Active prevention and control system and method for hydrogen leakage of hydrogen refueling station based on multiple feedback

Abstract

The invention relates to the technical field of hydrogen refueling station safety prevention and control, in particular to a hydrogen refueling station hydrogen leakage active prevention and control system and method based on multiple feedback. A hydrogen concentration detection unit monitors the hydrogen concentration in real time and generates a leakage alarm signal; the directional heat energy applying unit applies directional heat energy to a specific space area above the leakage source; the temperature sensing unit is configured to perform temperature measurement on a target space area and a potential leakage point; a hydrogen refueling station three-dimensional space structure model and a CFD simulation module are arranged in the central control unit; the environment parameter acquisition unit is used as a feed-forward compensation parameter; the emergency cut-off unit is connected with the central control unit and connected to a main gas inlet pipeline in series, early recognition and accurate positioning of hydrogen leakage are achieved through multiple information fusion and feedback of concentration, temperature and environmental parameters, a controllable thermal power channel is constructed through active thermal buoyancy guidance, the diffusion path after hydrogen leakage is changed, and the safety of hydrogen leakage is improved. And active guidance to a safe area is realized, so that randomness and low efficiency of passive ventilation dilution are avoided.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of hydrogen station safety prevention and control, and in particular to a hydrogen station hydrogen leakage active prevention and control system and method based on multiple feedbacks. BACKGROUND

[0002] With the rapid development of hydrogen energy industry, the safety and operation reliability of hydrogen stations, as the key infrastructure for hydrogen fuel cell vehicle energy supply, are increasingly concerned. Hydrogen is flammable, explosive, colorless and odorless. Its molecular weight is small, its viscosity is low, and its diffusion coefficient is large. Therefore, once it leaks, it is difficult to be detected in time and is prone to cause safety accidents. Therefore, building an efficient and intelligent hydrogen leakage prevention and control system has become a key direction for the development of hydrogen station safety technology.

[0003] In the prior art, in addition to passive monitoring relying on concentration sensors, there are also leakage detection schemes based on monitoring pipeline flow difference (for example, CN120332649A and CN120332692A). Although such schemes can achieve accurate positioning of the leakage point, they still belong to detection and isolation technology after leakage and cannot actively intervene and guide the diffusion behavior of hydrogen when leakage occurs, still having the problem of response lag.

[0004] Therefore, the prior art lacks an effective means that can actively intervene at the early stage of hydrogen leakage, rather than passive alarm, and can achieve early identification, accurate positioning and active intervention of hydrogen leakage, thereby improving the overall safety level and emergency response capability of the hydrogen station. SUMMARY

[0005] The technical problem to be solved by the present application is to provide a hydrogen station hydrogen leakage active prevention and control system and method based on multiple feedbacks, which can achieve early identification, accurate positioning and intelligent guidance of hydrogen, change passive alarm to active intervention, and significantly improve the safety protection level of the hydrogen station, in order to solve the problems of the prior art in the background.

[0006] The technical solution adopted by the present application to solve the technical problem is: a hydrogen station hydrogen leakage active prevention and control system based on multiple feedbacks, comprising a hydrogen concentration detection unit for real-time monitoring of hydrogen concentration in air in a preset monitoring area in the hydrogen station, and generating a leakage alarm signal when the hydrogen concentration exceeds a preset alarm threshold; a central control unit 4 reconstructs a three-dimensional concentration field by receiving the array data and using an algorithm based on spatial interpolation or gradient descent, accurately locates the initial position of the leakage source, and realizes the leap from area monitoring to point source positioning; The directional heat energy application unit is configured to apply controllable directional heat energy to a specific spatial area above the leakage source according to instructions. The core function is to quickly form a stable local high-temperature area above the leakage point through precise energy projection, and to use the thermal expansion of air to generate a density difference, thereby building a thermal updraft channel that guides the upward diffusion of hydrogen. This channel is essentially a man-made thermal chimney effect. The temperature sensing unit uses a non-contact infrared thermal imager and is configured to measure the temperature of the target spatial area acted on by the directional heat energy application unit and the potential leakage point, and output real-time temperature field data. This unit is the eye of the closed-loop feedback, providing a direct basis for evaluating the effect of the thermal updraft channel. The central control unit, as the intelligent core of the system, is in communication connection with the hydrogen concentration detection unit, the directional heat energy application unit, and the temperature sensing unit, to coordinate the linkage of the units. It has a built-in hydrogen filling station three-dimensional spatial structure model and a CFD simulation module. The central control unit 4 is responsible for coordinating the linkage of the units: first, locating the leakage source based on the concentration data; then, calling the CFD simulation module, combining real-time environmental parameters, predicting and simulating the hydrogen diffusion and thermal updraft effect, and optimizing the generation of the optimal guide path and initial heating strategy; finally, based on the feedback data of the temperature sensing unit, running the built-in PID control algorithm to dynamically adjust the power and direction of the heating unit, achieving precise closed-loop control of the stability of the thermal updraft channel. The environmental parameter acquisition unit is used to obtain real-time environmental wind speed, wind direction, and environmental temperature data, which are used by the central control unit as feedforward compensation parameters to dynamically correct the output of the control algorithm, effectively offsetting the disturbance of external wind fields and low-temperature environments on the thermal updraft channel, and significantly improving the adaptability and robustness of the system under different environmental conditions. The emergency shutdown unit is connected with the central control unit and installed in series on the main air inlet pipeline of the hydrogen filling station. When the system comprehensively determines that the leakage rate is too fast, the concentration continues to rise, or the thermal channel control fails, the unit serves as the final safety barrier to execute a fast shutdown instruction and block the gas source.

[0007] Further, the hydrogen concentration detection unit includes multiple hydrogen sensors distributed in a matrix around the potential leakage points in the hydrogen filling station. The central control unit accurately locates the initial position of the leakage source by analyzing the concentration data gradient of the multiple hydrogen sensors.

[0008] Further, the directional heat energy application unit is a focus-adjustable infrared heating array or a phased array microwave transmitter, and the energy focus and radiation direction can be accurately controlled by the central control unit.

[0009] Further, the central control unit is installed with The data acquisition module reads real-time hydrogen concentration, three-dimensional temperature field, and environmental parameter data. a leakage positioning module for accurately positioning an initial position of a leakage source by analyzing concentration data gradients of a plurality of hydrogen sensors; a path planning module for automatically planning an optimal safety path from the leakage point to the high-altitude safety zone and actively avoiding key dangerous sources in a hydrogen refueling station three-dimensional space structure model according to the position of the leakage source; a heating control module for accurately and stably constructing and maintaining a thermal buoyancy channel for guiding the upward rising of the leaked gas and actively resisting wind field interference; an emergency response module for monitoring the concentration rising rate and absolute concentration value in real time, automatically executing emergency shutdown and triggering all-around alarm when determining that the leakage level exceeds the safety threshold, so as to achieve rapid containment and early warning of the accident. An active prevention and control method for hydrogen leakage in a hydrogen refueling station based on multiple feedbacks, comprising an active prevention and control system for hydrogen leakage in a hydrogen refueling station based on multiple feedbacks according to any one of the above, and the specific steps are as follows: Step 1: In daily operation, the central control unit continuously collects and analyzes concentration and temperature data, and dynamically establishes and updates the baseline temperature model of each monitoring point in the hydrogen refueling station through machine learning algorithm, which lays a solid foundation for subsequent accurate identification of abnormalities; Step 2: The hydrogen concentration detection unit is used to monitor the hydrogen concentration in the air of the preset monitoring area in the hydrogen refueling station in real time, and generate concentration data; Step 3: Determine whether the concentration data exceeds the preset alarm threshold. If yes, trigger the leakage alarm and go to Step 4; if no, return to Step 2; Step 4: After receiving the leakage alarm signal, start the leakage source positioning module, calculate the spatial concentration field gradient based on the concentration data of a plurality of hydrogen sensors reported by the hydrogen concentration detection unit, accurately position the three-dimensional coordinates of the leakage source, and determine the target spatial area of thermal energy application, which realizes the key breakthrough from "finding leakage" to "knowing where the leakage is"; Step 5: The path planning module calls the pre-stored hydrogen refueling station three-dimensional space structure model, and plans a thermal buoyancy channel guide path from the leakage point to the high-altitude safety zone according to the three-dimensional coordinates of the leakage source, which avoids the key dangerous sources and equipment marked in the pre-stored hydrogen refueling station three-dimensional space structure model; Step 6: The heating control module activates and controls the directional thermal energy application unit to perform directional heating on the specific spatial area, so that the energy focus of the directional heating is scanned or statically directed along the starting segment of the guide path, and a local high-temperature area is formed above the leakage source, thereby constructing a thermal buoyancy channel for guiding the upward diffusion of the leaked hydrogen gas by using the thermal buoyancy effect; Step 7: Based on the real-time temperature data fed back by the temperature sensing unit, the temperature field distribution of the thermal buoyancy channel is constructed and compared with the preset ideal channel temperature model. The output power and direction of the directional thermal energy application unit are dynamically adjusted based on the PID control algorithm to maintain the stability and effectiveness of the thermal buoyancy channel. Step 8: If the leakage level is determined to exceed the safety threshold based on the comprehensive monitoring of the concentration rise rate, absolute concentration value, and temperature field stability, the emergency shutdown unit is triggered to generate a shutdown instruction to control the emergency shutdown unit to perform shutdown operations, shut off the hydrogen supply, and enter the system safety locking period. If not, it is determined whether the leak has stopped and the concentration has dropped to a safe range. If yes, return to step 1; if not, return to step 7.

[0010] Further, after the leakage alarm is triggered in step 3, the central control unit calculates the residual error between the real-time temperature data and the reference temperature model, and based on the spatial distribution of the monitoring points with residual error exceeding the preset threshold, it assists in verifying the leakage location and evaluating the leakage severity by performing the following quantitative calculation and logical judgment steps: Step 31: The system will calculate the residual error field between the real-time temperature field data and the reference temperature model under the corresponding spatio-temporal conditions ; Step 32: First, identify the maximum residual error region in the residual error field and calculate the three-dimensional centroid coordinates of this region ; Then, calculate the Euclidean distance D between the centroid and the initial positioning coordinates of the hydrogen concentration detection unit ; Finally, if D is less than the preset threshold, it is considered that the temperature anomaly and the concentration anomaly source are highly consistent, the verification is passed, and the location is reliable. If it exceeds the threshold, a verification failure flag is triggered, and the system will prompt that there may be sensor false alarms, multiple leakage sources, or positioning errors, which need to be urgently reviewed manually. Step 33: Calculate the total projection area S of the three-dimensional region formed by all pixels / volumes with residual error values exceeding the dynamic threshold in the residual error field. In the above abnormal region, calculate the average residual error of all points ; based on the two dimensions of S and , the system preliminarily divides the leakage severity into three levels: low, medium, and high.

[0011] Further, the algorithm steps for precisely positioning the three-dimensional coordinates of the leakage source in step 4 are as follows: Step 41: Use the position of the sensor with the highest concentration as the initial iteration point ; Step 42: In Around a given point, a local three-dimensional concentration field is constructed based on the readings of each sensor, and the concentration gradient vector at that point is calculated. ; Step 43: Move along the gradient direction with a preset step size Update location: ; Step 44: Repeat steps 42 and 43 until the magnitude of the gradient vector is less than the convergence threshold. ,Right now If the algorithm converges to the concentration maximum point, then the position at this point is considered to be... This refers to the three-dimensional coordinates of the identified leak source.

[0012] Furthermore, the central control unit is equipped with a feedforward compensator based on a physical model. In step 6, when the heating control module dynamically adjusts the directional heat energy application unit, it also introduces the real-time environmental parameters collected by the environmental parameter acquisition unit as feedforward compensation parameters to correct the output of the PID control algorithm, so as to offset the disturbance of the environmental wind field and low temperature environment on the stability of the heat floating channel. The central control unit receives the real-time environmental parameters obtained by the environmental parameter acquisition unit, and calculates the feedforward compensation amount based on the above real-time environmental parameters through the feedforward compensator based on the physical model. The specific compensation steps are as follows: Step 61: Compensate for the environmental wind field, including power compensation. and angle pre-compensation amount Among them, power compensation amount The calculation formula is: ; In the formula, The wind farm power compensation coefficient was determined experimentally. air density (kg / m³); The specific heat capacity of air at constant pressure (J / (kg·K)); The effective cross-sectional area (m²) of the thermal buoyancy channel. V Real-time wind speed (m / s); The target temperature of the channel (K); The ambient temperature (K); Angle pre-compensation amount The calculation formula is: ; In the formula, The angle compensation coefficient was determined experimentally. V Real-time wind speed (m / s); Real-time wind direction (in radians); The initial direction angle of the channel (in radians); Typical floating velocity (m / s) of the hot floating channel, determined by the temperature difference of the channel; Step 62: Ambient temperature compensation for power reference compensation , the calculation formula is: ; In the formula, is the temperature power compensation coefficient; is the reference ambient temperature; is the actual ambient temperature (K); the reference heating power (W) at the reference ambient temperature; Step 63: The above feedforward compensation amount is superimposed with the output of the PID feedback control algorithm to generate the final control command, which is: Final power command: ; Final angle command: ; Send the above final control command to the directional thermal energy application unit to real-time correct its output power and action angle; Step 64: Limit the final control command in step 63: ; ; Among them, and are the allowed working ranges of power and angle, respectively.

[0013] Further, the ideal channel temperature model is essentially a temperature distribution matrix in a three-dimensional space, which defines the temperature characteristics that an effective "hot floating channel" should have above the leakage point, and is also the reference target of closed-loop dynamic adjustment. The dynamic generation and optimization of the ideal channel temperature model in step 7 are as follows: Step 7A: In the system design stage, the CFD simulation module integrated in the central control unit will simulate the temperature distribution that the most effective hot floating channel should have under different leakage positions, different leakage rates, different environmental wind fields and temperature conditions according to the specific three-dimensional structure of the hydrogen filling station. The results of the above simulation are converted into a series of ideal channel temperature models and stored in the model library; Step 7B: When a real leakage occurs, the system will call the CFD simulation module to combine the current real-time leakage source position and environmental parameters to perform rapid prediction simulation and calculate the ideal temperature field distribution corresponding to the optimal hot floating channel under the current specific conditions. This real-time calculated distribution is the ideal channel temperature model under the current scenario; Step 7C: In the long-term operation of the system, the central control unit records the actual temperature field data in the successful boot cases and the corresponding control parameters, and continuously fine-tunes and optimizes the parameters of the ideal channel temperature model through machine learning algorithms.

[0014] Further, in step 7, the dynamic adjustment steps are as follows: Step 7a: Take N key points on the axis in the ideal channel model, calculate the root mean square error of the temperature deviation of the corresponding points in the real-time temperature field, denoted as the axial temperature deviation , and its dynamic allowable range is ; In the formula: is the basic tolerance coefficient, is the wind speed influence coefficient, is the ideal axial average temperature, is the real-time wind speed; Step 7b: Take M sampling points on the typical cross-sectional boundary of the channel, calculate the standard deviation of the temperature of these points , and its dynamic allowable range is ; In the formula: is the ideal boundary standard deviation threshold under windless conditions, is the wind speed influence coefficient; Step 7c: Divide the thermal buoyancy channel into upper and lower segments, respectively calculate the average temperatures and , and calculate the vertical temperature gradient ratio , and its dynamic allowable range is , which needs to meet ; In the formula: H is the height difference between the upper and lower segments, is the minimum gradient ratio necessary to maintain effective buoyancy, is the maximum gradient ratio set to prevent the channel from being unstable due to excessive thermal stratification; If is exceeded, it indicates that the heating intensity is insufficient or excessive, and the system will prioritize adjusting the heating power; If is exceeded, it indicates that the channel is deformed due to lateral interference, and the system will prioritize fine-tuning the heating angle to resist interference; If is exceeded, it indicates that the channel's thermal structure is abnormal, and the system will comprehensively adjust the power and angle for collaborative optimization; Step 7d: If a deviation is found, immediately adjust the output power and radiation angle of the directional thermal energy application unit through the built-in PID control algorithm, where the power output instruction in the Kth control cycle is : ; angle output instruction : ; wherein: , : initial reference power and angle given by the path planning module; , , : proportional, integral, derivative coefficients of the power loop; , , : proportional, integral, derivative coefficients of the angle loop; : control period; , , : proportional, integral, derivative term deviation of the Kth control period; , : discrete integral approximation of the historical deviation.

[0015] Advantages of the present application: (1) The present application realizes early identification and accurate positioning of hydrogen leakage by constructing a complete perception, decision-making and execution closed loop through multiple information fusion and feedback of concentration, temperature and environmental parameters, solving the problems of response lag and fuzzy positioning of traditional methods; (2) A controllable thermal channel is constructed by active thermal buoyancy guidance, fundamentally changing the diffusion path after hydrogen leakage, actively guiding it to a safe area, avoiding the randomness and inefficiency of passive ventilation dilution, and eliminating the explosion hazard from the source; (3) It has high intelligence and self-adaptive ability, adopts a closed-loop dynamic adjustment strategy, and adjusts the heating power and direction according to real-time temperature data and environmental parameters, ensuring the stability and effectiveness of the thermal buoyancy channel; (4) The classification of leakage level is an intelligent decision combining dynamic calculation and regulatory benchmarks, which is an intelligent judgment based on multi-dimensional real-time data fusion calculation, which surpasses fixed thresholds and realizes dynamic risk assessment; (5) A complete safety protection chain is formed from trace leakage warning, early active intervention to serious leakage emergency shutdown, greatly improving the intrinsic safety level and emergency response capability of hydrogen refueling stations. BRIEF DESCRIPTION OF DRAWINGS

[0016] The present application will be further described below in conjunction with the drawings and examples.

[0017] Figure 1is a structural block diagram of the present application; Figure 2 is a workflow diagram of the present application; Figure 3 is a schematic diagram of the hot buoyancy channel construction and stability control of the present application; In the figure: 1. Hydrogen concentration detection unit, 11. Hydrogen sensor, 2. Directional heat energy application unit, 3. Temperature sensor unit, 4. Central control unit, 41. Data acquisition module, 42. Leakage positioning module, 43. Path planning module, 44. Heating control module, 45. Emergency response module, 5. Environmental parameter acquisition unit, 6. Emergency shutdown unit. DETAILED DESCRIPTION

[0018] The present application will now be further described in detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams and only show the basic structure of the present application in a schematic manner, and therefore only show the components related to the present application.

[0019] As Figure 1 shown in a kind of hydrogen station hydrogen leakage active prevention and control system based on multiple feedback, the system innovatively introduces "hot buoyancy guide" active safety concept, through multi-source information fusion and closed loop control mechanism, a set of detection, positioning, guide and emergency treatment in one intelligent prevention and control system is built, including following unit, each unit cooperates with central control unit 4 through industrial communication network: Hydrogen concentration detection unit 1, including multiple high-precision hydrogen sensors 11 (such as electrochemical sensors) distributed in a spatial matrix along the periphery of potential leakage points in the hydrogen station (such as hydrogen dispensers, hydrogen storage tanks, pipeline valves, etc.), to monitor the hydrogen concentration in the air of the preset monitoring area in the hydrogen station in real time, and generate a leakage alarm signal when the hydrogen concentration exceeds the preset alarm threshold; The spacing of hydrogen sensor 11 is optimized according to the size of the hydrogen station space, usually 2-5 meters, to ensure coverage of all monitoring areas. All hydrogen sensors 11 are connected to the system through industrial Ethernet, and transmit data to the central control unit 4 through the data acquisition module 41. Based on sensor data, the central control unit 4 calculates the spatial concentration field gradient through algorithm to accurately locate the three-dimensional coordinates of the leakage source.

[0020] Directional heat energy application unit 2 is a tunable infrared heating array or phased array microwave transmitter, which is set to apply controllable directional heat energy to a specific spatial area above the leakage source according to instructions. It is designed in a modular way, which can flexibly expand the size and power of the heating array according to the actual size and safety level requirements of the hydrogen station, to meet the different application scenario requirements from small hydrogen dispensers to large stations. The power adjustment of the directional thermal energy application unit 2 preferably adopts a PWM (pulse width modulation) driving mode combined with a constant voltage DC power supply to realize rapid, stable and good linearity power control. Installed on an explosion-proof gimbal, the gimbal integrates a stepper motor or a servo motor + gear reduction mechanism, which can realize dual-axis linkage of azimuth (horizontal 360° continuous rotation) and pitch (±45° range), and the effective action distance is 5-15 meters. The central control unit issues angle correction instructions to the servo driver through an industrial communication interface, and the actuator changes the energy focus direction in the form of small angle stepping or continuous fine tuning, thereby realizing angle correction of the leakage channel. The angle positioning accuracy can be preferably ±0.5°, ensuring that the channel guide direction is highly consistent with the actual offset direction of the leakage plume.

[0021] In addition, the directional thermal energy application unit 2 can also use a phased array microwave transmitter. The heating efficiency of microwave energy on air is higher, and the penetration is better, and it is not easily affected by the water mist in the station. However, its core control logic and execution purpose are completely consistent with the above.

[0022] The temperature sensing unit 3 is arranged to measure the temperature of the target space area and potential leakage point acted on by the directional thermal energy application unit 2, and output real-time temperature field data; A non-contact infrared thermal imager (such as the FLIR series) is used, with a resolution of not less than 320x240 pixels, a temperature measurement range of -20°C to 500°C, and an accuracy of ±1°C. The deployment of the thermal imager follows the principle of "no blind area, overlapping, and key coverage", and is arranged at key high points, potential leakage areas, and key equipment. The field of view of the thermal imager is designed to form an overlapping coverage in the key three-dimensional space (such as the expected hot buoyancy channel path), effectively avoiding monitoring failure caused by single angle being blocked. It feeds back the temperature data stream to the central control unit 4 in real time through an industrial Ethernet, constituting an indispensable feedback link in the closed loop control.

[0023] The central control unit 4 is in communication connection with the hydrogen concentration detection unit 1, the directional thermal energy application unit 2 and the temperature sensing unit 3, respectively, to coordinate the linkage of the units, and has a built-in hydrogenation station three-dimensional space structure model and a CFD simulation module; An industrial PLC or an embedded industrial computer running Linux or Windows IoT system is used. It has a built-in hydrogenation station three-dimensional space structure model (i.e. hydrogenation station three-dimensional BIM digital twin model), and integrates a CFD (computational fluid dynamics) simulation module (such as ANSYS Fluent or OpenFOAM) and a PID control algorithm. It receives data from all sensing units, performs fusion analysis, executes leakage source positioning and prevention strategy formulation, and sends accurate control instructions to each execution unit.

[0024] The environmental parameter acquisition unit 5 is used to acquire environmental wind speed, wind direction and environmental temperature data in real time. The above data is used by the central control unit 4 as feedforward compensation parameters. This unit consists of an integrated miniature weather station capable of real-time acquisition of ambient wind speed (range 0-30 m / s, accuracy ±0.3 m / s), wind direction, and ambient temperature data. The weather station's installation location follows the relevant principles of the "Ground Meteorological Observation Specifications," typically installed in an open area upwind of the hydrogen refueling station area. Its sensor height is more than 10 meters above the ground (e.g., on the roof of the station building or on a freestanding pole) to avoid interference from equipment and buildings within the station, ensuring that the acquired data represents the overall ambient wind field and temperature conditions of the hydrogen refueling station area. This data is used as feedforward compensation parameters by the central control unit 4. The central control unit 4 internally incorporates a feedforward compensator based on a physical model.

[0025] Emergency shut-off unit 6, which serves as the system's final safety barrier, is an explosion-proof emergency shut-off valve (ESD valve) connected to the central control unit 4 and installed in series on the main gas inlet pipeline of the hydrogen refueling station. Upon receiving the shut-off signal from the central control unit 4, its full shut-off time is ≤200ms, ensuring rapid gas supply cut-off in emergencies.

[0026] The system's active guidance method for thermal buoyancy channels involves constructing a continuous and stable high-temperature air column above the leak point after confirming a hydrogen leak. Utilizing the low density of hydrogen, the system actively guides the leaking hydrogen upwards along a preset safe path through thermal buoyancy, preventing it from accumulating on the ground or in the equipment area to form an explosive mixture.

[0027] The central control unit 4 monitors the temperature distribution in the channel in real time through the temperature sensing unit 3 and compares it with the preset ideal channel temperature model generated in real time based on the current scenario. Based on the PID control algorithm, it dynamically adjusts the power and pointing angle of the heating unit to achieve closed-loop control of the stability of the thermal floating channel.

[0028] Simultaneously, the system incorporates real-time wind speed, wind direction, and ambient temperature data acquired by the environmental parameter acquisition unit 5 as feedforward compensation signals into the control loop. Specifically, the central control unit 4 predicts the cooling and shear effects of the wind field on the thermal buoyancy channel based on wind speed and direction, and assesses the background thermal buoyancy conditions based on ambient temperature. This allows for proactive dynamic correction of the power output and operating angle of the directional thermal energy application unit, enabling early suppression of environmental disturbances and ensuring the morphological stability and guiding effect of the thermal buoyancy channel under various environmental conditions.

[0029] The thermal buoyancy channel is entirely based on the principle of thermodynamic buoyancy, forming a channel by non-contactly heating the air with directional thermal energy, without relying on any mechanical forced ventilation devices such as fans, blowers, or deflectors. This purely thermal guidance method fundamentally avoids the risk of electrical sparks that may be generated by the operation of mechanical parts, achieving a higher level of intrinsic safety.

[0030] The central control unit 4 employs digital twin technology to construct a high-fidelity three-dimensional digital twin model (i.e., a three-dimensional spatial structure model of the hydrogen refueling station) that is synchronized in real time with the physical hydrogen refueling station. This model not only accurately reproduces the geometric structure and equipment layout within the station, but also continuously integrates real-time data from hydrogen concentration, temperature, and environmental parameter acquisition units through data interfaces, forming a vivid virtual mirror.

[0031] When a hydrogen leak occurs, the digital twin model provides the built-in CFD simulation module with accurate initial and boundary conditions, including the leak source coordinates, estimated leak rate, real-time wind speed and direction, and ambient temperature. The CFD simulation module performs rapid transient simulations by solving the Navier-Stokes equations, energy equations, and component transport equations. The output of this simulation not only provides quantitative and forward-looking data support for control decisions, such as optimizing the guidance path and initial heating parameters of the thermal flotation channel, but also visually presents the predicted hazardous area (uninterventional situation) and the planned safe guidance path (post-intervention situation) to operators through a 3D visualization interface, thus achieving a fundamental shift from passive monitoring to proactive prediction and intelligent decision-making.

[0032] The three-dimensional spatial structure model of the hydrogen refueling station marks the locations of safe evacuation routes and key hazard sources; when planning the path of the thermal buoyancy channel, the central control unit 4 ensures that the path avoids key hazard sources.

[0033] like Figure 1 As shown, the central control unit 4 is equipped with a software system, which consists of five modules: data acquisition module 41, leak location module 42, path planning module 43, heating control module 44, and emergency response module 45. The hardware side corresponds to the sensor array, directional thermal energy device, thermal imager, environmental station, and emergency shut-off valve, respectively.

[0034] The central control unit 4 is responsible for connecting "sensing—decision-execution—feedback": first, data acquisition sends concentration / temperature / environment data; then, positioning and path planning provide the "point" and "path"; heating control issues power and direction accordingly; and thermal imager returns results for continuous correction. Emergency response operates independently; once trigger conditions are met, it can shut down cascadingly to ensure safety as the priority. Specifically: The data acquisition module 41 is responsible for real-time reading of hydrogen concentration, three-dimensional temperature field, and environmental parameter data. The sampling frequency is uniformly set to 10Hz to ensure data synchronization. After being timestamped, the data is stored in a local time-series database for real-time processing, historical tracing, and offline training of machine learning models. The leak location module 42 accurately locates the initial position of the leak source by analyzing the concentration data gradient of multiple hydrogen sensors 11. When the concentration of any hydrogen sensor 11 exceeds a preset alarm threshold (e.g., 1000 ppm), a leak alarm is triggered. The leak location module 42 immediately retrieves the short-time series concentration data of all hydrogen sensors 11 before the alarm time and uses an optimization algorithm based on concentration gradient (such as gradient descent) to perform an iterative search in three-dimensional space to accurately locate the coordinates of the leak source. The path planning module 43 automatically plans an optimal safe path from the leak point to the high-altitude safety zone, actively avoiding key hazards, within the 3D spatial structure model of the hydrogen refueling station, based on the location of the leak source. Specifically, after receiving the leak source coordinates, the path planning module 43 calls the pre-stored 3D BIM model of the hydrogen refueling station. This model is abstracted as a 3D graph network, where spatial connectivity forms edges, and key locations (leak point, safety zone, center of hazard zone, etc.) form nodes. The module uses Dijkstra's algorithm to calculate an optimal guiding path from the leak point to the high-altitude safety zone (usually defined as open space above 10 meters above the ground). When calculating the path cost, the algorithm applies a large penalty weight to path segments that pass through or are close to "key hazard source" areas, thereby ensuring that the final path geometrically avoids these areas. This geometric path is then passed to the CFD simulation module as the centerline target of the thermal buoyancy channel for subsequent heating strategy simulation and optimization.

[0035] The heating control module 44 precisely and stably constructs and maintains a thermal buoyancy channel that guides the leaking gas upwards, and actively resists wind field interference. This module is the specific executor of PID closed-loop control and feedforward control. It receives the initial heating parameters (power, angle, mode) optimized by the path planning module 43 and controls the directional heat energy application unit 2 to start constructing the thermal channel. Simultaneously, it continuously receives three-dimensional temperature field data from the temperature sensing unit 3 and compares it with the ideal temperature model generated by the CFD simulation model to calculate the deviation. The PID control algorithm dynamically adjusts the heating power and direction based on this deviation to maintain channel stability. Furthermore, this module also receives wind speed data from the environmental parameter acquisition unit 5, and uses the feedforward compensation amount to superimpose with the output of the PID feedback control algorithm to jointly generate the final control command to control the directional heat energy application unit to actively counteract wind field disturbances.

[0036] The emergency response module 45 monitors the rate of increase in concentration and the absolute concentration value in real time (e.g., when the concentration exceeds 20,000 ppm or the rate exceeds 500 ppm / s). When it determines that the leakage level exceeds the safety threshold, it automatically performs an emergency shutdown and triggers an all-around alarm to achieve rapid containment and early warning of the accident.

[0037] The system also includes an alarm device, which is used to trigger an audible and visual alarm and send alarm information to the remote monitoring center when a leak alarm is triggered. All alarm facilities shall comply with the relevant provisions of the current national standard "Design Code for Hydrogen Stations" GB50177. In particular, it should be noted that the system involves two key temperature models, which have completely different functions: Baseline Temperature Model: This is a model developed by the system through learning, characterizing the background temperature values ​​at various locations during normal, leak-free operation of the hydrogen refueling station. It is primarily used to identify temperature anomalies. Through long-term data accumulation and the system's learning capabilities, the baseline temperature, initially a fixed value, is optimized into a predictive model. For example, the system will know that "at a certain time, the normal temperature at the back of hydrogen refueling unit A should be ambient temperature +5°C to +8°C," rather than a simple fixed value. When the real-time temperature deviates from this predicted range, the system can trigger alarms or assist in location assistance earlier and more accurately.

[0038] Ideal channel temperature model: After leakage is confirmed, the target temperature field distribution that needs to be achieved in the channel is calculated in real time by the CFD simulation module in order to construct an effective thermal buoyancy channel. It is mainly used for closed-loop control.

[0039] like Figures 2-3 As shown, a method for proactive prevention and control of hydrogen leakage at hydrogen refueling stations based on multiple feedback includes a proactive prevention and control system for hydrogen leakage at hydrogen refueling stations based on multiple feedback as described in any of the above-mentioned methods, with the specific steps as follows: Step 1: In daily operation, instead of passively waiting for alarms, it is in an active monitoring state. The central control unit 4 continuously collects and analyzes concentration and temperature data, and through machine learning algorithms, it autonomously learns the temperature fluctuation pattern of each monitoring point under normal operating conditions, and dynamically establishes and updates the benchmark temperature model of each monitoring point in the hydrogen refueling station. Step 2: Using hydrogen concentration detection unit 1, monitor the hydrogen concentration in the air of the preset monitoring area within the hydrogen refueling station in real time and generate concentration data; Step 3: Determine whether the concentration data exceeds the preset alarm threshold. If yes, trigger the leak alarm and proceed to Step 4; otherwise, return to Step 2. At the same time as triggering the leak alarm, the audible and visual alarm device is also triggered and an alarm message is sent to the remote monitoring center. The central control unit 4 is also used to dynamically establish and update the reference temperature model (i.e., the temperature distribution reference under normal operating conditions) of each monitoring point in the hydrogen refueling station based on the historical data of the temperature sensing unit 3 during normal system operation. It also includes learning and adaptive capabilities, continuously optimizing the reference temperature model and leakage judgment level through long-term operation and data accumulation, thereby improving the system's adaptability and reliability. In step 3, after the leak alarm is triggered, the central control unit 4 calculates the residual between the real-time temperature data and the reference temperature model. Based on the spatial distribution of monitoring points where the residual exceeds a preset threshold (i.e., the Euclidean distance D between the centroid coordinates of the region with the maximum residual and the concentration location result), it assists in verifying the leak location. Based on the total area of ​​abnormal regions where the residual exceeds the dynamic threshold and the average residual intensity, it assesses the severity of the leak. After the leak alarm is triggered, the central control unit compares the real-time temperature field with the reference temperature model and performs the following quantitative calculation and logical judgment steps to assist in verifying the leak location and assessing the severity: Step 31: Data Preprocessing and Residual Calculation: The system will process the real-time temperature field data... Reference temperature model under corresponding spatiotemporal conditions Accurate registration is performed in three-dimensional space; subsequently, the spatial residual field is calculated. ; Step 32: Auxiliary verification of leak location: First, identify the region with the largest residual in the residual field and calculate the three-dimensional centroid coordinates of this region. ; Then, the coordinates of the centroid and the leak source initially located by the hydrogen concentration detection unit are calculated. The Euclidean distance D between them; Finally, if D is less than a preset threshold (e.g., 1.0 meter), then the temperature anomaly and the concentration anomaly are considered to be highly consistent, the verification is successful, and the location is highly reliable. If the threshold is exceeded, a verification failure flag will be triggered. The system will indicate that there may be false alarms from sensors, multiple leakage sources, or positioning errors, and manual review needs to be initiated immediately. Step 33: Leakage Severity Auxiliary Assessment: Calculate the total projected area S of the 3D region formed by all pixels / voxels in the residual field whose residual values ​​exceed the dynamic threshold (determined by the historical fluctuation standard deviation of the baseline model). Within the aforementioned abnormal region, calculate the average residual value of all points. The system is based on S and Based on two dimensions, the severity of the leak is initially divided into three levels: low, medium, and high. The higher these two values ​​are, the more dangerous the leak is.

[0040] The leakage level determination logic is as follows: The central control unit 4 integrates real-time data from three core dimensions to perform calculations, classifying leakage risks into multiple levels (such as low, medium, high, and critical), and triggering corresponding prevention and control strategies: 1. Concentration Dimension: This includes the hydrogen concentration value and the rate of increase in concentration. The concentration value reflects the current level of danger, while the rate of increase predicts the trend of risk development. The system's initial basic alarm threshold is: when the hydrogen content in the air reaches 0.4% (volume fraction), an alarm should be triggered and recorded (《Technical Specification for Hydrogen Refueling Stations》GB50516).

[0041] 2. Channel stability dimension: This refers to the degree of matching between the actual temperature field of the thermal buoyancy channel and the ideal temperature model (quantified by calculating indicators such as root mean square error (RMSE)). The higher the degree of matching, the better the active guidance effect and the stronger the risk controllability.

[0042] 3. Leakage source characteristics dimension: The intensity (estimated leakage rate) and location (whether it is close to a critical hazard source) of the leakage source determined by the location algorithm.

[0043] Step 4: Upon receiving a leak alarm signal, the system immediately switches from normal operation to emergency response mode. The leak source location module 42 is activated. Based on the concentration data from multiple hydrogen sensors 11 reported by the hydrogen concentration detection unit 1, a high-precision location algorithm is used to calculate the spatial concentration field gradient to accurately locate the three-dimensional coordinates of the leak source. The location accuracy can reach within 0.5 meters, and the target spatial area for thermal energy application is determined. The algorithm steps for accurately locating the three-dimensional coordinates of the leak source in step 4 are as follows: Step 41: Use the sensor location with the highest concentration as the initial iteration point. ; Step 42: In Around a given point, a local three-dimensional concentration field is constructed based on the readings of each sensor, and the concentration gradient vector at that point is calculated. ; Step 43: Move along the gradient direction with a preset step size Update location: ; Step 44: Repeat steps 42 and 43 until the magnitude of the gradient vector is less than the convergence threshold. ,Right now If the algorithm converges to the concentration maximum point, then the position at this point is considered to be... This refers to the three-dimensional coordinates of the identified leak source.

[0044] Step 5: After successful location, the system's decision-making logic begins to work. The path planning module 43 calls the pre-stored three-dimensional spatial structure model of the hydrogen refueling station and plans a thermal flotation channel guide path from the leak point to the high-altitude safety zone based on the three-dimensional coordinates of the leak source. This path avoids the key hazardous sources and equipment marked in the pre-stored three-dimensional spatial structure model of the hydrogen refueling station. Specifically, the central control unit 4 utilizes its built-in CFD simulation module, combining real-time leak location and environmental wind field data, to rapidly predict and simulate the effects of hydrogen diffusion and thermal buoyancy. Based on the simulation results, the system automatically calculates and plans an optimal guidance path that intelligently avoids personnel activity areas, electrical equipment, and other hazards, ensuring that no new risks are introduced.

[0045] Step 6: The heating control module 44 activates and controls the directional thermal energy application unit 2 to perform directional heating on a specific spatial area, so that its energy focus is scanned or statically directionally heated along the starting section of the guide path, forming a local high temperature zone above the leakage source, thereby constructing a thermal buoyancy channel that guides the leaked hydrogen to diffuse upward using the thermal buoyancy effect. Specifically: When the control command is issued, the directional heat energy application unit 2 is activated, and its working process is as follows: 1. Precise Energy Projection: The directional thermal energy application unit 2 precisely focuses energy at a point approximately 0.5-1.5m above the leak source, according to instructions. The emitted infrared or microwave energy is absorbed by air molecules in that area, causing a rapid increase in air temperature. 2. Formation of localized high-temperature zones and buoyancy: As temperature rises, the activity of air molecules increases, leading to a decrease in the number of air molecules per unit volume, thus significantly reducing density. This portion of low-density hot air, surrounded by high-density cold air, generates strong net buoyancy and begins to accelerate upwards. 3. Ejection and Entrainment Effect: As the rising column of hot air (the core of the thermal buoyancy channel) moves, it ejects and entrains leaking hydrogen below, as well as the surrounding air. Because the density of hydrogen (0.089 kg / m³) is much lower than that of hot air (e.g., the density of air at 80°C is about 1.0 kg / m³), the hydrogen is more effectively "entrained" into the rising airflow. 4. Channel Maintenance: The directional thermal energy application unit 2 maintains a stable local high-temperature zone above the leak point through continuous energy input, thereby forming a continuous buoyancy source and ensuring the stable existence of the thermal buoyancy channel. The entire construction process is essentially a man-made and maintained, controllable "thermal chimney" effect, independent of any mechanical moving parts.

[0046] The central control unit 4 is equipped with a feedforward compensator based on a physical model. In step 6, when the heating control module 44 dynamically adjusts the directional heat energy application unit 2, it also introduces the real-time environmental parameters collected by the environmental parameter acquisition unit 5 as feedforward compensation parameters to correct the output of the PID control algorithm, so as to offset the disturbance of the environmental wind field and low temperature environment on the stability of the heat floating channel. The central control unit 4 receives the real-time environmental parameters obtained by the environmental parameter acquisition unit 5, and calculates the feedforward compensation amount based on the above real-time environmental parameters through the feedforward compensator based on the physical model. The specific compensation steps are as follows: Step 61: For environmental wind field compensation, the interference of the wind field on the thermal buoyancy channel is modeled as two effects: convective heat loss and momentum shear. The compensation objects and quantities include power compensation. and angle pre-compensation amount Among them, power compensation amount This is the additional heating power required to offset the convective cooling effect of the wind field. The compensation amount is calculated based on the square of the wind speed; the higher the wind speed, the more additional heating power is required. The calculation formula is as follows: ; In the formula, The wind farm power compensation coefficient was determined experimentally. air density (kg / m³); The specific heat capacity of air at constant pressure (J / (kg·K)); The effective cross-sectional area (m²) of the thermal buoyancy channel. V Real-time wind speed (m / s); The target temperature of the channel (K); The ambient temperature (K); Angle pre-compensation amount This is used to resist the shearing effect of the wind field and maintain the verticality of the channel. The compensation amount is calculated based on the angle between the wind direction and the channel axis. The directional thermal energy application unit is controlled to pre-shift its energy focus to the windward side by a small angle. The calculation formula is as follows: ; In the formula, The angle compensation coefficient was determined experimentally. V Real-time wind speed (m / s); Real-time wind direction (in radians); The initial direction angle of the channel (in radians); The typical buoyancy (m / s) of the thermal buoyancy channel is determined by the channel temperature difference; Step 62: For ambient temperature compensation, the influence of ambient temperature is considered as a change in the basic temperature difference threshold required to generate equivalent buoyancy. The compensation object and quantity are power reference compensation. In low-temperature environments, the system automatically increases the baseline setting of the heating power to ensure sufficient thermal buoyancy is generated in denser cold air. The calculation formula is as follows: ; In the formula, This is the temperature power compensation coefficient; For reference ambient temperature (e.g., 293K, i.e., 20°C); The actual ambient temperature (K); Reference heating power (W) at the reference ambient temperature; Step 63: The above feedforward compensation amount ( The output of the PID feedback control algorithm is superimposed on the output of the PID control algorithm to generate the final control command, which is dynamically sent to the directional thermal energy application unit to make real-time corrections to its output power and application angle, thereby significantly improving the system's ability to resist environmental interference. Specifically: Final power command: ; Final angle command: ; The final control command is sent to the directional thermal energy application unit (2) to make real-time corrections to its output power and angle of action. Step 64: To prevent the integral term from accumulating excessively (integral saturation) during startup or large disturbances, the algorithm adjusts the integral term... Upper and lower limits were set. At the same time, the final control command in step 63 and The amplitude is also limited according to the physical limits of the actuator: ; ; in, and These are the allowable operating ranges for power and angle, respectively.

[0047] Step 7: Based on the real-time temperature data fed back by the temperature sensing unit 3, construct the temperature field distribution map of the thermal floating channel and compare it with the preset ideal channel temperature model. Based on the PID control algorithm, dynamically adjust the output power and direction of the directional heat energy application unit 2 to maintain the stability and effectiveness of the thermal floating channel. The dynamic generation and optimization steps of the ideal channel temperature model in step 7 are as follows: Step 7A: Theoretical basis (CFD simulation): During the system design phase, the CFD simulation module integrated in the central control unit 4 will simulate the temperature distribution that the most effective thermal flotation channel should have under different leakage locations, different leakage rates, different ambient wind fields and temperature conditions, based on the specific three-dimensional structure of the hydrogen refueling station. The simulation results are converted into a series of ideal channel temperature models and stored in the model library. Step 7B: Real-time calculation (path planning): When a real leak occurs, the system will call the CFD simulation module, combine the current real-time leak source location and environmental parameters (wind speed, wind direction, ambient temperature) to perform rapid prediction simulation, and calculate the ideal temperature field distribution corresponding to the optimal thermal buoyancy channel under the current specific conditions. This real-time calculated distribution is the ideal channel temperature model under the current scenario. Step 7C: Self-learning optimization (machine learning): During long-term system operation, the central control unit 4 will record the actual temperature field data and corresponding control parameters in the successful boot cases, and continuously fine-tune and optimize the parameters of the ideal channel temperature model through machine learning algorithms.

[0048] In step 7, after the thermal buoyancy channel is established, the system's focus shifts to maintaining its stability. Temperature sensing unit 3 acquires the temperature distribution along the channel's central axis and surrounding area in real time. The central control unit 4 compares this data point-by-point with a preset ideal channel temperature model, and the dynamic adjustment steps are as follows: Step 7a: Is the central axis temperature high enough? Take N key points on the central axis of the ideal channel model, calculate the root mean square error of the temperature deviation at the corresponding points in the real-time temperature field, and denot it as the central axis temperature deviation. Its dynamic allowable range is ; In the formula: This is the basic tolerance factor. The wind speed influence coefficient, For the ideal central axis average temperature, Real-time wind speed; Step 7b: Continuity of channel boundary temperature: Take M sampling points on the typical cross-sectional boundary of the channel and calculate the standard deviation of the temperature at these points. The smaller this value, the smoother and more continuous the boundary. Its dynamic allowable range is... ; In the formula: The ideal boundary standard deviation threshold under windless conditions. This refers to the wind speed influence coefficient. Step 7c: Does the temperature of the upper and lower sections show a stable increasing trend? Divide the thermal buoyancy channel into upper and lower sections, and calculate the average temperature for each. and Calculate the vertical temperature gradient ratio Its dynamic allowable range is Must meet ; In the formula: H is the height difference between the upper and lower segments. It is the minimum gradient ratio necessary to maintain effective buoyancy. This is the maximum gradient ratio set to prevent excessive thermal stratification from causing channel instability; this range can be fine-tuned according to the ambient temperature, and R can be appropriately increased in cold environments. min To ensure sufficient buoyancy; As stated above, if If the value exceeds the limit, it indicates that the heating intensity is insufficient or excessive, and the system will prioritize adjusting the heating power. like Exceeding the limit indicates that the channel has been deformed by lateral interference (such as wind shear). The system will prioritize fine-tuning the heating angle to counteract the interference. like If the value exceeds the limit, it indicates an abnormality in the channel's thermal structure. The system will then comprehensively adjust the power and angle for coordinated optimization. Step 7d: If a deviation is detected (such as uneven temperature distribution in the channel due to crosswinds), the output power and radiation angle of the directional heat application unit 2 are immediately adjusted dynamically using the built-in PID (proportional-integral-derivative) algorithm to "repair" and stabilize the channel, controlling the input: P proportional term deviation ( Primarily used for rapid response, its core is the central axis temperature deviation. ; I. Integral term deviation ( It is mainly used to eliminate steady-state errors, and its core is the boundary continuity deviation. ; D. Differential term deviation ( It is mainly used to predict and suppress oscillations, and its core is the rate of change of the vertical gradient ratio. .

[0049] Control output is mapped to actuator to adjust power and angle: The power regulation loop mainly responds to temperature intensity deviation (strongly correlated with the P term) and gradient stability (correlated with the D term). Angle adjustment loop: mainly responds to the continuity of channel morphology (strongly correlated with item I) and morphological changes caused by wind field (correlated with item D).

[0050] Among them, the power output command in the Kth control cycle : ; Angle output command : ; In the formula: , The initial reference power and angle are given by the path planning module 43; , , The proportional, integral, and differential coefficients of the power circuit; , , : Proportional, integral, and differential coefficients of the angle loop; Control cycle; , , : Proportional, integral, and derivative deviations in the Kth control cycle; , Discrete integral approximation of historical deviation.

[0051] Step 8: If the leakage level is determined to exceed the safety threshold based on the comprehensive monitoring of the concentration rise rate, absolute concentration value, and temperature field stability, then the emergency shut-off unit 6 is triggered, generating a shutdown command to control the emergency shut-off unit 6 to perform the shutdown operation, shut off the hydrogen supply, and enter the system safety lockout period. If not, determine whether the leak has stopped and whether the concentration has dropped to a safe range. If yes, return to step 1; if no, return to step 7.

[0052] Safety is the paramount principle throughout the entire proactive prevention and control process. An independent safety assessment logic is executed in parallel throughout the process. It continuously and comprehensively analyzes the rate of increase in leaked hydrogen concentration, absolute concentration value, and stability indicators of the thermal channel. Once the leak level is determined to be high-risk or critical, it means the risk is approaching the edge of a controllable range. The central control unit 4 will immediately activate the final safety plan—immediately triggering the emergency shut-off unit 6 to shut down the hydrogen supply and simultaneously sending the highest-level accident alarm to the remote monitoring center, thus constructing a dual safety defense line of "proactively guiding failure and passively shutting down for protection."

[0053] For example, if a hydrogen leak occurs at a hydrogen refueling station, the system will make the following judgment: Level 1 (Low Risk): Concentration exceeds the initial alarm threshold but rises slowly, and the channel remains stable. Only audible and visual alarms are triggered; continuous monitoring is required. Level 2 (Medium Risk): Concentration continues to rise, and the channel shows slight fluctuations. The system activates proactive prevention and control logic to construct and stabilize the thermal buoyancy channel; Level 3 (High Risk / Critical): The absolute concentration value is close to the safety limit, the concentration rise rate is extremely rapid, or the thermal buoyancy channel is out of control due to strong interference and cannot be maintained. The system determines that the active guidance has failed, the risk is uncontrollable, and immediately triggers the emergency shutdown unit.

[0054] like Figure 2 As shown, the system logically constitutes a complete closed loop of "perception-decision-execution-feedback".

[0055] Based on the above-described preferred embodiments of the present invention, and through the foregoing description, those skilled in the art can make various changes and modifications without departing from the inventive concept. The technical scope of this invention is not limited to the contents of the specification, but must be determined according to the scope of the claims.

Claims

1. A multi-feedback-based active prevention and control system for hydrogen leakage in a hydrogen refueling station, characterized by: The application relates to a hydrogen leakage detection and emergency response system for hydrogen refueling stations. The system comprises a hydrogen concentration detection unit (1) for monitoring the hydrogen concentration in the air in a preset monitoring area in the hydrogen refueling station in real time and generating a leakage alarm signal when the hydrogen concentration exceeds a preset alarm threshold; a directional thermal energy application unit (2) for applying controllable directional thermal energy to a specific space area above a leakage source according to instructions; a temperature sensing unit (3) for measuring the temperature of the target space area acted on by the directional thermal energy application unit (2) and the potential leakage point and outputting real-time temperature field data; a central control unit (4) in communication connection with the hydrogen concentration detection unit (1), the directional thermal energy application unit (2) and the temperature sensing unit (3) respectively for coordinating the linkage of the units, wherein a three-dimensional space structure model of the hydrogen refueling station and a CFD simulation module are built in the central control unit (4); an environmental parameter acquisition unit (5) for acquiring environmental wind speed, wind direction and environmental temperature data in real time, wherein the data are used as feedforward compensation parameters by the central control unit (4); and an emergency shutdown unit (6) connected with the central control unit (4) and installed in series on the main air inlet pipeline of the hydrogen refueling station. The hydrogen concentration detection unit (1) comprises a plurality of hydrogen sensors (11) distributed in a matrix form along the space around the potential leakage point in the hydrogen refueling station, and the central control unit (4) can accurately locate the initial position of the leakage source by analyzing the concentration data gradient of the plurality of hydrogen sensors (11). The directional thermal energy application unit (2) is an adjustable-focus infrared heating array or a phased-array microwave transmitter, and the energy focus and radiation direction of the directional thermal energy application unit (2) can be accurately controlled by the central control unit (4). The central control unit (4) is provided with a data acquisition module (41) for reading the hydrogen concentration, three-dimensional temperature field and environmental parameter data in real time; a leakage positioning module (42) for accurately locating the initial position of the leakage source by analyzing the concentration data gradient of the plurality of hydrogen sensors (11); a path planning module (43) for automatically planning an optimal safety path from the leakage point to a high-altitude safety area and actively avoiding key dangerous sources in the three-dimensional space structure model of the hydrogen refueling station according to the position of the leakage source; a heating control module (44) for accurately and stably constructing and maintaining a thermal buoyancy channel for guiding the upward movement of the leaked gas and actively resisting wind field interference; and an emergency response module (45) for monitoring the concentration rising rate and absolute concentration value in real time and automatically executing emergency shutdown and triggering all-around alarm to realize the rapid containment and early warning of accidents when the leakage level exceeds the safety threshold. The specific steps are as follows: Step 1: In daily operation, the central control unit (4) continuously collects and analyzes the concentration and temperature data and dynamically establishes and updates the baseline temperature model of each monitoring point in the hydrogen refueling station by means of machine learning algorithm and self-learning of the temperature fluctuation law of each monitoring point under normal working conditions; 2. The active prevention and control system for hydrogen leakage of a hydrogen refueling station based on multiple feedback according to claim 1, characterized in that: Step 2: The hydrogen concentration in the air in the preset monitoring area in the hydrogen refueling station is monitored in real time by the hydrogen concentration detection unit (1), and concentration data are generated; 3. The active prevention and control system for hydrogen leakage of a hydrogen refueling station based on multiple feedback according to claim 1, characterized in that: Step 3: Whether the concentration data exceed the preset alarm threshold is judged, if yes, the leakage alarm is triggered, and step 4 is entered; if not, step 2 is returned to; 4. The active prevention and control system for hydrogen leakage of a hydrogen refueling station based on multiple feedback according to claim 1, characterized in that: ​ ​ ​ ​ ​ ​ 5. A method for active prevention and control of hydrogen leakage in a hydrogen refueling station based on multiple feedbacks, comprising a system for active prevention and control of hydrogen leakage in a hydrogen refueling station based on multiple feedbacks according to any one of claims 1 to 4, characterized in that: ​ ​ ​ ​ Step 4: After receiving the leakage alarm signal, start the leakage source positioning module (42), based on the concentration data of multiple hydrogen sensors (11) reported by the hydrogen concentration detection unit (1), calculate the spatial concentration field gradient by high-precision positioning algorithm to accurately locate the three-dimensional coordinates of the leakage source and determine the target space area of thermal energy application; Step 5: The path planning module (43) calls the pre-stored three-dimensional space structure model of the hydrogen refueling station, and plans a thermal buoyancy channel guide path from the leakage point to the high-altitude safety zone according to the three-dimensional coordinates of the leakage source, which avoids the key danger sources and equipment marked in the pre-stored three-dimensional space structure model of the hydrogen refueling station; Step 6: The heating control module (44) activates and controls the directional thermal energy application unit (2) to perform directional heating on the specific space area, so that the energy focus of the directional thermal energy application unit (2) performs scanning or static directional heating along the starting segment of the guide path, forming a local high-temperature area above the leakage source, thereby constructing a thermal buoyancy channel that guides the upward diffusion of leaked hydrogen gas using the thermal buoyancy effect; Step 7: According to the real-time temperature data fed back by the temperature sensing unit (3), construct the temperature field distribution map of the thermal buoyancy channel, and compare it with the pre-set ideal channel temperature model, based on the PID control algorithm to dynamically adjust the output power and direction of the directional thermal energy application unit (2), in order to maintain the stability and effectiveness of the thermal buoyancy channel; Step 8: If the leakage level exceeds the safety threshold based on the comprehensive monitoring of the concentration rise rate, absolute concentration value and temperature field stability, trigger the emergency shutdown unit (6) to generate a shutdown instruction to control the emergency shutdown unit (6) to perform shutdown operation, shut down the hydrogen supply, and enter the system safety locking period; If not, determine whether the leakage has stopped and the concentration has dropped to a safe range, if so, return to step 1; if not, return to step 7.

6. The method of claim 5, wherein the method further comprises: In step 3, after the leakage alarm is triggered, the central control unit (4) calculates the residual error between the real-time temperature data and the reference temperature model, and based on the spatial distribution of the monitoring points whose residual error exceeds the pre-set threshold, assists in verifying the leakage position and evaluating the leakage severity, performs the following quantitative calculation and logical judgment steps: Step 31 : The system will calculate the real-time temperature field data with the reference temperature model under the corresponding spatio-temporal condition Accurate registration in three-dimensional space; subsequently, calculate the spatial residual field ; Step 32: First, identify the region of maximum residual in the residual field and calculate the three-dimensional centroid coordinates of this region ; Then, the Euclidean distance D between the center of mass and the coordinates of the leak source preliminarily positioned by the hydrogen concentration detection unit is calculated between the center of mass and the coordinates of the leak source preliminarily positioned by the hydrogen concentration detection unit is calculated Finally, if D is less than the pre-set threshold, it is considered that the temperature anomaly and the concentration anomaly source are highly consistent, the verification is passed, and the position is highly reliable; If it exceeds the threshold, a verification failure flag is triggered, and the system will prompt that there may be sensor false alarm, multiple leakage sources or positioning error, which needs to be urgently started for manual review; Step 33: Calculate the total projection area S of the three-dimensional region formed by all the pixels / voxels whose residual values in the residual field exceed the dynamic threshold value, and calculate the average residual value of all points in the above abnormal region ; the system preliminarily divides the leakage severity into three levels of low, medium and high according to S and two dimensions.

7. The method of claim 5, wherein the method further comprises: The algorithm steps for accurately locating the three-dimensional coordinates of the leakage source in step 4 are as follows: Step 41 : Use the sensor location with the highest concentration as the initial iteration point ; Step 42: At each point, construct a local three-dimensional concentration field from the sensor readings and calculate the concentration gradient vector at that point ; and ; Step 43: move along the gradient direction with a preset step size Update location: ; Step 44: repeat step 42 and step 43 until the modulus of the gradient vector is less than a convergence threshold i.e. then the algorithm is considered to have converged to a concentration maximum, and the position i.e. the determined leak source three-dimensional coordinates.

8. The method of claim 5, wherein the method further comprises: The central control unit (4) is provided with a feedforward compensator based on a physical model. In step 6, the heating control module (44) introduces the real-time environmental parameters collected by the parameter collection unit (5) as feedforward compensation parameters when dynamically adjusting the directional heat energy application unit (2), and modifies the output of the PID control algorithm to offset the disturbance of the environmental wind field and low-temperature environment on the stability of the thermal floatation channel. The central control unit (4) receives the real-time environmental parameters obtained by the environmental parameter collection unit (5), calculates the feedforward compensation amount based on the above real-time environmental parameters through the feedforward compensator based on the physical model, and the specific compensation steps are as follows: Step 61: Compensation for the ambient wind field, including a power compensation amount and an angle pre-compensation amount wherein the power compensation amount The calculation formula is: ； wherein, is the wind farm power compensation coefficient, calibrated by experiment; is the air density (kg / m³); is the air constant pressure specific heat capacity (J / (kg·K)); is the effective cross-sectional area of the thermal buoyancy channel (m²); V is the real-time wind speed (m / s); is the channel target temperature (K); is the ambient temperature (K); Angle pre-compensation amount The calculation formula is: ； In the formula, is an angle compensation coefficient, calibrated by experiment; V is real-time wind speed (m / s); is real-time wind direction (radian); is the initial direction angle of the channel (radian); is the typical floating speed of the thermal floating channel (m / s), determined by the temperature difference of the channel; Step 62: Compensate the power reference for ambient temperature The calculation formula is: ； In the formula, is a temperature power compensation coefficient; is a reference ambient temperature; is an actual ambient temperature (K); is a reference heating power at the reference ambient temperature (W); Step 63: The above feedforward compensation amount is superimposed with the output of the PID feedback control algorithm to generate the final control instruction, which is: Final power command: ; Final angle command: ; The above final control instruction is sent to the directional heat energy application unit (2) to real-time correct its output power and action angle; Step 64: Limit the final control instruction in step 63: ； ； wherein, and are the allowed operating ranges for power and angle, respectively.

9. The method of claim 5, wherein the method further comprises: The dynamic generation and optimization steps of the ideal channel temperature model in step 7 are as follows: Step 7A: In the system design stage, the CFD simulation module integrated in the central control unit (4) will simulate the temperature distribution that the most effective thermal floatation channel should have under different leakage positions, different leakage rates, different environmental wind fields and temperature conditions according to the specific three-dimensional structure of the hydrogen filling station. The simulation results are converted into a series of ideal channel temperature models and stored in the model library; Step 7B: When a real leakage occurs, the system will call the CFD simulation module to combine the current real-time leakage source position and environmental parameters to perform rapid prediction simulation and calculate the ideal temperature field distribution corresponding to the optimal thermal floatation channel under the current specific conditions. This real-time calculated distribution is the ideal channel temperature model under the current scenario; Step 7C: In the long-term operation of the system, the central control unit (4) will record the actual temperature field data and corresponding control parameters in the successful guidance cases, and continuously fine-tune and optimize the parameters of the ideal channel temperature model through machine learning algorithms.

10. The method of claim 5, wherein the method further comprises: The dynamic adjustment steps in step 7 are as follows: Step 7a: Take N key points on the axis in the ideal channel model, calculate the root mean square error of the temperature deviation of the corresponding points in the real-time temperature field, and record it as the middle axis temperature deviation , and the dynamic allowable range is ; In the formula: is a basic tolerance coefficient, is a wind speed influence coefficient, is an ideal central axis average temperature, is a real-time wind speed; Step 7b: Take M sampling points on the typical cross-sectional boundary of the channel, calculate the standard deviation of the temperature of these points with a dynamic allowable range of ; In the formula: is the ideal boundary standard deviation threshold value under windless conditions, is the wind speed influence coefficient; Step 7c: Divide the hot updraft into two sections, top and bottom, and calculate the average temperature in each and Calculate the vertical temperature gradient ratio with a dynamic allowable range of Must satisfy ; where H is the height difference between the upper and lower sections, is the minimum gradient ratio necessary to maintain the effective buoyancy, is the maximum gradient ratio set to prevent the channel from becoming unstable due to excessive thermal stratification. If Exceeding the standard indicates that the heating intensity is insufficient or excessive, and the system will adjust the heating power first. If If the value is over the standard, it indicates that the channel is deformed by lateral interference, and the system will preferentially fine-tune the heating angle to resist the interference. If If the temperature exceeds the threshold, it indicates that the channel thermal structure is abnormal, and the system will comprehensively adjust the power and angle for collaborative optimization. Step 7d: If a deviation is found, the output power and radiation angle of the directional heat energy application unit (2) are immediately dynamically adjusted through the built-in PID control algorithm, wherein, in the Kth control cycle, the power output instruction : ； Angle output command : ； In the formula: , : initial reference power and angle given by the path planning module (43) , , : power loop proportional, integral, derivative coefficients; , , : angle loop proportional, integral, derivative coefficients; : control period; , , : deviation of proportional, integral, and derivative terms for the Kth control cycle , : Discrete integral approximation of historical bias.

Images