A method, program product and device for unmanned aerial vehicle hydrogen storage safety early warning

Abstract

The application provides a kind of unmanned aerial vehicle hydrogen storage safety early warning method, program product, device, in the method, not only based on hydrogen concentration in hydrogen storage equipment carries out early warning, also comprehensively considers the abnormal degree of medium working condition such as hydrogen storage equipment temperature parameter and hydrogen storage equipment pressure parameter of current hydrogen storage equipment, the working condition abnormal situation of hydrogen storage equipment is evaluated comprehensively, and the hydrogen storage working condition comprehensive abnormal degree is obtained.And, in the process of abnormal degree evaluation of hydrogen concentration parameter, hydrogen concentration abnormality identification criterion for evaluating the abnormal degree of hydrogen concentration parameter is also adjusted according to the flight condition, ambient temperature and hydrogen storage equipment pressure in medium working condition, so as to obtain more accurate, comprehensive evaluation of the working condition abnormal situation of hydrogen storage equipment, so that the obtained hydrogen storage working condition comprehensive abnormal degree can better adapt to the dynamic flight condition of unmanned aerial vehicle, so as to carry out effective early warning, and the method does not need to rely on artificial periodic detection, can respond immediately, and can be linked with airborne hydrogen storage system.

Description

Technical Field

[0001] This application relates to the field of hydrogen energy aviation safety technology, and in particular to a method, program product, and device for early warning of hydrogen storage safety in unmanned aerial vehicles. Background Technology

[0002] With the widespread application of hydrogen fuel cells in the field of drones, their advantages of high energy density and rapid hydrogen refueling have shown great application potential in complex operation scenarios such as power line inspection and security patrol. However, under the dynamic flight conditions of drones, hydrogen storage devices face many safety challenges. First, hydrogen is colorless and odorless, with an extremely low density (only 1 / 14 the density of air), making it difficult to quickly identify leaks. Second, various components of hydrogen storage devices are prone to damage under sudden pressure increases or decreases, leading to hydrogen leaks. Third, the ambient temperature range during drone flight is large (-20℃ to 50℃), and the temperature of hydrogen storage tanks, pipelines, and other hydrogen storage devices will also change significantly. Under extreme temperature conditions, damage is also likely to occur, further inducing hydrogen leaks.

[0003] Currently, hydrogen storage safety relies heavily on periodic manual inspections, which results in significant delays in fault identification and prevents real-time perception and timely handling of abnormal conditions. Some technologies issue alarms based on preset fixed hydrogen concentration thresholds, but they ignore the fact that hydrogen risks can change dynamically with the flight conditions of drones. For example, hydrogen storage safety suffers from high false alarm rates and poor adaptability.

[0004] Therefore, the industry urgently needs a hydrogen storage safety early warning solution that can respond to hydrogen storage safety risks in real time, adapt to the dynamic flight conditions of drones, and can work in conjunction with airborne hydrogen storage systems. Summary of the Invention

[0005] This application provides a method, program, and device for early warning of hydrogen storage safety in unmanned aerial vehicles (UAVs). Firstly, a method for early warning of hydrogen storage safety in unmanned aerial vehicles (UAVs) is provided, the method comprising: Acquire multi-sensor data from multiple measurement points collected by a multi-modal sensor array for hydrogen storage devices; The data from multiple sensor points are preprocessed to obtain the hydrogen storage state parameters of the corresponding hydrogen storage device under the current flight state of the UAV. The hydrogen storage state parameters include hydrogen concentration parameters and medium operating condition parameters. The preprocessing includes denoising the data from multiple sensor points according to the denoising processing method corresponding to the parameter change type. The parameter change type is divided according to the time constant corresponding to the hydrogen storage state parameter. Based on the flight parameters of the UAV, the ambient temperature parameters, and the pressure parameters of the hydrogen storage equipment in the medium parameters, the hydrogen concentration anomaly identification benchmark in the hydrogen storage anomaly identification benchmark of the hydrogen storage equipment is adjusted to obtain the adjusted hydrogen storage anomaly identification benchmark. Based on the hydrogen storage status parameters and the adjusted hydrogen storage anomaly identification benchmark, the overall anomaly degree of the hydrogen storage equipment's operating condition is obtained. The appropriate warning type is determined based on the overall degree of abnormality in hydrogen storage conditions, and safety warnings are issued according to the corresponding warning type.

[0006] Understandably, acquiring multi-point sensor data from a multi-modal sensor array for hydrogen storage devices and preprocessing this data involves denoising each sensor data point according to the denoising method corresponding to its respective parameter change type. The parameter change type is categorized based on the time constant corresponding to the hydrogen storage state parameter. This allows for a relatively accurate determination of the hydrogen storage state parameters corresponding to the drone's current flight state, while maintaining high resource utilization efficiency. These parameters include hydrogen concentration and medium operating conditions. Furthermore, when issuing warnings based on these hydrogen storage state parameters, the warnings are not only based on the hydrogen concentration in the storage device but also comprehensively consider the degree of anomaly in the medium operating conditions, such as the storage device's temperature and pressure parameters, to comprehensively assess the anomaly situation and determine the overall degree of anomaly in the hydrogen storage condition. Furthermore, during the assessment of the degree of anomaly in hydrogen concentration parameters, the hydrogen concentration anomaly identification benchmark used to assess the degree of anomaly in hydrogen concentration parameters is adjusted according to flight conditions, ambient temperature, and hydrogen storage equipment pressure in the medium conditions. This results in a more accurate comprehensive assessment of the abnormal conditions of the hydrogen storage equipment, enabling the obtained comprehensive anomaly level of hydrogen storage conditions to better adapt to the dynamic flight conditions of UAVs, thereby providing effective early warning. Moreover, this method does not rely on periodic manual detection, can provide an immediate response, and can be linked and coordinated with the airborne hydrogen storage system.

[0007] In one possible implementation of the first aspect, the medium operating condition parameters also include: hydrogen storage equipment temperature parameters; the hydrogen storage anomaly identification benchmark also includes equipment temperature anomaly identification benchmark and equipment pressure anomaly identification benchmark.

[0008] In one possible implementation of the first aspect, the parameter change type includes a first parameter change type and a second parameter change type, and the first parameter change type corresponds to a first time constant, the second parameter change type corresponds to a second time constant, the value of the first time constant is greater than a first preset time constant boundary value, and the value of the second time constant is less than the first preset time constant boundary value. Furthermore, the multi-point sensor data is preprocessed to obtain the hydrogen storage state parameters of the corresponding hydrogen storage device under the current flight state of the drone, including: The data belonging to the first parameter change type in the multi-point sensor data are denoised by moving average filtering, and the data belonging to the second parameter change type in the multi-point sensor data are denoised by Kalman filtering, thus obtaining the denoised multi-point sensor data. The denoised multi-point sensor data is spatiotemporally aligned and calibrated to obtain calibrated multi-point sensor data. The hydrogen storage state parameter is obtained by fusing sensor data belonging to the same parameter from the calibrated multi-point sensor data.

[0009] In one possible implementation of the first aspect, based on the flight condition parameters of the UAV, the ambient temperature parameter, and the hydrogen storage device pressure parameter among the medium condition parameters, the hydrogen concentration anomaly identification benchmark in the hydrogen storage device's hydrogen storage anomaly identification benchmark is adjusted to obtain the adjusted hydrogen storage anomaly identification benchmark, including: Obtain the flight parameters corresponding to the current flight status of the UAV; Based on the flight condition parameters and the flight condition mapping relationship, the flight condition correction coefficient is obtained. The flight condition mapping relationship is the correspondence between the flight condition parameters and the flight condition correction coefficient. Based on the ambient temperature parameters and the ambient temperature mapping relationship obtained from the flight control system, the ambient temperature correction coefficient is obtained. The ambient temperature mapping relationship includes the correspondence between the ambient temperature parameters and the ambient temperature correction coefficient. Based on the pressure parameters of the hydrogen storage equipment and the pressure mapping relationship, the pressure correction coefficient is obtained. The pressure mapping relationship includes the correspondence between the pressure parameters of the hydrogen storage equipment and the pressure correction coefficient. The hydrogen concentration anomaly identification benchmark is adjusted by using a comprehensive coefficient obtained by multiplying the flight condition correction coefficient, the ambient temperature correction coefficient, and the pressure correction coefficient to obtain the adjusted hydrogen concentration anomaly identification benchmark.

[0010] In one possible implementation of the first aspect, the overall degree of abnormality in the hydrogen storage condition of the hydrogen storage device is obtained based on the hydrogen storage state parameters and the adjusted hydrogen storage anomaly identification benchmark, including: Based on the hydrogen concentration parameter and the adjusted hydrogen concentration anomaly identification benchmark, the hydrogen concentration deviation is calculated. The hydrogen concentration deviation is the percentage of the hydrogen concentration parameter that exceeds the hydrogen concentration anomaly identification benchmark. Based on the temperature parameters of the hydrogen storage equipment and the temperature anomaly identification benchmark, the temperature deviation is calculated. The temperature deviation represents the percentage of the hydrogen storage equipment temperature parameters that exceed the temperature anomaly identification benchmark. Based on the pressure parameters of the hydrogen storage equipment and the equipment pressure anomaly identification benchmark, the pressure deviation is calculated. The pressure deviation represents the percentage of the hydrogen storage equipment pressure parameters that deviate from the equipment pressure anomaly identification benchmark. The overall degree of anomaly in hydrogen storage conditions is determined based on deviations in hydrogen concentration, temperature, and pressure.

[0011] In one possible implementation of the first aspect, the overall degree of anomaly in the hydrogen storage operating conditions is obtained based on the deviations in hydrogen concentration, temperature, and pressure, including: The deviations in hydrogen concentration, temperature, and pressure are weighted and summed according to their respective preset coefficients to obtain the hydrogen storage safety risk score. The hydrogen storage safety risk score represents the comprehensive degree of abnormality in the hydrogen storage operation. Among them, the coefficient corresponding to the deviation in hydrogen concentration is the largest, followed by the coefficient corresponding to the deviation in pressure, and the coefficient corresponding to the deviation in temperature is the smallest.

[0012] In one possible implementation of the first aspect, the type of early warning to be met is determined based on the overall degree of anomaly in the hydrogen storage operating conditions, including: Obtain the scoring range corresponding to each warning type; Determine the scoring range to which the hydrogen storage safety risk score belongs; The corresponding warning type is determined based on the scoring range to which it belongs.

[0013] In one possible implementation of the first aspect, the early warning types include Level 1, Level 2, and Level 3 early warnings. The early warning level is proportional to the overall degree of abnormality in the hydrogen storage operating conditions. Furthermore, safety warnings are issued according to the corresponding early warning type, including: If the corresponding warning type is a Level 1 warning, record the warning log and illuminate the yellow indicator light; When the corresponding warning type is a level 2 warning, an audible and visual alarm is triggered, and the flight control system is used to advise the pilot to adjust the flight attitude or prepare for landing. In the event of a Level 3 warning, the main hydrogen supply valve is first shut off, then the fuel cell is stopped, the safety pressure relief valve is triggered to perform controlled pressure relief to a safe range, and at the same time, the highest priority command is sent to the flight control system to request the immediate execution of the emergency landing procedure.

[0014] Secondly, a drone hydrogen storage safety early warning device is provided, comprising: The multi-parameter sensing module is used to acquire multi-point sensor data collected by the multi-modal sensor array for hydrogen storage devices; The main controller module is used to preprocess multi-point sensor data to obtain the hydrogen storage status parameters of the corresponding hydrogen storage device under the current flight state of the UAV. The hydrogen storage status parameters include hydrogen concentration parameters and medium operating condition parameters. The preprocessing includes denoising the multi-point sensor data according to the denoising processing method corresponding to the parameter change type. The parameter change type is classified according to the time constant corresponding to the hydrogen storage status parameter. Based on the current flight state of the UAV and the medium operating condition parameters of the hydrogen storage device, the hydrogen storage anomaly identification benchmark of the hydrogen storage device is adjusted. Based on the hydrogen storage status parameters and the hydrogen storage anomaly identification benchmark, the comprehensive degree of hydrogen storage anomaly of the hydrogen storage device is obtained. Based on the comprehensive degree of hydrogen storage anomaly, the required warning type is determined. The warning output and execution module is used to issue safety warnings according to the corresponding warning type.

[0015] Thirdly, an electronic device is provided, including a processor and a memory, the memory storing a computer program / instruction, which, when executed by the processor, implements the unmanned aerial vehicle hydrogen storage safety early warning method as described in the first aspect and any of the various possible implementations of the first aspect.

[0016] Fourthly, a computer program product is provided, including a computer program / instruction, which, when executed by a processor, implements the unmanned aerial vehicle hydrogen storage safety early warning method as described in the first aspect and any of the various possible implementations of the first aspect.

[0017] Fifthly, a computer-readable storage medium is provided, on which instructions are stored, which, when executed on a computer, cause the computer to implement the unmanned aerial vehicle hydrogen storage safety early warning method as described in the first aspect and any of the various possible implementations of the first aspect. Attached Figure Description To more clearly illustrate the technical solutions in this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 According to some embodiments of this application, a schematic diagram of a method for early warning of hydrogen storage safety using unmanned aerial vehicles is shown.

[0019] Figure 2 According to some embodiments of this application, a hydrogen concentration threshold graph is shown.

[0020] Figure 3According to some embodiments of this application, a schematic diagram of the process for obtaining the corresponding early warning type under different hydrogen storage safety risk scores is shown.

[0021] Figure 4 According to some embodiments of this application, a schematic diagram of a noise reduction process using Kalman filtering is shown.

[0022] Figure 5 According to some embodiments of this application, a safety early warning device for hydrogen storage in unmanned aerial vehicles is shown. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0024] Therefore, this application proposes a method for early warning of hydrogen storage safety on unmanned aerial vehicles (UAVs). This method involves acquiring multi-point sensor data collected by a multi-modal sensor array from the hydrogen storage device, and preprocessing this data. During preprocessing, the multi-point sensor data is denoised according to the denoising method corresponding to the parameter change type. The parameter change type is classified based on the time constant corresponding to the hydrogen storage state parameter. This allows for a more accurate determination of the hydrogen storage state parameters corresponding to the UAV's current flight state, while maintaining high resource utilization efficiency. The hydrogen storage state parameters include hydrogen concentration parameters and medium operating condition parameters. Furthermore, when issuing early warnings based on the hydrogen storage state parameters, the method considers not only the hydrogen concentration in the hydrogen storage device but also the degree of abnormality in the medium operating conditions, such as the hydrogen storage device's temperature and pressure parameters, to comprehensively assess the abnormality of the hydrogen storage device's operating conditions and obtain the overall degree of abnormality in the hydrogen storage conditions. Furthermore, during the assessment of the degree of anomaly in hydrogen concentration parameters, the hydrogen concentration anomaly identification benchmark used to assess the degree of anomaly in hydrogen concentration parameters is adjusted according to flight conditions, ambient temperature, and hydrogen storage equipment pressure in the medium conditions. This results in a more accurate comprehensive assessment of the abnormal conditions of the hydrogen storage equipment, enabling the obtained comprehensive anomaly level of hydrogen storage conditions to better adapt to the dynamic flight conditions of UAVs, thereby providing effective early warning. Moreover, this method does not rely on periodic manual detection, can provide an immediate response, and can be linked and coordinated with the airborne hydrogen storage system.

[0025] It is understood that the embodiments of this application are particularly applicable to medium and large long-endurance UAVs with a rated power of 3kW, a maximum takeoff weight of 40kg, and an endurance of 10 hours.

[0026] Figure 1 According to some embodiments of this application, a schematic diagram of a method for early warning of hydrogen storage safety using unmanned aerial vehicles (UAVs) is shown. Taking a hydrogen storage system as the implementing entity, the specific steps are as follows: S101, acquire multi-point sensor data collected by the multi-modal sensor array for hydrogen storage devices.

[0027] Understandably, multimodal sensors are deployed at multiple locations, such as the surface of the hydrogen storage tank and pipeline inlets, to measure sensor data such as temperature, pressure, and hydrogen concentration of the hydrogen storage device. The multi-point sensor data can include temperature data, pressure data, and hydrogen concentration data from multiple locations on the corresponding hydrogen storage tank.

[0028] In some embodiments, the multimodal sensor array is distributed. Specifically, an infrared absorption sensor can be used for hydrogen concentration monitoring. A pressure sensor is used for pressure monitoring to obtain pressure data. A temperature sensor is used for temperature monitoring.

[0029] S102, preprocess the multi-point sensor data to obtain the hydrogen storage state parameters of the corresponding hydrogen storage device under the current flight state of the UAV. The hydrogen storage state parameters include hydrogen concentration parameters and medium operating condition parameters. The preprocessing includes denoising the multi-point sensor data according to the denoising processing method corresponding to the parameter change type. The parameter change type is classified based on the time constant corresponding to the hydrogen storage state parameter.

[0030] The parameter change types include a first parameter change type and a second parameter change type. The first parameter change type corresponds to a first time constant, and the second parameter change type corresponds to a second time constant. The value of the first time constant is greater than a first preset time constant boundary value, and the value of the second time constant is less than the first preset time constant boundary value.

[0031] For example, the operating parameters of the medium include: temperature parameters of the hydrogen storage device and pressure parameters of the hydrogen storage device.

[0032] In some embodiments, the data belonging to the first parameter change type in the multi-point sensor data are denoised using a moving average filter, and the data belonging to the second parameter change type in the multi-point sensor data are denoised using a Kalman filter, thereby obtaining denoised multi-point sensor data; the denoised multi-point sensor data are then spatiotemporally aligned and calibrated to obtain calibrated multi-point sensor data; and the sensor data belonging to the same parameter in the calibrated multi-point sensor data are fused to obtain the hydrogen storage state parameter.

[0033] It's understandable that flight vibration noise, electromagnetic interference noise, and environmental noise in drones can affect the data monitored by sensors. It's also understandable that moving average filtering is simple to calculate, requiring only addition and division operations, without complex matrix or floating-point operations; it has low memory usage, requiring only an array of length N to store historical sample values, and N is usually no more than 20, so memory consumption is negligible. However, it has drawbacks such as response delay and insensitivity to abrupt changes. Regarding response delay, when the actual temperature changes abruptly (e.g., from 25°C to 30°C), the moving average needs N sampling points to fully keep up with this change. For N=5, the delay is 0.5 seconds, which is fatal in control systems requiring rapid response; it's also insensitive to abrupt changes, unable to distinguish between "real temperature changes" and "noise-induced temperature changes." For example, if the sensor experiences an abnormal value of 20°C due to strong electromagnetic interference, the moving average will smooth it out (which is good); however, if it's a real temperature change (e.g., a sudden temperature rise due to a fire), it will also smooth it out, leading to alarm delay. Kalman filtering can use state equations for prior estimation, then correct using sensor measurement data, and dynamically allocate the weights of both through Kalman gain to continuously provide a state estimate with minimum uncertainty, thus obtaining the optimal result. However, this Kalman filtering is highly complex and consumes a large amount of memory and computing resources. In this embodiment, parameters with different variation amplitudes are processed using different filtering methods, thereby minimizing the consumption of memory and computing resources while ensuring denoising accuracy.

[0034] In some implementations, the time constant τ of a physical quantity (i.e., the time required for the output to reach 63.2% of its steady-state value after a disturbance) is the fundamental basis for classifying the first parameter change type (i.e., slow variable) and the second parameter change type (i.e., fast variable). For example, the first preset time constant threshold is 1 second (i.e., 1s). When the time constant τ is greater than or equal to 1 second (i.e., 1s), the change amplitude within a unit time (millisecond level) under normal operating conditions is extremely small, and even a brief delay will not affect the safety judgment; this is called a slow variable. When the time constant τ is less than 1s, the change amplitude is considered large.

[0035] For example, slow variables could be temperature. Fast variables could be pressure and hydrogen concentration. Understandably, temperature changes depend on heat conduction / convection processes, resulting in a large time constant (in this system, the hydrogen storage tank is made of metal with a large heat capacity, and the temperature time constant is approximately 10-30 seconds). During normal flight, temperature changes within 0.5 seconds will not exceed 0.5°C, and millisecond-level abrupt changes will not occur. Even in the event of a leak or thermal runaway, temperature anomalies are gradual processes on a minute-by-minute basis, allowing for monitoring delays within 0.5 seconds. The temperature sensor in this system samples at a frequency of 5Hz (once every 200ms), which is inherently a low-rate sampling rate. Pressure is a fluid dynamics parameter with an extremely small time constant (approximately 10-100ms). Valve seal failure, pipeline blockage, or rupture can cause pressure to rise / fall rapidly within milliseconds. Hydrogen diffuses extremely quickly (its diffusion coefficient is four times that of air), and after a leak, the local concentration can reach a dangerous level within one second; therefore, hydrogen concentration is also a fast variable. Anomalies in fast variables require real-time responses on a second- or even millisecond-by-second basis; any delay could lead to a risk of combustion and explosion. In this system, the pressure sensor has a sampling frequency of 100Hz (sampling once every 10ms), and the hydrogen concentration sensor has a sampling frequency of 10Hz (sampling once every 100ms), which is considered high-rate sampling.

[0036] Understandably, after denoising each sensor (moving average filtering for temperature, Kalman filtering for pressure and hydrogen concentration), denoised multi-point sensor data is obtained. At this point, sensor data from different sampling frequencies and physical units are unified in time and space to improve the accuracy and robustness of state estimation. This process will be described in detail below.

[0037] S103, based on the flight parameters of the UAV, the ambient temperature parameters, and the pressure parameter of the hydrogen storage equipment in the medium parameters, adjust the hydrogen concentration anomaly identification benchmark in the hydrogen storage anomaly identification benchmark of the hydrogen storage equipment to obtain the adjusted hydrogen storage anomaly identification benchmark.

[0038] Understandably, the aircraft's flight conditions (e.g., flight speed V, typically between 0-40 m / s) also affect the threshold for identifying abnormal hydrogen concentrations. Specifically, at higher flight speeds, the high-speed airflow disperses leaked hydrogen, thus requiring more sensitive detection, which necessitates lowering the threshold for identifying abnormal hydrogen concentrations. Ambient temperature (T) env Temperatures (generally between -20℃ and 50℃) can affect the physical properties of hydrogen and increase the risk of leakage. Therefore, under extreme ambient temperature conditions, it is necessary to lower the threshold for identifying abnormal hydrogen concentrations. Pressure (P) base (Generally between 90-110 kPa), pressure affects the likelihood of leakage and detection requirements. When pressure is abnormal, the threshold for identifying abnormal hydrogen concentration needs to be lowered. In some embodiments of this application, the ambient temperature can be obtained based on the flight control system.

[0039] In some embodiments, the flight condition parameters corresponding to the current flight state of the UAV are obtained; based on the flight condition parameters and the flight condition mapping relationship, a flight condition correction coefficient is obtained, wherein the flight condition mapping relationship is the correspondence between the flight condition parameters and the flight condition correction coefficient; based on the ambient temperature parameters and the ambient temperature mapping relationship, an ambient temperature correction coefficient is obtained, wherein the ambient temperature mapping relationship includes the correspondence between the ambient temperature parameters and the ambient temperature correction coefficient; based on the hydrogen storage device pressure parameters and the pressure mapping relationship, a pressure correction coefficient is obtained, wherein the pressure mapping relationship includes the correspondence between the hydrogen storage device pressure parameters and the pressure correction coefficient; using the comprehensive coefficient obtained by multiplying the flight condition correction coefficient, the ambient temperature correction coefficient, and the pressure correction coefficient, the hydrogen concentration anomaly identification benchmark is adjusted to obtain the adjusted hydrogen concentration anomaly identification benchmark.

[0040] For example, flight condition parameters may include, but are not limited to, flight speed and flight acceleration. Taking a flight condition parameter that only includes flight speed as an example to illustrate the flight condition mapping relationship: Specifically, for a flight speed of 0-10 m / s, the corresponding flight condition correction factor is 1.15, indicating a 15% relaxation; for a flight speed range of 10 m / s ≤ V < 30 m / s, the flight condition correction factor is 1, indicating no adjustment; for a flight speed range of V ≥ 30 m / s, the flight condition correction factor is 0.75, indicating a 25% reduction. It is understandable that the boundary values ​​for each flight speed range, such as 10 m / s and 30 m / s, can be configured according to the type of UAV; the specific values ​​are not required here.

[0041] Understandably, when the flight speed is 0-10 m / s, it indicates a hovering / low-speed flight condition with stable airflow, where leaked hydrogen easily accumulates, and the threshold can be appropriately relaxed. When the flight speed is 10-30 m / s, it is generally in a cruise state with moderate airflow, and the initial benchmark should be maintained. When the flight speed is 30-40 m / s, it indicates a high-speed flight condition where strong airflow rapidly disperses the hydrogen, and the threshold for identifying abnormal hydrogen concentration must be lowered.

[0042] For example, consider the mapping relationship for ambient temperature. Specifically, when the ambient temperature T < -10℃, the ambient temperature correction factor is 0.85, which means a reduction of 15%. When the ambient temperature -10℃ ≤ T ≤ 30℃, the ambient temperature correction factor is 1.0, which means no adjustment. When the ambient temperature T > 30℃, the ambient temperature correction factor is 0.9, which means a reduction of 10%.

[0043] Understandably, when the ambient temperature T < -10℃ (low temperature), hydrogen molecules move slowly, altering leakage characteristics and necessitating a reduction in the hydrogen concentration anomaly detection threshold. When the ambient temperature T is between -10℃ and 30℃ (close to room temperature), hydrogen characteristics are stable, and no adjustment to the hydrogen concentration anomaly detection threshold is required. When the ambient temperature T > 30℃ (high temperature), the risk of leakage increases, requiring a reduction in the hydrogen concentration anomaly detection threshold from the original level.

[0044] For example, regarding the pressure mapping relationship: Specifically, when the hydrogen storage device pressure P < 95 kPa, the pressure correction factor is 0.95, which means a 5% reduction. When the hydrogen storage device pressure P is in the range of 95 ≤ P ≤ 105, the pressure correction factor is 1, which means no adjustment. When the hydrogen storage device pressure P is in the range of P > 105 kPa, the pressure correction factor is 0.9, which means a 10% reduction.

[0045] For example, a preset hydrogen concentration threshold is used as the benchmark for identifying abnormal hydrogen concentrations, and the adjusted hydrogen concentration threshold is used as the benchmark for identifying abnormal hydrogen concentrations.

[0046] The following formula shows the expression for obtaining the adjusted hydrogen concentration threshold.

[0047] Cthreshold=Cbase˙Kv˙Kt˙Kp Where Cthreshold represents the adjusted hydrogen concentration threshold, and Cbase represents the original hydrogen concentration threshold. Kv represents the flight condition correction factor, Kt represents the ambient temperature correction factor, and Kp represents the pressure correction factor. For example, the original hydrogen concentration threshold is 0.8%.

[0048] The following section will elaborate on this in specific scenarios.

[0049] (1) For high-speed, low-temperature cruise scenarios. Assume the current flight speed V = 35 m / s, the ambient temperature T = -15℃, and the pressure of the hydrogen storage tank P = 105 kPa. Based on the mapping relationships, the correction coefficients are determined as Kv = 0.75, Kt = 0.85, and Kp = 0.9. Substituting these values ​​into the expression for the adjusted hydrogen concentration threshold, i.e., calculating 0.8% × 0.75 × 0.85 × 0.9, we get Cthreshold = 0.46%. The final adjusted hydrogen concentration threshold is reduced from the original hydrogen concentration threshold of 0.8% to 0.46%, a decrease of 0.34%, 0.34% / 0.8% = 42.5%, an overall decrease of 42.5%. It is understandable that high-speed airflow will quickly disperse leaked hydrogen, and low ambient temperature will change the hydrogen diffusion characteristics. Therefore, the system automatically improves the monitoring sensitivity, and can provide early warning even for minor leaks, avoiding missed detection.

[0050] (2) For hovering scenarios at normal temperature. Assume the current flight speed V = 3 m / s, the ambient temperature T = 20℃, and the pressure of the hydrogen storage tank P = 100 kPa. Based on the mapping relationships, the correction coefficients are determined as Kv = 1.15, Kt = 1.0, and Kp = 1.0, resulting in Cthreshold = 0.92%. Understandably, the final hydrogen concentration anomaly detection benchmark is improved from 0.8% to 0.92%, an increase of 0.12%, 0.12% / 0.8% = 15%, an overall improvement of 15%. In hovering, the airflow is stable, hydrogen does not easily diffuse, and the environmental conditions are mild. Therefore, the system appropriately relaxes the threshold to reduce false alarms caused by environmental fluctuations or slight sensor drift.

[0051] (3) High-temperature, medium-speed, normal-pressure scenario. Assume flight speed V = 25 m / s, ambient temperature T = 35℃, and hydrogen storage tank pressure P = 100 kPa. Based on the mapping relationships, the correction coefficients are determined as Kv = 1.0, Kt = 0.9, and Kp = 1.0, resulting in Cthreshold = 0.72%. The final adjusted hydrogen concentration threshold is reduced from the original 0.8% to 0.72%, a decrease of 0.08%, 0.08% / 0.8% = 10%, and an overall decrease of 10%. It is understandable that high-temperature environments increase the risk of hydrogen leakage and material aging. Therefore, the system appropriately increases monitoring sensitivity to identify potential hazards in advance without false alarms.

[0052] Specifically, Figure 2 This diagram shows the hydrogen concentration thresholds adjusted for different values ​​of ambient temperature and flight speed under a constant pressure.

[0053] Understandably, the embodiments of this application can automatically adjust the warning sensitivity based on flight status parameters, ambient temperature, and pressure: in high-speed / low-temperature / high-pressure environments, the threshold is lowered, making the system more sensitive; in ambient temperature hovering, the threshold is raised to reduce false alarms; in high-temperature and medium-speed environments, the threshold is appropriately lowered to balance safety and stability, truly achieving the adaptive safety warning goal of no missed alarms under complex operating conditions and no false alarms under stable operating conditions.

[0054] S104. Based on the hydrogen storage status parameters and the adjusted hydrogen storage anomaly identification benchmark, the comprehensive degree of anomaly in the hydrogen storage condition of the hydrogen storage equipment is obtained.

[0055] In some embodiments, the hydrogen storage anomaly identification benchmarks also include equipment temperature anomaly identification benchmarks and equipment pressure anomaly identification benchmarks. Based on the hydrogen concentration parameter and the adjusted hydrogen concentration anomaly identification benchmark, a hydrogen concentration deviation is calculated, which represents the percentage by which the hydrogen concentration parameter exceeds the hydrogen concentration anomaly identification benchmark. Based on the hydrogen storage equipment temperature parameter and the equipment temperature anomaly identification benchmark, a temperature deviation is calculated, which represents the percentage by which the hydrogen storage equipment temperature parameter exceeds the equipment temperature anomaly identification benchmark. Based on the hydrogen storage equipment pressure parameter and the equipment pressure anomaly identification benchmark, a pressure deviation is calculated, which represents the percentage by which the hydrogen storage equipment pressure parameter deviates from the equipment pressure anomaly identification benchmark. The overall anomaly degree of the hydrogen storage operating condition is obtained based on the hydrogen concentration deviation, temperature deviation, and pressure deviation.

[0056] In some implementation methods, the overall abnormality of hydrogen storage conditions is obtained based on the deviation of hydrogen concentration, temperature, and pressure. This includes: weighting and summing the deviations of hydrogen concentration, temperature, and pressure according to their respective preset coefficients to obtain a hydrogen storage safety risk score. The hydrogen storage safety risk score represents the overall abnormality of hydrogen storage conditions, with the coefficient corresponding to the hydrogen concentration deviation being the largest.

[0057] The following formula shows a calculation formula for a hydrogen storage safety risk score.

[0058] R = a1 × Dc + b1 × Dp + c1 × Dt; Where R represents the hydrogen storage safety risk score, Dc represents the hydrogen concentration deviation, Dp represents the pressure deviation, and Dt represents the temperature deviation. a1 is the weighting coefficient corresponding to the hydrogen concentration deviation, b1 is the weighting coefficient corresponding to the pressure deviation, and c1 is the weighting coefficient corresponding to the temperature deviation. It can be understood that hydrogen concentration is the most direct source of danger, pressure deviation is an indirect hazard indicator, and temperature deviation is an auxiliary judgment indicator. For example, a1 is 60%, b1 is the weighting coefficient corresponding to the hydrogen concentration deviation, b1 is the weighting coefficient corresponding to the pressure deviation, and c1 is 10%. Hydrogen concentration deviation has the highest weight (60%): because exceeding the hydrogen concentration standard is directly related to the risk of combustion and explosion, it is the most direct source of danger and a key indicator for early warning; pressure deviation has the second highest weight (30%): abnormal pressure may cause leakage or container failure, but the change is relatively slow. Temperature deviation has the lowest weight (10%): temperature mainly affects the long-term performance of materials and changes slowly.

[0059] wherein, the hydrogen concentration deviation degree Dc=(Cs-Cthreshold) / Cthreshold×100%, where Cs is the currently obtained hydrogen concentration parameter, that is, the actual current hydrogen concentration value, and Cthreshold is the adjusted hydrogen concentration threshold, that is, the adjusted hydrogen concentration anomaly identification benchmark.

[0060] Pressure deviation degree Dp=|Ps-Po| / Po*100%, where Ps is the currently obtained pressure parameter of the hydrogen storage equipment, that is, the actual current pressure value, and Po is the equipment pressure anomaly identification benchmark, for example, Po is 100.

[0061] Temperature deviation degree Dt=max(0, (Ts-To) / To*100%), where Ts is the currently obtained temperature parameter of the hydrogen storage equipment, that is, the actual current temperature value of the hydrogen storage equipment, and To is the equipment temperature anomaly identification benchmark, for example, To is 40.

[0062] S105, determining the satisfied early warning type according to the comprehensive anomaly degree of the hydrogen storage working condition, and performing safety early warning according to the corresponding early warning type.

[0063] In some embodiments, acquiring the score range corresponding to each early warning type; determining the score range to which the hydrogen storage safety risk score value belongs; obtaining the corresponding early warning type according to the belonging score range. It can be understood that the early warning types may include first-level early warning, second-level early warning, and third-level early warning, and the early warning level is proportional to the comprehensive anomaly degree of the hydrogen storage working condition. When the corresponding early warning type is first-level early warning, an early warning log is recorded and a yellow indicator light is turned on; when the corresponding early warning type is second-level early warning, an audible and visual alarm is triggered, and the flight control system is used to advise the driver to adjust the flight attitude or prepare for landing; when the corresponding early warning type is third-level early warning, the main hydrogen supply valve is closed first, then the fuel cell is stopped, the safety relief valve is triggered to perform controllable pressure relief until the pressure reaches a safe range, and meanwhile, a highest-priority instruction is sent to the flight control system to request immediate execution of the emergency landing procedure.

[0064] for example, reference may be made to Figure 3 , when the hydrogen storage safety risk score value R<R1 (for example, R1 is 30 points), the early warning type is determined to be first-level early warning, indicating a slight anomaly. When the hydrogen storage safety risk score value R1≤R<R2 (for example, R2 is 70 points), the early warning type is determined to be second-level, indicating a moderate anomaly. When the hydrogen storage safety risk score value R2≤R, the early warning type is determined to be third-level, indicating a severe anomaly. Then early warning is executed according to the early warning mode corresponding to the early warning type. In addition, the execution state can also be transmitted back to the main controller, and the early warning type and disposal result can be synchronized to the flight control system and the ground station.

[0065] Understandably, the embodiments of this application innovatively integrate multimodal perception, dynamic adjustment of hydrogen concentration anomaly identification benchmark, and flight control coordination, thereby achieving a leap from passive protection to active intelligent early warning of safety risks of hydrogen storage systems for hydrogen-electric drones, providing systematic protection for the safe operation of core components and the overall flight safety of the aircraft.

[0066] The following is a detailed description of the noise reduction process using moving average filtering in S102 above. As can be understood, the principle of moving average filtering is a simple time-domain smoothing algorithm. Essentially, it replaces the original sampled value at the current moment with the arithmetic mean of the most recent N sampled values, thus offsetting the fluctuations of random noise through "averaging".

[0067] For example, refer to the following moving average filtering formula.

[0068] y[n]=(x[n]+x[n-1]+...+x[n-N+1]) / N; Where N is the sliding window size, y[n] is the filtered output at the nth point; x[·] is the original input discrete signal.

[0069] Application in temperature monitoring: The window size is set to N=5. This means that assuming a system sampling frequency of 10Hz (i.e., collecting temperature data once every 0.1 seconds), the time span corresponding to 5 sampling points is 5 × 0.1s = 0.5s, or 0.5 seconds. The reason for choosing N=5 is that temperature is a slowly changing physical quantity; under normal circumstances, the temperature will not change drastically within 0.5 seconds. A larger window can more effectively compensate for random noise from the sensor (such as ADC quantization noise and instantaneous fluctuations caused by electromagnetic interference). If N is too small (e.g., N=2), the smoothing effect is poor; if N is too large (e.g., N=20), the delay will be unacceptably severe. In practical situations, for example, the original temperature data may fluctuate randomly between 25.0℃ and 25.5℃. After a moving average of N=5, the output will stabilize at around 25.2℃, accurately reflecting the overall trend of a slow rise or fall in temperature.

[0070] The following is combined with Figure 4 The denoising process using Kalman filtering in S102 above will be described in detail. As can be understood, Kalman filtering is a linear optimal estimation algorithm that obtains a more accurate state estimate by fusing sensor measurements and system model predictions. It automatically calculates the confidence levels of these two information sources (through Kalman gain), and then weights and fuses them to obtain the optimal state estimate (i.e., the final denoising result). Simply put: if the sensor is accurate, trust the measured values ​​more; if the system model is accurate, trust the predicted values ​​more.

[0071] Figure 4According to some embodiments of this application, a schematic diagram illustrating a specific process for noise reduction using Kalman filtering is shown. The specific steps are as follows: S401, Initialize state estimation and covariance.

[0072] Initial state estimation This refers to the best guess at the initial state of the hydrogen storage state parameters, such as the first frame of sensor readings.

[0073] Initialize covariance This represents the uncertainty of the initial guess, and is usually set to a large value (such as 10) to indicate that the measurement is trusted more initially.

[0074] S402, Get the current control input.

[0075] Get control input uk: If the system has active control, read the current control input uk. If it is only passive monitoring, then uk = 0.

[0076] Pressure example: uk controls the inflation valve. When inflating, uk is greater than 0, and when deflation, uk is less than 0.

[0077] The following section introduces the prediction phase, namely the time update.

[0078] S403, State Prediction.

[0079] State prediction: Using the system model, the optimal prior estimate from the previous time step is used to predict the state. The prior estimate of the current state is derived from the control input uk. .

[0080] ; Where A is the state transition matrix, with an example of A=1 (i.e., pressure remains constant without intervention), and uk is the control input at the current moment (e.g., valve opening or hydrogen charge / discharge flow rate).

[0081] B is the input matrix, i.e., the control coefficients, representing the state change caused by a unit control input. For example, the prediction of the pressure increase per second in the tank under a unit hydrogen charging flow rate is entirely dependent on the model.

[0082] S404, Covariance Prediction.

[0083] Understandably, the model is not perfect. The process noise covariance matrix, denoted as Q, represents the model's uncertainty (such as pressure fluctuations). Covariance An increase in the value indicates a decrease in the reliability of estimations based solely on the model. The process noise covariance matrix, denoted by Q, is a core parameter in Kalman filtering and state estimation, describing the statistical characteristics of system process noise.

[0084] ; in, This represents the prior covariance at time k. Let A represent the posterior covariance, i.e., the optimal covariance of the previous period, and let A be the state transition matrix. This is the transpose of the state transition matrix.

[0085] S405, Get the current measurement value.

[0086] In some embodiments, a measured value zk is acquired. For example, the current reading of the pressure sensor is read, which is a noise-based estimate of the true pressure.

[0087] The following section describes the update phase, namely measurement update.

[0088] S406, calculate the Kalman gain.

[0089] Kalman gain, which is the core of the algorithm, automatically balances the reliability of model predictions with that of sensor measurements.

[0090] ; in, Let K be the Kalman gain at time k. H represents the observation matrix, i.e., the measurement matrix, describing the mapping relationship between the actual state and the sensor's measured values. For example, in a hydrogen storage pressure scenario: the sensor directly measures the pressure, H=1 (scalar). R represents the measurement noise covariance matrix, i.e., the variance of the random measurement noise introduced by the sensor itself. For example, in a hydrogen storage scenario, this includes pressure sensor circuit jitter, electromagnetic interference, and sampling error. When the measurement noise is very small (small R) → the sensor is very accurate → Kalman gain. Larger values ​​indicate greater confidence in the measured values. When process noise is very small (small Q), the system model is very accurate, leading to higher predicted prior covariance. Small → Kalman gain Smaller size → Greater confidence in model predictions.

[0091] S407, Status Update.

[0092] In some embodiments, the prior estimate is corrected using the measured value zk to obtain the posterior optimal estimate. .

[0093] ; in, This represents the posterior state estimate at time k. This represents the prior estimate of the state at time k.

[0094] For example, the measured pressure from the sensors is fused, the model predictions are corrected, and the optimal pressure estimate for the current moment is output.

[0095] S408, Covariance Update.

[0096] In some embodiments, the uncertainty of the updated estimate is used to output the optimal covariance after filtering in this round, which serves as the input for the next round of prediction, thus preparing for the next iteration.

[0097] ; in, This represents the posterior estimate of the covariance at time k. This represents the prior covariance at time k. Typically, after updating... Will be more A smaller value indicates that the estimate is more reliable after fusion measurement.

[0098] S409 outputs the optimal state estimate. That is, it outputs the posterior state estimate at time k. That is, the data after denoising.

[0099] S410: Determine whether to continue. If yes, proceed to S402; otherwise, proceed to S411 and terminate.

[0100] In some embodiments, the posterior state estimation at the current time This can serve as the basis for control or display, and then proceed to the next moment k+1, repeating the entire process, and proceeding to S402. Upon confirming receipt of the stop command, the process ends and proceeds to S411.

[0101] S411, End.

[0102] Understandably, the entire process embodies the essence of Kalman filtering as an "optimal weighted fusion device": it uses the state equation (model) for prior inference, then uses the measurement equation (sensor) for correction, and dynamically allocates the weights of the two through Kalman gain, thereby continuously providing a state estimate with the minimum uncertainty.

[0103] The following describes in detail the process in S102 of performing spatiotemporal alignment calibration on the denoised multi-point sensor data to obtain calibrated multi-point sensor data; and fusing sensor data belonging to the same parameter in the calibrated multi-point sensor data to obtain the hydrogen storage state parameter, which is the method of fusing multi-source asynchronous data.

[0104] In some embodiments, sensor data from multiple measuring points may collect data for the same parameter. For example, five digital temperature sensors (one each at the top, middle, and bottom, and two symmetrically arranged circumferentially) arranged on the surface of a hydrogen storage tank, due to their different installation positions, are all used to collect sensor data for the corresponding temperature parameter of the hydrogen storage device. The measured values ​​collected by sensors deployed at different locations will differ.

[0105] The equivalent average value can be obtained by weighting the data from multiple sensors corresponding to the same parameter according to the reliability of the measurement points.

[0106] For example, to obtain the equivalent average temperature of a hydrogen storage tank, the following formula shows a calculation expression based on a weighted fusion formula.

[0107] ; in, This represents the equivalent average temperature of the hydrogen storage tank, i.e., the temperature parameter of the hydrogen storage device obtained after fusion, with weights. The sensor's correlation with the reference temperature inside the bottle is determined based on historical data; sensors in the middle and bottom of the can have higher reliability and therefore higher weights (e.g., 0.25), while those in the upper and sides have lower weights (e.g., 0.15).

[0108] Furthermore, due to the different sampling frequencies of each sensor (100Hz for pressure, 10Hz for hydrogen concentration, and 5Hz for temperature), there is a slight delay in the time it takes for data to reach the main controller. This embodiment employs a method combining timestamp alignment and state preservation. Each sensor data point is timestamped with a system timestamp (provided by the RTOS of the early warning controller, with an accuracy of ±50μs) upon acquisition. The main controller runs the fusion algorithm with a basic cycle of 10ms (100Hz). Within each cycle: if a sensor has new data arriving in the most recent cycle, that data is used; if no new data is available (e.g., a new value for the temperature sensor arrives only every 200ms), the valid filtered value from the previous cycle is used. For pressure data (100Hz), there is exactly one new value every 10ms, serving as the core driving signal; hydrogen concentration data arrives every 100ms, and temperature data arrives every 200ms. The fusion algorithm updates asynchronously according to its respective cycle, without waiting for all sensors to synchronize.

[0109] Figure 5 According to some embodiments of this application, a hydrogen storage safety early warning device for unmanned aerial vehicles (UAVs) is shown. It is understood that the UAV hydrogen storage safety early warning device adopts a modular design, mainly composed of the following five modules, interconnected via an internal bus and interacting with other systems of the UAV through a communication interface module. Specifically, the device includes a multi-parameter sensing module, a main controller module, an early warning output and execution module, a communication module, and a power management module.

[0110] A multi-parameter sensing module is used to acquire sensor data from multiple measurement points collected by a multi-modal sensor array for hydrogen storage equipment. In some embodiments of this application, the multi-parameter sensing module is used to acquire hydrogen concentration data using an infrared hydrogen sensor array arranged at key sealing points of the hydrogen storage tank and pipeline flanges to achieve leak location; to monitor the pressure of the hydrogen supply pipeline using a MEMS pressure sensor; and to monitor the tank temperature using a digital temperature sensor. All sensor data is uploaded in CAN message format, with the pressure data sampling frequency at 100Hz and the hydrogen concentration at 10Hz.

[0111] The main controller module is used to preprocess multi-point sensor data to obtain the hydrogen storage status parameters of the corresponding hydrogen storage device under the current flight state of the UAV. The hydrogen storage status parameters include hydrogen concentration parameters and medium operating condition parameters. The preprocessing includes denoising the multi-point sensor data according to the denoising processing method corresponding to the parameter change type. The parameter change type is divided according to the time constant corresponding to the hydrogen storage status parameter. Based on the UAV's flight operating condition parameters, the ambient temperature parameter (the ambient temperature is from the flight control system), and the hydrogen storage device pressure parameter in the medium operating condition parameters, the hydrogen concentration anomaly identification benchmark in the hydrogen storage device's hydrogen storage anomaly identification benchmark is adjusted to obtain the adjusted hydrogen storage anomaly identification benchmark. Based on the hydrogen storage status parameters and the adjusted hydrogen storage anomaly identification benchmark, the comprehensive anomaly degree of the hydrogen storage device's hydrogen storage condition is obtained. The appropriate warning type is determined based on the comprehensive anomaly degree of the hydrogen storage condition. In some embodiments of this application, the main controller module uses a low-power embedded processor to run the dynamic risk assessment algorithm, manage the warning logic, and handle communication with the flight control system. Its built-in algorithm can update the overall anomaly level of hydrogen storage conditions in real time based on the received flight operating parameters (such as speed and altitude).

[0112] The early warning output and execution module is used to issue safety warnings according to the corresponding warning type. In some embodiments of this application, the early warning output and execution module uses multi-color LED indicators and buzzers for local early warning; it integrates a relay drive circuit, which can directly control the emergency shut-off valve and safety pressure relief valve on the hydrogen supply pipeline; and it can be equipped with an optional 4G / 5G wireless module for remotely sending alarm information and key data to the ground station.

[0113] The communication module provides a standard CAN bus interface to ensure seamless connection and data exchange with the UAV flight control computer and fuel cell controller, which is the foundation for achieving flight-safe collaborative control.

[0114] The power management module is responsible for converting the power supply of the UAV platform to the operating voltage required by each module. It is designed with overvoltage and overcurrent protection circuits to ensure stable power supply of the system in complex electromagnetic environments, with the overall power consumption being less than 10W.

[0115] In some embodiments, the UAV hydrogen storage safety early warning device can be deployed as part of the early warning system within the UAV's hydrogen storage system. This early warning system constructs a four-layer closed loop of "perception-analysis-decision-execution". The early warning system is based on a low-power embedded main controller and is deeply integrated with the UAV flight control system and the 3kW proton exchange membrane fuel cell controller via a CAN2.0 (250K) bus.

[0116] This application also provides a computer-readable medium storing instructions that, when executed on a server, cause the server to perform the unmanned aerial vehicle hydrogen storage safety early warning method mentioned in any of the above embodiments of this application.

[0117] This application also provides a computer program product, including: a computer program / instruction, which, when executed by a processor, implements the unmanned aerial vehicle hydrogen storage safety early warning method mentioned in any of the above embodiments of this application.

[0118] It should be noted that all units / modules mentioned in the device embodiments of this application are logical units / modules. Physically, a logical unit / module can be a physical unit / module, a part of a physical unit / module, or a combination of multiple physical units / modules. The physical implementation of these logical units / modules themselves is not the most important factor; the combination of functions implemented by these logical units / modules is the key to solving the technical problems proposed in this application. Furthermore, to highlight the innovative aspects of this application, the above-described device embodiments of this application have not introduced units / modules that are not closely related to solving the technical problems proposed in this application. This does not mean that the above-described device embodiments do not contain other units / modules.

[0119] It should be noted that, in the examples and description of this application, the terms "comprising," "including," or any other variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising one" does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0120] The above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A method for early warning of hydrogen storage safety in unmanned aerial vehicles (UAVs), characterized in that, The method includes: Acquire multi-sensor data from multiple measurement points collected by a multi-modal sensor array for hydrogen storage devices; The multi-sensor data is preprocessed to obtain the hydrogen storage state parameters of the hydrogen storage device under the current flight state of the UAV. The hydrogen storage state parameters include hydrogen concentration parameters and medium operating condition parameters. The preprocessing includes denoising the multi-sensor data according to the denoising processing method corresponding to the parameter change type. The parameter change type is divided according to the time constant corresponding to the hydrogen storage state parameter. Based on the flight parameters of the UAV, the ambient temperature parameters, and the hydrogen storage equipment pressure parameters in the medium parameters, the hydrogen concentration anomaly identification benchmark in the hydrogen storage anomaly identification benchmark of the hydrogen storage equipment is adjusted to obtain the adjusted hydrogen storage anomaly identification benchmark. Based on the hydrogen storage status parameters and the adjusted hydrogen storage anomaly identification criteria, the overall degree of anomaly in the hydrogen storage condition of the hydrogen storage equipment is obtained. The required warning type is determined based on the overall abnormality of the hydrogen storage conditions, and a safety warning is issued according to the corresponding warning type. The medium operating parameters also include: hydrogen storage equipment temperature parameters; the hydrogen storage anomaly identification criteria also include equipment temperature anomaly identification criteria and equipment pressure anomaly identification criteria. Furthermore, the step of adjusting the hydrogen concentration anomaly identification benchmark in the hydrogen storage anomaly identification benchmark of the hydrogen storage device based on the flight condition parameters of the UAV, the ambient temperature parameters, and the hydrogen storage device pressure parameter in the medium condition parameters to obtain the adjusted hydrogen storage anomaly identification benchmark includes: The system acquires flight condition parameters corresponding to the current flight state of the UAV; based on the flight condition parameters and the flight condition mapping relationship, it obtains flight condition correction coefficients, where the flight condition mapping relationship is the correspondence between the flight condition parameters and the flight condition correction coefficients; based on the ambient temperature parameters obtained by the flight control system and the ambient temperature mapping relationship, it obtains ambient temperature correction coefficients, where the ambient temperature mapping relationship includes the correspondence between the ambient temperature parameters and the ambient temperature correction coefficients; based on the pressure parameters of the hydrogen storage device and the pressure mapping relationship, it obtains pressure correction coefficients, where the pressure mapping relationship includes the correspondence between the hydrogen storage device pressure parameters and the pressure correction coefficients; using the comprehensive coefficient obtained by multiplying the flight condition correction coefficients, the ambient temperature correction coefficients, and the pressure correction coefficients, it adjusts the hydrogen concentration anomaly identification benchmark to obtain the adjusted hydrogen concentration anomaly identification benchmark; Furthermore, the step of obtaining the comprehensive degree of abnormality in the hydrogen storage condition of the hydrogen storage device based on the hydrogen storage state parameters and the adjusted hydrogen storage anomaly identification benchmark includes: Based on the hydrogen concentration parameters and the adjusted hydrogen concentration anomaly identification benchmark, a hydrogen concentration deviation is calculated, whereby the hydrogen concentration deviation represents the percentage by which the hydrogen concentration parameter exceeds the hydrogen concentration anomaly identification benchmark. Based on the hydrogen storage equipment temperature parameters and the equipment temperature anomaly identification benchmark, a temperature deviation is calculated, whereby the temperature deviation represents the percentage by which the hydrogen storage equipment temperature parameter exceeds the equipment temperature anomaly identification benchmark. Based on the hydrogen storage equipment pressure parameters and the equipment pressure anomaly identification benchmark, a pressure deviation is calculated, whereby the pressure deviation represents the percentage by which the hydrogen storage equipment pressure parameter deviates from the equipment pressure anomaly identification benchmark. Based on the hydrogen concentration deviation, the temperature deviation, and the pressure deviation, the overall anomaly degree of the hydrogen storage operating condition is obtained. The step of obtaining the overall abnormality level of the hydrogen storage operating condition based on the hydrogen concentration deviation, the temperature deviation, and the pressure deviation includes: The hydrogen concentration deviation, temperature deviation, and pressure deviation are weighted and summed according to their respective preset coefficients to obtain a hydrogen storage safety risk score. The hydrogen storage safety risk score represents the comprehensive abnormality of the hydrogen storage condition. Among them, the coefficient corresponding to the hydrogen concentration deviation is the largest, the coefficient corresponding to the pressure deviation is the second largest, and the coefficient corresponding to the temperature deviation is the smallest.

2. The method according to claim 1, characterized in that, The parameter change type includes a first parameter change type and a second parameter change type. The first parameter change type corresponds to a first time constant, and the second parameter change type corresponds to a second time constant. The value of the first time constant is greater than a first preset time constant boundary value, and the value of the second time constant is less than the first preset time constant boundary value. Furthermore, the preprocessing of the multi-point sensor data to obtain the hydrogen storage state parameters corresponding to the hydrogen storage device under the current flight state of the UAV includes: The data belonging to the first parameter change type in the multi-point sensor data are denoised by moving average filtering, and the data belonging to the second parameter change type in the multi-point sensor data are denoised by Kalman filtering, thereby obtaining the denoised multi-point sensor data. The denoised multi-point sensor data is spatiotemporally aligned and calibrated to obtain the calibrated multi-point sensor data. The hydrogen storage state parameter is obtained by fusing sensor data belonging to the same parameter from the calibrated multi-point sensor data.

3. The method according to claim 1, characterized in that, The determination of the early warning type based on the comprehensive abnormality of the hydrogen storage operating conditions includes: Obtain the scoring range corresponding to each warning type; Determine the scoring range to which the hydrogen storage safety risk score belongs; The corresponding warning type is determined based on the scoring range to which it belongs.

4. The method according to claim 3, characterized in that, The warning types include Level 1, Level 2, and Level 3 warnings. The warning level is directly proportional to the overall abnormality of the hydrogen storage operating conditions. Furthermore, the safety warnings issued according to the corresponding warning type include: If the corresponding warning type is a Level 1 warning, record the warning log and illuminate the yellow indicator light; When the corresponding warning type is a level 2 warning, an audible and visual alarm is triggered, and the flight control system is used to advise the pilot to adjust the flight attitude or prepare for landing. In the event of a Level 3 warning, the main hydrogen supply valve is first shut off, then the fuel cell is stopped, the safety pressure relief valve is triggered to perform controlled pressure relief to a safe range, and at the same time, the highest priority command is sent to the flight control system to request the immediate execution of the emergency landing procedure.

5. A hydrogen storage safety early warning device for unmanned aerial vehicles (UAVs), characterized in that, include: The multi-parameter sensing module is used to acquire multi-point sensor data collected by the multi-modal sensor array for hydrogen storage devices; The main controller module is used to preprocess the data from the multi-measuring point sensors to obtain the hydrogen storage state parameters corresponding to the hydrogen storage device in the current flight state of the UAV. The hydrogen storage state parameters include hydrogen concentration parameters and medium operating condition parameters. The preprocessing includes denoising the multi-point sensor data according to the denoising processing method corresponding to the parameter change type; the parameter change type is divided according to the time constant corresponding to the hydrogen storage state parameter; based on the UAV's flight condition parameters, ambient temperature parameters, and the hydrogen storage equipment pressure parameter in the medium condition parameters, the hydrogen concentration anomaly identification benchmark in the hydrogen storage equipment is adjusted to obtain the adjusted hydrogen storage anomaly identification benchmark; based on the hydrogen storage state parameter and the adjusted hydrogen storage anomaly identification benchmark, the comprehensive degree of anomaly in the hydrogen storage condition of the hydrogen storage equipment is obtained. The type of early warning required is determined based on the overall degree of abnormality in the hydrogen storage operating conditions. The warning output and execution module is used to issue safety warnings according to the corresponding warning type; The medium operating parameters also include: hydrogen storage equipment temperature parameters; the hydrogen storage anomaly identification criteria also include equipment temperature anomaly identification criteria and equipment pressure anomaly identification criteria. Furthermore, the step of adjusting the hydrogen concentration anomaly identification benchmark in the hydrogen storage anomaly identification benchmark of the hydrogen storage device based on the flight condition parameters of the UAV, the ambient temperature parameters, and the hydrogen storage device pressure parameter in the medium condition parameters to obtain the adjusted hydrogen storage anomaly identification benchmark includes: The system acquires flight condition parameters corresponding to the current flight state of the UAV; based on the flight condition parameters and the flight condition mapping relationship, it obtains flight condition correction coefficients, where the flight condition mapping relationship is the correspondence between the flight condition parameters and the flight condition correction coefficients; based on the ambient temperature parameters obtained by the flight control system and the ambient temperature mapping relationship, it obtains ambient temperature correction coefficients, where the ambient temperature mapping relationship includes the correspondence between the ambient temperature parameters and the ambient temperature correction coefficients; based on the pressure parameters of the hydrogen storage device and the pressure mapping relationship, it obtains pressure correction coefficients, where the pressure mapping relationship includes the correspondence between the hydrogen storage device pressure parameters and the pressure correction coefficients; using the comprehensive coefficient obtained by multiplying the flight condition correction coefficients, the ambient temperature correction coefficients, and the pressure correction coefficients, it adjusts the hydrogen concentration anomaly identification benchmark to obtain the adjusted hydrogen concentration anomaly identification benchmark; Furthermore, the step of obtaining the comprehensive degree of abnormality in the hydrogen storage condition of the hydrogen storage device based on the hydrogen storage state parameters and the adjusted hydrogen storage anomaly identification benchmark includes: Based on the hydrogen concentration parameters and the adjusted hydrogen concentration anomaly identification benchmark, a hydrogen concentration deviation is calculated, whereby the hydrogen concentration deviation represents the percentage by which the hydrogen concentration parameter exceeds the hydrogen concentration anomaly identification benchmark. Based on the hydrogen storage equipment temperature parameters and the equipment temperature anomaly identification benchmark, a temperature deviation is calculated, whereby the temperature deviation represents the percentage by which the hydrogen storage equipment temperature parameter exceeds the equipment temperature anomaly identification benchmark. Based on the hydrogen storage equipment pressure parameters and the equipment pressure anomaly identification benchmark, a pressure deviation is calculated, whereby the pressure deviation represents the percentage by which the hydrogen storage equipment pressure parameter deviates from the equipment pressure anomaly identification benchmark. Based on the hydrogen concentration deviation, the temperature deviation, and the pressure deviation, the overall anomaly degree of the hydrogen storage operating condition is obtained. The step of obtaining the overall abnormality level of the hydrogen storage operating condition based on the hydrogen concentration deviation, the temperature deviation, and the pressure deviation includes: The hydrogen concentration deviation, temperature deviation, and pressure deviation are weighted and summed according to their respective preset coefficients to obtain a hydrogen storage safety risk score. The hydrogen storage safety risk score represents the comprehensive abnormality of the hydrogen storage condition. Among them, the coefficient corresponding to the hydrogen concentration deviation is the largest, the coefficient corresponding to the pressure deviation is the second largest, and the coefficient corresponding to the temperature deviation is the smallest.

6. A computer program product comprising a computer program / instructions, wherein when the computer program / instructions are executed by a processor, the method for early warning of hydrogen storage safety in unmanned aerial vehicles as described in any one of claims 1-4 above is implemented.

Images