.

[0067] The mineralization amount is calculated using a proportional relationship. The change in CO2 mass caused by the mineralization reaction is proportional to the reaction rate constant, humidity factor, amount of gaseous CO2, and the ratio of remaining reaction capacity to initial reaction capacity. During the injection phase, the injected mass flow rate is converted into an increase in the number of CO2 moles and added to the total mass. During the extraction phase, the number of moles discharged is converted into the current average molecular weight of the mixed gas and deducted from the total mass.

[0068] When the peak pressure exceeds the maximum allowable injection pressure or the pressure rise rate exceeds the threshold, the system calculates the allowable injection mass flow rate range based on the effective volume of the storage, temperature, current amount of gaseous CO2, mineralization reaction consumption, and pumping rate, and feeds the calculated allowable injection mass flow rate back to the process verification and closed-loop coupling control module.

[0069]

[0070] The software outputs the cumulative injection volume curve, cumulative mineralization volume curve, residual gas phase CO2 mass curve, pressure time history curve, porosity change trend curve, pH change trend curve, and capacity utilization index. When the cumulative injection volume continues to rise while the cumulative mineralization volume grows slowly, it indicates insufficient reaction utilization. When the residual gas phase CO2 mass remains at a high level for a long time, it indicates excessive free CO2 and the risk of leakage should be monitored.

[0071] In one specific implementation, the process verification includes: constructing a time-series dependency diagram comprising the construction stage, filling stage, curing stage, airtightness testing stage, injection and storage stage, and extraction and pressure regulation stage; verifying, based on the time-series dependency diagram, whether the preconditions, ventilation status, airtight door status, curing strength of the filling body, airtightness testing results, injection / extraction switches, and flow parameters of each stage meet the process logic of the goaf CO2 storage facility; when there are process parameters that violate the time-series dependency, the system locates conflict nodes and conflict constraint edges, and generates a parameter correction scheme based on the injection reaction and storage simulation results.

[0072] In this embodiment, the system constructs a time-series dependency diagram of the CO2 storage process in the goaf. This time-series dependency diagram includes the construction stage, filling stage, solidification and curing stage, airtightness testing stage, injection and sealing stage, and extraction and pressure regulation stage; preconditions, prohibition conditions, and parameter constraints are set between each stage.

[0073] During the construction phase, the system verifies whether the construction passage, ventilation status, and airtight door opening status meet the construction safety requirements; during the filling phase, the system verifies whether the filling material layout, filling flow rate, filling height, and lower roadway airtight door status meet the filling process requirements; during the curing phase, the system estimates the curing strength of the filling body based on the filling material type, curing time, temperature, and humidity, and determines whether it has reached the minimum strength threshold before injection.

[0074] During the airtightness testing phase, the system determines whether the goaf reservoir is ready for injection based on the pressure holding test pressure, test duration, pressure decay rate, and allowable leakage rate; the leakage rate can be estimated using the following formula:

[0075] Among them, Q leak P represents the leakage rate, P0 represents the initial pressure of the pressure holding test, and P t The pressure at which the pressure holding test ends, Δt d This indicates the duration of the test. When the airtightness test fails, the system will prevent the injection and sealing phase from proceeding and will output a corrective plan, such as re-pressure holding test, repairing the airtight door, performing partial sealing, or reducing the injection pressure.

[0076] When the process parameters configured by the user violate the timing dependency, the system automatically locates the conflict nodes and conflict constraint edges; when the injected mass flow rate causes the expected peak pressure to exceed the maximum allowable injection pressure, the system calculates the allowable range of injected mass flow rate based on the maximum allowable injection pressure and outputs the recommended injection flow rate.

[0077] In one specific implementation, the risk assessment includes: identifying risks of structural instability in the goaf, sealing failure, local CO2 accumulation, fault fracture leakage, pipeline leakage, and ecological impact; mapping the probability and severity of each risk type to a risk matrix to determine the risk level; when any risk level is high or extremely high, the system feeds back the corresponding risk item to the closed-loop coupled control process to trigger local multiphysics analysis, injection flow adjustment, drainage and pressure regulation, local sealing, or design parameter rollback adjustment.

[0078] In this embodiment, the likelihood level and severity level of each risk item are determined based on engineering parameters, design verification results, process validation results, and injection reaction sealing simulation results. The risk level is then obtained through risk matrix mapping. The risk assessment results serve as feedback input for closed-loop coupled control. When any risk item is high-risk or extremely high-risk, the system automatically determines a fallback path based on the risk type: for structural instability risk, it falls back to design verification and adjusts the span, support width, filling strength, or maximum allowable injection pressure; for sealing failure risk, it falls back to process validation and adds sealing detection or plugging measures; for fault fracture leakage risk, it triggers local multiphysics analysis and corrects the risk matrix and injection boundary conditions based on the leakage flux.

[0079] In one specific implementation, the local multiphysics analysis includes: constructing a two-dimensional mesh model for a local area automatically determined by high-risk items, setting boundary conditions and geological features, performing thermal-hydraulic-mechanical coupling analysis or thermal-hydraulic-mechanical-chemical coupling analysis, and outputting at least one of pressure field, component concentration field, reaction progress field, porosity field, permeability field, instability index field and leakage flux.

[0080] The specific implementation is as follows: The system automatically constructs a two-dimensional mesh model for the local area determined by high-risk items; the user or system sets the size of the computational domain and the number of mesh divisions, and then sets boundary conditions and defines geological features such as faults, weak zones, injection points or drainage points; the pressure field output by the local multiphysics analysis is used to determine the pressure concentration factor, the instability index field is used to determine the support reduction factor, the leakage flux is used to correct the probability level of the risk of sealing failure or the risk of leakage from faults and fractures, and the permeability field is used to correct the equivalent permeability parameters in the injection reaction sealing simulation.

[0081]

[0082] In this embodiment, when local multiphysics analysis identifies a dominant CO2 migration pathway in the fault fracture zone, the system triggers an adaptive dynamic plugging and pressure oscillation synergistic strategy. Based on fracture aperture, permeability, and CO2 phase changes, the system automatically calculates the recommended injection volume and curing time of the gel plugging agent. During plugging operations, it generates periodic pressure oscillation pulses by adjusting the wellhead injection pressure. These pressure fluctuations promote capillary capture and mineral dissolution of CO2 in the matrix pores, while simultaneously expanding the CO2 sweep volume in low-permeability areas. After pressure oscillation adjustment, the system dynamically evaluates the plugging effect based on real-time feedback results from the permeability and pressure fields, verifying whether the leakage risk after plugging has been reduced to a preset safety level.

[0083] In one specific implementation, the digital twin analysis includes: when the deviation of the pressure curve, gas phase CO2 mass curve, or cumulative mineralization curve from the preset theoretical trend exceeds a set threshold, importing sensor observation data, performing quality control and time-series alignment on the observation data, mapping the sensor data to model observation variables, using ensemble Kalman filtering or least squares methods for parameter assimilation, and performing prediction and inference based on the assimilated model.

[0084] The specific implementation method is as follows: The system imports one or more CSV format sensor observation data files, completes outlier removal, missing value interpolation, and unified step long-term alignment, and then maps the sensor data to the model observation target location; the system uses ensemble Kalman filtering or least squares method for parameter assimilation. After assimilation, it not only generates the comparison curve between predicted and observed values, residual curve, parameter update trajectory, and alarm event list, but also performs prediction and extrapolation for a future preset time based on the assimilated model. It simulates the impact of reducing injection flow, starting depressurization and drainage, extending the sealing and settling time, and local sealing measures on pressure changes, gas phase CO2 enrichment, mineralization, and leakage risk level, and outputs recommended disposal strategies.

[0085] In this embodiment, digital twin analysis introduces a real-time error compensation and state awareness mechanism based on a Physical Information Neural Network (PINN). The system not only employs ensemble Kalman filtering for parameter assimilation but also constructs a dynamic error compensation field between model predictions and actual sensor observations. The system utilizes the PINN to capture the nonlinear relationship between the residuals of the physical equations and the residuals of the observed data, correcting constitutive equation errors and boundary condition errors in the simulation calculations in real time, and generating a dynamically calibrated twin-corrected solution. Based on this, the system fuses heterogeneous data from multiple sources, such as pressure sensors, microseismic monitoring, and seepage flow sensors, into a unified digital twin information state matrix. Based on this matrix, it infers the CO2 enrichment trend and leakage risk evolution path over a preset time period, outputting a probabilistic risk warning containing confidence intervals, rather than a single deterministic indicator, to improve the accuracy and robustness of on-site warnings.

[0086] In one specific implementation, the exported evaluation results include: generating an evaluation report that includes at least an engineering overview, design verification results, scheme comparison table, process verification events, airtightness test results, response trends, risk conclusions, closed-loop iterative process, and key control parameters, and / or exporting one or more of the following according to the user: a one-page decision summary, detailed technical report, field process operation card, diagnostic data package, grid snapshot file, time series data file, and twin result file.

[0087] The specific implementation is as follows: After completing the main process evaluation, enhanced analysis trigger judgment, and closed-loop iteration, the system provides a unified result export interface; when the user clicks the export report button, the system generates an evaluation report, which includes at least the following: project overview information, design verification results, scheme comparison table, process verification event list, airtightness test results, response trend curve, risk conclusions, closed-loop iteration records, key control parameters, and recommended adjustment schemes; for decision-makers, the system outputs a one-page decision summary; for design engineers, the system outputs a detailed technical report; and for on-site operators, the system outputs process operation cards.

[0088] In one specific implementation, the method further includes: performing formula verification and parameter sensitivity analysis steps, comparing the calculation results of the preset reference case with the expected value, outputting the verification pass rate and error information, and calculating the sensitivity of each input parameter to the support safety factor, peak pressure and final mineralization amount, and outputting the sensitivity ranking and key control parameters; the key control parameters are used to determine the priority adjustment order of automatic adjustment and scheme comparison selection.

[0089] The specific implementation method is as follows: The system has several built-in reference cases, each of which includes input parameters, expected output values ​​and allowable error ranges; when the user runs the formula verification, the system substitutes the input parameters in the current project or built-in test cases into the design verification formula, injection reaction sealing formula, process constraint rules and risk assessment rules to obtain the actual output value, and compares it with the expected output value to calculate the absolute error, relative error and verification pass rate.

[0090] After completing the formula verification, the system automatically performs parameter sensitivity analysis, selecting the reservoir span, reservoir length, top plate thickness, uniformly distributed load, volumetric weight, burial depth, plastic zone parameters, compressive strength of the support, effective reservoir volume, injection mass flow rate, reaction rate constant, humidity factor, and initial reaction capacity as candidate input parameters, and calculating the sensitivity of each input parameter to the support safety factor, peak pressure, and final mineralization amount.

[0091]

[0092] When a disturbance in a certain input parameter causes the change in the support safety factor, peak pressure, or final mineralization to exceed a preset threshold, the system marks the input parameter as a critical control parameter; the parameter sensitivity analysis results are used to determine the priority order of automatic adjustment and scheme comparison.

[0093] This application also provides a CO2 storage filling and sealing simulation system, including: The project management module is used to create or open projects and configure project parameters; The main process evaluation module includes a design verification unit, a process validation unit, an injection reaction sealing simulation unit, and a risk assessment unit. The enhanced analysis module includes a local multiphysics analysis unit and a digital twin analysis unit; The closed-loop coupling control module is used to transmit feedback parameters between the design verification unit, process verification unit, injection reaction sealing simulation unit, risk assessment unit, and enhancement analysis module, and automatically adjust design parameters, injection boundary conditions, or process control parameters based on the feedback results. At the same time, the closed-loop coupling control module is equipped with a real-time rapid deduction and multi-objective optimization unit based on an artificial intelligence agent model, which is used to perform digital twin self-correction and error feedback closed-loop correction. The results export module is used to output evaluation reports, scheme comparison tables, process operation cards, and data files. The system is used to perform a CO2 storage filling and sealing simulation method.

[0094] The specific implementation is as follows: The system includes an engineering management module, a main process evaluation module, an enhanced analysis module, a closed-loop coupling control module, and a results export module. The engineering management module is used for unified configuration of geological parameters, structural parameters, material parameters, process parameters, risk configuration, local model configuration, and digital twin configuration.

[0095] The main process evaluation module includes a design verification unit, a process validation unit, an injection-reaction-sealage simulation unit, and a risk assessment unit. The design verification unit outputs the plastic zone radius, bearing width, equivalent bearing pressure, support safety factor, top-fill-bottom plate collaborative bearing results, and maximum allowable injection pressure; the process validation unit outputs time-dependent validation results, airtightness test results, conflict nodes, and correction schemes; the injection-reaction-sealage simulation unit outputs peak pressure, pressure rise rate, cumulative injection volume, cumulative mineralization, residual gaseous CO2, and capacity utilization rate; and the risk assessment unit outputs the risk level of each risk item and the overall risk level.

[0096] The closed-loop coupling control module corrects the effective volume and effective transport channel parameters of the storage tank based on the radius of the plastic zone and the bearing width, and determines the maximum allowable injection pressure based on the safety factor of the support and the results of the cooperative bearing. When the safety factor of the support is too low, the closed-loop coupling control module reduces the maximum allowable injection pressure in the injection reaction sealing simulation to avoid top plate instability induced by the injection pressure.

[0097]

[0098] When the peak pressure output by the injection reaction sealing simulation exceeds the maximum allowable injection pressure, or the pressure rise rate exceeds the preset threshold, the closed-loop coupling control module automatically reduces the injection mass flow rate in subsequent periods, starts pumping and pressure regulation, or extends the sealing and settling time, and feeds back the adjusted process parameters to the process verification unit and the injection reaction sealing simulation unit. When the structural instability risk output by the risk assessment is high risk or extremely high risk, the closed-loop coupling control module automatically reverts to the design verification unit and adjusts the storage span, the width of the coal pillar or filling support, the compressive strength of the filling body, or the maximum allowable injection pressure.

[0099] The closed-loop coupled control module performs iterations in the order of design verification, process validation, injection reaction sealing simulation, risk assessment, enhancement analysis, parameter correction, and reassessment until all risk levels are reduced to an acceptable range or the preset maximum number of iterations is reached. If high-risk or extremely high-risk items still exist after reaching the maximum number of iterations, the system outputs the conclusion that the solution is unacceptable and lists the key parameters, risk items, and recommended adjustment directions that lead to the unacceptability.

[0100] Specifically, in this embodiment, the closed-loop coupled control module also performs real-time rapid deduction and multi-objective optimization based on an artificial intelligence proxy model. After each main process evaluation, the system records the parameter-result mapping dataset and trains a lightweight proxy model using a deep neural network or Gaussian process regression. During parameter adjustment or emergency control, the system prioritizes calling the proxy model to complete the combined deduction of multiple sets of injection flow-extraction strategies-reaction parameters within milliseconds. Subsequently, with the multi-objective functions of maximizing the safety factor, minimizing pressure fluctuations, and maximizing cumulative mineralization, the system automatically outputs the globally optimal real-time control strategy using particle swarm optimization or Bayesian optimization algorithms. Simultaneously, this module is also used to reverse-correct the equivalent permeation parameters, reaction rate constant, and effective reservoir volume in the main process evaluation module based on the dynamic error compensation field and twin correction solution output by the digital twin analysis unit, enabling the simulation system to possess self-learning and self-evolution capabilities.

[0101] The specific applications of this application will be explained in detail below with specific examples: Example 1 In teaching demonstrations of mining engineering or carbon sequestration-related courses, teachers use a built-in basic teaching example project for demonstration. After the teacher starts the software, the software automatically opens the example project named "Basic Teaching." The teacher then reviews the default configurations of geological parameters, structural parameters, and material parameters in this project, and clicks the "Run" button to execute the main process evaluation. The software sequentially completes design verification, process validation, injection reaction sequestration simulation, and risk assessment. The teacher switches to the design verification tab to show students the calculation results of the safety factor, recommended span, and equivalent bearing pressure, explaining that a safety factor greater than or equal to 2.5 is a relatively ideal design target. The teacher then switches to the process validation tab to demonstrate... The presentation includes a timeline table and verification messages for the construction phase, waste disposal phase, and injection / sealage phase, outlining the door status and ventilation requirements for each phase. The instructor then switches to the reaction / sealage tab, displaying the cumulative injection volume curve, cumulative mineralization curve, and residual gaseous carbon dioxide mass curve, explaining the relationship between curve shapes and sealage effectiveness. Finally, the instructor switches to the risk tab, displaying the risk level and overall acceptability conclusions for each risk item. After completing the main process assessment, the instructor clicks the export report button to generate a PDF report containing all the above information as teaching material. The entire demonstration requires no coding and is completed entirely through a graphical interface, making it suitable for classroom instruction and initial student understanding.

[0102] Example 2 In the early design phase of a carbon dioxide storage facility project in a goaf area, engineers need to quickly compare and select different design schemes. Engineers open a sample project titled "Typical Design" or replicate multiple versions based on this sample. Engineers modify key parameters in different versions, for example, setting the storage span to 15 meters in the first version, 20 meters in the second, and 25 meters in the third. Engineers also adjust parameters such as the compressive strength of the support structure, burial depth, and injection flow rate in different versions. For each version, engineers perform a main process evaluation, recording the safety factor and recommended span on the design verification tab. The peak pressure and final mineralization amount in the sealed tab, and the overall risk level in the risk tab; engineers compared the results of different versions and found that the safety factor was 3.2 when the span was 15 meters but the storage volume was small, the safety factor was 2.7 when the span was 20 meters and the storage volume was moderate, and the safety factor was 1.8 when the span was 25 meters, which was below the safety threshold; therefore, the engineers chose the 20-meter span scheme as the benchmark for further detailed design; if it is necessary to evaluate the impact of different gating strategies, engineers can also modify the gate status parameters in the process verification and rerun the main process for evaluation; this application example realizes the rapid and unified comparison of multiple design schemes, which significantly improves the efficiency of initial screening of schemes.

[0103] Example 3 In goaf storage projects with faults or weak zones, engineers need to assess local leakage sensitivity and failure risk. Engineers open a sample project titled "Risk Sensitivity," which pre-defines fault locations, weak zone extents, and the layout of injection and drainage points. Engineers first execute the main workflow assessment to obtain the overall risk level. Then, they switch to the Local Multiphysics Analysis tab, confirming that the local 2D model configuration is enabled. Engineers check parameters such as computational domain size, grid number, and time step, selecting the coupling mode as thermal, hydrological, mechanical, and chemical coupling analysis, while also considering carbon dioxide transport and reaction. Engineers click the "Run Local Model" button; the software constructs a 2D mesh and maps faults and weak zones onto the mesh cells, correcting local leakage. The engineers calculated the permeability and strength parameters. After solving the problem, they switched between different field variables in the interface. In the pressure field, they observed a significant increase in pressure near the fault. In the carbon dioxide mass fraction field, they observed a trend of carbon dioxide migrating upwards along the fault. In the instability index field, they found that the instability index was high in the weak zone region. Based on this, the engineers determined that the fault and weak zone were potential leakage channels and instability areas, and these locations needed to be given priority consideration when deploying monitoring points in the field. The engineers also incorporated the maximum leakage flux and failure zone area ratio output by the local model as supplementary information into the risk assessment and reviewed the corresponding risk item prompts in the risk tab. Finally, the engineers exported the VTK snapshot file for reporting and expert review.

[0104] Example 4 During the trial operation phase of the carbon dioxide storage facility in the goaf, pressure sensors, carbon dioxide volume fraction sensors, microseismic velocity sensors, seepage flow sensors, and pH sensors were deployed on-site, accumulating several weeks of monitoring data. Engineers used digital twin analysis to replay the monitoring data and calibrate parameters. The engineer opened a sample project named "Twin Playback," which contained model parameter configurations corresponding to the site conditions. The engineer switched to the Digital Twin Analysis tab, confirming that software dependencies were installed and digital twin configurations were enabled in the project. The engineer loaded multiple CSV format sensor data files, and the software automatically performed quality control and time-series alignment. The engineer selected ensemble Kalman filtering as the assimilation method and set the update interval to one hour. After clicking the "Run Twin Playback" button, the software mapped the sensor data to the model observation locations, step by step. The process involves parameter assimilation. After completion, engineers review the sensor and prediction comparison charts, observing a good match between the assimilated prediction and observation curves. The residual plot shows that most residuals fluctuate around the zero line. The parameter trajectory plot shows that the permeability and reaction rate constants gradually converge to stable values. The alarm event table indicates that the standardized residual exceeded the threshold once in the pressure channel on a certain day. Engineers, combining this with field logs, confirmed that injection flow adjustment was performed that day, which is normal operation. The root mean square error (RMSE) output by the software after assimilation is significantly smaller than the uncorrected RMSE, proving the effectiveness of parameter assimilation. Engineers import the twin results summary into a PDF report, forming an analysis document for the trial operation phase. This application example supports the import, playback, model calibration, and anomaly warning of field monitoring data, providing quantitative evidence for trial operation diagnosis.

[0105] In this embodiment, digital twin analysis also integrates real-time error compensation and probabilistic risk warning functions based on Physical Information Neural Network (PINN). Based on parameter assimilation, the system constructs a dynamic error compensation field and automatically corrects the constitutive error of the model, generating a twin-corrected solution. The final output is not just a single-point prediction result, but a probabilistic risk warning containing a confidence interval, providing solid data support for dynamic emergency response during the trial operation phase.

[0106] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the invention can be implemented in other specific forms without departing from its spirit or essential characteristics. Therefore, the embodiments should be considered in all respects as exemplary and non-limiting, and the scope of the invention is defined by the appended claims rather than the foregoing description. Thus, all variations falling within the meaning and scope of equivalents of the claims are intended to be included within the present invention. No reference numerals in the claims should be construed as limiting the scope of the claims.

[0107] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. A method for simulating the filling and sealing of a CO2 storage facility, characterized in that, include: Establish a project and configure project parameters, which include at least the geometric parameters of the goaf, the mechanical parameters of the roof, the bearing capacity parameters of the floor, the parameters of the filling body, the airtightness testing parameters, the injection and drainage process parameters, and the reaction and storage parameters. The main process assessment for the goaf is conducted, which includes design verification, process validation, injection reaction sealing simulation, and risk assessment. The design verification determines the injection pressure boundary in a chain-like sequence based on span ratio, allowable span, plastic zone radius, bearing width, equivalent bearing pressure, support safety factor, and maximum allowable injection pressure. Furthermore, the design verification and risk assessment incorporate uncertainty quantification methods based on Monte Carlo simulation and geostatistics, setting key geological and mechanical input parameters as random variables following probability distributions. Multiple sets of random sampling are performed in parallel to output failure probability and reliability indices, and a comprehensive risk probability index replaces the single risk level evaluation. The process validation determines whether injection sealing is permitted to commence based on the time-series dependency diagram of the construction stage, filling stage, solidification and curing stage, airtightness testing stage, injection sealing stage, and drainage and pressure regulation stage. The maximum allowable injection pressure determined by the design verification and the process verification results are used as constraints for the injection reaction storage simulation. During the injection reaction storage simulation, an adaptive time step dynamic control strategy based on rate of change feedback is introduced to monitor the storage pressure change rate, gas phase CO2 concentration change rate, and mineralization reaction rate in real time. The calculation time step Δt is dynamically adjusted according to the maximum relative change rate among the three, and a macro-micro cross-scale coupled model is constructed. A macro equivalent continuous medium model is used in the far field region, and the model is automatically switched to a micro discrete model in key areas such as the injection point and fault zone. The storage pressure, gas phase CO2 mass, and cumulative mineralization are calculated, and the risk level is determined through risk assessment. When the support safety factor is lower than the preset safety threshold, the airtightness test fails, the peak pressure exceeds the maximum allowable injection pressure, or any risk item reaches the preset high-risk level, the system automatically triggers parameter rollback adjustment and / or enhanced analysis. The enhanced analysis includes local multiphysics analysis and digital twin analysis. When local multiphysics analysis identifies a dominant migration channel in the fault fracture zone, the system triggers an adaptive dynamic plugging and pressure oscillation synergistic strategy, automatically calculating gel plugging agent parameters based on fracture aperture and permeability and adjusting the wellhead injection pressure to generate periodic pressure oscillation pulses. When digital twin analysis is initiated, a real-time error compensation mechanism based on a physical information neural network (PINN) is introduced. A dynamic error compensation field is established between the residuals of the physical equations and the residuals of the observation data. A dynamically calibrated twin correction solution is generated, and a probabilistic risk warning containing confidence intervals is output. The adjustment results are fed back to the design verification, process validation, injection reaction sealing simulation, and risk assessment iteration. At the same time, a real-time rapid inference and multi-objective optimization unit based on an artificial intelligence agent model is added to the closed-loop iteration process. The parameter-result mapping dataset is recorded and a lightweight agent model is trained using a deep neural network. The multi-objective functions are maximizing the safety factor, minimizing pressure fluctuations, and maximizing the cumulative mineralization. The particle swarm optimization algorithm is used to output the globally optimal real-time control strategy and feed it back to each evaluation stage. The evaluation results are then derived.

2. The method according to claim 1, characterized in that, The design verification includes calculating the span ratio, theoretical allowable span of the slab, theoretical allowable span of the beam, recommended span, plastic zone radius, bearing width, equivalent bearing pressure, support safety factor, modified support safety factor, and maximum allowable injection pressure based on the following formulas in sequence: Where a represents the actual span of the goaf reservoir, L represents the reservoir length, λ represents the span ratio, h represents the roof thickness, ν represents Poisson's ratio, and σ t α represents the tensile strength of the top plate, q represents the uniformly distributed load, and α represents the tensile strength of the top plate. s a represents the theoretical parameters of the beam related to the strength and thickness of the top slab. plate The plate theory allows for a span, a beam Indicates the span allowed by beam theory, a rec Indicates the recommendation span, β a r represents the span safety reduction factor. p H represents the radius of the plastic region. c Indicates the height of the coal pillar or backfill support, α p φ represents the plastic zone coefficient, K represents the internal friction angle, K represents the lateral pressure coefficient, γ represents the bulk density, H represents the burial depth, θ represents the stress reduction angle, c represents the cohesion, b represents the bearing width, and σ represents the internal pressure coefficient. c P represents the compressive strength of the support. eq FS represents the equivalent bearing pressure, and FS represents the safety factor of the support. ′ χ represents the corrected safety factor of the support structure. s P represents the local stress concentration factor. inj,max Indicates the maximum permissible injection pressure, η p P represents the injection pressure reduction factor. ref Indicates the reference allowable pressure, FS th The safety factor threshold is represented by the span ratio, allowable span, plastic zone radius, bearing width, equivalent bearing pressure, support safety factor and maximum allowable injection pressure being sequentially related so that the design verification results can constrain the pressure boundary conditions of the injection reaction sealing simulation. When the safety factor of the corrected support is lower than the preset safety threshold, the system automatically generates multiple design adjustment schemes. The design adjustment schemes include one or more of the following: reducing the span of the storage tank, increasing the width of the coal pillar or backfill support, increasing the compressive strength of the backfill, extending the solidification time of the backfill, and reducing the maximum allowable injection pressure. The system calculates the safety factor of the support, the effective volume of the storage tank, the volume loss rate, and the maximum allowable injection pressure under each design adjustment scheme, and sorts each design adjustment scheme according to the degree of safety factor compliance, the effective volume retention rate, and the adjustment range. The design verification also includes a joint load-bearing verification of the top plate, filling body, and bottom plate. Based on the equivalent bearing pressure, mechanical parameters of the top plate, elastic modulus of the filling body, height of the filling body, and bearing parameters of the bottom plate, the system estimates the deflection of the top plate, the compression of the filling body, and the additional deformation of the bottom plate, and determines whether the joint deformation condition is met between the deflection of the top plate, the compression of the filling body, and the additional deformation of the bottom plate. When the joint deformation condition is not met, the system lowers the structural stability evaluation level and generates adjustment schemes to increase the strength of the filling body, increase the support width, or reduce the injection pressure.

3. The method according to claim 1, characterized in that, The injection reaction sequestration simulation includes calculating pressure and mineralization based on the following formulas, and using the maximum allowable injection pressure obtained from the design verification as the pressure boundary condition: Where P represents the storage pressure, n represents the total amount of gas in the storage, R represents the ideal gas constant, T represents the temperature, and V represents the volume of gas. eff Δn represents the effective volume of the storage facility. min k represents the change in the amount of CO2 caused by the mineralization reaction. r f represents the reaction rate constant. h Represents the humidity factor, n g C represents the amount of gaseous CO2. rem C0 represents the initial reaction capacity, Δt represents the calculation time step, and P represents the remaining reaction capacity. inj,max It indicates the maximum allowable injection pressure for design verification feedback; and outputs the pressure time history, gas phase CO2 mass time history, cumulative mineralization time history, porosity and pH change trends. When the peak pressure exceeds the maximum allowable injection pressure or the pressure rise rate exceeds the threshold, the system calculates the allowable injection mass flow rate range based on the effective volume of the storage, temperature, current amount of gaseous CO2, mineralization reaction consumption, and pumping rate, and feeds the calculated allowable injection mass flow rate back to the process verification and closed-loop coupling control process. 。 4. The method according to claim 1, characterized in that, The process verification includes: constructing a time-series dependency diagram encompassing the construction phase, filling phase, curing phase, airtightness testing phase, injection and storage phase, and extraction and pressure regulation phase; verifying, based on the time-series dependency diagram, whether the preconditions, ventilation status, airtight door status, curing strength of the filling body, airtightness testing results, injection / extraction switches, and flow parameters of each phase meet the process logic of the goaf CO2 storage facility; when there are process parameters that violate the time-series dependency, the system locates conflict nodes and conflict constraint edges, and generates parameter correction schemes based on the injection reaction and storage simulation results. The airtightness testing phase includes pressure holding testing and leakage rate assessment. The system determines whether the airtightness test is qualified based on the test pressure, test duration, pressure decay rate, and allowable leakage rate. The leakage rate is estimated using the following formula: Among them, Q leak P represents the leakage rate, P0 represents the initial pressure of the pressure holding test, and P t The pressure at which the pressure holding test ends, Δt d Indicates the detection duration; when Q leak If the leakage rate is not greater than the allowable leakage rate and the pressure decay rate is less than the preset threshold, the airtightness test is deemed qualified; otherwise, the injection and sealing stage is prohibited, and a correction plan is generated to extend the pressure holding test time, repair the airtight door, perform local sealing, or reduce the injection pressure.

5. The method according to claim 1, characterized in that, The risk assessment includes: identifying risks of structural instability in goaf areas, sealing failure, localized CO2 accumulation, fault fracture leakage, pipeline leakage, and ecological impact; mapping the likelihood and severity of each risk type to a risk matrix to determine the risk level; when any risk level is high or extremely high, the system... The corresponding risk items are fed back to the closed-loop coupled control process to trigger local multiphysics analysis, injection flow adjustment, pumping and draining pressure regulation, local blocking, or design parameter rollback adjustment.

6. The method according to claim 1, characterized in that, The local multiphysics analysis includes: constructing a two-dimensional mesh model for a local area automatically determined by high-risk terms, setting boundary conditions and geological features, performing thermal-hydraulic-mechanical coupling analysis or thermal-hydraulic-mechanical-chemical coupling analysis, and outputting at least one of the following: pressure field, component concentration field, reaction progress field, porosity field, permeability field, instability index field, and leakage flux; wherein, the pressure field is used to determine the pressure concentration factor, the instability index field is used to correct the support safety factor, the leakage flux is used to correct the risk level, and the permeability field is used to correct the equivalent permeability parameters in the injection reaction sealing simulation.

7. The method according to claim 1, characterized in that, The digital twin analysis includes: when the deviation of the pressure curve, gas phase CO2 mass curve, or cumulative mineralization curve from the preset theoretical trend exceeds a set threshold, importing sensor observation data, performing quality control and time-series alignment on the observation data, mapping the sensor data to model observation variables, using ensemble Kalman filtering or least squares methods for parameter assimilation, and performing prediction and inference based on the assimilated model to output the pressure change trend, gas phase CO2 enrichment trend, leakage risk change trend, and the handling effect under different emergency measures within a preset time period.

8. The method according to claim 1, characterized in that, The exported evaluation results include: generating an evaluation report that includes at least an engineering overview, design verification results, scheme comparison table, process verification events, airtightness test results, response trends, risk conclusions, closed-loop iterative process and key control parameters, and / or exporting one or more of the following according to the user: one-page decision summary, detailed technical report, field process operation card, diagnostic data package, grid snapshot file, time series data file, and twin result file.

9. The method according to claim 1, characterized in that, It also includes formula verification and parameter sensitivity analysis steps; the system compares the calculation results of the preset reference case with the expected value, outputs the verification pass rate, absolute error and relative error, and calculates the sensitivity of each input parameter to the support safety factor, peak pressure and final mineralization, and outputs the sensitivity ranking and key control parameters; the key control parameters are used to determine the priority adjustment order of automatic adjustment and scheme comparison selection.

10. A CO2 storage filling and sealing simulation system, characterized in that, include: The project management module is used to create or open projects and configure project parameters; The main process evaluation module includes a design verification unit, a process validation unit, and an injection reaction sealing module. Units and risk assessment units; The enhanced analysis module includes a local multiphysics analysis unit and a digital twin analysis unit; The closed-loop coupling control module is used to receive the plastic zone radius, bearing width, support safety factor and maximum allowable injection pressure output by the design verification unit, the time dependence verification results and airtightness test results output by the process verification unit, the peak pressure, pressure rise rate, residual gaseous CO2 and cumulative mineralization output by the injection reaction and storage simulation unit, and the risk level output by the risk assessment unit. Based on the above results, it corrects the effective volume of the reservoir, equivalent permeability parameters, maximum allowable injection pressure, injection mass flow rate and pumping strategy. The results export module is used to output evaluation reports, scheme comparison tables, process operation cards, and data files. The closed-loop coupling control module is also equipped with a real-time rapid deduction and multi-objective optimization unit based on an artificial intelligence agent model, which is used to perform digital twin self-correction and error feedback closed-loop correction. The system is used to perform the method according to any one of claims 1 to 9.

Images