Intelligent alert and control system for carbon dioxide leaks throughout the oxygen-enriched air combustion and geological storage process

Abstract

The invention discloses a method for intelligent alerting and regulating carbon dioxide leaks throughout the oxygen-enriched air combustion and geological storage process, which falls within the technical field of safety monitoring for carbon capture and storage. In order to solve the problems of traditional monitoring methods, such as the high rate of false alarms related to the single threshold, the insufficient accuracy in locating leaks, and the imperfect conditions for lifting regulation, the method proposes the following solutions: deploying a gridded network of multiparameter sensors on the surface to monitor the carbon dioxide concentration and soil pH in real time, and defining differentiated triggering thresholds adapted respectively to agricultural, industrial, and alpine zones;Locate the leak source based on the alert signal by combining the three-dimensional geological model and the inverse diffusion algorithm; select the adjacent control well group, simultaneously reduce the pressure of the injection well and start the pumping well; continuously monitor the carbon dioxide concentration and soil pH in the area surrounding the leak, and lift the alert only when both parameters meet predefined criteria. This process achieves closed-loop management integrating precise alerting, intelligent localization, and safe control of carbon dioxide leaks from geological storage.

Description

1 Intelligent alert and control method for carbon dioxide leaks throughout the oxygen-enriched combustion and geological storage process. Technical Field 5 The invention relates to the technical field of monitoring the safety of carbon capture and storage. More specifically, the present invention relates to an intelligent alert and control method for carbon dioxide leaks throughout the oxygen-enriched combustion and geological storage process. 10 Prior Technology In the context of implementing geological carbon dioxide storage, real-time leak monitoring and rapid response are key elements guaranteeing long-term storage safety. Existing technologies rely mainly on a single-variable alarm mechanism with a fixed threshold, for example, triggering an alert only when the concentration The carbon dioxide level on the surface exceeds the permitted limit. However,These methods have significant limitations: differences in background environmental values ​​between different regions make a unified threshold poorly suited (for example, background concentration is higher around industrial areas, 20 while agricultural areas are frequently disturbed by biological activities), and a single threshold is likely to cause false alarms or missed alerts. In parallel, traditional localization methods mainly use static geological models for reversing the leak source, without taking into account the dynamic characteristics of the 25 storage layer permeability which evolves with injection pressure, resulting in a significant discrepancy between the localization result and the actual leak point, delaying subsequent treatment measures. Post-leak control strategies also have drawbacks. Most solutions aim to inhibit the leak solely by reducing the pressure in the injection wells.without having means of active intervention on the movement of gases in the leak zone, which prevents the rapid dissipation of locally accumulated carbon dioxide. Furthermore, the conditions for lifting the regulations are too simplistic, often based on a temporary drop in concentration, without taking into account the delay in the recovery of the soil's chemical properties. In practice, it has been observed that the rate of increase in soil pH is much slower than the diffusion of the gas concentration, and a premature cessation of regulation can lead to a risk of secondary leakage. These combined shortcomings limit the reliability of existing monitoring systems, increase operating and maintenance costs due to frequent false alarms, and the speed of response to events. real leaks struggle to meet industrial safety standards.10 The fundamental reasons for these problems lie in: Firstly, the spatial heterogeneity of environmental parameters has not been quantitatively assessed,The definition of thresholds lacks an adjustment mechanism adapted to the regions. Secondly, the updating of the geological model lags behind the dynamic changes in the storage layer, and the permeability field correction depends on periodic borehole exploration data, which lack responsiveness. Thirdly, the control means have not formed a coordination mechanism between the injection and suction ends, resulting in low efficiency in controlling gas movement. Fourthly, the judgment dimension for raising the alert is single, not allowing for continuous verification by coupling physical, chemical, and other indicators. Do these technical bottlenecks make it difficult for existing systems to balance sensitivity and reliability in complex site conditions?limiting the safe deployment of large-scale storage projects. Description of the invention 25 The present invention provides an intelligent method for preventing, alerting and regulating carbon dioxide leaks throughout the entire process of geological storage after oxygen-enriched combustion, aimed at solving the following problems: In geological storage after oxygen-enriched combustion, 30 traditional single-variable threshold monitoring is liable to cause false alarms or omissions of alerts; leak location is based on a static geological model, resulting in insufficient accuracy; the conditions for lifting regulation after intervention lack continuous multi-parameter verification, leading to a delayed response or a premature cessation of regulation. To solve the problem where a single sensor node is likely to trigger a false alarm due to local disturbances (such as vehicle exhausts, fertilization work),It is necessary to eliminate instantaneous anomalies by verifying the spatial distribution and temporal synchronization. To solve the problem where point anomalies in sparsely monitored areas cannot be verified by multi-node coordination, it is necessary to actively activate leak response features to distinguish real leaks from equipment failures. To solve the problem where micro-leaks are difficult to detect by concentration threshold or pulsed suction, it is necessary to excite low-frequency oscillation signals by periodic pressure excitation and extract their frequency characteristics. To solve the problem where the static geological model cannot adapt to dynamic changes in the permeability of the storage layer, resulting in excessive discrepancies in leak source localization,It is necessary to integrate the depth pressure anomalies to correct the model in real time. To solve the problem where pressure gradient anomalies can originate from factors unrelated to leaks (such as injection-recovery fluctuations), it is necessary to perform a secondary verification by associating the specific cooling characteristic with the leaks. To solve the problem where, when both pressure and temperature verification fail, it is necessary to inject a tracer gas to directly visualize the leak path and avoid missing hidden channels. To solve the problem where frequency analysis can generate false positive signals due to electrical disturbances, it is necessary to verify the physical mechanism by coupling the heat absorption and local cooling characteristics caused by the leak. 2025 / 5812 BE2025 / 5812 4 To solve the problem where the dilution of water in soils with high humidity delays the response of H,It is necessary to extend the monitoring period to capture the slow acidification process and avoid missing small leaks. In order to achieve these objectives and other advantages of the present invention, an intelligent alert and control method for carbon dioxide leaks is provided throughout the oxygen-enriched combustion and geological storage process, comprising the following steps: Step 1: Deploy a grid of fiber optic sensors distributed on the surface of the target storage area, with a spacing of 10 to 50 meters; said network includes carbon dioxide concentration sensors, soil pH sensors, surface temperature sensors, and relative humidity sensors. Step 2: Obtain real-time monitoring data from each sensor in Step 1 via a data collection unit; said monitoring data includes the percentage volume concentration of carbon dioxide, the soil pH,the surface temperature in degrees Celsius and the relative humidity value. Step 3: Select a trigger threshold according to the type of target storage area: agricultural area: use a threshold where the percentage of volume concentration of carbon dioxide exceeds 1.0% and the soil pH is less than 5.5; industrial area: use a threshold where the percentage of volume concentration of carbon dioxide exceeds 0.6% and the soil pH is less than 5.2; high altitude area: use a threshold where the percentage of volume concentration of carbon dioxide exceeds 0.8%, the soil pH is less than 5.3 and the surface temperature in degrees Celsius is greater than less than ten degrees Celsius; trigger a fault alert signal when any sensor node continuously detects parameters satisfying the threshold corresponding to the zone for 3 to 10 minutes. Step 4: Based on the spatial coordinates of a sensor node that triggered the alert signal, combined with the three-dimensional geological model of the target storage layer,generate a probability map of the leak source position using the inverse diffusion algorithm. Step 5: in the probability map of the leak source position, select the areas with a probability greater than 70% and filter the target control well groups whose spatial distance is less than 300 meters; said target control well group includes at least 1 to 3 carbon dioxide injection wells and 1 to 2 monitoring pumping wells. Step 6: Send a pressure regulation command to the carbon dioxide injection wells of the target control well group defined in Step 5, reducing the opening of the wellhead pressure regulation valve from 40% to 10-70% of its initial working; simultaneously, send a pumping command to the monitoring pump wells of said target control well group, start the vacuum pump and extract gas from the wells at a rate of 5 to 15 cubic meters per minute. Step 7: After the execution of Step 6,Continuously obtain the percentage of 15 volumetric carbon dioxide concentration and the soil pH at all sensor nodes within a 1 kilometer radius of the leak source. Step 8: When the percentage of volumetric carbon dioxide concentration at all sensor nodes mentioned in Step 7 is less than 0.5% continuously for 30 to 120 minutes and the soil pH is greater than 5.0 for all, lift the leak alert signal and restore the carbon dioxide injection well pressure relief valve to its initial position. Preferably, in the intelligent carbon dioxide leak warning and control process throughout the oxygen-enriched combustion and geological storage process of the present invention, in Step 3, the triggering of the leak warning signal must simultaneously satisfy the following additional conditions: The rate of temperature variation in degrees Celsius at the surface of a sensor node is less than 0.5 degrees Celsius per hour; Within a radius of 20 to 50 meters centered on the triggering node,at least 330 adjacent sensor nodes simultaneously detect that the percentage of volume concentration of carbon dioxide exceeds 0.5% to 1.2% and that the pH 2025 / 5812 BE2025 / 5812 6 of the soil is less than 5.0 to 5.8; The time gap of the detections of said adjacent sensor nodes does not exceed 5 minutes. Preferably, in the intelligent carbon dioxide leak warning and control process throughout the oxygen-enriched combustion and geological storage process of the present invention, if only one sensor node meets the triggering conditions in Step 3, and fewer than 3 adjacent sensor nodes trigger simultaneously within a radius of 20 to 50 meters, the following operations are performed: Send a pulsed pumping command to the nearest monitoring pump well, lowering the wellhead pressure from normal atmospheric pressure to -5 to -20 kilopascals (kPa) in 30 to 60 seconds; After the pulsed pumping starts,collect continuously the percentage of volumetric concentration of carbon dioxide of decenode and adjacent sensor nodes in 8 surrounding directions; 15 If rise rates of decenode concentration reach 1.5 to 3 times the average rise rates of adjacent nodes, and the peak concentration exceeds 15% to 30% of the baseline value before pulse pumping, then it is determined as a valid fault warning signal. Preferably, in the intelligent alert and control method for carbon dioxide leaks throughout the oxygen-enriched combustion and geological storage process of the present invention, after pulsed pumping, if the rate of rise in concentration at one node does not reach 1.5 to 3 times the average rate of rise of the adjacent nodes, or the peak concentration does not exceed 15% to 30% of the base value before pulsed pumping, then the following 25 operations are performed: Send a triple-level pressure wave excitation instruction to the same monitoring pumping well, successively adjusting the wellhead pressure - 10 to -15 kilopascals (kPa), - 15 to -20 kPa and - 20 to -25 kPa,each pressure level being maintained for 120 to 240 seconds; 30 During the maintenance of each pressure level, continuously collect the percentage 2025 / 5812 BE2025 / 5812 7 of volumetric concentration of carbon dioxide of a target with a sampling interval of 1 to 5 seconds; Perform a sliding FFT calculation or a wavelet transformation on the sequence of concentration data collected, and extract the signal energy in the frequency band of 0.01 to 0.1 hertz (Hz); 5 If, for any pressure level, the energy value of this frequency band exceeds 1.8 to 2.5 times the background energy value for 60 to 120 seconds continuously, then it is determined to be a valid fault alarm signal. Preferably, in the intelligent early warning and control process for carbon dioxide leaks throughout the oxygen-enriched combustion and geological storage process of the present invention, the following operations are added between Step 4 and Step 5: Obtain real-time data from deep pressure sensors for priority monitoring wells,which represent at most twenty percent of a total number of wells in the target regulating well group;15 For wells without a deep pressure gauge, calculate the pressure at the top of the storage layer using the wellhead pressure transmitter combined with the hydrostatic pressure formula of the liquid column: Ptop= Pwellhead−ρ⋅g⋅H Where: ρ is the fluid density in the well, H is the depth of the storage layer20, g is the acceleration due to gravity and Pwellhead is the wellhead pressure; Calculate the change in pressure gradient of each well at the depth of the upper interface of the underground storage layer, and mark as an abnormal pressure source any well whose change in pressure gradient exceeds 0.5 to 1.2 kilopascals per meter (kPa / m); Take the position of the abnormal pressure source as the stress point,and update the permeability distribution field of the three-dimensional geological model via the Kriging interpolation algorithm; Introduce the updated permeability distribution field into the inverse diffusion algorithm, and generate the probability map of the position of the leak source. Principle for defining the pressure gradient threshold: Based on the pressure rupture model: ΔPcrit = [2E] / [(1-ν²)H] × ln(R / rw), where: E is the Young's modulus of the opening (in MPa), ν is Poisson's ratio, H is the thickness of the opening (in m), R is the radius of influence, rw is the diameter of the well. The threshold is set at 0.3ΔPcrit (with a safety factor of 3.3). Example: when E = 5000 MPa, ν = 0.25 and H = 100 m, the threshold = 0.82 kPa / m. Preferably, in the intelligent alert and regulation process for carbon dioxide leaks throughout the oxygen-enriched combustion and geological storage process of the present invention, when calculating the variation of the pressure gradient of each well,the following operations are executed simultaneously: Obtain the axial temperature distribution of the well section from 0 to 50 meters under the upper interface of the storage layer via the distributed temperature detection system 15 of the carbon dioxide injection wells in the target regulation well group; When the pressure gradient variation at a certain depth point exceeds 0.5 to 1.2 kilopascals per meter (kPa / m), detect the temperature variation rates of the 10-meter well section above and below centered on that point; if the temperature variation rates of this well section exceed 1.0 to 2.0 degrees Celsius per 100 meters and the direction of temperature variation is negative (i.e., a temperature drop), retain the abnormal pressure source label; otherwise, remove the abnormal pressure source label and do not update the permeability distribution field. Preferably,in the intelligent carbon dioxide leak warning and control process throughout the oxygen-enriched combustion and geological storage process of the present invention, when the abnormal pressure source label is removed, the following operations are performed: 2025 / 5812 BE2025 / 5812 9 Inject sulfurhexafluoride (SF6) tracer gas at ambient temperature into the monitoring pump well corresponding to the abnormal pressure point, with an injection rate of 0.2 to 0.5 cubic meters per minute and an injection duration of 120 seconds; Within 60 to 180 seconds following the end of the injection, detect the concentration of sulfurhexafluoride at all nodes within a radius of 100 meters centered on the surface projection point; If the concentration of sulfurhexafluoride at any node within a 100-meter radius exceeds 1 ppm and lasts for more than 10 seconds continuously, mark the leak channel and trigger an alert. Preferably,in the Intelligent Carbon Dioxide Leak Alert and Control Process throughout the Oxygen-Enriched Combustion and Geological Storage Process of the present invention, when the energy of the frequency band E exceeds 2.0 times the background energy E0 for any pressure level, the following operations are performed simultaneously: 15 Obtain the temperature data in degrees Celsius on the surface of a target during the corresponding pressure holding phase, and calculate the temperature variation rates β (in degrees Celsius per hour) with an interval of 10 to 30 seconds; If β remains in the range of -0.8 to -0.3 degrees Celsius per hour for 90 to 12020 seconds continuously, confirm the triggering of the leak alert signal; Otherwise, it is identified as a disturbance signal, and the subsequent regulation process is not triggered. Preferably, in the intelligent alert and regulation process for carbon dioxide leaks throughout the oxygen-enriched combustion and geological storage process of the present invention,when the temperature variation rate β is not in the range of -0.8 to -0.3 degrees Celsius per hour, when the frequency band energy E exceeds 2.0 times the background energy E0, and when the target area belongs to a high humidity climate zone with an average annual relative humidity greater than 70%, the following operations are carried out simultaneously: 2025 / 5812 BE2025 / 5812 10 The present invention has at least the following beneficial effects. 1. Early leak detection is achieved through a grid of multi-parameter sensors, and the design of thresholds differentiated by region adapts to various environmental contexts; the inverse diffusion algorithm combined with the three-dimensional geological model improves localization accuracy; after regulation, continuous monitoring of the two parameters (CO₂ concentration and pH) ensures the reliable raising of the alert. This process forms a closed loop of monitoring-localization-regulation-verification.significantly reducing the rate of false alarms and the risk of leaks. 2. It requires that at least 3 nodes within the radius simultaneously trigger the alert with a controlled time gap, effectively filtering out local instantaneous disturbances (such as mechanical vibrations); the temperature change rate threshold eliminates slowly varying environmental influences such as sunlight. The dual spatial and temporal constraints reduce false alarms in agricultural areas and allow industrial areas to distinguish factory emissions from actual leaks. 3. Negative pressure pulsed aspiration stimulates accelerated gas exudation from the leak point; thanks to the difference in gradient between the rising levels of concentration and the peak concentration of a target node relative to the surrounding nodes, confirmation of a point leak can be achieved in 10 minutes.20 Compared to manual washing, the efficiency is considerably improved.and losses due to well shutdown are avoided. 4. Triple-level pressure wave excitation can stimulate gas oscillation in micro-fissures, and energy extraction from the 0.01 to 0.1 Hz frequency band allows capture of weak signals undetectable by traditional methods. Frequency characteristics can be separated from the mechanical noise spectrum, and micro-leak verification is completed in 18 minutes. 5. The dynamic correction of the permeability field is carried out based on the marking of abnormal pressure gradients (>0.5 kPa / m), overcoming the lack of adaptability of static models to heterogeneous formations. The application of the hydrostatic pressure formula improves the usability of data from wells lacking depth sensors and reduces the localization error. 2025 / 5812 BE2025 / 5812 11 6. A real leak is accompanied by the dissolution of CO₂ in the soil pore water to form carbonic acid, and the endothermic ionization reaction leads to local cooling by heat absorption.While injection-recovery fluctuations generally exhibit an upward temperature trend, double-checking the direction and magnitude of the temperature variation can reduce the error rate in judgments of pressure anomalies. Sulfur hexafluoride has an extremely low background concentration; combined with high-sensitivity laser detection, it allows direct visualization of the gas migration path. The 60- to 180-second migration window corresponds to the diffusion rate in fractured media, providing a final judgment method for scenarios where pressure and temperature verification fails. 8. After the frequency signal is triggered, the local surface cooling characteristic (β is negative, due to heat absorption by the dissolution of CO₂) is verified simultaneously; coupling the physical mechanism can eliminate false positives caused by electrical disturbances. This process uses existing temperature sensors.without additional costs. 9. Extending the monitoring duration of H to 200 seconds and calculating the negative slope γ can solve the problem of signal attenuation due to ionic diffusion delay in soils with high humidity. Adaptive judgment to the 20 climatic zones ensures that the method is used only in the necessary regions, balancing efficiency and reliability. Other advantages, objectives, and features of the present invention will become partly apparent from the following description and partly will be understood by those skilled in the art through research and practice of the present invention. 25 Description of embodiments With reference to the drawings below, the present invention will be described in more detail so that technical personnel skilled in the art can implement it with reference to the text of the description. It should be understood that terms such as “have”,“include” and “comprehension” used here30 do not exclude the presence or addition of one or more other elements or combinations thereof. 2025 / 5812 BE2025 / 5812 12 According to one embodiment of the present invention, an intelligent alert and control method for carbon dioxide leaks is provided throughout the oxygen-enriched combustion and geological storage process: Deploy a grid of fiber optic sensors distributed on the surface of the target storage area, with a spacing of 30 meters. This grid comprises four types of sensors: the carbon dioxide concentration sensor operates on the non-dispersive infrared principle, the soil pH sensor has a glass electrode, the surface temperature sensor has a platinum resistance, and the relative humidity sensor is capacitive. All sensors are connected to a data collection unit by a fiber optic composite cable, which is 10 buried at a depth of 0.3 meters to prevent mechanical damage. The data collection unit reads the sensor data every 5 seconds. For agricultural areas,A leak alert signal is triggered when a node continuously detects for 5 minutes that the percentage volume concentration of carbon dioxide exceeds 1.0% and that the soil pH is less than 5.5; for the 15 industrial zones, the threshold is set at 0.6% percentage volume concentration of carbon dioxide and 5.2 pH; for the high altitude zones, an additional constraint on the surface temperature is added (must be above minus ten degrees Celsius), with the percentage volume concentration of carbon dioxide exceeding 0.8% and the soil pH less than 5.3. After the alert is triggered, the system calls the three-dimensional geological model of the target storage layer (built from previous seismic exploration data and boreholes) and generates a probability map of the leak source's position using the inverse diffusion algorithm. The algorithm's input parameters include sensor coordinates, monitoring data, and the stratum permeability field.and areas with a probability greater than 70% are marked as high-risk areas. (Note on the parameters of the inverse diffusion algorithm: the calculation of the probability of the leak source uses the formula: P(x,y,z)=30 exp(-||cobs-csim|| / σ), where: cobs is the concentration vector measured by the sensor (%), csim is the concentration simulated by the direct diffusion model (%), σ is the observation error coefficient, calculated from the fluctuations of the background concentration, cback is the background concentration, μ is the average, and N is the number of sensors). Value used in the Bohai aqueduct project: σ=0,12. Filter target regulation well groups whose spatial distance is less than 300 meters in high-risk areas. A typical well group includes 25 carbon dioxide injection wells and 1 monitoring pumping well. Send an instruction to the injection wells to reduce the opening of the wellhead pressure control valve to 50% of its initial setting; simultaneously start the vacuum pump at the monitoring pump wells and extract the gas at a rate of 10 cubic meters per minute. After starting the control, continuously obtain the carbon dioxide volume percentage concentration and the soil pH at all nodes within a 1-kilometer radius of the leak source. When the carbon dioxide volume percentage at all nodes is less than 0.5% and the soil pH is greater than 5.0 continuously for 60 minutes, lift the alert and restore the injection well pressure control valve to its initial setting. Multi-level lifting mechanism: 1. The basic conditions must be met (CO2<0.5% and pH>5,0);2. If the following checks have been triggered, additional conditions are required: frequency analysis → Eleak < 0.5E₀ for 1 hour continuously; temperature check → |β| < 20 0.3°C / h; humidity compensation → γ > -0.02pH / h. In the present technology, the application of the hydrostatic pressure formula is as follows: for wells without a deep pressure gauge, calculate the pressure at the top of the storage layer using data from the wellhead pressure transmitter combined with the hydrostatic pressure formula for the liquid column: Ptop25 = Pwellhead − ρ ⋅ g ⋅ H, where the fluid density in the well ρ is taken to be 800 kg / m³, the The depth of the storage layer is 1500 meters, and the acceleration due to gravity is 9.8 m / s². Real-time calibration of fluid density: for wells without a manometer at depth, obtain a fluid sample daily using a wellhead sampler and measure the density ρ; when it is impossible to take a sample,use the time inversion model of 2025 / 5812 BE2025 / 5812 14 acoustic propagation: ρ=0.23×Δt-0.35 (Δtest in μs / ft, applicable to formation water with a minerality <50000 ppm); the system automatically marks wells whose rate of variation of ρ exceeds 2% and triggers a manual check. This calculation can control the pressure error at the peak within a range of ±2%. The fiber optic sensor array can be a standard industrial-grade distributed sensing system; the vacuum pump must meet an extraction flow rate of 5 to 15 cubic meters per minute and a negative pressure capacity of -25 kPa; the pressure control valve is an electric ball valve with an opening control accuracy of ±5%. The surface sensor array is fixed to the surface layer of soil according to the grid coordinates; the wellhead pressure control valve is installed downstream of the main valve of the injection wellhead; the vacuum pump is connected to the monitoring pump wellhead piping by a flange clamp. In the present technology,The definition of thresholds is based on the correlation model between the lithology of the cover and the response to leaks: agricultural areas are mostly covered with clay (permeability < 10⁻⁷ m²), and a high threshold (CO₂ > 1.0%, pH < 5.5) is defined to avoid false alarms due to biodegradation; industrial areas are underlain by sandstone 20 (permeability > 10⁻⁶ m²), and a low threshold (CO₂ > 0.6%, pH < 5.2) is adopted to improve sensitivity; high-altitude areas add a temperature constraint (>-10°C) to prevent monitoring failure due to permafrost pore closure. Supporting data: statistics from a storage site show that the surface CO2 peak during a leak through a 25-degree shale cover is 40% higher than in lower-degree areas. Antifreeze design of sensors for high-altitude areas: for environments with a temperature <-10°C: the pH electrode is filled with the antifreeze electrolyte ethylene glycol (freezing point -35°C); the cable is fitted with a polyurethane sheath (brittleness temperature -60°C); when the temperature is ≤-30 15°C, automatically switch to wellhead monitoring mode (close the surface sensors and activate the wellhead gas analyzer). Compared to existing technologies, traditional methods rely solely on an alert mechanism based on a single carbon dioxide concentration threshold.with a false alarm rate of up to 30% to 40%. The present technical solution, thanks to two-parameter thresholds adapted to the regions (for example, 1.0% CO2 and pH < 5.5 for agricultural areas) combined with continuous verification of the maintenance of both indicators in compliance after regulation, reduces the false alarm rate to less than 10%, while avoiding the risk of a secondary leak due to a premature lifting of the alert. Implementation verification: In a storage project in a shale gas field, 400 sensor nodes were deployed with a spacing of 30 meters. The results of the simulated leak experiment show that: when carbon dioxide is artificially released to a concentration of 1.2% and a pH reduced to 5.3, the system triggers an alert signal in 6 minutes; the localization error of the reverse diffusion algorithm is less than 35 meters; after the response of the regulating well group, the carbon dioxide concentration within a radius of 1 kilometer decreases to 0.4% in 45 minutes and the pH rises back to 5.2.15. Meeting the alert lifting conditions. No false alarms or missed alerts were recorded throughout the process. Additional alert lifting conditions: In addition to the basic conditions (CO2 < 0.5%, pH > 5.0), the following conditions are added: if frequency analysis has been triggered, Eleak(f) < 0.5E₀ must be met for 1 continuous hour; if humidity compensation has been activated, γ > -0.02 pH / h must be met; temperature variation rates β > -0.3°C / h. Closed-loop management: In the case of Bohai Bay, the risk of secondary leakage has been reduced to less than 0.5%. According to another embodiment of the present invention, an intelligent alert and control method is provided for carbon dioxide leaks throughout the oxygen-enriched combustion and geological storage process: Deploy a grid of fiber optic sensors distributed across the surface of the target storage area, with a spacing of 40 meters. In the grid, the carbon dioxide concentration sensor detects, according to the non-dispersive infrared principle,with a measurement range of 0-5%; the soil pH sensor is a 30 solid electrode, with a measurement range of 3-9 pH; the surface temperature sensor is a PT100 platinum resistance sensor, with an accuracy of ±0.2°C; the 2025 / 5812 BE2025 / 5812 16 relative humidity sensor is capacitive, with a measurement range of 0-100%RH. All sensors are connected to a data collection unit by a shielded fiber optic cable. When a sensor node (for example, located in an agricultural area) continuously detects for 6 minutes that the percentage volume concentration of carbon dioxide exceeds 1.0% and that the soil pH is less than 5.5, the system starts a supplementary verification process: first, calculate the temperature variation in degrees Celsius at the surface of that node, requiring that the temperature fluctuation during the last hour be less than 0.3°C / hour; then, centered on that node, scan the adjacent nodes within a radius of 35 meters, requiring that at least 3 nodes simultaneously detect that the percentage volume concentration of carbon dioxide exceeds 0.8% and that the soil pH is less than 5.5. The temporal gap of the detections of adjacent nodes is controlled to less than 4 minutes. Formula for linking thresholds: Threshold of adjacent nodes = k × Threshold of a main node; k = {0.8 (agricultural / industrial zones), 0.7 (high altitude zones)}. Example: For the main node in an agricultural zone, CO₂ > 1.0% → For adjacent nodes, CO₂ > 0.8%. To meet the time synchronization requirement, the data collection unit adopts the precise IEEE 1588 time protocol, and the clock synchronization error between the different nodes is less than 10 milliseconds. The temperature change rate is calculated using the linear regression method, with a sampling interval of 5 minutes. Verification of the simultaneous detection of adjacent nodes uses a spatial grouping algorithm to automatically identify all valid sensors within a radius of 35 meters. In the present technical solution, the temperature sensor probe can be a PT100 platinum resistance probe.installed in the soil layer 5 centimeters below the surface; the data collection unit can be equipped with a GPS time synchronization module to ensure accurate time synchronization; the fiber optic communication cable can be a 30 shielded GYTA53 type fiber optic cable, buried at a depth of 0.4 meters. (The temperature variation rate threshold of 0.5°C / hour is defined based on daily temperature variation amplitude data; the radius range of 20 to 50 meters covers the initial diffusion diameter of a typical leakage plume; the minimum requirement of 3 knots ensures spatial correlation; the time deviation limit of 5 minutes corresponds to the gas diffusion velocity in the (underground cracks.) The temperature sensor probe is inserted vertically into the ground, with the sensitive part 5±1 centimeters from the surface; the data collection unit is installed in an explosion-proof junction box.which is fixed 10 meters from the nearest well; the fiber optic cable junctions are treated by welding and sealing, and the burial route avoids areas of mechanical work. Technical principle verification: A simulated experiment was conducted at a storage site in an abandoned oil field: when carbon dioxide was artificially injected until the concentration of a central node reached 1.2% and the pH dropped to 5.4, approximately 3 nodes were successively detected. Concentrations of 0.9% to 1.1% and a pH of 5.3 to 5.5 were detected in 3 minutes. The rate of temperature change during the same period was 0.2°C / hour. The system triggered an alert signal after satisfying all the additional conditions, without being affected by the temperature rise due to afternoon sunshine (0.8°C / hour). Compared to existing technologies,The traditional single-node threshold alert mechanism has a high false alarm rate when vehicles pass or fertilization work is carried out. The present technical solution eliminates short-term environmental disturbances by verifying temperature elasticity and local anomalies by verifying multi-node spatial coordination, reducing false alarms by two-thirds in agricultural areas. Applications in industrial areas show that it can effectively distinguish factory emissions from actual leaks. Implementation case: In a storage project in a coastal industrial area, a node detected a CO2 concentration of 0.65% (exceeding the industrial threshold of 0.6%) and a pH of 5.1 (below the threshold of 5.2). System verification revealed that only 2 adjacent nodes had been triggered within a radius of 30 meters (insufficient to reach the minimum of 3) and that the temperature variation rate of one node reached 0.6°C / hour (exceeding the limit of 0.5°C). It was determined that this was due to a temporary factory emission, and 2025 / 5812 BE2025 / 5812 18 no warning signal was triggered. Subsequent follow-ups confirmed the accuracy of this judgment, avoiding an unnecessary stoppage of injection operations. According to another embodiment of the present invention, an intelligent alert and control method is provided for carbon dioxide leaks throughout the oxygen-enriched combustion and geological storage process: 5 When a single sensor node in the target storage area (for example, a high-altitude area) continuously detects for 8 minutes that the volumetric concentration percentage of carbon dioxide exceeds 0.8%, that the soil pH is less than 5.3, and that the surface temperature is greater than -8°C, if fewer than 3 adjacent nodes trigger simultaneously within a radius of 40 meters, the system automatically locates the nearest monitoring pumping well to that node. Sends a pulsed pumping command to that well,controlling the wellhead pressure to bring it down from normal atmospheric pressure of 101kPa to -15kPa in 45 seconds. After starting the pulsed pumping, continuously collect the percentage of the volumetric carbon dioxide concentration of a target node and adjacent nodes in 8 surrounding directions (i.e., 9 nodes total) at a frequency of once per second. Calculate the rate of rise of the concentration of a target node, which must reach at least twice the average rate of rise of the 8 adjacent nodes, and the peak concentration must exceed 20% of the average value of 20 bases 60 seconds before pumping. (The average base value is the arithmetic mean of the concentration of a node during the 60 seconds preceding the sending of the pumping command.) In the present technical solution, the pulsed pumping equipment can be a two-stage vane vacuum pump, equipped with a regulating valve. electrical to achieve a linear pressure drop; gas concentration sampling can use a high-response infrared sensor,with a detection period ≤1 second; the arrangement of adjacent nodes follows a regular octagonal distribution, with a spacing consistent with the sensor grid (40 meters). (The pressure drop range of -5 to -20kPa covers the typical stratum rupture pressure gradient; the pressure drop time of 30 to 60 seconds corresponds to the technical specifications of the vacuum pump; the rate threshold of 1.5 to 3 times distinguishes the leak response from the non-leak response; the peak threshold of 15% to 30% avoids disturbances due to background fluctuations.) The vacuum pump inlet line is connected to the side valve of the production head of the monitoring pumping station; the pressure transmitter is installed on the right-hand section of the wellhead downstream of the main valve; the concentration sensor probe is 10±2 centimeters from the surface, avoiding obstruction by vegetation. Verification of the technical principle: A point leak was simulated on a permafrost storage site: the concentration of a central node increased from 0.82% to 1,15% in 120 seconds after the start of pulsed pumping (the peak exceeded the base value of 28%), with a rate of increase of 1.8% / minute; the average rate of the surrounding nodes was only 0.7% / minute. The ratio between the rate of the target node and the average of the surrounding nodes was 2.57 (>threshold of 1.5 times), and the system determined that it was a valid leak. Comparative experiments showed that with an artificial mechanical vibration perturbation, the peak of a target node exceeded the base value by only 12% and the rate ratio was 1.2 times, not triggering an alert. Compared to existing technologies, traditional methods require manual verification for isolated anomalies, with a response delay exceeding 2 hours. The present technical solution stimulates the characteristic leak response by pulsed negative-pressure pumping, combined with verification by spatial gradient comparison.allowing automatic discrimination to be determined in 10 minutes and increasing the detection rate of point leaks to over 90%. Implementation case: Node no. 203 of a platform storage project detected a CO2 concentration of 0.85% and a pH of 5.2, without triggering other 25 nodes within a 50-meter radius. After the execution of pulsed pumping (drop to -18kPa in 45 seconds): the concentration of one node increased from 0.84% ​​to 1.03% (the peak exceeded the base value of 19%); the rate of increase was 1.5% / minute, compared to 0.6% / minute on average for the 8 surrounding nodes (a ratio of 2.5 times); The system confirmed the leak and started regulation. Subsequent verification by drilling revealed a leak point in a deep crack, with a horizontal error of 23 meters relative to node #203. (Data from core filtration experiments: the amplitude response threshold of sandstone is 15%, which is the safety limit; the typical value of the measured pulsed response amplitude of sandstone is 18% to 25%, with a response time of 90 to 120 seconds; the pulsed response amplitude of clay is 8% to 12%.)with a response time of 180 to 240 seconds.) According to another embodiment of the present invention, an intelligent carbon dioxide leak warning and control method is provided throughout the oxygen-enriched combustion and geological storage process: When pulsed pumping does not reach the expected conditions, the system sends a triple-level pressure wave excitation instruction to the monitoring pump well. This process controls the wellhead pressure via an electrically controlled valve, and a pressure transmitter with an accuracy of 0.1% can be used to send back real-time signals. First, linearly regulate the pressure to -12 kPa in 10 seconds and maintain it for 180 seconds; then, gradually adjust it from -18 kPa to - 23 kPa, each pressure level being maintained for 180 seconds. The pressure regulation rate is controlled to less than 5 kPa / second to avoid a sudden change in stratum stress. During the maintenance of each pressure level,A non-dispersive infrared sensor with a response time of 0.5 seconds is used to continuously collect the volumetric percentage concentration of carbon dioxide from a target at 2-second intervals. The collected data is transmitted to an edge computing terminal, which performs a fast sliding Fourier transform (FFT) on the 90 concentration data points recorded every 180 seconds (60-second window width and 10-second increments). The analysis focuses on the low frequency band from 0.01 to 0.1 Hz, which corresponds to the characteristic frequency of gas flow in the storage layer. The background energy value of the frequency band E₀ is taken from the reference data of the 120 seconds preceding the start of the vacuum excitation. When the frequency band energy exceeds 2.0 times E₀ for 90 seconds continuously at any pressure level, it is determined to be a valid fault signal. In the present technical solution,The pressure regulating valve is installed on the straight section of the monitoring pump well inlet line, 1.5 meters from the wellhead flange collar, ensuring the elasticity of the flow field; the sensor design cable is a double-shielded cable, with a grounding resistance of less than 4Ω to suppress disturbances; the selection of the frequency band of 0.01 to 0.1Hz is based on data from previous core filtration experiments, this band being the most sensitive to leaks in micro-cracks; the three defined pressure levels (-12 / -18 / -23 kPa) cover the critical value of the typical rupture pressure gradient of the storage layer. In the present technical solution, a noise-reduction design for frequency analysis can also be implemented: 1. Simultaneously collect the seismograph data (frequency band of 0,0.1 to 10 Hz) to construct a background noise spectrum; 2. Apply the following formula to the CO2 concentration signal: Eleak(f) = Etotal(f) - k•Eoise(f) (where k is dynamically calibrated by the background noise spectrum, constraint: k ∈ [0, 8, 1, 2]); 3. Determine the signal as valid only when Eleak > 1.8E₀ in the frequency band from 0.05 to 0.1 Hz. (Effect: On-site tests of a shale field show that the rate of false alarms due to mechanical vibrations decreased by 85%.) Alternative solution (low-cost scenario): Reuse the distributed acoustic detection system (DAS) to collect 20 the vibration signals, and apply the disturbance elimination formula: , where x(t) is the pump vibration signal, y(t) is the concentration signal; a signal is considered valid when Rxy(t) < 0.3. Operational case: On a storage layer site, after the failure of the 25 pulsed pump verification, triple-level pressure wave excitation is started: 1. Level -12kPa: The concentration fluctuation of a target fluid has increased,but the peak energy of the frequency band from 0.01 to 0.1 Hz only reached 1.8 times E₀; 2. Level -18 kPa: The energy of the frequency band exceeded 2.0 times E₀ at the 40th second and was maintained until the 130th second (compliant duration of 90 seconds); 3. The system determined the leak as valid and automatically triggered the regulation process. Subsequent drilling verification revealed a crack 50.2 mm wide at a depth of 1020 meters. The excitation by the pressure wave periodically modifies the pressure at the bottom of the well, stimulating the oscillatory movement of the gases in the potential leak channels. The concentration signal near the leak point exhibits specific low-frequency fluctuations (0.01 to 0.1 Hz).While environmental disturbances (such as mechanical vibrations) are mostly concentrated in the frequency band >1Hz, extracting energy from the characteristic frequency band by FFT effectively distinguishes real leaks from noise. Traditional static pressure tests require the well to be shut down for more than 24 hours, whereas the present method completes the verification in 18 minutes, with a significant improvement in efficiency. In the present technical solution, the holding time of 120 to 240 seconds per level guarantees that the pressure wave propagates to strata above 500 meters; the sampling interval of 1 to 5 seconds corresponds to the gas diffusion relaxation time; the energy threshold of 1.8 to 2.5 times is defined based on the statistical distribution of background noise. The on-board calculation terminal is deployed in an explosion control room.with an ambient temperature range of -20°C to +60°C; the installation direction of the pressure transmitter is perpendicular to the direction of fluid flow to avoid dynamic pressure error; the concentration sensor probe has a protection level of IP68, suitable for rainy and snowy environments. According to another embodiment of the present invention, an intelligent alert and control method for carbon dioxide leaks is provided throughout the oxygen-enriched combustion and geological storage process: Before generating the probability map of the leak source position, deep pressure monitoring is implemented on the priority monitoring wells of the target control well group, which represent, as of BE2025 / 5812 23, more than twenty percent of the total number of wells. Taking as an example a group of 10 wells: 2 priority wells are selected for the installation of deep pressure sensors, which may be piezoresistive sensors resistant to hydrogen sulfide corrosion,with a measurement range of 0-50 MPa, installed 10 meters above the upper interface of the storage layer. For the 85 other non-priority wells, data is collected by a wellhead pressure transmitter (class 0.1 accuracy), and the pressure at the top of the layer is calculated by combining the hydrostatic pressure formula of the liquid column: Ptop = Pwellhead − ρ ⋅ g ⋅ H, where ρ (density of the fluid in the well) is taken as 820 kg / m³ (calibrated according to the composition of the fluid in the well), H₁₀ (depth of the storage layer) is 1500 meters, g (acceleration due to gravity) is 9.8 m / s², and Pwellhead (wellhead pressure) is obtained. actual by the transmitter. After calculation, the pressure error at the top of the test is controlled within a range of ±0.3 MPa. The variation of the pressure gradient of each well at the depth of the upper interface of the storage layer is calculated every 30 minutes. When the variation of the pressure gradient of a well exceeds 0.8 kilopascals per meter (for example, going from 12.3 kPa / m to 13.5 kPa / min 24 hours),It is marked as an abnormal pressure source. Take the coordinates of this abnormal point as a constraint, and update the permeability distribution field of the three-dimensional geological model using the Kriging interpolation algorithm. Interpolation parameters: range of 500 meters, nugget effect of 0.01, plateau value of 0.15. The updated permeability field is fed into the inverse diffusion algorithm to regenerate the probability map of the leak source. In the present technical solution, the deep pressure sensors can be lowered to the design depth by means of a shielded cable, with a cable seal equipped with a double O-ring; the wellhead pressure transmitter is installed on the production head pressure interface, with a pressure tap tube length of less than 1 meter; the calculation Kriging interpolation can be deployed on a cloud server,calling the historical borehole permeability data as a priori field. Operational case: During the location phase of a storage project in a 2025 / 5812 BE2025 / 5812 24 saline aquifer layer, the initial model indicated a probability of 72% for zone A. After the execution of the claim process5: 1. The priority monitoring well B23 measured a pressure gradient variation of 1.05 kPa / m (exceeding the threshold of 0.8); 2. For the non-priority well B07, Ptop=28.7 MPa was calculated by the hydrostatic pressure formula5, with a gradient variation of 0.92 kPa / m; 3. Using B23 and B07 as constraint points to update the permeability field: the initially low-permeability zone (5 mD) was corrected to 12 mD; 4. The inverse diffusion algorithm regenerated the probability map, and the 10th fault source was corrected to zone C (75% probability). Subsequent borehole verification revealed the presence of faults in zone C, with a horizontal error of 17 meters. Compared to traditional static geological models,whose average location error reaches 40 to 60 meters due to the lack of real-time permeability updates, the present technical solution reduces the location error to less than 20 meters thanks to the dynamic correction of the model by marking abnormal pressure gradients (>0.8 kPa / m). The application of the hydrostatic pressure formula has made it possible to double the data utilization rates for wells lacking depth sensors. (The basis of the 20 key parameters includes: the threshold of 0.5 to 1.2 kPa / m is based on mechanical verification of the storage stratum; the ratio of 20% of priority wells balances the cost and reliability of the data; the 500-meter range of the Kriging interpolation corresponds to the density distribution of distances between wells.) In the present technical solution, the fluid density is calibrated by 25 monthly samplings, with a temperature compensation range of -10°C to 80°C; the pressure gradient calculation uses a sliding time window (window width of 24 hours,(not 1 hour); the difference between the result of the permeability update module and the log interpretation data is <15%.30 According to another embodiment of the present invention, an intelligent alert and control method for carbon dioxide leaks is provided throughout the oxygen-enriched combustion and geological storage process: During the calculation of the well pressure gradient variation, the temperature data of the target well section are obtained simultaneously by the distributed temperature detection (DTS) system. Taking the upper interface of the storage layer as a reference point, fiber optic temperature sensors are deployed at 1-meter intervals in the section of 50-meter wells below the upper interface; the DTS system can be used with a temperature measurement accuracy of ±0.5°C. When the variation of the pressure gradient at a certain depth point exceeds 0.9 kilopascals per meter (for example, increasing from an initial 12.1 kPa / m to 13.3 kPa / m),immediately detect the 10 rate of temperature change of the 10-meter well section above and below centered on this point. The linear regression algorithm is used to calculate the axial temperature gradient of this 20-meter section; if the detected rate of temperature change exceeds 1.5°C / 100 meters and the temperature shows a downward trend (for example, going from 35.2°C to 34.1°C), retain the abnormal pressure source label; otherwise, remove the label. (Criteria for determining the direction of temperature variation: a negative variation is considered when the regression slope is less than -0.015°C / meter.) Implementation case: In a coal seam storage project, well B05 recorded a pressure gradient variation of 1.1 kPa / m at a depth of 1020 meters. Temperature detection showed that the temperature gradient of the well section 1020 ± 10 meters (centered on this point) was -1.8°C / 100 meters (i.e., a temperature decrease of 1.8°C per 100 meters).satisfying the negative variation condition. The system retained the abnormal label and updated the permeability field; subsequent localization revealed a 0.3 mm crack at this location. For comparison well B12, the pressure gradient variation was 0.95 kPa / m, but the temperature gradient was +0.7°C / 100 meters (upward temperature trend); the system removed this abnormal point to avoid an error in judgment. 30 Technical principle of temperature verification: A real leak causes CO2 to dissolve in the soil pore water to form carbonic acid (H2CO3), and the endothermic ionization reaction (ΔH = -24.3 kJ / mol) causes a 2025 / 5812 BE2025 / 5812 26 Local cooling. According to thermal conduction simulations, the typical surface temperature variation rate β = -0.5 ± 0.3°C / hour, a characteristic absent from disturbances such as mechanical failures. The distributed temperature sensors are deployed along the well's shielded fiber optic cable, which is fixed to the external wall of the production tubing.with a depth error of less than 0.5 meters. Temperature data are updated every 30 seconds, and the pressure gradient calculation uses a 24-hour moving window. In the present technical solution, the threshold of 1.0 to 2.0°C / 100 meters is based on leak simulation experiment data, with a minimum detectable temperature variation of 0.3°C; the temperature sensor's fiber optic cable is fixed by pipe centralizers to ensure its adhesion to the pipe wall; The slope of -0.015°C / meter for determining the negative variation corresponds to an inverse fluctuation of the geothermal background gradient of 0.03°C / meter.15 Compared to traditional methods that rely on a single pressure indicator, where approximately 35% of abnormal readings are false alarms, the present technical solution, after adding verification of the direction of temperature variation, reduces the error rate to less than 5%. On-site applications show that: when the variation of the pressure gradient20 exceeds the threshold,The negative temperature variation can effectively distinguish actual leaks (92% probability of cooling) from injection-recovery fluctuations (85% probability of heating). (The key equipment used is an industrial-grade fiber optic distributed temperature measurement system with a maximum operating temperature of 150°C.) 25 According to another embodiment of the present invention, an intelligent carbon dioxide leak warning and control method is provided throughout the oxygen-enriched combustion and geological storage process: When the abnormal pressure source label is removed because the temperature check has not been validated, the system injects 30% sulfur hexafluoride (SF6) tracer gas at ambient temperature into the pumping well. corresponding monitoring. The injection process is controlled by the wellhead injection valve, with a fixed flow rate of 0.3 cubic meters per minute and a continuous duration of 120 seconds. After mixing the tracer gas with the residual well gases,This migrates towards the surface along potential leak channels. A timer is started at the end of the injection, and surface detection is initiated within a time window of 60 to 180 seconds: centered on the surface projection of the abnormal pressure point, all sensor nodes 5 within a radius of 100 meters switch to sulfur hexafluoride monitoring mode, using a laser spectrometry sensor to perform continuous sampling, with a detection limit of 0.1 ppm. If the concentration of sulfur hexafluoride detected by any of the nodes within the radius exceeds 1 ppm and lasts more than 10 seconds continuously, the leak channel 10 is marked immediately and an alert signal is triggered. Operational example: In a salty cavity storage site, the C18 well recorded a variation in the pressure gradient of 1.05 kPa / m, but the temperature check showed a positive gradient variation (+1.2°C / 100 meters),The abnormal label was therefore removed. The trace gas injection was then carried out: a flow rate of 0.3 cubic meters per minute was maintained for 120 seconds. At 85 seconds after the end of the injection, a node located 80 meters in the northeast direction detected a sulfurhexafluoride concentration of 1.7 ppm for 15 seconds, and the system confirmed the existence of the leak channel. Verification by directional drilling revealed that this path was due to the activation of an abandoned fault. In the present technical solution, the wellhead injection valve is installed on the test orifice of the production head, with a pressure resistance level of at least 25 MPa; the injection line is equipped with a static mixer to improve the mixing of the gas source; the laser spectrometry sensor 25 can use tunable diode laser absorption spectrometry (TDLAS) technology, with a spectral range of 10.55 μm,adapted to an environment of -30°C to 50°C; the window of 60 to 180 seconds is defined according to the typical migration speed of the tracer in the fractured medium; the condition of 1 ppm for 10 seconds avoids instantaneous environmental disturbances; the 30 sensor has an IP67 protection level and is buried directly 20 centimeters below the surface. Technical principle: Sulfur hexafluoride (SF6), as an inert gas, has an extremely low background concentration (<0.01 ppm), and its detection can clearly indicate the gas migration path. Compared to traditional pressure tests, this method allows direct visualization of the leak channel, and detection by laser spectrometry avoids designation delays due to adsorption by the soil. During implementation, it is necessary to note: before injection,5. The vacuum pump of the monitoring pumping station must be closed to prevent reverse gas flow; the surface nodes must block carbon dioxide analysis during the detection period and focus on tracer monitoring; multi-node time-series analysis can reverse the direction of the leak channel. 10. This technical solution is an ultimate means of verification for anomaly discrimination. When dual verification by pressure and temperature fails, active injection of tracer gas allows visualization of the leak channel. On-site applications show that this method can effectively identify fracture channels at the millimeter level, without being affected by the stratum temperature field. 15. Compared to borehole exploration, operational efficiency is significantly improved. When the project budget is limited: environmentally friendly tracer gases can be adopted: perfluorotert-butylamine (C4F9N,LD50=5200 mg / kg) for ecologically protected areas; sulfurhexafluoride (equipped with a 20-gas recovery device) for industrial areas, its background concentration being close to 0; existing PID sensors can be reused (measuring range 0 to 50 ppm) with a detection limit of 0.5 ppm; the determination threshold can be adjusted to >1.5 ppm for 10 seconds continuously. The use of the gamma-ray densitometer complies with the safety standard GB18871-2002.25 According to another embodiment of the present invention, an intelligent alert and control method for carbon dioxide leaks is provided throughout the oxygen-enriched combustion and geological storage process: When the frequency band energy E exceeds 2.0 times the background energy E₀ for any pressure level,The system simultaneously starts the temperature verification process. The corresponding surface temperature data of a target during the pressure holding phase are obtained by the platinum resistance temperature sensors already deployed (accuracy ±0.1°C), with a sampling interval of 20 seconds. The temperature change rate β (unit: °C / hour) is calculated by the linear regression method, according to the formula: β = ΔT / Δt × 3600, where ΔT is the temperature difference between adjacent sampling points (°C) and ΔT is the 20-second sampling interval. If the value of β remains constant for 100 seconds in the range of -1.0 to -0.6°C / hour (e.g., fluctuating from -0.8°C / hour to -1.2°C / hour), the leak alert signal is confirmed as valid; otherwise, it is considered a disturbance signal. Implementation case: In a shale storage site, during the 10-minute pressure maintenance phase at -18 kPa, the frequency band energy reached 2.3 times E₀. Synchronous temperature monitoring showed: Minute 10-30: β = -0,5°C / hour (continuous duration of 20 minutes); Minutes 30-50: β = -0.7°C / hour (continuous duration of 20 minutes); as the value of β remained continuously for 100 seconds within the range of -0.7 to -0.5°C / hour, the system confirmed the leak. It was subsequently observed that the cooling was due to heat absorption by the CO2 leak. In the present technical solution: 1. Homogeneous data collection: Use of surface temperature sensors, avoiding the addition of extra equipment; 2. Rate of change algorithm: A linear regression is performed on each series of 3 consecutive sampling points (60-second data window), with an update frequency of the value of β of 1 time / minute; 3. Time synchronization mechanism: The timestamp difference between the temperature data and the frequency analysis is less than 1 second; 25 4. Determination logic: The period during which β continuously exceeds the threshold must be completely included in the pressure maintenance phase. In the present technical solution,The temperature sensor probe is exposed to the surface, 5±1 centimeters from the ground-air interface; linear regression eliminates outliers (for example, the points of change where |ΔT|>0.5°C / 2030 seconds); Filtering of background temperature disturbances: when the wind speed > 5m / s, the verification time is automatically extended to 120 seconds. 2025 / 5812 BE2025 / 5812 30 Technical Principle: A real leak is accompanied by the dissolution of CO2 in the soil's pore water to form carbonic acid (H2CO3), and the endothermic ionization reaction (ΔH = -24.3 kJ / mol) leads to local cooling of the soil through heat absorption. According to thermal conduction simulations, the typical value of β is -0.5 ± 0.3 °C / hour. Factors unrelated to leaks, such as electrical disturbances,do not exhibit this characteristic of temperature variation. This step reduces false alarms from frequency analysis to an acceptable level thanks to the double verification of the direction and amplitude of the temperature variation. Cooling principle: 10 1. Formula for calculating the quantity of heat absorbed by the ground per unit volume: Qdiss=n×ΔH. 2. Formula for calculating the rate of temperature change: ΔT=(Qdiss×t) / (Cp×ρsoil×d) Where: Qdiss: Quantity of heat absorbed by the soil per unit volume (J / m3); n: Quantity of CO2 dissolved under (taken at 0.1 mol / m³·s, estimated on the basis of a leakage rate of 1 kg / s); ΔH: Enthalpy change of the reaction, taken at -24.3 kJ / mol; Cp: Specific heat capacity of the soil, taken at 1.5 kJ / kg·K; ρsoil: Density of the soil, taken at 1600 kg / m3; d: Depth of thermal impact, taken at 0.2 m; Calculation result: ΔT ≈ -0.6°C / hour (the corrected threshold of β is -0.8~-0.3°C / hour). Typical application case: In a storage site in an industrial area, a frequency band energy 20 triggered the threshold of 2.1 times E₀,However, the measured value of the temperature change rate β was +0.3°C / hour (increasing trend). The system determined that this was due to a disturbance caused by vibration of the nearby compressor, which blocked the false alarm. (The basic equipment parameters include: the 20-second sampling interval corresponds to the thermal inertia response of the surface; the continuous duration requirement of 90 to 120 seconds guarantees the entire thermal conduction process; the threshold of -0.8 to -0.3°C / hour is based on thermodynamic simulations of leak conditions.) According to another embodiment of the present invention, an intelligent warning and control method for carbon dioxide leaks is provided throughout the oxygen-enriched combustion and geological storage process: 2025 / 5812 BE2025 / 5812 31 When the temperature variation rate β is not within the range of -1.5 to -0.5°C / hour but the frequency band energy E > 2.0 times E₀,If the target area belongs to a high-humidity climate zone with an annual relative humidity greater than 70% (e.g., a coastal marsh area), the system starts the moisture compensation process. The relative humidity of a target soil (H5) is continuously monitored by a surface capacitive moisture sensor with a sampling interval of 5 minutes. If H5 is greater than 85% continuously for 30 minutes (e.g., fluctuating from 88% to 92%), the monitoring duration of the soil pH is extended from 60 seconds (standard) to 200 seconds (extended to 10 minutes for clay / humus). During the extended period, the pH is recorded every 10 seconds, and the least squares method is used to calculate the pH. slope of variation γ (unit: pH / hour). When γ is in the range of -0.10 to -0.08 (i.e., a decrease in pH of 0.10 to 0.08 units per hour), the leak alert signal is confirmed as triggered. (Adaptive adjustment to soil type: Monitoring duration = {10 minutes (clay / humus), 200 seconds (sandy soils)}; Case of clay zone 15: In a marsh, the pH has decreased from 5.40 to 5.32 (γ = -0.12 pH / h,(detected = 8 minutes).) Implementation case: In a mangrove storage site, the frequency band energy triggered 2.4 times E₀, but β = -0.3°C / hour (not compliant with criteria). The system detected an annual humidity of 82% in the area and a real-time humidity of 86% to 91% continuously for 35 minutes. Monitoring of the pH was extended to 200 seconds, and it was measured that the pH decreased linearly from 5.35 to 5.28 (calculated γ = -0.09 pH / hour). As the value of γ fell within the valid range, the system confirmed the leak. It was subsequently observed that the slow decrease in pH was due to the accelerated dissolution of CO2 in the surface humus. In the present technical solution: 1. Humidity monitoring: Use of a relative humidity sensor, with a measurement range of 0 to 100% RH and an accuracy of ±2% RH; 2. Enhanced measurement mechanism: Standard mode: Output of the average RH value once per minute; Enhanced mode: Acquisition of raw data every second.generation of 20 2025 / 5812 BE2025 / 5812 32 valid values ​​in 200 seconds; 3. Calculation of the slope: γ=ΔpH / Δt×3600 Where ΔpH is the variation of pH in 200 seconds, and Δt=200 seconds; 4. Determination of the climate zone: Called upon the basis of historical meteorological data; a zone is marked as a high humidity zone if the average daily humidity over the last 365 days is >70%. In the present technical solution, the pH sensor electrode is calibrated monthly with pH buffer solutions of 7.0 / 4.0; an anti-condensation heating film is automatically activated in high-humidity environments (after rain); outliers are eliminated from the γ calculation (e.g., a sudden variation >0.1 pH in a single measurement). Technical principle: The dilution effect of water in soil at high humidity delays the pH response, and extending the monitoring period allows the slow acidification process to be captured. The range of negative γ values ​​is defined as follows: -0.15 to -0.05pH / hour corresponds to the typical dissolution rate of CO2 leakage; the extended duration of 200 seconds covers the ionic diffusion cycle of soil interstitial water; the moisture threshold of 85% is based on correlation experiments between the water content of clay and electrical conductivity. Typical case: In a rainforest storage project, a frequency band energy anomaly was detected (E=2.1E₀), but β=-0.4°C / hour. The system detected a real-time humidity of 89% for 40 minutes continuously and started extended monitoring of pH. In 200 seconds, pH decreased from 5.41 to 5.38 (γ=-0.12pH / hour), outside the valid range. It was determined that this was due to a disturbance by humic acid decomposition.avoiding false triggering. (Key equipment parameters: The pH sensor uses an antimony electrode with an anti-mud coating; the humidity compensation algorithm is integrated into the onboard computing platform; the API response time for the historical weather data call is <2 seconds.) Implementation method: Complete management of a leak event at a storage site in a saline aquifer in the Gulf of Bohai. 30 Project context: 2025 / 5812 BE2025 / 5812 33 An oil company in the Gulf of Bohai stores carbon dioxide generated by oxygen-rich combustion in a saline aquifer at a depth of 1500 meters. The surface of the storage area is an interleaved zone of soils saline-alkaline and industrial zones. 512 monitoring nodes were deployed at 45-meter intervals, each node integrating a carbon dioxide concentration sensor (measurement range 0 to 10%), a soil pH electrode (measurement range 3 to 10 pH), as well as temperature and humidity probes. Three groups of regulating wells were deployed,Each group comprises 2 injection wells and 1 pumping well, with a distance between wells of 280 to 320 meters. Chronology of the leak event: 1. Initial alert triggered: The industrial node S15 detected a CO2 concentration of 0.68% (exceeding the industrial threshold of 0.6%) and a pH of 5.18 (below the threshold of 5.2a), for 7 minutes continuously; The adjacent nodes S14, S16 and S22 within a radius of 40 meters were triggered simultaneously: S14: Concentration 0.72% / pH 5.15. S16: Concentration 0.65% / pH 5.22. S22: Concentration 0.71% / pH 5.19. The temperature change rate of all nodes was <0.4°C / hour, and the time difference ≤4 minutes; The system determined that this was a valid fault alert and marked S15 as the initial point. 2. Dynamic localization and verification: Initial localization: The inverse diffusion algorithm indicated a probability of 65% for zone A (not reaching the 70% threshold); Model update: The depth gauge at priority well IW-3 showed a pressure gradient change of +1.12 kPa / m (exceeding the 0,8); For the non-priority well IW-5, a gradient variation of +0.95kPa / ma was calculated by the hydrostatic pressure formula; Taking the abnormal points as constraints to update the permeability field, the probability of the Ba zone increased to 76%; Spot check: Only node S39 in the Ba zone triggered (concentration 0.83% / pH 5.24); 5 Pulsed pumping execution on the nearest pumping well E-2: the wellhead pressure was reduced to -17kPa in 48 seconds; The peak concentration of S39a reached 1.05% (exceeding the base value of 19%), and the ratio between the rising levels and the average of the surrounding nodes was 2.3 times; 10. Confirmation of the leak and start of regulation: the opening of the injection well valve IW-4 was reduced to 55%, and the pumping well E-2 operated at a flow rate of 12 m³ / min. 3. Intervention / additional verification: After regulation, the frequency band energy of a node S41a reached 2.2 times the 15 background value; Temperature verification: The surface temperature variation rate β = -0,6°C / hour (continuous duration of 105 seconds), satisfying the determination criteria; Humidity compensation: 20 The energy of the frequency band of a node S33 was 2.1 times the background value, but β=-0.2°C / hour; The system detected an annual humidity of 78% in the area and a real-time humidity >87% continuously for 38 minutes; Extension of the pH monitoring to 200 seconds: the pH decreased from 5.42 to 5.3825 (γ=-0.11pH / hour), outside the valid range, determined as a disturbance. Regulation lifted and verification: 56 hours after regulation: The CO2 concentration of all nodes within a 1-kilometer radius of zone B was <0.45%; the pH was stable between 5.1 and 5.3; the system triggered the alert, and the injection well was returned to its initial opening. Verification by drilling: Drilling to a depth of 1482 meters at the coordinate point of zone B: 5. Discovery of a 0.3 mm wide crack,with a horizontal error of 21 meters; gas detection in the crack showed a CO2 proportion of 89%. Demonstration of technical advantages: 1. Adaptability of thresholds by zone: The industrial thresholds (0.6% CO2 + 5.2 pH) effectively distinguish industrial emissions (temporary exceedance of the concentration but no pH variation) from actual leaks. 2. Dynamic localization accuracy: The pressure gradient marking controls the updating of the permeability field, reducing the localization error from the initial 42 meters to 21 meters. 3. Reliable multi-level washing: Pulsed pumping confirms the localized leak (S39); coupled temperature-frequency washing (S41) blocks electrical disturbances; moisture compensation (S33) prevents false positives in saline-alkaline soils. 4. Efficient closed-loop control: Fully automated process from alert to recovery,without resorting to tracer gas washing.20 The number of pieces of equipment and the scale of treatment described herein are used to simplify the description of the present invention. The applications, modifications and variations of the present invention are obvious to persons competent in the field. Although embodiments of the present invention have been described above, they are not limited to the applications listed in the specification and implementations, and can be fully applied to various fields suitable for the present invention, and further modifications can be easily implemented by a person skilled in the art. Consequently, the present invention is not limited to the specific details and illustrations shown and described herein without departing from the general concept defined by the claims and the equivalent scope.