Gas leak monitoring system, gas leak monitoring method, and computer program product

Abstract

The invention relates to a gas leakage monitoring system and method, computer equipment and a computer program, and the system comprises a pressure sensor which is disposed on the inner wall of a gas storage shaft, is in communication connection with a controller, and is used for detecting the gas pressure in the gas storage shaft; the gas sensors are arranged on the inner wall and the outer wall of the gas storage shaft, are in communication connection with the controller and are used for detecting the gas concentration inside the gas storage shaft and the gas concentration outside the gas storage shaft; the ultrasonic sensor is arranged on the inner wall of the gas storage shaft, is in communication connection with the controller and is used for detecting a sound wave signal generated by gas leakage; and the controller is used for judging whether gas leakage occurs in the gas storage shaft or not through a preset risk grade judgment model according to the gas pressure, the gas concentration and the sound wave signal. According to the utility model, the monitoring accuracy and sensitivity are improved, real-time monitoring and early warning can be realized, the leakage position can be accurately positioned, and the safety and reliability of the monitoring system are improved.

Description

Technical Field

[0001] This disclosure relates to the field of gas leak detection technology, and in particular to a gas leak monitoring system, a gas leak monitoring method, and a computer program product. Background Technology

[0002] As one of the main alternatives for future energy transition, safe hydrogen energy terminal storage is a key scientific issue affecting its rapid development and widespread application. Compared with traditional equipment such as hydrogen storage cylinders and tanks, underground well hydrogen storage (hydrogen storage wells) is an emerging high-pressure hydrogen storage technology that has attracted much attention due to its advantages in saving land area, shortening safety distances, controlling failure range, anti-static properties, and reducing subsequent operating costs, and has become one of the hot topics in research and industrial application.

[0003] However, due to the flammability, explosiveness, wide combustion range (4%–75%), low ignition energy, and high diffusion coefficient of hydrogen, the casing of gas storage wells faces intense high-pressure cycling and temperature fluctuations under long-term complex service environments. This further leads to material damage and equipment failure, resulting in serious accidents such as leaks, fires, and even explosions. On the other hand, the high design hydrogen pressure (up to 90 MPa) within the gas storage well is one of the main reasons for the risk of hydrogen leakage in the wellbore, wellhead, and other components during operation.

[0004] However, research on safety precautions (leakage) for hydrogen storage wells is generally lacking in the industry. Most studies focus on hydrogen diffusion and leakage behavior in vehicle-mounted hydrogen supply systems (hydrogen tanks, hydrogen cylinders), buried hydrogen transportation underground storage well systems, underground injection and production tubing leakage monitoring systems, hydrogen storage facilities, and hydrogen refueling stations. Existing monitoring technologies have the following limitations:

[0005] 1. On-board hydrogen supply system monitoring technology. This involves leak detection of hydrogen storage tanks and cylinders, primarily relying on a combination of mechanical and hydraulic sensors, or hydrogen sensors, to monitor hydrogen concentration and airtightness. Its advantages lie in its simplicity, speed, and reliability, enabling timely detection and handling of hydrogen leaks. For example, invention patent CN114060719A, "A Hydrogen Leakage Detection Device for a Hydrogen-Powered Bus," describes a hydrogen leak detection device for a hydrogen-powered bus. By setting up a main hydraulic component and a secondary hydraulic component, in the event of a leak, the device uses hydraulic transmission to quickly determine whether a leak has occurred at the connection between the output pipe and the branch pipe of the hydrogen storage cylinder's output port by judging whether an indicator rod has moved out of the cover. However, this technology is only suitable for relatively simple on-board environments and can only detect leaks at connections, lacking in real-time monitoring of hydrogen leaks and precise location of leaks.

[0006] 2. Monitoring Technology for Buried Hydrogen Transportation Underground Gas Storage Wellbores. Monitoring technology for buried hydrogen transportation underground gas storage wellbore systems is relatively mature, primarily employing fiber optic sensing for leak detection. However, these technologies are mostly used for monitoring long-distance underground gas storage wellbores and have poor adaptability to high-pressure environments. For example, invention patent CN115751206A, "A Device and Method for Detecting Hydrogen Leakage in Underground Hydrogen Transportation Wellbores," describes a device for detecting hydrogen leakage in underground hydrogen transportation wellbores and a method for using this device to detect hydrogen leakage. It mainly includes an accelerometer sensor and a computer. The accelerometer sensor is installed at both ends of the underground hydrogen transportation wellbore. By detecting voltage fluctuation signals, the computer locates the leak point based on the signal time difference and the distance between the sensors. It relies on the accuracy of the accelerometer sensor to capture the voltage fluctuation signal. The accuracy and sensitivity of the sensor directly affect the accuracy of leak detection; any sensor malfunction or decrease in accuracy will affect system performance. Accelerometers are sensitive to vibration and electromagnetic interference. Environmental noise, mechanical vibration, and electromagnetic interference around underground gas storage wells used for long-distance hydrogen transportation can affect the accuracy of signal detection, leading to false alarms or missed alarms. Accelerometers must be installed at both ends of the underground gas storage well, which limits their practical application.

[0007] 3. Underground Injection and Production Tubing Leak Detection Technology. Underground injection and production tubing systems are used for leak detection during underground gas injection or production processes, primarily relying on devices such as temperature and pressure sensors. For example, invention patent CN116517527A, "An Underground Injection and Production Tubing Leak Monitoring Device and Method," describes an underground injection and production tubing leak monitoring device and method. This method monitors the temperature of the annular fluid at the leak point using a temperature monitoring module, monitors the density change of the annular fluid using a density monitoring module, and monitors the fluid pressure of the annular fluid at the leak point using a fluid pressure monitoring module, determining the first leakage rate at the leak point. However, this method relies heavily on monitoring the temperature, density, and fluid pressure changes of the annular fluid to determine the leak point, and cannot effectively detect leaks in other parts of the tubing, such as leaks at the interface of underground gas storage wells. Furthermore, the leak monitoring device may have insufficient response speed, failing to detect small or rapid leaks in real time, thus delaying the response to leak events. This method is more suitable for underground injection and production tubing systems. For other types of underground gas storage well systems or gas storage wells, it is not possible to fully monitor the hydrogen diffusion inside the well and in the surrounding environment, and improvements are needed.

[0008] 4. Hydrogen Storage Monitoring Technology. Hydrogen storage monitoring technology typically involves the combined application of temperature, pressure, and hydrogen concentration sensors to monitor hydrogen leaks over a wide area. For example, invention patent CN117028839A, "An Intelligent Monitoring Method and Early Warning System for Hydrogen Storage Leaks," discloses an intelligent monitoring method and early warning system for hydrogen storage leaks. This system primarily uses a temperature and pressure demodulator, a temperature-measuring optical transceiver, a pressure sensor, and an industrial control computer to achieve real-time monitoring of temperature and pressure within the storage facility. The system determines the storage facility's sealing performance by comparing the difference between the actual and calculated pressure. However, relying solely on temperature and pressure changes for leak detection may not comprehensively reflect all leaks within the storage facility. Environmental factors such as temperature and pressure fluctuations, both internal and external to the storage facility, can affect measurement results, leading to false alarms or missed alarms. Widespread deployment of such equipment in large storage facilities may face implementation difficulties and high maintenance costs.

[0009] 5. Hydrogen Refueling Station Leakage Monitoring Technology. Leakage monitoring technology for hydrogen refueling stations is primarily used for monitoring surface facilities. For buried gas storage wells, monitoring methods are limited and insufficient to meet the leakage detection needs under high-pressure environments. For example, invention patent CN114882669A provides a hydrogen refueling station safety monitoring and early warning system, mainly including a hazard source monitoring unit, a transmission analysis unit, and a monitoring alarm unit. This system monitors environmental data at preset monitoring points within the hydrogen refueling station in real time, analyzes this data to predict operational safety, and issues an alarm when data exceeds preset thresholds. However, the system alarm relies on preset thresholds. If the thresholds are not set appropriately, false alarms or missed alarms may occur. Furthermore, the hazard source monitoring unit requires preset monitoring points; if the distribution of monitoring points is not reasonable or the coverage is insufficient, some critical areas may be missed, resulting in incomplete monitoring of the hydrogen refueling station's safety. Simultaneously, it lacks advanced intelligent analysis algorithms, relying solely on simple threshold judgments, and cannot fully utilize monitoring data for in-depth analysis and prediction.

[0010] 6. Experimental Simulation Technology. This refers to experimental apparatus and testing methods used to simulate hydrogen leakage and diffusion in hydrogen storage wells. For example, invention patent CN117571554A, "An Experimental Apparatus and Testing Method for Simulating Hydrogen Leakage and Diffusion in Hydrogen Storage Wells," discloses a testing method for simulating hydrogen leakage and diffusion in hydrogen storage wells, used to detect the hydrogen concentration in different formations after a hydrogen leak. It is mainly used for simulation testing in a laboratory environment and does not involve real-time monitoring of hydrogen-contaminated structures such as the casing and cement sheath of actual gas storage wells, thus reducing its practical guiding significance.

[0011] Because hydrogen leakage from underground hydrogen storage wells differs from that of other storage and transportation facilities, the possibility of hydrogen leaking into cement, the first and second interfaces of the cement sheath, and even into the surrounding soil layers requires special attention.

[0012] Therefore, existing gas monitoring devices generally suffer from defects such as low monitoring accuracy, poor real-time performance, poor adaptability, and lack of comprehensive monitoring. Summary of the Invention

[0013] This disclosure provides a gas leak monitoring system, method, computer equipment, and computer program product. Through multimodal data fusion and intelligent analysis, it achieves efficient, accurate location, and real-time monitoring of leaks in underground gas storage wells, enabling timely detection, assessment, and handling of leak risks. This improves the safety and reliability of hydrogen storage systems, reduces the need for human intervention, and lowers maintenance costs and operational risks. It addresses the shortcomings of existing technologies in real-time monitoring and accuracy in assessing leak risks in gas storage well leak monitoring.

[0014] In a first aspect, this disclosure provides a gas leak monitoring system for use in gas storage wellbores, the system comprising:

[0015] A pressure sensor is installed on the inner wall of the gas storage well and is connected in communication with the controller to detect the gas pressure inside the gas storage well.

[0016] A gas sensor is installed on the inner and outer walls of the gas storage well and is communicatively connected to the controller to detect the first gas concentration inside the gas storage well and the second gas concentration outside the gas storage well.

[0017] An ultrasonic sensor is installed on the inner wall of the gas storage well and is communicatively connected to the controller to detect the sound wave signal generated by gas leakage.

[0018] The controller is used to determine whether a gas leak has occurred in the gas storage wellbore based on the gas pressure, the first gas concentration, the second gas concentration, and the acoustic signal, using a preset risk level determination model.

[0019] In some embodiments, the controller is also configured to determine the location of a gas leak and / or issue a warning message in the event of a gas leak.

[0020] In some embodiments, the number of pressure sensors is multiple, and they are respectively disposed at the top and bottom of the inner wall of the gas storage well.

[0021] The number of gas sensors is multiple, and they are respectively set at the wellhead, the bottom of the gas storage well, and a preset position at a first preset well depth away from the wellhead; wherein, the preset position includes: the inner wall, the outer wall, and the outer wall of the covering material; wherein, the covering material is wrapped around the periphery of the gas storage well.

[0022] The ultrasonic sensors are multiple and are respectively installed on the top of the inner wall, the bottom of the inner wall, and on the inner wall at a second preset well depth from the wellhead.

[0023] In some embodiments, it also includes:

[0024] The display device is communicatively connected to the controller and is used to display one or more of the monitoring data, the early warning information, and the gas leak location; wherein the monitoring data includes the gas pressure, the first gas concentration, the second gas concentration, and the acoustic signal.

[0025] In some embodiments, it also includes:

[0026] The early warning subsystem is communicatively connected to the controller and is used to perform early warning operations and / or issue instructions to perform corresponding countermeasures based on the location of the gas leak in response to receiving the early warning information.

[0027] Secondly, this disclosure provides a gas leak detection method, implemented based on the gas leak detection system described above, the method comprising:

[0028] The gas concentration, gas pressure, and acoustic signals at multiple locations within the gas storage wellbore are acquired; wherein the gas concentration includes a first gas concentration inside the gas storage wellbore and a second gas concentration outside the wellbore.

[0029] The gas pressure, the gas concentration, and the acoustic signal are preprocessed respectively, and the dataset composed of the preprocessed gas pressure, gas concentration, and acoustic signal is used as the data to be judged.

[0030] Based on the data to be determined, a preset risk level determination model is used to determine whether a gas leak has occurred.

[0031] In some embodiments, determining whether a gas leak has occurred based on the dataset to be determined using a preset risk level determination model includes:

[0032] The leakage risk level corresponding to the data to be judged is determined by the preset risk level judgment model.

[0033] Based on the leakage risk level, a preset judgment strategy is used to determine whether a gas leak has occurred.

[0034] In some embodiments, it also includes:

[0035] In the event of a gas leak, the location of the gas leak is determined by a cross-correlation algorithm based on the preprocessed acoustic signal.

[0036] In some embodiments, it also includes:

[0037] When the leakage risk level meets the preset warning conditions, a warning message is issued; wherein the warning message includes one or more of the following: preset information of the gas storage well, leakage risk level, and gas leakage location.

[0038] In some embodiments, the preset risk level determination model includes:

[0039]

[0040] in, To determine the result, x i For the data to be determined, φ(x) i To use a kernel function to determine the data x to be judged i The feature vector after mapping to a high-dimensional space, where b is the bias term and w is the weight vector.

[0041] Thirdly, this disclosure provides a computer program product, including a computer program, characterized in that, when executed by a processor, the computer program implements the steps of the method described in the second aspect above.

[0042] This disclosure provides a gas leak monitoring system, method, computer equipment, and computer program product. The beneficial effects include:

[0043] (1) Improved monitoring accuracy and sensitivity: By combining ultrasonic sensors, hydrogen sensors, and pressure sensors, this disclosure enables multimodal monitoring of the operating status of hydrogen storage wellbores. The ultrasonic sensor detects changes in sound waves within the wellbore, the hydrogen sensor monitors hydrogen concentration in real time, and the pressure sensor monitors pressure changes within the wellbore, thereby improving the accuracy and sensitivity of leak detection;

[0044] (2) Real-time monitoring and early warning: This disclosure introduces a support vector regression algorithm (SVM) to intelligently analyze the data collected by sensors. With the help of advanced artificial intelligence prediction models and real-time data processing technology, this disclosure can realize real-time risk assessment and rapid response to leakage events, thereby shortening the response time for leakage handling;

[0045] (3) Precise Leak Location: This disclosure enables precise location of leaks. An ultrasonic sensor determines the leak location by detecting changes in sound waves, while a hydrogen sensor further verifies the accuracy of the leak location by monitoring changes in hydrogen concentration at different locations. Pressure sensor-assisted monitoring provides additional clues and data support. This feature significantly improves the speed and accuracy of problem location, ensuring rapid response and repair.

[0046] (4) Enhance safety and reliability: By timely monitoring and risk assessment of hydrogen leaks, the prevention of safety accidents that may be caused by leaks is strengthened, thereby enhancing the safety and reliability of hydrogen storage wells.

[0047] (5) Reduced maintenance costs and operational risks: Automated and intelligent monitoring systems reduce reliance on manual monitoring, thus lowering maintenance costs. Real-time early warning functions can detect problems in advance, reducing potential economic losses and safety risks, thereby improving the overall operational efficiency of the system;

[0048] (6) Optimize management decisions: The visualization of real-time monitoring data and risk assessment can be used to formulate scientific safety management strategies, provide data support for management decisions, and improve the scientificity and accuracy of management decisions;

[0049] (7) Practical application guidance significance: This disclosure provides a complete method for monitoring leakage in hydrogen storage wells, which can be applied to different types of hydrogen / gas storage wells, filling the gap in this field. Attached Figure Description

[0050] The present disclosure will be described in more detail below based on embodiments and with reference to the accompanying drawings:

[0051] Figure 1 This is a schematic diagram of the structure of a gas leak monitoring system provided in an embodiment of the present disclosure;

[0052] Figure 2 This is a schematic diagram of another gas leak monitoring system provided in an embodiment of the present disclosure;

[0053] Figure 3 A flowchart illustrating a gas leak monitoring method provided in this embodiment of the disclosure;

[0054] Figure 4 A flowchart of another gas leak monitoring method provided in this disclosure embodiment.

[0055] In the accompanying drawings, the same parts are referred to by the same reference numerals, and the drawings are not drawn to scale. Detailed Implementation

[0056] To enable those skilled in the art to better understand the technical solutions of this disclosure, and to fully understand and implement the process of how this disclosure applies technical means to solve technical problems and achieve corresponding technical effects, the technical solutions in the embodiments of this disclosure will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this disclosure, not all embodiments. The embodiments of this disclosure and the various features within them can be combined with each other without conflict, and the resulting technical solutions are all within the protection scope of this disclosure. All other embodiments obtained by those skilled in the art based on the embodiments of this disclosure without creative effort should fall within the protection scope of this disclosure.

[0057] It should be noted that the terms "first," "second," etc., in the specification, claims, and accompanying drawings of this disclosure are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this disclosure described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0058] It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases the steps shown or described may be executed in a different order than that shown here.

[0059] This disclosure aims to present an artificial intelligence method for real-time monitoring and risk assessment of underground gas storage wellbore leaks based on diversified sensing and monitoring devices. These devices include an ultrasonic sensor array, a hydrogen sensor array, and a pressure sensor array. By combining these with an integrated controller (which can be a central controller) and a display device (i.e., a monitor), a highly centralized and comprehensive real-time monitoring network is formed. Furthermore, based on the processing and fusion of multimodal data, monitoring indicators are extracted, and artificial intelligence technologies, such as machine learning, are applied to achieve risk assessment and early warning of hydrogen leaks in underground gas storage wells.

[0060] This disclosure addresses the shortcomings of existing technologies in real-time monitoring and accuracy in assessing leakage risks when monitoring gas storage wellbore leaks. Through multimodal data fusion and intelligent analysis, it achieves efficient and accurate location and real-time monitoring of leaks in underground gas storage wells, enabling timely detection, evaluation, and handling of gas leak risks. This improves the safety and reliability of hydrogen storage systems, reducing the need for human intervention and lowering maintenance costs and operational risks.

[0061] Example 1

[0062] This embodiment provides a gas leak monitoring system applicable to gas storage wellbores. A covering is provided around the gas storage wellbore to cover and secure it. The outer edge of the covering is in contact with the formation interface.

[0063] Figure 1 This is a schematic diagram of a gas leak monitoring system provided in an embodiment of this disclosure. Figure 1 As shown, the system disclosed in this embodiment includes:

[0064] A pressure sensor is installed on the inner wall of the gas storage well and is connected in communication with the controller to detect the gas pressure inside the gas storage well.

[0065] A gas sensor is installed on the inner and outer walls of the gas storage well and is communicatively connected to the controller to detect the first gas concentration inside the gas storage well and the second gas concentration outside the gas storage well.

[0066] An ultrasonic sensor is installed on the inner wall of the gas storage well and is communicatively connected to the controller to detect the sound wave signal generated by gas leakage.

[0067] The controller is used to determine whether a gas leak has occurred in the gas storage wellbore based on the gas pressure, the first gas concentration, the second gas concentration, and the acoustic signal, using a preset risk level determination model.

[0068] In some embodiments, the number of pressure sensors is multiple, and they are respectively disposed at the top and bottom of the inner wall of the gas storage well.

[0069] The number of gas sensors is multiple, and they are respectively installed at the wellhead, the bottom of the gas storage well, and at a preset position at a first preset well depth from the wellhead; wherein, the preset position includes: the inner wall, the outer wall (i.e., the junction of the gas storage well and the covering material), and the outer wall of the covering material (i.e., the junction of the covering material and the formation interface); wherein, the covering material wraps around the periphery of the gas storage well;

[0070] The ultrasonic sensors are multiple and are respectively installed on the top of the inner wall, the bottom of the inner wall, and on the inner wall at a second preset well depth from the wellhead.

[0071] It should be noted that pressure sensors can also be installed in other places on the inner wall, depending on actual needs; ultrasonic sensors can also be installed in other places on the inner wall, such as at the upper and lower quarters of the well depth, depending on actual needs.

[0072] The first preset well depth can be set to half the well depth value; the second preset well depth can also be set to half the well depth value, or it can be set according to actual needs.

[0073] Among them, the gas storage well can be a hydrogen storage well, and the gas stored in the gas storage well can be hydrogen.

[0074] Understandably, the controller can acquire data collected by each sensor. The controller can acquire data from each sensor at certain time intervals, such as once every 1 second. The specific time interval can be set according to actual needs.

[0075] In this embodiment, the gas storage well is an underground hydrogen storage well, the covering material is a cement ring, the outer periphery of the cement ring is in contact with the formation interface, the gas sensor is a hydrogen sensor, and the example of monitoring whether hydrogen leakage occurs in the underground hydrogen storage well is used for illustration.

[0076] Please refer to the following. Figure 2 , Figure 2 This is a schematic diagram of another gas leak monitoring system provided in an embodiment of this disclosure.

[0077] Optionally, an ultrasonic sensor array is deployed at key locations such as the bottom and top of the hydrogen storage wellbore to detect minute changes in sound waves caused by leaks and locate the leak location; a hydrogen sensor array monitors the hydrogen concentration inside and outside the wellbore to determine the severity of the leak; and a pressure sensor array monitors pressure changes inside the wellbore to provide auxiliary information to improve the accuracy of leak detection.

[0078] The system deploys ultrasonic sensors (1, ultrasonic sensor ①, 2, ultrasonic sensor ②, 3, ultrasonic sensor ③, 4, ultrasonic sensor ④, 5, ultrasonic sensor ⑤), hydrogen sensors (8, hydrogen sensor (wellhead), 9, hydrogen sensor (inside the well), 10, hydrogen sensor (wellwall-cement sheath interface), 11, hydrogen sensor (cement sheath interface-formation interface), 12, hydrogen sensor (bottom of the well)) and pressure sensors (6, pressure sensor (wellhead), 7, pressure sensor (bottom of the well)) in high-risk areas (wellhead and wellbore-cement sheath-formation interface). This monitoring system can monitor the health status of the hydrogen storage wellbore in real time, promptly detect changes in hydrogen concentration, locate hydrogen leaks, and generate leak warnings, thereby ensuring the safe operation of the hydrogen storage system.

[0079] Alternatively, the sensors and devices can be arranged as follows:

[0080] (1) Ultrasonic sensor, the specific location of which is as follows:

[0081] An ultrasonic sensor ① is placed near the wellhead (upper part).

[0082] Location: Approximately 5 meters from the wellhead.

[0083] Objective: To monitor changes in acoustic waves near the wellhead and detect leak signals from the upper part of the well.

[0084] An ultrasonic sensor is installed in the middle section (upper middle part) of the wellbore.

[0085] Location: Approximately 1 / 4 of the well depth from the wellhead.

[0086] Objective: To enhance monitoring of the wellhead and central region to ensure the detection of leaks in the upper and middle parts.

[0087] An ultrasonic sensor is installed in the middle section of the wellbore.

[0088] Location: Middle of the wellbore, at half the well depth.

[0089] Objective: To cover the acoustic wave propagation characteristics in the central region of the wellbore and enhance the accuracy of detection.

[0090] An ultrasonic sensor is installed in the middle section (lower part) of the wellbore.

[0091] Location: Approximately 1 / 4 of the way down the well.

[0092] Objective: To enhance monitoring in the middle to bottom region and ensure the detection of leaks in the middle and lower parts.

[0093] An ultrasonic sensor is installed at the bottom (lower part) of the wellbore.

[0094] Location: Near the bottom of the well, about 5 meters from the bottom.

[0095] Objective: To monitor acoustic signals in the bottom area of ​​the well to ensure full coverage of the entire wellbore.

[0096] High-frequency ultrasonic sensors are used, which are suitable for detecting tiny cracks and leaks.

[0097] Model: Olympus V260-SU series, Honeywell UA50 series, or Siemens US series high-frequency ultrasonic sensor; Frequency range: 40kHz to 100kHz; Accuracy: ±1%.

[0098] Selection criteria: high sensitivity, high resolution, high pressure resistance, and corrosion resistance.

[0099] (2) Hydrogen sensor: A high-sensitivity, fast-response, and anti-interference hydrogen sensor is selected, with a response time of less than 1 second;

[0100] Available models include: MQ-8, SGX Sensortech MiCS-5524, or Honeywell Midas Gas Detector high-sensitivity hydrogen sensor; detection range: 0-10000ppm; accuracy: ±5%.

[0101] (3) The central controller model can be either Texas Instruments TMS320 series DSP or NVIDIA Jetson series embedded system.

[0102] (4) Display device (i.e., monitor), which may be an industrial-grade liquid crystal display screen, with a size of 10 to 15 inches.

[0103] (5) The specifications of the hydrogen storage wellbore can be:

[0104] Diameter: 339.7 mm and above; Length: 200 m to 500 m;

[0105] Materials: High-strength hydrogen corrosion resistant steel (e.g., 30CrMo / 4130X), cement ring thickness 15mm to 30mm.

[0106] The controller can be used to process data from various sensors; establish a leakage risk level based on the actual hydrogen leakage situation, combined with expert opinions and industry standards; establish an intelligent evaluation model based on the above sensor and leakage risk level data; use metaheuristic optimization algorithms to improve the model prediction accuracy and use diversified statistical indicators to evaluate the model prediction performance.

[0107] In some embodiments, it also includes:

[0108] In the event of a gas leak, the location of the gas leak is determined by a cross-correlation algorithm based on the preprocessed acoustic signal.

[0109] In some embodiments, it also includes:

[0110] When the leakage risk level meets the preset warning conditions, a warning message is issued; wherein the warning message includes one or more of the following: preset information of the gas storage well, leakage risk level, and gas leakage location.

[0111] In some embodiments, it also includes:

[0112] The display device is communicatively connected to the controller and is used to display one or more of the monitoring data, the early warning information, and the gas leak location; wherein the monitoring data includes the gas pressure, the first gas concentration, the second gas concentration, and the acoustic signal.

[0113] The gas concentration includes the gas concentration inside the gas storage well and the gas concentration outside the gas storage well.

[0114] The display device can intuitively show one or more of the monitoring data, early warning information, and gas leak locations.

[0115] In some embodiments, it also includes:

[0116] The early warning subsystem is communicatively connected to the controller and is used to perform early warning operations and / or take corresponding countermeasures based on the location of the gas leak in response to receiving the early warning information.

[0117] Optionally, the early warning operation includes issuing warnings via photoelectric and / or audible alerts, or providing feedback to relevant personnel via message push. Therefore, the early warning subsystem may include corresponding light warning modules or voice broadcast modules.

[0118] The gas leak monitoring system can also be equipped with a display device (e.g., a high-definition monitor) to display monitoring data, early warning information, and gas leak locations in real time, providing operators with an intuitive monitoring interface. Based on a developed predictive model (i.e., a preset risk level determination model), the system makes a real-time assessment of the hydrogen leak level in the hydrogen storage wellbore under the monitored data supply. An early warning subsystem then implements early warning and response measures. Upon receiving an early warning message, the early warning subsystem issues corresponding control commands based on the message, such as automatically closing valves or activating the ventilation system, thereby controlling and / or repairing the leak in the hydrogen storage wellbore.

[0119] Example 2

[0120] Based on the above embodiments, this embodiment provides a gas leak monitoring method, which is implemented based on the gas leak monitoring system described above.

[0121] The method provided in this embodiment aims to achieve efficient and accurate location and real-time monitoring of leaks in underground hydrogen storage wells, as well as reliable assessment of leak risks, through high-precision sensing technology and advanced data processing algorithms.

[0122] like Figure 3 As shown, the method disclosed in this embodiment includes:

[0123] Step 310: Obtain gas concentration, gas pressure, and acoustic signals at multiple locations within the gas storage wellbore; wherein the gas concentration includes a first gas concentration inside the gas storage wellbore and a second gas concentration outside the wellbore.

[0124] Acquire the gas pressure detected by all pressure sensors, the gas concentration detected by all gas sensors, and the sound wave signals detected by all ultrasonic sensors.

[0125] Specifically, the system acquires real-time acoustic data from the hydrogen storage wellhead, wellbore, and bottom via ultrasonic sensors; a hydrogen sensor array monitors hydrogen concentrations at different locations, including inside the wellbore, the wellbore-cement sheath bonding surface, and the cement sheath-formation bonding surface; and pressure sensors monitor pressure at the wellhead and bottom. All sensor data is transmitted to the controller via wired (cable) or wireless means.

[0126] Step 320: Preprocess the gas pressure, the gas concentration, and the acoustic signal respectively, and combine the data composed of the preprocessed gas pressure, gas concentration, and acoustic signal into data to be judged.

[0127] Optionally, the gas pressure, gas concentration, and acoustic signal data obtained at the same time can be stored in the same array and used as the data to be judged.

[0128] As an example, the data to be determined can be represented as [20, 0.9, 0.02, 0.01, 10, 12]; where the data of each element in this array, from first to last, represent: data detected by an ultrasonic sensor (frequency or maximum amplitude), data detected by another ultrasonic sensor (frequency or maximum amplitude), data detected by a hydrogen sensor (e.g., first hydrogen concentration), data detected by a hydrogen sensor (e.g., second hydrogen concentration), data monitored by a pressure sensor (e.g., hydrogen pressure), and data monitored by another pressure sensor. The number and / or position of each element value in the data to be determined can also be set according to actual needs.

[0129] The controller collects and integrates data from various sensors, and performs data processing and feature extraction.

[0130] Preprocessing includes Fourier transform and feature extraction to remove outliers from the initial data and extract relevant features. For example, data acquired by an ultrasonic sensor requires Fourier transform and feature extraction, typically involving the following steps: ① Noise denoising: Using filters or denoising algorithms to remove noise from the sensor data. ② Signal smoothing: Using techniques such as moving average or median filtering to smooth the sensor data. ③ Data correction: Correcting the data according to the sensor's calibration parameters to ensure accuracy and consistency. ④ Frequency feature analysis: Using Fourier transform (spectral analysis technique) to extract frequency distribution features such as the dominant frequency and bandwidth of the acoustic signal. ⑤ Amplitude feature analysis: Calculating amplitude parameters of the acoustic signal, such as maximum amplitude and root mean square amplitude, to reflect the energy distribution of the acoustic signal. ⑥ TDOA (Time Difference of Arrival) algorithm to determine the leak location: Using a cross-correlation algorithm to calculate the time delay between signals to determine the leak location. ⑦ Time-domain feature analysis: Extract statistical features of the acoustic signal from the time domain, such as mean, variance, kurtosis, skewness, etc., to reflect the temporal characteristics of the acoustic signal.

[0131] Alternatively, the above data processing can all be accomplished using Python code. For example, the SciPy library can be used to extract the frequency characteristics of ultrasonic sensor data. Specific code examples are shown below:

[0132] import numpy as np

[0133] from scipy.fft import fft

[0134] #Assume sensor_data is the raw data collected by the ultrasonic sensor, which is a time series.

[0135] sensor_data=[1.2,2.5,3.6,...]

[0136] # Perform a Fourier transform on the sensor data to obtain the spectrum.

[0137] spectrum = fft(sensor_data)

[0138] #Calculate frequency characteristics, such as the dominant frequency

[0139] freqs=np.fft.fftfreq(len(sensor_data))

[0140] main_freq=freqs[np.argmax(np.abs(spectrum))]

[0141] print("Main frequency:", main_freq)

[0142] Optionally, the Python code for the Time Difference of Arrival (TDOA) triangulation algorithm includes:

[0143]

[0144]

[0145] In addition, the following data processing was performed on each hydrogen sensor: analyzing changes in hydrogen concentration and recording the trend of hydrogen concentration changes over time; the following data processing was performed on each pressure sensor: monitoring pressure fluctuations, identifying abnormal changes, and extracting relevant features.

[0146] Step 330: Based on the data to be determined, use a preset risk level determination model to determine whether a gas leak has occurred.

[0147] In some embodiments, the preset risk level determination model is constructed based on the SVM algorithm.

[0148] Alternatively, the first consideration is to use the Support Vector Machine (SVM) algorithm, which has shown good performance and application potential when dealing with complex datasets and nonlinear relationships, to build a predictive model. SVM first maps the features in the input space to a high-dimensional space, and then finds the best-fit hyperplane in the feature space to achieve classification.

[0149] The basic structure of the SVM algorithm and its application in hydrogen storage wellbore leakage risk assessment can be seen as follows:

[0150] Step 1: Determine the input features (i.e., the data to be judged). Denoise the raw data from the ultrasonic sensor, hydrogen sensor, and pressure sensor, and extract statistical information such as peak values ​​and means to represent the corresponding features. Feature standardization and normalization help eliminate prediction errors caused by differences in data types. All features are grouped into the same dataset and denoted as X = {x1, x2, ..., x...}. n}, where x i It is the i-th data point.

[0151] Standardized formula:

[0152]

[0153] Where x: original data points; μ: mean, representing the average value of the data; σ: standard deviation, representing the dispersion of the data; mean μ: for a given feature x, the mean μ is the average value of all data points on that feature. Assume there are n samples, each with a value x1, x2, ..., xn on that feature. n The formula for calculating the mean is:

[0154]

[0155] The mean is calculated for the value of a certain feature across the entire dataset and is a value that reflects the central tendency of that feature.

[0156] Standard deviation σ: Indicates how close the distribution of data on this characteristic is to the mean. The formula for calculating the standard deviation is:

[0157]

[0158] The larger the standard deviation, the more dispersed the data is on that feature; the smaller the standard deviation, the more concentrated the data is on that feature.

[0159] The purpose of standardization is to transform data into a standard normal distribution with a mean of 0 and a standard deviation of 1. This allows data with different characteristics to be processed on the same scale, which helps improve the convergence speed and performance of the model.

[0160] Normalization formula:

[0161]

[0162] Where x: original data point; min(x): minimum value in the feature dataset; max(x): maximum value in the feature dataset.

[0163] For a given feature x, min(x) is the minimum value of that feature in the entire dataset, while max(x) is the maximum value of that feature in the entire dataset.

[0164] Calculation process: When calculating normalization, first iterate through all samples in the dataset to find the minimum and maximum values ​​of the feature. For example, if the values ​​of the feature are [2, 5, 7, 1, 9], then min(x) = 1 and max(x) = 9.

[0165] Normalization: By normalizing, the range of data is scaled to between [0,1] or [-1,1]. This is especially important for algorithms that use gradient descent, as it allows gradient descent to converge faster and avoids computational instability caused by excessively large values.

[0166] It is mainly used to scale data to a fixed range, ensuring that the data is compared within the same range:

[0167] To normalize multiple features, the min(x) and max(x) of each feature are typically calculated separately. The min(x) and max(x) of each feature are calculated independently, based only on the dataset for that feature.

[0168] In cases involving multiple datasets (such as training and test sets), to ensure consistency, min(x) and max(x) are typically computed on the training set and then applied to the test set.

[0169] Step 2: Output feature Y: Hydrogen storage wellbore leakage risk level, including high leakage risk, medium leakage risk, low leakage risk, and no leakage risk. The output set is denoted as Y = {y1, y2, ..., y}. n}, where y i It corresponds to the input feature x i The leakage risk level.

[0170] Step 3: Feature Selection and Optimization. Principal component analysis (PCA) or correlation analysis is used to select the features most relevant to the leakage risk level.

[0171] Step 4, Model Prediction. The prediction model proposed in this disclosure is constructed based on the SVM (Support Vector Machine) algorithm. Therefore, the Radial Basis Function (RBF) is selected as the kernel function according to the data characteristics. During the prediction process, the kernel function K(x) i ,x j The input feature X is mapped to a high-dimensional space, and the prediction result is obtained by linearly combining the weight vector w and the bias term b; specifically, the preset risk level determination model includes:

[0172]

[0173] in, For the data to be determined, x i The corresponding judgment result, x i For the data to be determined, φ(x) i To use a kernel function to determine the data x to be judged i The feature vector mapped to a high-dimensional space, where b is the bias term and w is the weight vector of the preset risk level determination model.

[0174] The preset risk level determination model uses input features x iThe data is mapped to a high-dimensional space, and its inner product with the model weights w is calculated. This product, along with the bias b, yields the prediction result. This process captures the non-linear relationships between input features and can be used for high-precision regression prediction.

[0175] Input feature x i The corresponding predicted value. Corresponding to the input sample x. i The output (i.e., the predicted leakage risk level). <w,φ(x i )> is the input feature x i The eigenvector φ(x) after mapping to a higher-dimensional space i The inner product (dot product) between the vectors and the weight vector w represents the "similarity" or "alignment" of the two vectors. A large inner product value means that the vectors are similar or related in a high-dimensional space. In this inner product, φ(x) = φ(w) + ... i ) is the input feature x i The result is mapped to a high-dimensional space through a kernel function, and w is the weight vector learned by the model.

[0176] φ(x i ) is the input feature x i The feature vector is mapped to a higher-dimensional space using a kernel function. This mapping process aims to capture non-linear relationships within the data. In the original space, the data may be difficult to separate or fit, but by mapping to a higher-dimensional space, a suitable hyperplane can be found for classification or regression. Kernel functions (such as radial basis functions, RBF) implicitly perform this mapping process when calculating the inner product, without explicitly calculating the mapped higher-dimensional vector.

[0177] For SVM, the commonly used kernel function is the radial basis function (RBF), where the RBF kernel function has the following form:

[0178] K(x i x j )=exp(-γ||x i -x j || 2 )

[0179] The kernel function is used to measure the two input features x. i and x j The similarity between them.

[0180] In support vector regression, the kernel function is used to compute the inner product between data points in a high-dimensional space, without explicitly computing the mapping of the data points in the high-dimensional space (i.e., φ(x)). i ) and φ(x jThe RBF kernel function controls the rate of similarity decay through the γ parameter. The output of the kernel function is a scalar representing the degree of similarity between two data points in a high-dimensional space. Using the kernel function, SVM can perform complex nonlinear regressions in the original low-dimensional space without explicit high-dimensional mapping.

[0181] In short, the kernel function K(x) i ,x j This is used to measure the similarity between different input samples and to help calculate the preset risk level determination model. High-dimensional inner product <w,φ(x i )>.

[0182] b serves as a bias term (or intercept term) used to adjust the baseline of the predicted values. Geometrically, b determines the position of the hyperplane in high-dimensional space, i.e., how it intersects the origin.

[0183] To achieve optimal prediction, it is necessary to minimize the regularized risk function:

[0184]

[0185] Where C is the regularization parameter, used to balance model complexity and training error.

[0186] In support vector regression, to achieve the best prediction results, we need to minimize the following regularization risk function:

[0187]

[0188] This represents a regularization term used to control the complexity of the model.

[0189] ||w|| represents the norm of the weight vector w, specifically the second-order norm of the weight vector (i.e., the sum of squares of the weights). By minimizing this term, SVM tends to choose smaller weights, thereby avoiding overcomplexity and preventing overfitting.

[0190] in, It is a constant factor, which facilitates mathematical processing during the optimization process.

[0191] C is the regularization parameter, used to balance the relationship between model complexity and training error. If the value of C is large, the model will pay more attention to the training error, that is, try to fit each training sample as well as possible, but this may lead to overfitting. If the value of C is small, the model will tend to keep it simple, allowing for larger prediction errors on some training samples, thus preventing overfitting.

[0192] Methods for determining the C value: Cross-validation: A common method is to select the value through cross-validation. In cross-validation, the dataset is divided into multiple subsets, and the model needs to be trained and validated on different subsets. The C value that performs best on the validation set is then selected.

[0193] Grid Search: Finds the C value that optimizes the model's performance on the validation set by trying a series of candidate values.

[0194] Rule of thumb: Sometimes, depending on the size and characteristics of the dataset, an initial C value can be selected using a rule of thumb or prior knowledge, and then fine-tuned based on the actual results.

[0195] in, It is the sum of the loss functions, representing the model's prediction error across all training data. L ∈ (y i f(x) i The loss function is called the loss function, and the loss function that is insensitive to the loss function is usually used. It can be defined as:

[0196] L ∈ (y i f(x) i ))=max(0,|y i -f(x i )|-∈)

[0197] When the predicted value f(x) i ) and the true value y i When the error between the two values ​​is less than ∈, the loss is 0, meaning the error is considered negligible; when the error is greater than ∈, the loss is the error minus the value of ∈.

[0198] By minimizing this loss function, the model minimizes the difference between the predicted and the true values, but does not over-penalize when the error is small.

[0199] f(x i ) is the prediction function of the model, i.e. For each input feature x i The model calculates the predicted value based on the parameters w and b obtained during training.

[0200] ∈ is a small positive number representing the range of error the model can tolerate. It determines the range of errors that can be ignored. By setting ∈, the model can tolerate small errors, thereby improving generalization ability and avoiding overfitting to the training data.

[0201] Step 5: Model Optimization. A metaheuristic optimization algorithm is used to select the optimal hyperparameters for the SVM model.

[0202] Step 6: Model Output. Using the data X collected by the sensor as input to the model, the leakage result is predicted by the preset risk level determination model.

[0203] Then, the leakage risk level is determined based on the leakage results corresponding to all inputs; finally, if the preset early warning conditions are met, the corresponding leakage early warning information is issued based on the predicted leakage risk level.

[0204] In some embodiments, determining whether a gas leak has occurred based on the dataset to be determined using a preset risk level determination model includes:

[0205] The leakage risk level corresponding to the data to be judged is determined by the preset risk level judgment model; wherein, the leakage risk level includes: high leakage risk, medium leakage risk, low leakage risk and no leakage risk;

[0206] Based on the leakage risk level, a preset judgment strategy is used to determine whether a gas leak has occurred.

[0207] Specifically, firstly, based on a preset risk level determination model, the determination result (i.e., the predicted value) of the data to be determined is obtained. Then, the leakage risk level is determined based on the predicted value. For example, the predicted value is compared with each preset threshold. If the predicted value is not less than the first preset threshold, it is determined to be a strong leakage risk; if the predicted value is between the second preset threshold and the first preset threshold, it is determined to be a medium leakage risk; if the predicted value is between the third preset threshold and the second preset threshold, it is determined to be a low leakage risk; and if the predicted value is less than the third preset threshold, it is determined to be no leakage risk. The first, second, and third preset thresholds decrease sequentially and can all be set according to actual needs.

[0208] Optionally, the preset judgment strategy can be set as follows: if the leakage risk level is high leakage risk or medium leakage risk, then it is judged that a gas leak has occurred.

[0209] Step 7: Performance Evaluation. The model's predictive ability is evaluated using metrics such as confusion matrix, accuracy, precision, recall, and receiver operating characteristic (ROC) curves, which are then used for further model tuning and optimization.

[0210] Furthermore, the optimal SVM model is used to intelligently analyze actual multi-sensor data, identify abnormal fluctuations, and determine the presence of leakage risks. Based on the real-time leakage risk assessment results, leakage early warning information is generated and fed back to the central controller. Upon detecting leakage signs and obtaining the corresponding leakage risk level, the system automatically adjusts operating parameters within the wellbore or suspends operation, promptly notifying operators and taking appropriate measures to ensure the safe operation of the hydrogen storage system.

[0211] In some embodiments, the step of determining the location of the gas leak includes:

[0212] The location of the gas leak is determined by a cross-correlation algorithm based on all preprocessed acoustic signals.

[0213] Optionally, based on all preprocessed acoustic signals, the time delay between all acoustic signals is calculated using a cross-correlation algorithm to determine the location of the gas leak.

[0214] Understandably, with multiple ultrasonic sensors, sound wave signals from multiple locations at the same time can be collected. Then, based on all the pre-processed sound wave signals, the time delay between all the sound wave signals can be calculated using a cross-correlation algorithm to determine the location of the gas leak.

[0215] In some embodiments, it also includes:

[0216] Based on the gas concentrations detected by the gas sensor at the wellhead, bottom, and a preset well depth, the gas leak location is verified to determine its accuracy; and / or,

[0217] The location of the gas leak is verified by using the wellhead pressure and bottom hole pressure detected by the pressure sensor to determine whether the location of the gas leak is accurate.

[0218] The location of a leak is determined by detecting changes in sound waves using ultrasonic sensors. The accuracy of the leak location is further verified by monitoring changes in hydrogen concentration at different locations using hydrogen sensors. Pressure sensors also provide additional clues and data support. This functionality significantly improves the speed and accuracy of leak location, ensuring rapid response and repair.

[0219] In some embodiments, it also includes:

[0220] In the event of a gas leak, the location of the gas leak is determined by a cross-correlation algorithm based on the preprocessed acoustic signal.

[0221] In some embodiments, it also includes:

[0222] When the leakage risk level meets the preset warning conditions, a warning message is issued; wherein the warning message includes one or more of the following: preset information of the gas storage well, leakage risk level, and gas leakage location.

[0223] The preset early warning conditions can be set to a leakage risk level of strong leakage risk or medium leakage risk; the preset information of the gas storage well can include: the number of the underground gas storage well, the location information of the underground gas storage well, and a brief information about the underground gas storage well, etc.

[0224] In addition, for easier understanding of the technical solution of this application, please refer to Figure 4 , Figure 4 A flowchart of another gas leak monitoring method provided in this disclosure embodiment.

[0225] Example 3

[0226] Based on the above embodiments, this disclosure provides an application example.

[0227] In this embodiment of the disclosure, specific steps for determining the risk level are given based on the specific data monitored by each sensor:

[0228] (1) Data standardization and normalization:

[0229] Assume the data obtained from the ultrasonic, hydrogen, and pressure sensors are as follows:

[0230] #Raw Data

[0231] X = np.array([[20,0.9,0.02,0.01,10,12],#No leakage)

[0232] [35,2.1,2.0,1.5,8,14]])# There is a leak.

[0233] Furthermore, the raw data is standardized and normalized:

[0234] Standardization ensures that each feature has a mean of 0 and a variance of 1, which helps improve model performance. Standardization can be performed using the following formula:

[0235]

[0236] In the standardized calculation formula, x is the original data, μ is the mean, and σ is the standard deviation.

[0237] The purpose of normalization is to ensure that the data falls within the range [0,1]. Normalization can be performed using the following formula:

[0238]

[0239] In the normalization formula, min(x) and max(x) are the minimum and maximum values ​​of the feature, respectively.

[0240] Optionally, the Python statements for standardizing and normalizing the data are as follows:

[0241] from sklearn.preprocessing import StandardScaler,MinMaxScaler

[0242] #standardization

[0243] scaler = StandardScaler()

[0244] X_standardized=scaler.fit_transform(X)

[0245] #Normalization

[0246] min_max_scaler=MinMaxScaler()

[0247] X_normalized=min_max_scaler.fit_transform(X_standardized)

[0248] The standardized data are:

[0249] X_standardized = [[-1.0,-1.0,-1.0,-1.0,1.0,-1.0], # No leakage

[0250] [1.0,1.0,1.0,1.0,-1.0,1.0]# Leakage detected

[0251] The normalized data are:

[0252] X_normalized = [[0.0,0.0,0.0,0.0,1.0,0.0], #No leakage

[0253] [1.0,1.0,1.0,1.0,0.0,1.0]] #Leakage detected

[0254] (2) Definition of output target value:

[0255] The output value (Y) corresponds to the leakage risk level, defined as 0 for no leakage and 1 for leakage.

[0256] (3) SVM model construction and training:

[0257] Kernel function selection:

[0258] For SVM, the commonly used kernel function is the radial basis function (RBF), where the RBF kernel function has the following form:

[0259] K(x i x j )=exp(-γ||x i -x j || 2 ) where γ is an adjustable parameter used to control the smoothness of the curve in the high-dimensional feature space.

[0260] Model training:

[0261] Train an SVM model using the RBF kernel function. Assuming a small dataset, you can use the entire dataset for training in Python.

[0262] from sklearn.svm import SVM

[0263] #Set up the SVM model and use the RBF kernel function

[0264] svm_model=SVM(kernel='rbf', C=1.0, epsilon=0.1, gamma='scale')

[0265] #Training the model

[0266] svm_model.fit(X_normalized,Y)

[0267] Where C is the regularization parameter, which balances model complexity and training error; epsilon(∈) is a parameter of the loss function of the SVM model, which controls the maximum difference between the predicted value and the actual value; gamma(γ) is a parameter in the RBF kernel function, which determines the measure of the influence on model complexity.

[0268] (4) Model prediction:

[0269] After the model training is completed, the trained model can be used to predict the leakage risk level of the data to be predicted.

[0270] #Data to be predicted (standardized and normalized):

[0271] X_new=np.array([[30,1.5,1.0,0.5,9,13]])

[0272] X_new_standardized=scaler.transform(X_new)

[0273] X_new_normalized=min_max_scaler.transform(X_new_standardized)

[0274] #Predicting data to be predicted:

[0275] Y_pred=svm_model.predict(X_new_normalized)

[0276] Assume the predicted result is Y_pred = [0.75].

[0277] (5) Judgment Results and Decisions:

[0278] Interpretation of results: The predicted value Y_pred = 0.75 indicates that the current leakage risk level is medium to high. The closer the predicted value is to 1, the higher the leakage risk level.

[0279] Decision: Based on the risk level, determine whether to issue an early warning. If the risk level is high, issue an early warning and remind relevant personnel to implement appropriate emergency response measures.

[0280] (6) Furthermore, the model can be evaluated and optimized:

[0281] Confusion matrix: Used to evaluate the model's accuracy, recall, precision, and other metrics, and to further adjust the model parameters;

[0282] Hyperparameter optimization: Grid search or random search can be used to optimize the parameters C, ∈, and γ in the SVM model.

[0283] from sklearn.model_selection import GridSearchCV

[0284] # Define parameter mesh

[0285] param_grid={'C':[0.1,1,10],'epsilon':[0.01,0.1,1],'gamma':['scale',0.1,1]}

[0286] #Grid Search

[0287] grid_search=GridSearchCV(SVM(kernel='rbf'),param_grid,cv=5)

[0288] grid_search.fit(X_normalized,Y)

[0289] # Output optimal parameters

[0290] best_params=grid_search.best_params_

[0291] The evaluation results can be used to guide further adjustments and optimizations of the model.

[0292] Example 4

[0293] Based on the above embodiments, this embodiment provides a computer device, including a memory, a processor, and a computer program stored in the memory, wherein the processor executes the computer program to implement the steps of the method described in the above embodiments.

[0294] In some embodiments of this example, a computer-readable storage medium is also provided, on which a computer program is stored, which, when executed by a processor, implements the steps of the method described in the above embodiments.

[0295] In some embodiments of this example, a computer program product is also provided, including a computer program that, when executed by a processor, implements the steps of the method described in the above embodiments.

[0296] The processor may include, but is not limited to, one or more processors or microprocessors. Each processor may be implemented as an Application Specific Integrated Circuit (ASIC), Digital Signal Processor (DSP), Digital Signal Processing Device (DSPD), Programmable Logic Device (PLD), Field Programmable Gate Array (FPGA), controller, microcontroller, microprocessor, or other electronic component, for executing the methods in the above embodiments.

[0297] Computer-readable storage media can be implemented by any type of volatile or non-volatile storage device or a combination thereof. Computer-readable storage media may include, but are not limited to, random access memory (RAM), read-only memory (ROM), flash memory, EPROM memory, EEPROM memory, registers, and computer storage media (e.g., hard disks, floppy disks, solid-state drives, removable disks, CD-ROMs, DVD-ROMs, Blu-ray discs, etc.).

[0298] Computer-readable storage media may also store at least one computer-executable program, such as computer-readable instructions. Computer-readable storage media include, but are not limited to, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and / or cache memory. Computer-readable storage media may include, for example, read-only memory (ROM), hard disk, flash memory, etc. For example, a non-transitory computer-readable storage medium may be connected to a computing device such as a computer, and then, when the computing device executes the computer-readable instructions stored on the computer-readable storage medium, the various methods described above can be performed.

[0299] In addition, the computer device may include, but is not limited to, a data bus, an input / output (I / O) bus, a monitor, and input / output devices (e.g., a keyboard, a mouse, speakers, etc.).

[0300] The processor can communicate with external devices via the I / O bus through wired or wireless networks.

[0301] In one embodiment, the at least one computer-executable instruction may also be compiled into or comprise a software product / computer program product, wherein one or more computer-executable instructions are executed by a processor to perform the steps of the various functions and / or methods in the embodiments described herein.

[0302] In the embodiments provided in this disclosure, it should be understood that the disclosed apparatus and methods can also be implemented in other ways. The apparatus embodiments described above are merely illustrative; for example, the flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of apparatus, methods, and computer program products according to various embodiments of this disclosure. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in a different order than those marked in the drawings. For example, two consecutive blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram and / or flowchart, and combinations of blocks in block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or action, or using a combination of dedicated hardware and computer instructions.

[0303] It should be noted that, in this disclosure, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element limited by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0304] While the embodiments disclosed herein are as described above, the foregoing content is merely for the purpose of facilitating understanding of this disclosure and is not intended to limit this disclosure. Any person skilled in the art to which this disclosure pertains may make any modifications and changes in form and detail of the implementation without departing from the spirit and scope of this disclosure; however, the scope of patent protection of this disclosure shall still be determined by the scope defined in the appended claims.

Claims

1. A gas leak monitoring system, characterized in that, The system is applied to gas storage wellbores and includes: A pressure sensor is installed on the inner wall of the gas storage well and is connected in communication with the controller to detect the gas pressure inside the gas storage well. A gas sensor is installed on the inner and outer walls of the gas storage well and is communicatively connected to the controller to detect the first gas concentration inside the gas storage well and the second gas concentration outside the gas storage well. An ultrasonic sensor is installed on the inner wall of the gas storage well and is communicatively connected to the controller to detect the sound wave signal generated by gas leakage. The controller is used to determine whether a gas leak has occurred in the gas storage wellbore based on the gas pressure, the first gas concentration, the second gas concentration, and the acoustic signal, using a preset risk level determination model.

2. The system according to claim 1, characterized in that, The pressure sensors are multiple and are respectively installed at the top and bottom of the inner wall of the gas storage well. The number of gas sensors is multiple, and they are respectively set at the wellhead, the bottom of the gas storage well, and a preset position at a first preset well depth away from the wellhead; wherein, the preset position includes: the inner wall, the outer wall, and the outer wall of the covering material; wherein, the covering material is wrapped around the periphery of the gas storage well. The ultrasonic sensors are multiple and are respectively installed on the top of the inner wall, the bottom of the inner wall, and on the inner wall at a second preset well depth from the wellhead.

3. The system according to claim 1, characterized in that, The controller is also used to determine the location of a gas leak and / or issue an early warning message in the event of a gas leak.

4. The system according to claim 3, characterized in that, Also includes: The early warning subsystem is communicatively connected to the controller and is used to perform early warning operations and / or issue instructions to perform corresponding countermeasures based on the location of the gas leak in response to receiving the early warning information.

5. The system according to claim 4, characterized in that, Also includes: The display device is communicatively connected to the controller and is used to display one or more of the monitoring data, the early warning information, and the gas leak location; wherein the monitoring data includes the gas pressure, the first gas concentration, the second gas concentration, and the acoustic signal.

6. A method for monitoring gas leaks, characterized in that, Based on the gas leak monitoring system according to any one of claims 1 to 5, the method includes: The gas concentration, gas pressure, and acoustic signals at multiple locations within the gas storage wellbore are acquired; wherein the gas concentration includes a first gas concentration inside the gas storage wellbore and a second gas concentration outside the wellbore. The gas pressure, the gas concentration, and the acoustic signal are preprocessed respectively, and the dataset composed of the preprocessed gas pressure, gas concentration, and acoustic signal is used as the data to be judged. Based on the data to be determined, a preset risk level determination model is used to determine whether a gas leak has occurred.

7. The method according to claim 6, characterized in that, The step of determining whether a gas leak has occurred based on the data to be determined using a preset risk level determination model includes: The leakage risk level corresponding to the data to be judged is determined by the preset risk level judgment model. Based on the leakage risk level, a preset judgment strategy is used to determine whether a gas leak has occurred.

8. The method according to claim 6, characterized in that, Also includes: In the event of a gas leak, the location of the gas leak is determined by a cross-correlation algorithm based on the preprocessed acoustic signal.

9. The method according to claim 8, characterized in that, Also includes: When the leakage risk level meets the preset warning conditions, a warning message is issued; wherein the warning message includes one or more of the following: preset information of the gas storage well, leakage risk level, and gas leakage location.

10. The method according to claim 6, characterized in that, The preset risk level determination model includes: in, To determine the result, x i For the data to be determined, φ(x) i To use a kernel function to determine the data x to be judged i The feature vector after mapping to a high-dimensional space, where b is the bias term and w is the weight vector.

11. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program performs the steps of the method described in any one of claims 5 to 10.

Images