A hydrogen micro-leakage risk grading response system under time series analysis

Abstract

The application discloses a hydrogen micro-leakage risk grading response system under time series analysis and relates to the related technical field of hydrogen leakage detection, which comprises an image data acquisition module, a leakage joint detection module and a risk control response analysis module.The image data acquisition module is used for collecting adhesive tape image data by planning and navigating the directional path of the mobile monitoring equipment within the monitoring range of the hydrogen production plant.The leakage joint detection module is used for triggering the leakage detection component, performing joint detection based on position offset and color variation, and generating a first leakage detection result.The risk control response analysis module is used for building a leakage alarm threshold based on the leakage mode and the hydrogen production condition, performing risk control alarm management under hydrogen micro-leakage through risk control response analysis under the time sequence coordinate system.The technical problems of limited monitoring range, delayed response, poor anti-interference ability, and difficulty in accurately identifying and early warning micro-leakage in the prior art are solved, and the technical effects of omnidirectional real-time monitoring, early accurate identification, dynamic risk assessment and graded efficient response are achieved.

Description

Technical Field

[0001] This invention relates to the field of hydrogen leak detection technology, specifically to a hydrogen micro-leakage risk classification and response system based on time series analysis. Background Technology

[0002] As a crucial link in the hydrogen energy supply chain, the safe operation of hydrogen production plants is paramount. However, hydrogen possesses unique physical and chemical properties, including low density, flammability, explosiveness, and high diffusivity. Even under micro-leakage conditions, it can accumulate in confined or semi-confined spaces, forming explosive mixtures and triggering serious safety accidents. Traditional gas leak detection methods are limited by monitoring range, response delays, and high costs, making it difficult to achieve real-time and accurate monitoring of critical components of hydrogen production plants. While image analysis-based leak detection is increasingly being used, typically employing visual sensors to capture image data of target areas and analyzing color, texture, or motion changes in the images to identify leak signs, the complex environment of hydrogen production plants, including factors such as lighting variations and equipment vibrations, makes image processing prone to false alarms or missed detections. Furthermore, micro-leakage of hydrogen often manifests as slow, minute physical changes, such as shifts in the position or color changes of tape at pipe connections, requiring high-precision time-series analysis for early warning.

[0003] Therefore, current technologies suffer from limitations such as limited monitoring range, delayed response, poor anti-interference capabilities, and difficulty in accurately identifying and warning of micro-leaks in their early stages. Summary of the Invention

[0004] This application provides a hydrogen micro-leakage risk classification and response system based on time series analysis, which solves the technical problems of limited monitoring range, response delay, poor anti-interference ability, and difficulty in accurate identification and early warning of micro-leakage in the existing technology, and achieves the technical effects of comprehensive real-time monitoring, accurate early identification, dynamic risk assessment, and efficient classification and response.

[0005] This application provides a hydrogen micro-leakage risk classification and response system based on time series analysis. The system includes: an image data acquisition module for defining a monitoring range within a hydrogen production plant and acquiring tape image data of a first key location through directional path planning and navigation using mobile monitoring equipment; a leak joint verification module for transmitting the tape image data back to the risk control platform, triggering a leak detection component, performing a joint verification based on position offset and color variation, and generating a first leak detection result, wherein a first control group is used for position offset verification based on sparse point clouds, and a second control group is generated for color variation verification based on dual differences; and a risk control response analysis module for establishing leak alarm thresholds based on leak patterns and hydrogen production conditions, performing risk control response analysis under time-series coordinate system updates on the first leak detection result, and performing risk control alarm management under hydrogen micro-leakage.

[0006] In a possible implementation, the acquisition of tape image data at the first critical location includes: defining a critical risk control zone based on the hydrogen production process, wherein the critical risk control zone includes at least an electrolysis zone, a separation zone, a purification zone, and a storage zone; deploying hydrogen leak detection tape at the critical risk control location for the critical risk control zone; uploading a first monitoring task to drive the mobile monitoring device to perform image data monitoring and acquisition, and determining the tape image data, wherein the first monitoring task is to perform leak monitoring at the first critical risk control location, and the first critical risk control location is any at least one critical risk control location.

[0007] In a possible implementation, before driving the mobile monitoring device to perform image data monitoring and acquisition, the method further includes: constructing a trajectory network based on feasible paths using the physical environment layout constraints of the key risk control area; using the deployment location of the hydrogen leak detection tape as multiple directional navigation targets, performing trajectory pre-planning in the trajectory network to generate a pre-made trajectory network; and storing the pre-made trajectory network in the data control center of the mobile monitoring device.

[0008] In a possible implementation, the joint verification based on position offset and color variation is performed. The verification based on position offset includes: acquiring standardized images as the hydrogen leak detection tape is deployed; determining the sparse point cloud distribution using the geometric deployment boundary, calibrating the standardized images as a first control group, wherein the sparse point cloud distribution includes key inflection point clouds and linear uniform point clouds; and performing position offset verification on the tape image data according to the first control group to determine the first verification data.

[0009] In a possible implementation, the position offset detection of the tape image data further includes: performing spatial phase mapping on the sparse point cloud distribution of the first control group in the tape image data to determine point cloud mapping pairs; traversing the point cloud mapping pairs to perform consistency comparison to generate the first detection data.

[0010] In a possible implementation, the joint verification based on position offset and color variation is performed. The verification based on color variation further includes: determining a first local image data and a second local image data in the tape image data, wherein the first local image data belongs to the non-misaligned part of the first verification data and the V value belongs to the standard range; calling the upper-level detection image data, combining the first local image data and the second local image data, and generating second verification data by performing a double difference analysis of the V value change.

[0011] In a possible implementation, generating the second verification data by performing a dual difference analysis of V-value changes further includes: matching the first upper-level local image data mapped from the first local image data and the second upper-level local image data mapped from the second local image data in the upper-level detection image data; measuring the V-value change between the first local image data and the first upper-level local image data to determine the first V-value data, wherein the first V-value data characterizes the environmental change effect; measuring the V-value change between the second local image data and the second upper-level local image data to determine the second V-value data, wherein the first V-value data characterizes the superposition state of the environmental change effect and the leakage effect; and calculating the difference between the second V-value data and the first V-value data to generate the second verification data, wherein the second verification data characterizes the net effect of leakage.

[0012] In a possible implementation, the hydrogen micro-leakage risk classification response system under time series analysis also performs the following processing: locating micro-leakage points by cross-validating the first and second verification data; performing leakage degree analysis based on the second verification data for each micro-leakage point to determine the leakage level; and integrating the micro-leakage points and the leakage level to generate a first leakage detection result.

[0013] In a possible implementation, the hydrogen micro-leakage risk classification and response system based on time series analysis further performs the following processing: classifying leakage modes, wherein the leakage modes include at least slow seepage, rapid injection, and intermittent leakage; classifying hydrogen production conditions, wherein the hydrogen production conditions include at least maintenance period and high-load operation period; and combining the leakage modes and the hydrogen production conditions to set classification response rules based on leakage risk and establish leakage alarm thresholds.

[0014] In a possible implementation, the hydrogen micro-leakage risk classification and response system under time series analysis further performs the following processing: constructing a time series coordinate system and updating the first leak detection result to the time series coordinate system; according to the leak alarm threshold, performing leak pattern matching based on the evaluation sequence within the time series coordinate system and generating risk control response information under hydrogen production operating conditions; and performing hydrogen micro-leakage risk response management according to the risk control response information.

[0015] This application proposes a hydrogen micro-leakage risk classification and response system based on time series analysis. The system comprises an image data acquisition module for acquiring tape image data within a defined monitoring area of ​​a hydrogen production plant using directional path planning and navigation via mobile monitoring equipment; a leak joint verification module for triggering leak detection components to perform joint verification based on position offset and color variation, generating a first leak detection result; and a risk control response analysis module for establishing leak alarm thresholds based on leak patterns and hydrogen production conditions, and performing risk control alarm management for hydrogen micro-leakage through risk control response analysis updated in a time-series coordinate system. This system solves the technical problems of limited monitoring range, response delay, poor anti-interference capability, and difficulty in accurately identifying and warning of early micro-leakage in existing technologies, achieving comprehensive real-time monitoring, accurate early identification, dynamic risk assessment, and efficient graded response. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of this disclosure, the accompanying drawings of the embodiments of this disclosure will be briefly described below. Flowcharts are used in this application to illustrate the operations performed by the system according to the embodiments of this application. It should be understood that the preceding or following operations are not necessarily performed precisely in sequence. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from these processes.

[0017] Figure 1 This is a schematic diagram of a hydrogen micro-leakage risk classification and response system based on time series analysis, provided as an embodiment of this application.

[0018] Figure 2 This is a schematic diagram of a location offset-based verification process in a hydrogen micro-leakage risk classification and response system based on time series analysis, provided as an embodiment of this application.

[0019] Figure labeling: Image data acquisition module 10, Leakage joint verification module 20, Risk control response analysis module 30. Detailed Implementation

[0020] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0021] This application provides a hydrogen micro-leakage risk classification and response system based on time series analysis, such as... Figure 1 As shown, the system includes: The image data acquisition module 10 is used to delineate the monitoring range within the hydrogen production plant and to acquire tape image data of the first key part through directional path planning and navigation of the mobile monitoring equipment.

[0022] Preferably, defining the monitoring scope refers to defining the area requiring monitoring within the hydrogen production plant. This includes deploying hydrogen leak detection tape on critical risk control points such as valves, flanges, and welds in electrolyzers, separators, purification units, and storage tanks. The hydrogen leak detection tape may curl, lift, or shift under the impact of extremely small hydrogen gas flows, and its surface is coated with a chemical substance that changes color upon contact with hydrogen. Mobile monitoring equipment may include inspection robots, drones, or automated guided vehicles equipped with cameras. Based on the physical environment constraints of the hydrogen production plant, such as passageways and obstacles, the marked critical risk control points serve as directional navigation targets. The mobile monitoring equipment performs directional path planning and navigation, pre-planning the optimal inspection trajectory to ensure that all critical risk control points are visited efficiently and without omission. The mobile monitoring equipment, equipped with this inspection trajectory, automatically travels along the path using SLAM, GPS, etc., and accurately stops at each critical risk control point, thereby collecting tape image data from the first critical point. The first critical point refers to any critical risk control point, and the tape image data consists of digital photographs or video frames containing the leak detection tape.

[0023] Furthermore, the specific configuration of the image data acquisition module 10 also includes defining a key risk control zone based on the hydrogen production process, wherein the key risk control zone includes at least an electrolysis zone, a separation zone, a purification zone, and a storage zone; deploying hydrogen leak detection tape at the key risk control location for the key risk control zone; uploading a first monitoring task to drive the mobile monitoring device to perform image data monitoring and acquisition, and determining the tape image data, wherein the first monitoring task is to perform leak monitoring at a first key risk control location, and the first key risk control location is any at least one key risk control location.

[0024] Preferably, based on the hydrogen production process and equipment layout of the hydrogen production plant, high-risk areas with the highest probability of hydrogen leakage or the most serious consequences are designated as key risk control zones. These zones should include at least an electrolysis zone, a separation zone, a purification zone, and a storage zone. The electrolysis zone is the core area for hydrogen production via water electrolysis, containing the electrolyzer and its associated high-voltage electrical equipment. The separation zone is where equipment performs preliminary separation of the electrolysis products hydrogen and oxygen, and is at risk of gas leakage. The purification zone is where devices for drying and removing impurities from crude hydrogen are located, including various processing equipment and pipeline connections. The storage zone is where compressed hydrogen storage tanks or liquid hydrogen storage tanks are located, used for centralized hydrogen storage. Then, hydrogen leak detection tape is installed at key risk control points in all key risk control zones. The tape's color and shape change upon contact with hydrogen, converting invisible hydrogen leaks into visible image signals. Key risk control points may include pipe flange connections, valve and pump seals, reactor or tank welds, and interfaces such as safety valves and pressure gauges. The first monitoring task is sent to the mobile monitoring device to drive it to collect image data. The first monitoring task is to perform leakage monitoring of the first key risk control point. The first key risk control point is any at least one key risk control point. Specifically, the mobile monitoring device navigates to the first key risk control point according to the pre-stored path plan, and uses its onboard image sensor to collect images of the hydrogen leakage detection tape deployed at the point to obtain tape image data.

[0025] Furthermore, the specific configuration of the image data acquisition module 10 also includes: constructing a trajectory network based on feasible paths, constrained by the physical environment layout of the key risk control area; using the deployment location of the hydrogen leak detection tape as multiple directional navigation targets, performing trajectory pre-planning in the trajectory network to generate a pre-made trajectory network; and storing the pre-made trajectory network in the data control center of the mobile monitoring device.

[0026] Preferably, the physical environment layout constraints of the key risk control area include all physical elements affecting the passage of mobile monitoring equipment, such as walls, equipment, pipes, the width of passages, the location of obstacles, and prohibited areas. Then, based on these physical environment layout constraints, a trajectory network based on feasible paths is constructed. Feasible paths refer to all paths that mobile monitoring equipment can safely and freely pass through under the constraints, such as workshop passages that robots can drive through or safe corridors that drones can fly through. These feasible paths are modeled into a trajectory network, where nodes represent intersections or points where devices can stop, and edges represent feasible paths connecting nodes. Then, all deployed hydrogen... Key risk control points of the gas leak detection tape are used as multiple directional navigation targets, such as the flange of tank No. 3 and the valve of electrolytic cell No. 5. These are set as target points that the mobile monitoring equipment must reach and acquire images from. Based on the locations of all directional navigation targets, trajectory pre-planning is performed in the trajectory network. That is, the optimal continuous driving path with the shortest total distance or the least total time is calculated as a pre-made trajectory network, which contains a complete inspection route plan with multiple ordered navigation instructions. Finally, the pre-made trajectory network is stored in the data control center of the mobile monitoring equipment. The data control center is the local control unit of the mobile monitoring equipment, thereby ensuring the completion of autonomous navigation and image monitoring and acquisition tasks.

[0027] The joint leakage detection module 20 is used to transmit the tape image data back to the risk control platform, trigger the leakage detection component, perform joint detection based on position offset and color variation, and generate a first leakage detection result. The first control group is used to perform position offset detection based on sparse point cloud, and a second control group is generated to perform color variation detection based on double difference.

[0028] Preferably, the collected tape image data is transmitted back to the risk control platform, which is a central processing unit. This activates the leak detection component, which then performs a joint verification based on positional offset and color variation in parallel. Specifically, a positional offset verification based on sparse point clouds is performed using a first control group. The first control group consists of standard tape images collected in a leak-free state. Multiple representative feature points of the tape are extracted from the first control group images, such as the corners of the tape edges and the centers of specific markings. All feature points form a sparse point cloud in three-dimensional space, which is used to describe the geometry and position of the leak detection tape in a normal state. The collected tape image data is then registered with the first control group to determine the same set of feature points. By analyzing the spatial positional differences between the point cloud and the sparse point cloud in the new image, it is determined whether the leak detection tape has undergone physical displacement, warping, or deformation, which may be caused by the impact of airflow from a micro-leak.

[0029] Preferably, color variation detection based on dual difference is performed by generating a second control group. The second control group refers to the reference image area selected to eliminate common interference, including normal areas in the current image and image data from a historical point in time. Specifically, the second control group undergoes a first time difference, i.e., a suspected leakage area is selected, and the color value change between the current image and the historical image is calculated, including leakage effect and environmental effect. At the same time, a normal area without positional shift is selected, and its color change between the current image and the historical image is also calculated, including only environmental effect. Then, a second inter-group difference is performed, i.e., the difference between the two color change values ​​is calculated to offset the influence of environmental change, thereby determining the color variation caused by leakage. Finally, the positional shift and color change values, along with their related positional information, are output as the first leakage detection result.

[0030] Furthermore, such as Figure 2 As shown, the specific configuration of the leak joint verification module 20 also includes: acquiring standardized images as the hydrogen leak detection tape is deployed; determining the sparse point cloud distribution based on the geometric deployment boundary; calibrating the standardized images as a first control group, wherein the sparse point cloud distribution includes key inflection point clouds and linear uniform point clouds; and performing position offset verification on the tape image data based on the first control group to determine the first verification data.

[0031] Preferably, after the hydrogen leak detection tape is initially applied at the hydrogen production plant, under ideal conditions such as no leaks and fixed shooting distance, angle, and lighting conditions, images of the tape are acquired using a mobile monitoring device to determine standardized images, ensuring their reliability and consistency as a comparison benchmark. The geometric deployment boundary refers to the precise geometric contour presented by the tape image, such as the outer edge line or the boundary of a specific internal pattern. Multiple key feature points representing the shape are selected on the geometric deployment boundary to form a sparse point cloud distribution, including key inflection point clouds and linear uniform point clouds. The key inflection point cloud refers to the set of feature points located at the corners, sharp angles, or points where the shape of the tape changes drastically. The linear uniform point cloud refers to the set of feature points at fixed intervals along the long straight edge or curve of the tape. The system first obtains a set of uniformly sampled feature points; then it calibrates the standardized image by precisely binding the sparse point cloud to the location of the leak detection tape in the standardized image and recording the precise coordinates of each feature point, thus obtaining a first control group for position comparison; next, it performs position offset detection on the tape image data. Specifically, it identifies and determines the feature points in the tape image data that correspond to the sparse point cloud in the first control group, calculates the spatial transformation relationship between the two sets of feature point clouds through image registration, analyzes and determines the residual positional deviation between the two sets of point clouds, so as to reveal the microscopic physical deformation of the leak detection tape, such as warping, stretching, and local peeling, and finally outputs the first detection data, which may include, but is not limited to, the overall offset vector, the maximum offset, and the average offset error.

[0032] Furthermore, the specific configuration of the joint leakage detection module 20 also includes: determining point cloud mapping pairs by performing spatial phase mapping on the sparse point cloud distribution of the first control group in the tape image data; traversing the point cloud mapping pairs for consistency comparison to generate the first detection data.

[0033] Preferably, the sparse point cloud distribution of the first control group is spatially phase-mapped in the tape image data. That is, the optimal spatial transformation model is determined through image registration, and the correspondence between the two sets of point clouds is established, such as rigid body transformation, affine transformation, etc., so that when the point cloud of the first control group is mapped to the tape image data, it can achieve the best alignment with the actual feature position of the tape in the current image. Then, the theoretical matching position point on the current tape image is determined for each feature point in the first control group, and multiple sets of point cloud mapping pairs are generated. The residual or deviation vector between the source point and the target point is calculated by traversing the point cloud mapping pairs to quantify the positional movement of the feature point. Finally, the first verification data is generated and output, such as the average value of the deviation vector of all point pairs, the root mean square error, the largest positional offset among all point pairs, the number of feature points whose positional offset exceeds the preset tolerance threshold, and the distribution of the deviation vector is analyzed to determine whether it is an overall translation, rotation, or local warping.

[0034] Furthermore, the specific configuration of the leak joint verification module 20 also includes determining a first local image data and a second local image data in the tape image data, wherein the first local image data belongs to the non-misaligned part of the first verification data and the V value belongs to the standard condition range; calling the upper-level detection image data, combining the first local image data and the second local image data, and generating the second verification data by performing a double difference analysis of the V value change.

[0035] Preferably, a first local image data and a second local image data are determined from the tape image data based on the first verification data. The first local image data refers to the tape area determined to be without positional misalignment, and the V value of this area is within the standard range. The V value is the color value, which refers to the brightness of a color or the intensity of a specific color channel, and is mainly affected by ambient light and not by hydrogen leakage. The second local image data refers to the area suspected of leaking, such as the area delineated based on positional deviation or preliminary color anomaly, which is simultaneously affected by both ambient light and hydrogen leakage.

[0036] Preferably, the upper-level detection image data refers to historical inspection tape images collected during the previous inspection. Combining the first local area image data and the second local area image data, a dual difference analysis of V-value changes is performed. Specifically, firstly, a first time difference is performed, calculating the difference between the current mean V-value of the second local area image and the mean V-value of its corresponding region in the upper-level detection image, including environmental changes and leakage effects; secondly, the difference between the current mean V-value of the first local area image and the mean V-value of its corresponding region in the upper-level detection image is calculated, including only environmental changes; then, a second inter-group difference is performed, that is, the difference between the two time difference results is calculated to cancel out environmental interference, and finally determine the amount of color change caused by leakage, as the second verification data output.

[0037] Furthermore, the specific configuration of the leakage joint detection module 20 also includes: matching the first upper-level local image data mapped from the first local image data and the second upper-level local image data mapped from the second local image data in the upper-level detection image data; measuring the V-value change between the first local image data and the first upper-level local image data to determine the first V-value data, wherein the first V-value data characterizes the environmental change effect; measuring the V-value change between the second local image data and the second upper-level local image data to determine the second V-value data, wherein the first V-value data characterizes the superposition state of the environmental change effect and the leakage effect; and calculating the difference between the second V-value data and the first V-value data to generate second detection data, wherein the second detection data characterizes the net effect of the leakage.

[0038] Preferably, based on image coordinates or feature matching, a region in historical upper-level detection image data that is exactly the same as the location of the first local image data is matched as the first upper-level local image data. Similarly, a region in historical images that is exactly the same as the location of the current second local image is located as the second upper-level local image data. The change in V value between the first local image data and the first upper-level local image data is calculated to determine the first V value data, which is caused by environmental factors such as light intensity, angle change, and camera parameter drift. That is, the first V value data represents the environmental change effect. The change in V value between the second local image data and the second upper-level local image data is calculated to determine the second V value data, which is caused by the environmental change effect and the color variation effect caused by the leakage itself. That is, the second V value data represents the superposition state of the environmental change effect and the leakage effect. Finally, the difference between the second V value data and the first V value data is calculated to cancel out the environmental change effect and generate the second verification data. The second verification data represents the net effect of the leakage, and finally outputs a quantitative value that accurately represents the degree of color change caused by the leakage itself.

[0039] Furthermore, the specific configuration of the joint leakage detection module 20 also includes locating micro-leakage points by cross-validating the first and second detection data; performing leakage degree analysis based on the second detection data for each micro-leakage point to determine the leakage level; and integrating the micro-leakage points and the leakage level to generate a first leakage detection result.

[0040] Preferably, the first and second verification data are collaboratively verified, i.e., spatial correlation analysis is performed to verify spatial consistency and identify overlapping or adjacent areas that exhibit both significant spatial displacement and significant net color variation, thereby accurately locating micro-leak points. Then, for each micro-leak point, a leakage degree analysis based on the second verification data is performed. Specifically, the intensity and area of ​​color change are positively correlated with the concentration or exposure of hydrogen, which can more intuitively quantify the severity of the leak than physical location displacement. The values ​​of the second verification data are then read and compared with a preset threshold range to output the leakage level, such as minor leak, moderate leak, and severe leak. Finally, the location information of each leak point and the corresponding leakage level are integrated into a structured first leak detection result.

[0041] The risk control response analysis module 30 is used to establish a leakage alarm threshold based on the leakage mode and hydrogen production conditions. By performing risk control response analysis on the first leakage detection result under the time-series coordinate system update, it performs risk control alarm management under hydrogen micro-leakage.

[0042] Furthermore, the specific configuration of the risk control response analysis module 30 also includes classifying leakage modes, wherein the leakage modes include at least slow seepage, rapid injection, and intermittent leakage; classifying hydrogen production conditions, wherein the hydrogen production conditions include at least maintenance period and high-load operation period; and combining the leakage modes and the hydrogen production conditions to set graded response rules based on leakage risk and establish leakage alarm thresholds.

[0043] Preferably, leakage modes are defined based on historical data and physical models, including at least slow seepage, rapid injection, and intermittent leakage, with different risks for each mode. Hydrogen production conditions refer to the current operating status of the plant, such as maintenance periods or high-load operation periods; the actual risk for the same leakage level differs under different conditions. Then, leakage alarm thresholds are established based on the leakage modes and hydrogen production conditions, i.e., dynamic threshold rules are created. For example, during high-load operation and rapid injection modes, even with lower leakage levels, high-level alarms are triggered; during maintenance periods and slow seepage modes, for the same lower leakage level, only low-level warnings or alerts are triggered. The risk control response analysis is performed on the first leak detection results under a time-series coordinate system update. Each first leak detection result is treated as a data point and updated to the time-series coordinate system along with its timestamp. This allows for observation of the dynamic development trend of the leak. Specifically, this includes determining whether the level of a leak point is increasing over time, matching the leak development trend with the leak pattern, and then, combined with the current hydrogen production conditions and the identified leak pattern, invoking dynamic alarm threshold rules to conduct a final assessment of the current time-series risk status. This leads to the execution of risk control alarm management for micro-leaks in hydrogen. Based on the risk response analysis results, corresponding response measures are initiated. For example, for low-risk warnings, the information is recorded in the control room log, and inspection personnel are notified to pay attention during the next routine inspection. For medium-risk warnings, an audible and visual alarm is issued, prompting operators to immediately conduct remote confirmation. For high-risk emergency alarms, the highest-level alarm is triggered, emergency ventilation is activated, related equipment is isolated, and emergency response personnel are immediately notified.

[0044] Furthermore, the specific configuration of the risk control response analysis module 30 also includes: constructing a time-series coordinate system and updating the first leakage detection result to the time-series coordinate system; performing leakage pattern matching based on the evaluation sequence within the time-series coordinate system and generating risk control response information under hydrogen production conditions based on the leakage alarm threshold; and performing hydrogen micro-leakage risk response management based on the risk control response information.

[0045] Preferably, a time-series coordinate system is constructed with time as the horizontal axis and leak detection results as the vertical axis data points. The leak status of each detection cycle is recorded, and the first leak detection result generated each time is used as a new data point, which is added to the time-series coordinate system along with a timestamp, thereby obtaining the historical status record of each leak point. Then, the changing trend of the data of a certain leak point in the time-series is analyzed and matched with a preset leak mode. For example, by analyzing the change curve of the leak level over time, it is determined whether it conforms to a slow seepage mode with a linear and slow increase, a rapid injection mode with a sudden increase in level, or an intermittent leak mode. At the same time, the current real-time hydrogen production conditions are obtained, and the identified leak mode and the current hydrogen production conditions are used as input and matched with the leak alarm threshold rules. Based on the matching results, specific risk control response information is generated. Finally, hydrogen micro-leakage risk response management is executed based on the risk control response information, such as automatically triggering audible and visual alarms, popping up alarms on the control room screen, starting the emergency ventilation system, or isolating related equipment, and simultaneously pushing work orders to the mobile terminals of inspection personnel. Ultimately, comprehensive real-time monitoring, early dynamic and accurate identification, and hierarchical and efficient response to hydrogen micro-leakage are achieved.

[0046] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A hydrogen micro-leakage risk classification and response system based on time series analysis, characterized in that, The system includes: The image data acquisition module is used to delineate the monitoring area within the hydrogen production plant and to collect image data of the tape at the first critical location by using the directional path planning and navigation of the mobile monitoring equipment. The joint leakage detection module is used to transmit the tape image data back to the risk control platform, trigger the leakage detection component, perform joint detection based on position offset and color variation, and generate a first leakage detection result. The first control group is used to perform position offset detection based on sparse point cloud, and a second control group is generated to perform color variation detection based on double difference. The risk control response analysis module is used to establish leakage alarm thresholds based on leakage modes and hydrogen production conditions. By performing risk control response analysis on the first leakage detection results under time-series coordinate system updates, it performs risk control alarm management under hydrogen micro-leakage.

2. The hydrogen micro-leakage risk classification and response system based on time series analysis as described in claim 1, characterized in that, The steps performed by the image data acquisition module include: Based on the hydrogen production process, a key risk control area is defined, wherein the key risk control area includes at least an electrolysis area, a separation area, a purification area, and a storage area; For the aforementioned key risk control areas, hydrogen leak detection tape is deployed at key risk control locations; Upload the first monitoring task to drive the mobile monitoring device to monitor and collect image data, and determine the tape image data. The first monitoring task is to perform leakage monitoring of the first key risk control part, and the first key risk control part is any at least one key risk control part.

3. The hydrogen micro-leakage risk classification and response system based on time series analysis as described in claim 2, characterized in that, The steps performed by the image data acquisition module also include: Based on the physical environment layout constraints of key risk control areas, a trajectory network based on feasible paths is constructed; The deployment location of the hydrogen leak detection tape is used as multiple directional navigation targets, and trajectory pre-planning is performed in the trajectory network to generate a pre-made trajectory network; The pre-made trajectory network is stored in the data control center of the mobile monitoring device.

4. The hydrogen micro-leakage risk classification and response system based on time series analysis as described in claim 1, characterized in that, The steps performed by the joint leakage detection module include: Standardized images were collected as hydrogen leak detection tape was deployed. The sparse point cloud distribution is determined by the geometric deployment boundary, and the standardized image is calibrated as a first control group. The sparse point cloud distribution includes key inflection point cloud and linear uniform point cloud. Based on the first control group, the positional offset of the tape image data is checked to determine the first check data.

5. The hydrogen micro-leakage risk classification and response system based on time series analysis as described in claim 4, characterized in that, The steps performed by the joint leakage detection module include: By performing spatial phase mapping on the sparse point cloud distribution of the first control group in the tape image data, point cloud mapping pairs are determined. The point cloud mapping pairs are traversed and a consistency comparison is performed to generate the first verification data.

6. The hydrogen micro-leakage risk classification and response system based on time series analysis as described in claim 5, characterized in that, The steps performed by the joint leakage detection module include: In the tape image data, a first local image data and a second local image data are determined, wherein the first local image data belongs to the part of the first verification data where the position is not misaligned, and the V value belongs to the standard range; The upper-level detection image data is called, and combined with the first local image data and the second local image data, the second verification data is generated by performing a double difference analysis of the V value change.

7. The hydrogen micro-leakage risk classification and response system based on time series analysis as described in claim 6, characterized in that, The steps performed by the joint leakage detection module include: In the upper-level detected image data, match the first upper-level local image data mapped by the first local image data, and the second upper-level local image data mapped by the second local image data; The V-value change is measured between the first local image data and the first upper-level local image data to determine the first V-value data, wherein the first V-value data characterizes the effect of environmental change. The V-value change is measured on the second local image data and the second upper-level local image data to determine the second V-value data, wherein the first V-value data represents the superposition state of environmental change effect and leakage effect; The difference between the second V value data and the first V value data is calculated to generate the second verification data, wherein the second verification data characterizes the net effect of leakage.

8. The hydrogen micro-leakage risk classification and response system based on time series analysis as described in claim 7, characterized in that, By cross-checking the first and second verification data, the micro-leakage point can be located; For each micro-leak point, a leakage degree analysis based on the second verification data is performed to determine the leakage level; By integrating the micro-leakage points with the leakage level, a first leakage detection result is generated.

9. The hydrogen micro-leakage risk classification and response system based on time series analysis as described in claim 1, characterized in that, Leakage modes are classified, wherein the leakage modes include at least slow seepage, rapid jetting, and intermittent leakage; Hydrogen production operating conditions are divided, wherein the hydrogen production operating conditions include at least the maintenance period and the high-load operation period; By combining the aforementioned leakage mode with the aforementioned hydrogen production conditions, a graded response rule based on leakage risk is established, and a leakage alarm threshold is constructed.

10. The hydrogen micro-leakage risk classification and response system based on time series analysis as described in claim 9, characterized in that, Construct a time-series coordinate system and update the first leakage detection result to the time-series coordinate system; Based on the leakage alarm threshold, perform leakage mode matching and generate risk control response information under hydrogen production operating conditions constraints based on the evaluation sequence within the time coordinate system; Based on the risk control response information, perform hydrogen micro-leakage risk response management.

Images