A hydrogen leakage intelligent detection and automatic plugging system and method for a hydrogen production station by hydrogenation

Abstract

This invention relates to an intelligent detection and automatic sealing system and method for hydrogen leaks in hydrogen refueling and production stations, belonging to the field of hydrogen leak detection and sealing. The system includes a hydrogen leak detection module, a sealing execution module, a control module, a sealing effect monitoring module, and a mobile robot and communication module. The system can monitor potential leak points in hydrogen refueling and production stations in real time, automatically calculate and execute adaptive sealing operations based on leak information. The sealing material adopts a multi-level composite structure, combining a rapid-condensing hydrogen-absorbing gel, a shape memory alloy skeleton, a liquid metal micropump, and a high-pressure wear-resistant outer layer to achieve precise and efficient sealing of leaks with different geometric characteristics. The system has a high degree of automation, intelligent monitoring of sealing effects, and adaptive optimization, improving the safety and emergency response capabilities of hydrogen refueling and production stations.

Description

Technical Field

[0001] This invention belongs to the field of hydrogen leak detection and sealing, and more specifically relates to an intelligent detection and automatic sealing system and method for hydrogen leaks in hydrogen refueling and hydrogen production stations. Background Technology

[0002] Hydrogen, as a clean and efficient energy source, poses a significant threat to personnel and property during its preparation, storage, transportation, and use due to its flammable, explosive, rapidly diffusing, colorless, and odorless properties, especially in the event of a leak. This invention relates to the field of hydrogen energy, specifically to the safety management of hydrogen leaks during hydrogen production, storage, transportation, and refueling, and particularly to a device and method for rapidly and accurately detecting and automatically sealing hydrogen leaks.

[0003] 1. Traditional fixed-point sensors have spatial blind spots, low positioning accuracy (can only locate areas), difficulty in quantifying leakage, and are easily affected by environmental interference.

[0004] 2. Leakage response relies mainly on manual labor, resulting in slow response times and personnel being exposed to high-risk environments, posing serious safety risks.

[0005] 3. Manual sealing methods lack specificity, are difficult to adapt to complex and ever-changing types of leaks, and the sealing effect depends on experience, resulting in an unstable success rate.

[0006] 4. Leak detection and sealing execution systems are often fragmented, lacking information linkage, real-time evaluation and optimization capabilities, and decision-making relies on manual intervention and cannot learn on its own. Summary of the Invention

[0007] This invention addresses the shortcomings of existing hydrogen leak detection technologies, which struggle to achieve precise three-dimensional spatial localization of leak points and accurate real-time quantification of leak volume. Existing hydrogen leak sealing methods primarily rely on manual operation, resulting in slow response times, high operational risks, and a lack of targeted and adaptable sealing solutions. Furthermore, existing leak emergency response systems lack integrated, closed-loop, and adaptive capabilities encompassing detection, precise quantification, intelligent decision-making, automated execution, real-time effect evaluation, and strategy adjustment.

[0008] To solve the above problems, the present invention is implemented using the following technical solution: The system includes:

[0009] The hydrogen leak detection module is deployed near various potential leak points at the hydrogen refueling station to detect whether there is a hydrogen leak at the hydrogen refueling and hydrogen production station;

[0010] The sealing execution module achieves adaptive deformation and precise fit of the sealing material based on the complex geometric features of the leak point;

[0011] The control module integrates leakage information and assesses risks, intelligently calculates and optimizes plugging parameters, and calculates the optimal plugging force.

[0012] The blocking effect monitoring module features multimodal real-time monitoring and data fusion, intelligent evaluation, and adaptive strategy adjustment.

[0013] The mobile robot is powered by a high-capacity lithium battery pack, and the communication module establishes a real-time data link with the central control room and emergency command center of the hydrogen refueling / production station.

[0014] In one embodiment, the adaptive blocking material of the blocking execution module adopts a multi-level composite structure, including:

[0015] The inner layer is in direct contact with the leak: a fast-condensing hydrogen-absorbing gel / polymer that rapidly polymerizes upon contact with hydrogen or a specific catalyst, forming a gel with a certain degree of elasticity and viscosity, and then quickly solidifies.

[0016] Intermediate layer: Shape memory alloy (SMA) framework network, a precision three-dimensional mesh framework made of nickel-titanium alloy with shape memory effect and superelasticity, achieves precise shape recovery and deformation by precisely controlling electric heating or laser heating to achieve phase transformation between austenite and martensite; Liquid metal micropump, a micro liquid metal pump and channel integrated inside or adjacent to the SMA framework.

[0017] Outer layer: High-pressure wear-resistant polyurethane / fluororubber, protecting the internal structure; micro magnetic adsorption units, designed for uneven surfaces, with an integrated array of micro magnetic adsorption units at the edges of the sealing material.

[0018] In one embodiment, the mobile robot is powered by a high-capacity lithium battery pack, supporting long-term operation. The robotic arm and sensors are powered by the mobile robot. It is equipped with wireless charging or automatic charging stations. When the battery level is below a threshold, the mobile robot autonomously navigates to the charging station for charging.

[0019] Mobile robots have the following devices:

[0020] a) Tunable diode laser absorption spectroscopy (TDLAS) hydrogen sensor: After the laser beam passes through the hydrogen plume, the hydrogen concentration integral along the path is calculated by scanning the attenuation of the laser absorption intensity, and a three-dimensional hydrogen concentration field is quickly constructed.

[0021] b) A high-resolution stereo vision camera is used to acquire three-dimensional images of the leak area and capture the geometric information of the leak point;

[0022] c) Ultrasonic / infrasonic sensors: Hydrogen leakage under high pressure produces unique acoustic characteristics, and multiple acoustic sensors are used for sound source localization.

[0023] d) Infrared thermal imager: The rapid expansion of hydrogen gas can cause a sudden drop in local temperature. Thermal imaging can help locate the leak point.

[0024] The communication module has a built-in high-performance wireless communication module, including Wi-Fi, 5G, and satellite communication modules.

[0025] Furthermore, a method for intelligent detection and automatic sealing of hydrogen leaks in hydrogen refueling and production stations is provided. This method is applicable to the aforementioned system and includes the following steps:

[0026] Step 1: Configure and fix the sensors, collect all sensor data, and transmit the data to the CCD module via wireless network for coarse and fine leak location, as well as leak volume estimation and leak vent shape identification.

[0027] Step 2: Plan the collision-free trajectory of the robotic arm from its current position to the leak point and execute the sealing operation;

[0028] Step 3: The control module calculates the sealing material parameters, calculates the optimal sealing force, and performs path planning and obstacle avoidance calculations for the sealing robot.

[0029] Step 4: Monitoring the sealing effect, multimodal real-time monitoring and data fusion, intelligent evaluation of sealing effect and adaptive strategy adjustment, establishment of machine learning model, simulation of different leakage scenarios, continuous optimization of sealing strategy, and improvement of the ability to respond to unknown leakage in the future.

[0030] In one approach, the coarse leak location specifically includes: after receiving an alarm signal, the control module combines the location information and concentration readings of multiple alarm sensors; when the hydrogen concentration detected by any hydrogen sensor exceeds a preset first threshold, the area to which the sensor belongs is marked as a potential leak area; when multiple fixed sensors trigger alarms, the centroid method or weighted average method is used for preliminary area location, to initially delineate the approximate range of the leak, and to quickly estimate the rough location area of ​​the leak point.

[0031] In one approach, the precise leak location involves the control module immediately dispatching a mobile robot equipped with high-precision sensors to the area for detailed scanning after detecting the initial leak area.

[0032] Construct a three-dimensional hydrogen concentration field C(x,y,z) in the leak area. The leak point is usually located in the region of maximum concentration gradient, since the leak point (x... l ,y l ,z l The maximum value of the three-dimensional gradient vector is consistent with the direction of the leakage source. By calculating the three-dimensional gradient vector ▽C(x,y,z) of the concentration field, the starting point (edge ​​of the high concentration region) and direction of the three-dimensional gradient vector indicate the leakage source.

[0033] Based on the highest hydrogen concentration, concentration diffusion range, diffusion rate (using time series data) measured near the leak point after precise location, as well as ambient temperature and pressure parameters.

[0034] In one approach, the leakage estimation and leak outlet shape identification include:

[0035] (1) Based on the three-dimensional coordinates of the leak point and the hydrogen concentration values ​​detected by multiple sensors, combined with environmental parameters, flow rate inversion was performed using TDLAS.

[0036] Specifically, the hydrogen concentration distribution C(r,θ,z) at different cross-sections was obtained by scanning the leaking plume using TDLAS; the plume velocity v(r,θ,z) and leakage flow rate Q were obtained with the assistance of laser Doppler velocimetry or particle image velocimetry. leak It is obtained by integrating the mass flux over the cross-section of the plume.

[0037] Where A is the cross-sectional area of ​​the plume, v z It is the velocity component of the plume along the z-axis (or the main diffusion direction), ρ H2 It is the density of hydrogen gas;

[0038] (3) The high-resolution stereo vision camera on the mobile robot performs high-density scanning of the leak point in a spiral or zigzag scanning manner within the coarse positioning area to generate high-density point cloud data.

[0039] After removing noise points from the point cloud, the RANSAC algorithm is used to identify geometric primitives, such as planes, spheres, and cylinders (corresponding to pipes). Then, on the identified equipment surface, the edges of the anomaly—i.e., the outline of the leak—are identified using normal estimation and curvature analysis methods.

[0040] Circular hole: Fit a circle and extract the center coordinates (x) c ,y c ,z c ) and diameter D leak ;

[0041] Crack: Fit a straight line segment or a curve segment and extract the length L. leak Average width W leak and direction angle θ leak ;

[0042] Irregular damage: Calculate its minimum bounding rectangle or convex hull and extract its area A. leak and perimeter P leak .

[0043] In one approach, the blocking steps are as follows:

[0044] S201, Path Planning and Autonomous Navigation: The control module sends the calculated optimal path and target coordinates to the sealing execution module; the mobile robot uses its onboard LiDAR, visual odometry, and inertial measurement unit sensors, combined with a pre-loaded high-precision map of the hydrogen refueling / hydrogen production station, to navigate autonomously, avoid obstacles, and quickly reach the vicinity of the leak point.

[0045] S202. Fine Positioning and Attitude Adjustment: After the mobile robot reaches the target area, the robotic arm extends, and the end effector, equipped with a high-precision vision sensor, performs a secondary scan of the leak point to obtain its precise three-dimensional geometric information (hole shape, crack direction). The control module combines the scan data with a preset model to calculate the precise position and attitude of the robotic arm's end effector, ensuring it is perfectly aligned with the leak point. Based on the fine positioning results, the robotic arm adjusts its own attitude and the position and angle of the end effector to perfectly align it with the leak point and place it in the optimal operating position.

[0046] S203 Surface Pretreatment: For sealing materials that require surface cleanliness for good adhesion, the robotic arm switches to the pretreatment tool to lightly clean the area around the leak point and remove loose contaminants; if it is necessary to enhance adhesion, apply a non-sparking primer or activator.

[0047] S204, Deployment of sealing materials.

[0048] In one approach, the calculation of the sealing material parameters includes: based on the geometric parameters of the leak opening (diameter D) leak Length L leak Average width W leak Direction angle θ leak Area A leak and perimeter P leak (etc.), select the most suitable sealing material template;

[0049] Circular hole: Desired sealing material diameter D seal =k D* D leak ;

[0050] Crack: Expected sealing material length L seal =k L* L leak Width W seal =k W* W leak ;

[0051] Where k D k L k W A redundancy factor of 1.1-1.2 is used to ensure edge coverage;

[0052] According to the required D seal Lseal W seal The desired deformation of the SMA skeleton is deduced by reverse calculation. The heating parameters are determined by combining the SMA phase transition curve, and then the required heating current, laser power and heating time are calculated.

[0053] In one approach, calculating the optimal sealing force includes: considering the pressure at the leak point, the elastic modulus of the material being sealed, and the compressive modulus of the sealing material, to calculate the target sealing force F. 目 Mark;

[0054] F 目标 =P leak* A leak +K 压缩 *A 接触

[0055] Where K 压缩 The compression coefficient of the seal is used to ensure that the seal deformation fills even minor unevenness. A 接触 This refers to the contact area of ​​the seal.

[0056] Beneficial effects of this invention:

[0057] 1. Significantly improves the accuracy and comprehensiveness of leak detection:

[0058] By fusing multimodal sensors (such as infrared, acoustic, thermal imaging, and distributed fiber optic sensing), the limitations of single sensors are overcome, enabling precise three-dimensional spatial coordinate localization of hydrogen leaks, including micro-leaks and high-pressure jet leaks.

[0059] By combining high-precision sensors carried by mobile robots / drones, autonomous inspections can be carried out in complex and hard-to-reach areas, further eliminating blind spots and achieving full-coverage, high-resolution leak detection.

[0060] It can quantify the leakage amount and leakage characteristics (such as hole size, shape, injection pressure, etc.) in real time and accurately, providing precise data support for subsequent sealing decisions and making emergency response more targeted.

[0061] 2. Greatly improves emergency response speed and ensures personnel safety.

[0062] The device can operate autonomously around the clock. Once a leak is detected, it can complete information transmission and decision calculation within seconds and control the sealing execution module to quickly reach the leak point, reducing the time required for manual response from tens of minutes or even hours to minutes, and significantly reducing the risk of escalation of the accident.

[0063] The entire sealing process was completed independently by an intelligent robotic system, eliminating the need for personnel to enter areas with high concentrations of hydrogen or come into contact with the leak point. This completely eliminated the direct risk of personnel being exposed to flammable and explosive environments and effectively protected the lives of operators.

[0064] 3. Achieve intelligent customization and efficient, precise execution of containment solutions.

[0065] By using fluid dynamics and materials mechanics simulation (CFD / FEA), the optimal plugging strategy and parameters can be intelligently calculated based on precise leakage parameters, including the type and amount of plugging material, the applied force, and the posture of the robotic arm, ensuring the scientific validity and effectiveness of the plugging solution.

[0066] By introducing AI-driven intelligent matching of sealing materials and on-site adaptive 3D printing of customized sealing components, the device can provide highly customized and perfectly fitting sealing solutions for leaks of different shapes, sizes, pressures, and media, significantly improving the success rate and durability of sealing.

[0067] The multi-functional robot platform is equipped with an automatic tool changing system and high-precision force / visual feedback control, enabling it to complete the entire process from surface cleaning and polishing to precise patch application and colloid injection, achieving unmanned and high-precision sealing operations.

[0068] 4. Construct an intelligent, adaptive closed-loop emergency response system

[0069] It achieves complete intelligent closed-loop control, from leak detection, precise quantification, intelligent decision-making, and automatic execution to real-time effect monitoring and dynamic strategy adjustment. The system can evaluate the sealing effect in real time, and if it is unsatisfactory, it will automatically initiate secondary optimization sealing or even switch to a more aggressive approach until the leak is completely controlled.

[0070] The system has the ability to learn and accumulate experience. Through machine learning algorithms, it uses the successful or failed experience of each leak handling, the parameters used, and the monitoring results as training data to continuously optimize its decision-making model and execution strategy. This makes the intelligence level of the device and the future sealing success rate continuously improve with the number of uses, and it has the characteristic of "getting smarter with use". Attached Figure Description

[0071] Figure 1 This is a system block diagram of the present invention;

[0072] Figure 2 This is a flowchart of the method of the present invention. Detailed Implementation

[0073] To facilitate understanding of the present invention, a more complete description will be given below with reference to the accompanying drawings. Typical embodiments of the invention are shown in the drawings. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete.

[0074] Unless otherwise defined, all technical and scientific terms used in this invention have the same meaning as understood by one of ordinary skill in the art to which this invention pertains. The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. To facilitate understanding, the invention will now be described more fully with reference to the accompanying drawings. Typical embodiments of the invention are shown in the drawings. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided to make the disclosure of the invention more thorough and complete.

[0075] like Figure 1 and Figure 2 As shown, the intelligent detection and automatic sealing system and method for hydrogen leakage at hydrogen refueling and hydrogen production stations includes:

[0076] 1. Hydrogen Leakage Detection Module

[0077] 1.1 Sensor Configuration and Layout:

[0078] (1) High-sensitivity electrochemical hydrogen sensor (EC-H2): Deployed near potential leak points in hydrogen refueling stations (such as the top of hydrogen storage tanks, valve groups, and above hydrogen refueling nozzles) for real-time monitoring of local hydrogen concentration. Its characteristic is fast response to trace amounts of hydrogen, capable of detecting hydrogen as low as a few ppm.

[0079] (2) Semiconductor hydrogen sensor (MOS-H2): It is low in cost, suitable for large-scale dense grid deployment, provides wider coverage, and is used to quickly locate the initial leak area.

[0080] Fixed hydrogen sensor arrays: High-density hydrogen sensor arrays are established in key areas of hydrogen refueling / production stations (such as electrolyzer areas, compressor rooms, hydrogen storage tank areas, high-pressure pipeline connections, valves, flanges, hydrogen refueling nozzles, etc.). Sensors can be distributed along pipelines in a grid, point, or linear pattern, with particular emphasis on areas above where hydrogen tends to accumulate (hydrogen is lighter than air). For large equipment or storage tanks, a surround or multi-layered three-dimensional layout can be used to ensure three-dimensional spatial coverage.

[0081] 1.2 Leakage Location

[0082] (1) Data collection: All sensors (hydrogen concentration, pressure, flow rate, ambient temperature and humidity, etc.) collect data at a high frequency and transmit the data to the CCD module via wireless network.

[0083] (2) Data cleaning and filtering: The CCD module performs outlier removal, noise filtering, calibration and redundancy processing on the raw data to improve data quality.

[0084] (3) Coarse Leakage Location: After receiving the alarm signal, the coarse location control module combines the location information and concentration readings of multiple alarm sensors. When the hydrogen concentration detected by any hydrogen sensor exceeds a preset first threshold, the area where that sensor is located is marked as a potential leak area. When multiple fixed sensors trigger alarms, the centroid method or weighted average method is used for preliminary area location, initially delineating the approximate range of the leak and quickly estimating the rough location of the leak point, as follows: Assuming there are M sensors alarming, the position of alarm sensor Si is denoted as (x i ,y i ,z i Its concentration reading is C. i The initial location of the leak center (x0, y0, z0) can be estimated as follows:

[0085] (4) Precise leak location:

[0086] After detecting a preliminary leak area, the control module immediately dispatches a mobile robot (or drone) equipped with high-precision sensors to the area for detailed scanning, constructing a three-dimensional hydrogen concentration field C(x,y,z) of the leak area. The leak point is typically located in the region of maximum concentration gradient, because the leak point (x...y...z)... l ,y l ,z l The maximum value of the three-dimensional gradient vector must be consistent with the direction of the leakage source. The leakage source can be indicated by calculating the three-dimensional gradient vector ▽C(x,y,z) of the concentration field. The starting point (edge ​​of the high-concentration region) and direction of the three-dimensional gradient vector can be used to identify the leakage source. This is based on the highest hydrogen concentration, concentration diffusion range, diffusion rate (using time-series data), and ambient temperature and pressure parameters measured near the leakage point after precise location.

[0087] The robot / drone payload includes the following devices: a) a tunable diode laser absorption spectroscopy (TDLAS) hydrogen sensor, which calculates the hydrogen concentration integral along the path by scanning and measuring the attenuation of laser absorption intensity after the laser beam passes through the hydrogen plume, and can quickly construct a three-dimensional hydrogen concentration field; b) a high-resolution stereo vision camera, used to acquire three-dimensional images of the leak area and capture the geometric information of the leak point; c) an ultrasonic / infrared sensor, as hydrogen leaks under high pressure produce unique acoustic characteristics, and multiple acoustic sensors are used for sound source localization; d) an infrared thermal imager, as the rapid expansion of hydrogen causes a sudden drop in local temperature, and thermal imaging can assist in locating the leak point.

[0088] 1.3 Leakage estimation and leak point shape identification

[0089] (1) Based on the three-dimensional coordinates of the leak point and the hydrogen concentration values ​​detected by multiple sensors, combined with environmental parameters (wind speed, wind direction, temperature, humidity, etc.), flow inversion is performed based on TDLAS, as follows: The hydrogen concentration distribution C(r,θ,z) at different cross sections is obtained by scanning the leak plume with TDLAS; the plume velocity v(r,θ,z) and leak flow rate Q are obtained with the assistance of laser Doppler velocimetry or particle image velocimetry. leak It can be obtained by integrating the mass flux over the cross-section of the plume.

[0090] Where A is the cross-sectional area of ​​the plume, v z It is the velocity component of the plume along the z-axis (or the main diffusion direction), ρ H2 It is the density of hydrogen gas.

[0091] (2) The high-resolution stereo vision camera mounted on the mobile robot performs a high-density scan of the leak point within the coarse localization area using a spiral or zigzag scanning method, generating high-density point cloud data. After removing noise points from the point cloud, the RANSAC (Random Sample Consensus) algorithm is used to identify geometric primitives from the point cloud, such as planes, spheres, and cylinders (corresponding to pipes). Then, on the identified equipment surface, the edges of the anomaly—i.e., the outline of the leak—are identified using methods such as normal estimation and curvature analysis.

[0092] Circular hole: Fit a circle and extract the center coordinates (x) c ,y c ,z c ) and diameter D leak ;

[0093] Crack: Fit a straight line segment or a curve segment and extract the length L. leak Average width W leak and direction angle θ leak ;

[0094] Irregular damage: Calculate its minimum bounding rectangle or convex hull and extract its area A. leak and perimeter P leak .

[0095] 2. Blocking Execution Module

[0096] The core of this module is its ability to adapt and precisely fit the sealing material based on the complex geometric features of the leak point.

[0097] 2.1 Adaptive plugging materials

[0098] 2.1.1 Multi-level composite structure

[0099] Inner layer (in direct contact with the leak): Rapidly solidifying hydrogen-absorbing gel / polymer, which rapidly polymerizes upon contact with hydrogen (or a specific catalyst) to form a gel with a certain degree of elasticity and viscosity, and then solidifies quickly. The gel contains nanoporous structures or metal-organic framework (MOF) particles, which can instantly adsorb a large amount of hydrogen, further reducing the leak pressure and concentration.

[0100] Intermediate Layer (Shape-Driven Core): 1) Shape Memory Alloy (SMA) Skeleton Network: A precision three-dimensional mesh skeleton made of nickel-titanium (NiTi) alloy possesses shape memory effect and superelasticity. Through precise control of electrothermal (Joule heating) or laser heating, it can undergo a phase transformation between austenite and martensite, achieving precise shape recovery and deformation. The skeleton is designed as a stretchable, bendable, and expandable unit; 2) Liquid Metal Micropumps: Micro-liquid metal (such as gallium-based alloy) pumps and channels are integrated inside or adjacent to the SMA skeleton. The liquid metal micropumps utilize the high thermal conductivity of liquid metal, and by controlling its flow and solidification / melting, localized heating or cooling of the SMA skeleton can be achieved, thereby enabling more precise control of its shape deformation.

[0101] Outer layer (protection and connection): 1) High-pressure wear-resistant polyurethane / fluororubber, providing excellent pressure resistance, wear resistance and corrosion resistance, protecting the internal structure; 2) Micro magnetic adsorption unit, for non-flat surfaces, the edge of the sealing material can be integrated with a micro magnetic adsorption unit array to assist in fixation in the early stage of sealing and improve stability.

[0102] 2.2 High-precision drive mechanism and force control

[0103] The seven-degree-of-freedom robotic arm, driven by a high-precision servo motor, offers extremely high flexibility and operating space, enabling it to mimic human hands in complex environments.

[0104] 2.2.1 Kinematic Model: The core of the model consists of forward kinematics (calculating the end effector position and orientation based on joint angles) and inverse kinematics (calculating joint angles based on the target position and orientation). Forward kinematics uses homogeneous transformation matrix multiplication; inverse kinematics is solved using numerical iterative methods (such as Jacobi iteration).

[0105] 2.2.2 Visual Servo: A high-resolution industrial camera is integrated into the end effector of the robotic arm to track leak points in real time through image processing. The control system adjusts the movement of the robotic arm based on image error signals, achieving high-precision visual feedback control. Control Law: v c =-λL + e v , where v c For camera speed, L + e is the pseudo-inverse of the image Jacobian matrix. v This is the visual error vector.

[0106] 2.2.3 End Effector: The robotic arm end effector integrates a multi-functional flexible gripper containing: 1) a force / torque sensor array, capable of sensing the pressure distribution on the contact surface and measuring the contact force between the sealing material and the leak point in real time, which is crucial for assessing the uniformity of the sealing; and an SMA skeleton drive controller, which precisely controls the current or laser energy applied to the SMA skeleton to achieve the desired shape deformation. The SMA deformation is related to temperature, ε=α*(TT). A (Austenitic phase transformation region) or ε=β*(TT) M (Martensitic transformation region), where ε is strain, α is the austenitic material constant, β is the martensitic material constant, and T A T is the austenitic phase transformation initiation temperature. M 3) Pneumatic system: used to control the suction force of the suction cup or to assist in the deployment of the sealing material.

[0107] 2.2.4 Motion Control Strategy: Based on the coordinates of the leak point, algorithms such as PRM and RRT are used to plan a collision-free trajectory for the robotic arm from its current position to the leak point. When the sealing material comes into contact with the leak point, the system switches to force control mode. Through force sensor feedback, the robotic arm's pose and the deformation of the SMA frame are adjusted in real time to ensure the sealing force reaches the preset target value F. 目标 This also prevents excessive force from damaging the equipment. Force control law τ = J T (K p (X d -X c )+K f (F d -F c ), where τ is the joint torque, J T K is the transpose of the Jacobian matrix. p and K f For position and force gain, X d and X c F represents the desired position and the current position. d and F c These are expected force and present force.

[0108] 2.3 Blocking Execution Steps

[0109] (1) Path planning and autonomous navigation: The control module sends the calculated optimal path and target coordinates to the sealing execution module. The mobile robot uses its onboard LiDAR, visual odometry, inertial measurement unit and other sensors, combined with a pre-loaded high-precision map of the hydrogen refueling / hydrogen production station, to navigate autonomously, avoid obstacles, and quickly reach the vicinity of the leak point.

[0110] (2) Precise Positioning and Attitude Adjustment: After the mobile robot reaches the target area, the robotic arm extends, and the end effector, equipped with high-precision vision sensors (depth camera, structured light sensor), performs a secondary scan of the leak point to obtain its precise three-dimensional geometric information (hole shape, crack direction). The control module combines the scan data with a preset model to calculate the precise position and attitude of the robotic arm's end effector, ensuring it is perfectly aligned with the leak point. Based on the precise positioning results, the robotic arm adjusts its own attitude and the position and angle of the end effector to perfectly align it with the leak point and place it in the optimal operating position.

[0111] (3) Surface pretreatment: For some sealing materials that require surface cleaning for good adhesion, the robotic arm can switch to a pretreatment tool to lightly clean the area around the leak point and remove loose contaminants. If enhanced adhesion is needed, a non-sparking primer or activator can be applied. (4) Deployment of sealing materials:

[0112] The robotic arm selects and retrieves the calculated sealing material from the internal storage system.

[0113] Patch-type sealing: The end effector precisely attaches a flexible patch or wrapping material to the surface of the leak point and applies a calculated force to press or roll evenly, ensuring that the material is in close contact with the leak surface.

[0114] Injection / spray sealing: The injector / jet injector is aimed at the leak point and, with preset flow rate, pressure and spray pattern, the fast-curing foam, gel or liquid sealant is precisely sprayed / injected into the leak hole or crack.

[0115] Clamping-type sealing: If the leak point is suitable for a mechanical clamp, the robotic arm operates a special clamp with a sealing gasket to apply a preset clamping force to firmly clamp it onto the leak point.

[0116] Inflatable plug / bladder: Insert or attach a plug or uninflated bladder to the leak, and then rapidly inflate it using an internal air pump or chemical reaction until the leak is completely blocked.

[0117] During deployment, force / torque sensors on the robotic arm monitor the applied force in real time to ensure that the sealing force is appropriate, effectively sealing the blockage without damaging the original components.

[0118] 3. Control Module

[0119] 3.1 Information Integration and Risk Assessment of Leaked Data

[0120] The multimodal data (location, leakage amount, shape, environmental parameters) provided by the leak detection module is synchronized in time and registered spatially. Then, based on a multi-factor fuzzy comprehensive evaluation or neural network model, the current leak hazard level is assessed. Input parameter: Leakage flow rate Q leak Leakage point pressure Pleak Leakage area A leak The risk level is determined by factors such as diffusion rate, distance from surrounding combustibles, wind speed / direction, and ambient temperature. Level I - Low risk, Level II - Medium risk, Level III - High risk, and Level IV - Emergency danger.

[0121] 3.2 Intelligent Calculation and Optimization of Blocking Parameters

[0122] 3.2.1 Selection of sealing materials and matching of shape and size

[0123] The system integrates a storage and automatic selection and supply system for various prefabricated sealing materials, ensuring that the required materials can be automatically selected on-site based on calculation results. Materials can be in roll form, boxed form, or liquid / paste containers. Following instructions from the control module, a robotic arm or automatic feeding mechanism selects the specified sealing material from the material bin and delivers it to the end effector. For materials requiring mixing or activation, the device has an automatic mixing or water spray activation function.

[0124] In addition, the system internally includes a database containing parameters such as different SMA skeleton structures, self-expansion characteristics, and mechanical properties. The matching algorithm is based on the geometric parameters of the leak (diameter D). leak Length L leak Average width W leak Direction angle θ leak Area A leak and perimeter P leak (etc.), select the most suitable sealing material template.

[0125] Circular hole: Desired sealing material diameter D seal =k D* D leak ;

[0126] Crack: Expected sealing material length L seal =k L* L leak Width W seal =k W* W leak ;

[0127] Where k D k L k W A redundancy factor (typically 1.1-1.2) is used to ensure edge coverage.

[0128] According to the required D seal L seal W seal The desired deformation of the SMA skeleton is deduced by reverse calculation. The heating parameters are determined by combining the SMA phase transition curve, and then the required heating current, laser power and heating time are calculated.

[0129] 3.2.2 Calculation of Optimal Blocking Force

[0130] Taking into account the pressure at the leak point, the elastic modulus of the sealing material, and the compressive modulus of the sealing material, the target sealing force F is calculated. 目标 F 目标 =P leak* A leak +K 压缩 *A 接触 K 压缩 The compression coefficient of the seal is used to ensure that the seal deformation fills even minor unevenness. A 接触 This represents the contact area of ​​the seal. The target sealing force must overcome the leakage pressure and ensure that the seal can form a reliable mechanical seal after deformation.

[0131] 3.2.3 Path Planning and Obstacle Avoidance

[0132] (1) Real-time 3D Environment Map Construction: A LiDAR (Light Detection and Ranging) sensor is deployed on the mobile robot to perform multi-line scanning of the hydrogen refueling station. Feature extraction (such as planar and edge features) is performed using LiDAR point cloud data. The robot's relative motion is estimated through point cloud registration, and simultaneously, a sparse or dense 3D point cloud map of the environment is constructed. This real-time 3D environment map of the hydrogen refueling station allows for obstacle identification. Point cloud registration (ICP) error function. Where p i It is a point in the source point cloud, q i Let R be the point in the target point cloud, R be the rotation matrix, and t be the translation vector. The goal is to find R and t that minimize the error function.

[0133] (2) Obstacle Recognition and Environment Representation: Deep learning (such as PointNet++, RandLA-Net) is used for semantic segmentation of point cloud data. This not only identifies "obstacles" but also their specific categories, such as "hydrogen storage tanks," "high-pressure pipelines," "hydrogen refueling machines," "personnel," and "ground." This facilitates smarter path planning; for example, it can prioritize paths farther from high-pressure hydrogen equipment or avoid paths near personnel. Simultaneously, the poses of the robot and robotic arm in the environment are represented as a high-dimensional space. The shapes of obstacles in the environmental space are expanded to include the robot's own size and safety margin.

[0134] (3) Collision-free path planning algorithm. The RRT series of algorithms is a sampling-based path planner, particularly suitable for high-dimensional spaces and complex environments. It expands the search tree through random sampling until the tree grows to the target area. The specific steps for planning the optimal collision-free path for the robotic arm from its current position to the leak point based on the RRT algorithm are as follows:

[0135] Initialize a, the space tree T contains the starting point q. 开始

[0136] A random sampling method is used to randomly sample a point q from free space. 随机 ;

[0137] b. Find the nearest neighbor by finding the distance q in the spatial tree T. 随机 The nearest node q 最近 ;

[0138] c from q 随机 To q 最近 Extend the direction by one step (step size Δq) to obtain the new node q. 新 ;

[0139] d-collision detection, q-collision detection 新 Whether it is within the obstacle space, and q 最近 to q 新 Is the path collision-free? If there is no collision, then q... 新 Add it to the spatial tree T and record its parent node as q. 最近 ;

[0140] e RRT optimization, adding q 新 After that, in q 新 Within the neighborhood, a better parent node is found for "backtracking rewiring" and the affected subtrees are "forward rewiring" to optimize path cost.

[0141] f repeats until q 新 close enough to q 目标 ;

[0142] g backtracking: from q 目标 Backtracking along the parent node link to q 开始 The planned path is obtained.

[0143] (4) Path optimization objectives: Optimization objectives include shortest path, shortest time, and lowest energy consumption. Details are as follows:

[0144] Shortest path: Reduces the movement distance of the robotic arm and decreases wear and tear. Path length.

[0145] b. Shortest Time: Improve emergency response speed. This involves kinematic and dynamic constraints. Time Where v s It is the speed along the path;

[0146] c. Minimum power consumption: Extends battery life and reduces operating costs. Power consumption Where τ i It is joint torque.

[0147] qj It is the joint angular velocity.

[0148] 4. Blocking effect monitoring module

[0149] 4.1 Multimodal Real-time Monitoring and Data Fusion

[0150] Hydrogen concentration re-detection: After sealing is completed, a mobile robot / drone immediately performs a high-precision TDLAS scan of the sealed area to obtain the hydrogen concentration distribution after sealing. Effectiveness assessment: Comparing the concentrations before and after sealing, the hydrogen concentration after sealing should rapidly decrease to below the safe threshold and remain stable.

[0151] Leakage acoustic fingerprint analysis: High-sensitivity acoustic sensors continuously monitor the sealing point. Effectiveness assessment: The characteristic high-frequency hissing sound of a leak should disappear or be significantly reduced. Through spectral analysis, if the energy of the characteristic frequency decreases by more than 90%, it is considered effective.

[0152] Pressure / flow sensor feedback (for pipelines): If a blockage is made in a pipeline, the pressure sensor inside the pipeline or the upstream and downstream flow meters can be used to monitor whether the pressure or flow rate has returned to normal.

[0153] Internal sensor feedback from the sealing material: A miniature pressure sensor array and fiber optic strain sensor are embedded within the sealing material. These sensors directly sense the pressure distribution and material deformation at the interface between the sealing material and the leak. Effectiveness assessment: The pressure distribution reported by the sensor array should be uniform and reach a preset value, indicating a tight seal. Strain sensor data indicates that the material is under stable pressure.

[0154] 4.2 Intelligent Evaluation and Adaptive Strategy Adjustment

[0155] (1) Intelligent Assessment: All monitoring data (hydrogen concentration, acoustic characteristics, temperature, pressure, internal stress, etc.) are input into an assessment model based on Bayesian networks or deep reinforcement learning. Bayesian networks are used to establish probabilistic relationships between each monitoring indicator and "blocking effectiveness," handling uncertainty. Deep reinforcement learning is used to try different blocking parameters (actions), observe the blocking effect (reward), and learn the optimal blocking strategy. Finally, the blocking success rate S and remaining leakage risk R are output in real time.

[0156] (2) Adaptive adjustment strategy:

[0157] Successful blocking (S>95%): Maintain the current state and continue monitoring. The system records the blocking parameters as empirical data.

[0158] Partial leakage (70%) <S<95%):

[0159] a. Fine-tuning the sealing force: The robotic arm slightly increases the pressure on the sealing material, and the force sensor provides feedback to ensure that it does not overload.

[0160] b. SMA skeleton fine-tuning: Based on feedback from internal strain sensors, the SMA skeleton is locally fine-tuned and heated to make the sealing material fit more tightly against the uneven parts of the leak.

[0161] c. Secondary curing: Activate the secondary curing mechanism inside the sealing material (e.g., release a small amount of secondary curing agent) to further enhance the sealing performance.

[0162] Blocking failure (S≤70%):

[0163] a. Relocation and Analysis: The leak detection module performs a high-precision scan again to confirm whether the leak point has changed or whether a new leak point exists.

[0164] b. Change the sealing strategy: 1) Try different sealing materials; 2) Adjust the contact angle or direction of the robotic arm to find the best sealing posture; 3) For complex or large leaks, consider multi-robotic arm collaborative sealing; 4) Dispatch multiple sealing robots to seal simultaneously.

[0165] c. High-level alarm and manual intervention: Issue the highest-level alarm to the operator, provide a detailed report on the leak, the attempted plugging schemes and the reasons for their failure, and guide manual handling. At the same time, the system maintains emergency hydrogen venting and shutdown status.

[0166] 4.3 Machine Learning

[0167] (1) Record all data of each sealing process (leakage characteristics, environmental conditions, sealing parameters, monitoring data, forming a huge dataset).

[0168] (2) Use machine learning algorithms (such as deep neural networks and support vector machines) to train these data and establish a mapping relationship between leakage characteristics and optimal plugging parameters.

[0169] (3) By using reinforcement learning to simulate different leakage scenarios, we can continuously optimize the plugging strategy and improve our ability to respond to unknown leakage in the future.

[0170] 5. Energy and Communication Module

[0171] 5.1 Energy Module

[0172] The mobile robot can be powered by a high-capacity lithium battery pack, supporting long-term operation. The robotic arm and sensors can be powered by the mobile robot itself. It can be equipped with wireless charging or automatic charging stations; when the battery level drops below a threshold, the device can autonomously navigate to the charging station for recharging.

[0173] 5.2 Communication Module

[0174] Wireless communication: The device has a built-in high-performance wireless communication module (such as Wi-Fi, 5G, satellite communication) to establish a real-time data link with the central control room of the hydrogen refueling / hydrogen production station, the emergency command center, etc.

[0175] Data transmission: Real-time transmission of sensor data, leak location, leak volume, sealing progress, device status, battery level, and other information.

[0176] Multi-level early warning: Based on the leakage level and sealing effect, different levels of early warning mechanisms are triggered, including audible and visual alarms, SMS notifications, telephone dialing, and email sending.

[0177] Remote control: Allows authorized personnel to remotely monitor the device, issue commands (such as pause, restart, forced evacuation), or adjust parameters through the control room.

[0178] Example:

[0179] Suppose the process for handling the failure of the high-pressure valve seal ring at a hydrogen refueling station.

[0180] Step 1, Detection and Identification: Fixed sensor network alarm. After the mobile robot arrives, TDLAS scanning detects a hydrogen plume at the valve connection, and infrared thermal imaging shows an abnormally low temperature at that location. Stereo vision camera 3D reconstruction combined with a deep learning model identifies a localized failure of the O-ring seal at the valve flange connection, forming a thin, elongated gap with a moderate leakage.

[0181] Step 2, Parameter Calculation: Based on the identified gap length, width, and pipeline pressure, the control module calculates the required size of the annular sealing material. Its SMA skeleton should be deformed into an ellipse to fit the gap shape, and the required sealing pressure is calculated.

[0182] Step 3, Precise Sealing: The robotic arm selects an end effector pre-installed with a ring-shaped sealing head. The SMA skeleton deforms into an elliptical shape under precise heating. Using vision servoing and force control, the robotic arm precisely applies the sealing material to the valve leakage gap and applies a preset pressure. The sealing material rapidly solidifies upon contact with hydrogen gel, filling the gap, while internal pressure sensors indicate uniform contact force distribution.

[0183] Step 4, Effect Monitoring and Adjustment: The robot scans again. TDLAS shows that the hydrogen concentration rapidly drops to the background value, and the leak sound disappears. However, the internal pressure sensor feedback shows that the pressure in the middle of the sealing material is slightly lower than expected. The control module determines this as a "partial leak" and immediately instructs the robotic arm to fine-tune its posture and perform localized secondary heating on the SMA skeleton, causing a slight expansion in the middle area and increasing the overall sealing force. Upon re-monitoring, the internal pressure distribution is uniform, and the system determines that the sealing is successful.

[0184] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. The storage medium can be a magnetic disk, optical disk, read-only memory (ROM), or random access memory (RAM), etc.

[0185] It should be understood that the above detailed description of the technical solutions of the present invention with reference to preferred embodiments is illustrative and not restrictive. Those skilled in the art can modify the technical solutions described in the embodiments or make equivalent substitutions for some of the technical features based on reading this specification; however, these modifications or substitutions do not cause the essence of the corresponding technical solutions to depart from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A smart detection and automatic sealing system for hydrogen leakage in a hydrogen refueling and hydrogen production station, characterized in that: The system includes: The hydrogen leak detection module is deployed near various potential leak points at the hydrogen refueling station to detect whether there is a hydrogen leak at the hydrogen refueling and hydrogen production station; The sealing execution module achieves adaptive deformation and precise fit of the sealing material based on the complex geometric features of the leak point; The control module integrates leakage information and assesses risks, intelligently calculates and optimizes plugging parameters, and calculates the optimal plugging force. The blocking effect monitoring module features multimodal real-time monitoring and data fusion, intelligent evaluation, and adaptive strategy adjustment. The mobile robot is powered by a high-capacity lithium battery pack, and the communication module establishes a real-time data link with the central control room and emergency command center of the hydrogen refueling / production station.

2. The intelligent detection and automatic sealing system for hydrogen leakage in a hydrogen refueling and hydrogen production station according to claim 1, characterized in that: The adaptive blocking material of the blocking execution module adopts a multi-level composite structure, including: The inner layer is in direct contact with the leak: a fast-condensing hydrogen-absorbing gel / polymer that rapidly polymerizes upon contact with hydrogen or a specific catalyst, forming a gel with a certain degree of elasticity and viscosity, and then quickly solidifies. Intermediate layer: Shape memory alloy (SMA) framework network, a precision three-dimensional mesh framework made of nickel-titanium alloy with shape memory effect and superelasticity, achieves precise shape recovery and deformation by precisely controlling electric heating or laser heating to achieve phase transformation between austenite and martensite; Liquid metal micropump, a micro liquid metal pump and channel integrated inside or adjacent to the SMA framework. Outer layer: High-pressure wear-resistant polyurethane / fluororubber, protecting the internal structure; micro magnetic adsorption units, designed for uneven surfaces, with an integrated array of micro magnetic adsorption units at the edges of the sealing material.

3. The intelligent detection and automatic sealing system for hydrogen leakage in a hydrogen refueling and hydrogen production station according to claim 1, characterized in that: The mobile robot is powered by a high-capacity lithium battery pack, supporting long-term operation. The robotic arm and sensors are powered by the mobile robot. It is equipped with wireless charging or automatic charging stations. When the battery level is lower than the threshold, the mobile robot will autonomously navigate to the charging station for charging. Mobile robots have the following devices: a) Tunable diode laser absorption spectroscopy (TDLAS) hydrogen sensor: After the laser beam passes through the hydrogen plume, the hydrogen concentration integral along the path is calculated by scanning the attenuation of the laser absorption intensity, and a three-dimensional hydrogen concentration field is quickly constructed. b) A high-resolution stereo vision camera is used to acquire three-dimensional images of the leak area and capture the geometric information of the leak point; c) Ultrasonic / infrasonic sensors: Hydrogen leakage under high pressure produces unique acoustic characteristics, and multiple acoustic sensors are used for sound source localization. d) Infrared thermal imager: The rapid expansion of hydrogen gas can cause a sudden drop in local temperature. Thermal imaging can help locate the leak point. The communication module has a built-in high-performance wireless communication module, including Wi-Fi, 5G, and satellite communication modules.

4. A method for intelligent detection and automatic sealing of hydrogen leaks in a hydrogen refueling and hydrogen production station, wherein the method is applicable to the system as described in any one of claims 1-3, characterized in that: The method includes: Step 1: Configure and fix the sensors, collect all sensor data, and transmit the data to the CCD module via wireless network for coarse and fine leak location, as well as leak volume estimation and leak vent shape identification. Step 2: Plan the collision-free trajectory of the robotic arm from its current position to the leak point and execute the sealing operation; Step 3: The control module calculates the sealing material parameters, calculates the optimal sealing force, and performs path planning and obstacle avoidance calculations for the sealing robot. Step 4: Monitoring the sealing effect, multimodal real-time monitoring and data fusion, intelligent evaluation of sealing effect and adaptive strategy adjustment, establishment of machine learning model, simulation of different leakage scenarios, continuous optimization of sealing strategy, and improvement of the ability to respond to unknown leakage in the future.

5. The intelligent detection and automatic sealing method for hydrogen leakage in a hydrogen refueling and hydrogen production station according to claim 4, characterized in that: The coarse leak location specifically includes: after receiving an alarm signal, the control module combines the location information and concentration readings of multiple alarm sensors; when the hydrogen concentration detected by any hydrogen sensor exceeds a preset first threshold, the area to which the sensor belongs is marked as a potential leak area; when multiple fixed sensors trigger alarms, the centroid method or weighted average method is used for preliminary area location, to initially delineate the approximate range of the leak, and to quickly estimate the rough location area of ​​the leak point.

6. The intelligent detection and automatic sealing method for hydrogen leakage in a hydrogen refueling and hydrogen production station according to claim 4, characterized in that: The precise leak location involves the control module immediately dispatching a mobile robot equipped with high-precision sensors to the area for detailed scanning after detecting the initial leak area. Construct a three-dimensional hydrogen concentration field C(x,y,z) in the leak area. The leak point is usually located in the region of maximum concentration gradient, since the leak point (x... l ,y l ,z l The maximum value of the three-dimensional gradient vector is consistent with the direction of the leakage source. By calculating the three-dimensional gradient vector ▽C(x,y,z) of the concentration field, the starting point (edge ​​of the high concentration region) and direction of the three-dimensional gradient vector indicate the leakage source. Based on the highest hydrogen concentration, concentration diffusion range, diffusion rate (using time series data) measured near the leak point after precise location, as well as ambient temperature and pressure parameters.

7. The intelligent detection and automatic sealing method for hydrogen leakage in a hydrogen refueling and hydrogen production station according to claim 4, characterized in that: The leakage estimation and leakage outlet shape identification include: (1) Based on the three-dimensional coordinates of the leak point and the hydrogen concentration values ​​detected by multiple sensors, combined with environmental parameters, flow rate inversion was performed using TDLAS. Specifically, the hydrogen concentration distribution C(r,θ,z) at different cross-sections was obtained by scanning the leaking plume using TDLAS; the plume velocity v(r,θ,z) and leakage flow rate Q were obtained with the assistance of laser Doppler velocimetry or particle image velocimetry. leak It is obtained by integrating the mass flux over the cross-section of the plume. Where A is the cross-sectional area of ​​the plume, v z It is the velocity component of the plume along the z-axis (or the main diffusion direction), ρ H2 It is the density of hydrogen gas; (2) The high-resolution stereo vision camera on the mobile robot performs high-density scanning of the leak point in a spiral or zigzag scanning manner in the coarse positioning area to generate high-density point cloud data. After removing noise points from the point cloud, the RANSAC algorithm is used to identify geometric primitives, such as planes, spheres, and cylinders (corresponding to pipes). Then, on the identified equipment surface, the edges of the anomaly—i.e., the outline of the leak—are identified using normal estimation and curvature analysis methods. Circular hole: Fit a circle and extract the center coordinates (x) c ,y c ,z c ) and diameter D leak ; Crack: Fit a straight line segment or a curve segment and extract the length L. leak Average width W leak and direction angle θ leak ; Irregular damage: Calculate its minimum bounding rectangle or convex hull and extract its area A. leak and perimeter P leak .

8. The intelligent detection and automatic sealing method for hydrogen leakage in a hydrogen refueling and hydrogen production station according to claim 4, characterized in that: The steps for performing the blocking are as follows: S201, Path Planning and Autonomous Navigation: The control module sends the calculated optimal path and target coordinates to the sealing execution module; the mobile robot uses its onboard LiDAR, visual odometry, and inertial measurement unit sensors, combined with a pre-loaded high-precision map of the hydrogen refueling / hydrogen production station, to navigate autonomously, avoid obstacles, and quickly reach the vicinity of the leak point. S202. Fine Positioning and Attitude Adjustment: After the mobile robot reaches the target area, the robotic arm extends, and the end effector, equipped with a high-precision vision sensor, performs a secondary scan of the leak point to obtain its precise three-dimensional geometric information (hole shape, crack direction). The control module combines the scan data with a preset model to calculate the precise position and attitude of the robotic arm's end effector, ensuring it is perfectly aligned with the leak point. Based on the fine positioning results, the robotic arm adjusts its own attitude and the position and angle of the end effector to perfectly align it with the leak point and place it in the optimal operating position. S203 Surface Pretreatment: For sealing materials that require surface cleaning for good adhesion, the robotic arm switches to a pretreatment tool to lightly clean the area around the leak point and remove loose contaminants. To enhance adhesion, apply a non-sparking primer or activator. S204, Deployment of sealing materials.

9. The intelligent detection and automatic sealing method for hydrogen leakage in a hydrogen refueling and hydrogen production station according to claim 4, characterized in that: The calculation of the sealing material parameters includes: based on the geometric parameters of the leak opening (diameter D) leak Length L leak Average width W leak Direction angle θ leak Area A leak and perimeter P leak Choose the most suitable sealing material template; Circular hole: Desired sealing material diameter D seal =k D* D leak ; Crack: Expected sealing material length L seal =k L* L leak Width W seal =k W* W leak ; Where k D k L k W A redundancy factor of 1.1-1.2 is used to ensure edge coverage; According to the required D seal L seal W seal The desired deformation of the SMA skeleton is deduced by reverse calculation. The heating parameters are determined by combining the SMA phase transition curve, and then the required heating current, laser power and heating time are calculated.

10. The intelligent detection and automatic sealing method for hydrogen leakage in a hydrogen refueling and hydrogen production station according to claim 4, characterized in that: The calculation of the optimal sealing force includes: considering the pressure at the leak point, the elastic modulus of the material being sealed, and the compressive modulus of the sealing material, to calculate the target sealing force F. 目标 ; F 目标 =P leak* A leak +K 压缩 *A 接触 Where K 压缩 The compression coefficient of the seal is used to ensure that the seal deformation fills even minor unevenness. A 接触 This refers to the contact area of ​​the seal.

Images