A fuel cell aircraft motor power limiting method, system and device

Abstract

The present application discloses a method, system and device for limiting the motor power of a fuel cell aircraft, which relates to the field of motor power control. The method includes a hydrogen management system (HMS) and a battery management system (BMS) that calculate and update the hydrogen SOC and the power battery SOC in real time; a dual-channel multi-task Transformer model is established based on the historical data of the hydrogen storage system and the power battery, and combined with the real-time flight conditions, the changing trends of the hydrogen SOC and the power battery SOC are predicted and whether a preset critical threshold is reached in real time, and then the motor is started to be progressively limited in power; according to the flight conditions of the fuel cell aircraft, a comprehensive analysis is performed to determine whether to give priority to reducing the motor torque or reducing the speed; the motor torque or speed is dynamically adjusted based on the hydrogen SOC consumption rate and the power battery SOC consumption rate, and a time smoothing factor is introduced to gradually reduce the torque or speed. The present application implements precise progressive limitation on the motor output power by comprehensively evaluating the hydrogen SOC and the power battery SOC.

Description

Technical Field

[0001] The present application relates to the field of motor power control, and in particular to a method, system and device for limiting the motor power of a fuel cell aircraft. Background Art

[0002] As fuel cell aircraft research and development progresses, the motor, as a core component of the aircraft's propulsion, directly impacts its power performance and energy efficiency. The motor is not only the terminal for fuel cell system energy output but also the key to achieving efficient and stable flight. To support the development of this emerging clean energy aircraft, optimizing fuel cell aircraft motor technology is crucial.

[0003] In the propulsion system of a fuel cell aircraft, fuel cells and power batteries jointly power the motors for efficient energy management and range control. Energy management of fuel cells and power batteries is even more critical in complex flight conditions, such as takeoff, hovering, descent, and taxiing. Fuel cells rely on hydrogen as fuel, but hydrogen reserves are strictly limited by the aircraft's design requirements for payload and volume. Power batteries are also subject to power degradation during frequent charging and discharging. Therefore, when hydrogen reserves are insufficient or the power battery charge is low, regulating the motor's power output to avoid excessive consumption of the fuel cell and power battery is crucial for improving the range and flight safety of fuel cell aircraft.

[0004] The current methods for limiting the motor power of fuel cell aircraft have the following shortcomings:

[0005] (1) Ignoring the hydrogen SOC and power battery SOC leads to unstable motor power output. When the hydrogen SOC or power battery SOC is low, the system cannot gradually reduce the motor power output, which affects the stable performance of the fuel cell aircraft under different operating conditions. (2) In addition, when predicting the hydrogen SOC and power battery SOC, ignoring the charging behavior of the fuel cell to the power battery causes dual SOC prediction deviation, which in turn cannot provide an accurate reference for motor power limit. Summary of the Invention

[0006] The purpose of this application is to provide a fuel cell aircraft motor power limitation method, system and device to improve the accuracy of motor power limitation.

[0007] To achieve the above objectives, this application provides the following solutions:

[0008] In a first aspect, the present application provides a method for limiting motor power of a fuel cell aircraft, comprising:

[0009] In the hydrogen storage system, parameters such as hydrogen pressure and flow are collected and transmitted to the hydrogen management system (HMS). Simultaneously, parameters such as current, voltage, and temperature are collected in the power battery and transmitted to the battery management system (BMS). The HMS and BMS calculate and update the hydrogen SOC and power battery SOC in real time based on this data. Based on this information, the HMS and BMS transmit the latest status information of the hydrogen SOC and power battery SOC to the aircraft control unit respectively.

[0010] A dual-channel, multi-task Transformer model is established based on historical data of the hydrogen storage system and power battery. Combined with real-time flight conditions, the model predicts the changing trends of hydrogen SOC and power battery SOC within the next 5-10 minutes.

[0011] If the hydrogen SOC is within the normal range and the power battery SOC is close to the critical threshold, the fuel cell will be used to power the motor first, and the power battery output will be gradually reduced. At the same time, if the flight stability phase allows, the power battery can be charged through the fuel cell;

[0012] If the power battery SOC is within the normal range and the hydrogen SOC is close to the critical threshold, the power battery will be used to power the motor first, and the fuel cell output will be gradually reduced;

[0013] The aircraft control unit instructs the motor control unit (MCU) to implement progressive power limits on the motors;

[0014] Based on the flight conditions of the fuel cell aircraft, a comprehensive analysis is conducted to determine whether to prioritize reducing motor torque or speed, including:

[0015] If you need to maintain a stable flight speed but the power demand is low, prioritize reducing the motor torque;

[0016] If the overall thrust requirement is low and the impact on flight stability is small, prioritize reducing the motor speed.

[0017] Optionally, it also includes:

[0018] If it is determined that the motor torque should be reduced first, then the hydrogen SOC consumption rate of the hydrogen storage system is obtained. and the consumption rate of the power battery SOC Finally, the motor torque is dynamically adjusted based on the dual SOC consumption rate:

[0019]

[0020] in, It is the basic torque value calculated based on the current load and flight conditions of the fuel cell aircraft; γ is the adjustment coefficient used to control the speed and smoothness of torque changes; and are the real-time consumption rates of hydrogen and power battery respectively; H max rate and E max rate are the maximum allowable consumption rates of hydrogen and power battery power, respectively;

[0021] When the hydrogen SOC or the power battery SOC approaches the critical threshold, a time smoothing factor is introduced to gradually reduce the torque to ensure that the fuel cell aircraft smoothly transitions to the low-energy consumption mode:

[0022]

[0023] Among them, α T To adjust the time smoothing factor of the torque to ensure that the torque adjustment is not too drastic and improve the stability of the fuel cell aircraft flight;

[0024] If it is determined that the motor speed should be reduced first, the hydrogen SOC consumption rate H(t) of the hydrogen storage system and the power battery SOC consumption rate are obtained. Finally, the motor speed is dynamically adjusted based on the dual SOC consumption rate:

[0025]

[0026] in, The motor target speed is set based on the current flight conditions and the driver's needs; δ is the adjustment coefficient used to control the change in the target speed;

[0027] When the hydrogen SOC or power battery SOC approaches the critical threshold, a time smoothing factor is introduced to gradually reduce the speed to ensure that the fuel cell aircraft smoothly transitions to low-energy consumption mode:

[0028]

[0029] Among them, α ω To adjust the time smoothing factor of the speed, ensure that the speed adjustment will not be too drastic, and improve the stability of fuel cell aircraft driving.

[0030] Optionally, it also includes:

[0031] After the MCU implements progressive power limiting on the motor, the motor's real-time output power is fed back to the aircraft control unit through the MCU. After receiving the motor's real-time output power, the aircraft control unit analyzes and evaluates the aircraft's power demand and energy balance based on the current hydrogen SOC and power battery SOC, as well as the specific power requirements of the flight phase. Based on this real-time data, the aircraft control unit dynamically adjusts the motor's output power.

[0032] Optionally, constructing a dual-channel multi-task Transformer model considering the behavior of a fuel cell charging a power battery specifically includes:

[0033] It has a dual-channel design, which is used to predict hydrogen SOC and power battery SOC respectively, and introduces the charging behavior of the fuel cell to the power battery as a dynamic adjustment factor;

[0034] The model input consists of two parts: the hydrogen SOC input channel and the power battery SOC input channel. The hydrogen SOC input channel is used to process characteristic data related to the hydrogen storage system, and the power battery SOC input channel is used to process characteristic data related to the power battery system.

[0035] The data output from the hydrogen SOC input channel and the power battery SOC input channel are input into the charging behavior feature processing module in their respective channels. This module extracts the key features of the time series changes during the charging process through steps such as feature normalization, time series feature extraction, feature interaction fusion, and dynamic weight adjustment, and effectively integrates the relationship between hydrogen consumption and power battery status.

[0036] At the output stage of the charging behavior feature processing module, the position encoding vector is added or concatenated with the charging behavior feature vector to incorporate position information into the charging features. This ensures that the model in the encoder layer can understand the sequential relationship between charging features at different time steps and capture the temporal dynamics of charging behavior.

[0037] The data output by the charging behavior feature processing module is position-encoded, integrated with the timing information, and then input into the encoder layer;

[0038] The final feature vectors output by the encoder layer in the hydrogen SOC channel and the power battery SOC channel are input into the shared layer to achieve information sharing and feature interaction between the hydrogen SOC channel and the power battery SOC channel. The feature vectors of the hydrogen SOC and the power battery SOC are combined using a linear weighted fusion method to form a shared feature representation containing the hydrogen SOC channel and the power battery SOC channel:

[0039]

[0040] Among them, α and β are the weight coefficients of each channel, which are learnable parameters and satisfy α + β = 1;

[0041] The task-specific layer further extracts features highly relevant to the specific task based on the shared features to meet the prediction requirements of each task. In the hydrogen SOC channel, the task-specific layer of this channel will strengthen the focus on features related to hydrogen SOC, while in the power battery SOC channel, the task-specific layer of this channel will strengthen the focus on features related to power battery SOC. The task-specific layer of each channel consists of a fully connected layer, a feedforward neural network, and a convolutional layer.

[0042] The task-specific layer of each channel outputs accurate prediction values ​​of hydrogen SOC and power battery SOC respectively.

[0043] Optionally, the dual-channel multi-task Transformer model introduces the charging behavior of the fuel cell to the power battery as a dynamic adjustment factor, specifically including:

[0044] When the fuel cell charges the power battery, the power battery SOC channel will feed back charging-related information (battery voltage, battery current, battery temperature, charging efficiency, charging time, and power battery SOC) to the hydrogen SOC channel. This feedback indicates that the hydrogen SOC will decrease due to the fuel cell charging the power battery. The information fed back by the power battery to the hydrogen SOC channel during charging can represent the reduction in hydrogen SOC, which can be expressed by the following formula:

[0045]

[0046] Among them, ΔSOC H2 is the change in hydrogen SOC, k is the proportional coefficient used to convert the charging power of the power battery into the consumption of hydrogen SOC, V bat is the voltage of the power battery, I bat is the current of the power battery, η charge is the charging efficiency of the power battery, t charge is the charging time;

[0047] When the hydrogen SOC drops to 30%, the hydrogen SOC channel will feed this information back to the power battery SOC channel to limit the charging behavior of the fuel cell and avoid further consumption of hydrogen. The following formula is used to limit the charging behavior:

[0048] When SOC H2 ≥SOC H2，threshold hour:

[0049]

[0050] When SOC H2 <SOC H2，threshold hour:

[0051]

[0052] in, is the fuel cell charging current after limitation; I FC charging current for the original fuel cell; is the fuel cell charging power after limitation; P FC The original fuel cell charging power; SOC H2 is the current hydrogen SOC; SOC H2，threshold To limit the hydrogen SOC threshold when the fuel cell charges the power battery, it is usually set to 30%; when SOC H2 When it is above the threshold, The charging behavior of the fuel cell should be kept unrestricted, allowing the charging current and power to be maintained at their original values; when SOC H2 When it is less than the threshold, The charging current and power should be gradually reduced to avoid further consumption of hydrogen.

[0053] Optionally, the encoder layer specifically includes:

[0054] After each channel's encoder layer captures changes in hydrogen SOC or power battery SOC through its respective multi-head attention mechanism, it then adds the original features output by the charging behavior feature processing module to the output features of the multi-head attention mechanism to form a residual connection to better preserve the original hydrogen SOC features. The output of each layer is then normalized to ensure a more even data distribution.

[0055] The features extracted by the multi-head attention mechanism are further transformed nonlinearly using a feedforward fully connected layer, enabling the model to learn more complex feature relationships between hydrogen SOC and power battery SOC, selectively retaining and suppressing features. The feedforward fully connected layer is also combined with residual connections and layer normalization to enhance feature expression and maintain model training stability.

[0056] N encoder layers are stacked, and each encoder layer receives and processes the output of the previous layer. Through operations such as multi-head attention mechanism, feedforward fully connected layers, residual connections and layer normalization, the feature information for predicting hydrogen SOC or power battery SOC is gradually extracted and enriched. In the process of stacking layer by layer, the output of each encoder layer will become the input of the encoder of the next layer. After the last encoder layer completes processing, a final feature representation will be obtained. This final feature representation contains the deep features of hydrogen SOC or power battery SOC refined by N layers of encoders.

[0057] Optionally, the multi-head attention mechanism specifically includes:

[0058] The multi-head attention mechanism of each channel performs a linear transformation on the position-encoded feature data;

[0059] In the hydrogen SOC channel, in order to focus on the changes in hydrogen SOC-related features, a hydrogen query vector, a hydrogen bond vector, and a hydrogen value vector are generated, and the features related to hydrogen SOC are weighted separately;

[0060] In the power battery SOC channel, in order to focus on the changes in the power battery SOC-related features, a power battery query vector, a power battery key vector, and a power battery value vector are generated, and the power battery SOC-related features are individually weighted.

[0061] For each channel, the attention score is calculated using the respective query vector and key vector, and the similarity between each time step is obtained by dot product;

[0062] The multi-head attention mechanism divides the input features of hydrogen SOC and power battery SOC into multiple subspaces by parallel computing multiple heads, and processes them from different perspectives. Each head independently calculates the attention output of the query vector, key vector, and value vector, and then splices the outputs of multiple heads together to form the feature aggregation vector of each channel.

[0063] Optionally, the collaborative control system specifically includes: a hydrogen storage system, an HMS, a fuel cell control unit (FCU), a fuel cell, a unidirectional DC / DC, a BMS, a power battery, a bidirectional DC / DC, an inverter, and an aircraft control unit;

[0064] The collaborative control system dynamically adjusts the power output of the motor according to the real-time status changes of hydrogen SOC and power battery SOC through information interaction among control units such as HMS, BMS, FCU, MCU and centralized control of the aircraft control unit.

[0065] In a second aspect, the present application provides a fuel cell aircraft motor power limiting system, comprising:

[0066] In the hydrogen storage system, parameters such as hydrogen pressure and flow are collected and transmitted to the HMS. Simultaneously, parameters such as current, voltage, and temperature are collected in the power battery and transmitted to the BMS. Based on this data, the HMS and BMS calculate and update the hydrogen SOC and power battery SOC in real time. Based on this information, the HMS and BMS transmit the latest status information of the hydrogen SOC and power battery SOC to the aircraft control unit, respectively.

[0067] A dual-channel, multi-task Transformer model is established based on historical data of the hydrogen storage system and power battery. Combined with real-time flight conditions, the model predicts the changing trends of hydrogen SOC and power battery SOC within the next 5-10 minutes.

[0068] If the hydrogen SOC is within the normal range and the power battery SOC is close to the critical threshold, the fuel cell will be used to power the motor first, and the power battery output will be gradually reduced. At the same time, if the flight stability phase allows, the power battery can be charged through the fuel cell;

[0069] If the power battery SOC is within the normal range and the hydrogen SOC is close to the critical threshold, the power battery will be used to power the motor first, and the fuel cell output will be gradually reduced;

[0070] The MCU implements progressive power limiting on the motor;

[0071] Based on the flight conditions of the fuel cell aircraft, a comprehensive analysis is conducted to determine whether to prioritize reducing the motor torque or speed;

[0072] If the flight speed needs to be kept stable but the power demand is low, the motor torque is reduced first. After obtaining the hydrogen SOC consumption rate of the hydrogen storage system and the power battery SOC consumption rate, the motor torque is dynamically adjusted based on the dual SOC consumption rates. When the hydrogen SOC or the power battery SOC approaches the critical threshold, a time smoothing factor is introduced to gradually reduce the torque to ensure that the fuel cell aircraft smoothly transitions to low-energy consumption mode.

[0073] If the overall thrust demand is low and the impact on flight stability is small, the motor speed is reduced first. After obtaining the hydrogen SOC consumption rate of the hydrogen storage system and the power battery SOC consumption rate, the motor speed is dynamically adjusted based on the dual SOC consumption rate. When the hydrogen SOC or the power battery SOC approaches the critical threshold, a time smoothing factor is introduced to gradually reduce the speed to ensure that the fuel cell aircraft smoothly transitions to low-energy consumption mode.

[0074] In a third aspect, the present application provides a computer device comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement any of the above-described fuel cell aircraft motor power limiting methods.

[0075] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-mentioned fuel cell aircraft motor power limiting methods.

[0076] In a fifth aspect, the present application provides a computer program product, comprising a computer program, which, when executed by a processor, implements any one of the above-mentioned fuel cell aircraft motor power limitation methods.

[0077] According to the specific embodiments provided in this application, this application has the following technical effects:

[0078] In the hydrogen storage system, parameters such as hydrogen pressure and flow are collected and the acquired data is transmitted to the HMS. At the same time, parameters such as current, voltage and temperature are collected in the power battery and the acquired data is transmitted to the BMS. The HMS and BMS calculate and update the hydrogen SOC and power battery SOC in real time based on these data. On this basis, the HMS and BMS transmit the latest status information of the hydrogen SOC and power battery SOC to the aircraft control unit respectively. A dual-channel multi-task Transformer model is established based on the historical data of the hydrogen storage system and power battery, and combined with real-time flight conditions, the hydrogen SOC in the next 5-10 minutes is predicted. and the changing trend of the power battery SOC; if the hydrogen SOC is within the normal range and the power battery SOC is close to the critical threshold, the fuel cell is used to power the motor first, and the power battery output is gradually reduced. At the same time, if the flight stability phase allows, the power battery can be charged by the fuel cell; if the power battery SOC is within the normal range and the hydrogen SOC is close to the critical threshold, the power battery is used to power the motor first, and the fuel cell output is gradually reduced; then, the aircraft control unit instructs the MCU to implement progressive power limitation on the motor; according to the flight conditions of the fuel cell aircraft, a comprehensive analysis is performed to determine whether to give priority to reducing the motor torque or reducing the speed. Low, also includes: if it is necessary to maintain a stable flight speed but the power demand is low, the motor torque is reduced first, and after obtaining the hydrogen SOC consumption rate of the hydrogen storage system and the power battery SOC consumption rate, the motor torque is dynamically adjusted based on the dual SOC consumption rate. When the hydrogen SOC or the power battery SOC approaches the critical threshold, a time smoothing factor is introduced to gradually reduce the torque to ensure that the fuel cell aircraft smoothly transitions to the low energy consumption mode; if the overall thrust demand is low and has little impact on flight stability, the motor speed is reduced first, and after obtaining the hydrogen SOC consumption rate of the hydrogen storage system and the power battery SOC consumption rate, the motor torque is adjusted based on the dual SOC consumption rate. The motor speed is dynamically adjusted according to the consumption rate. When the hydrogen SOC or the power battery SOC approaches the critical threshold, a time smoothing factor is introduced to gradually reduce the speed to ensure that the fuel cell aircraft smoothly transitions to the low-energy consumption mode. The method proposed in this application considers whether the hydrogen SOC and the power battery SOC have reached the preset critical threshold, and then gradually limits the motor output power to ensure the stable flight and endurance of the fuel cell aircraft. At the same time, the method takes into account the charging behavior of the fuel cell to the power battery, and can more accurately predict the dynamic change trend of the hydrogen SOC and the power battery SOC, thereby providing an accurate reference for the motor output power limit. BRIEF DESCRIPTION OF THE DRAWINGS

[0079] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0080] Figure 1 This is a flow chart of a method for limiting motor power of a fuel cell aircraft according to the present application;

[0081] Figure 2 This is the architecture diagram of the dual-channel multi-task Transformer model that considers the fuel cell charging behavior of the power battery in this application;

[0082] Figure 3 A diagram of the motor coordinated control system for the fuel cell aircraft in this application taking into account the hydrogen SOC and power battery SOC;

[0083] Figure 4 A schematic diagram of the structure of a computer device provided in one embodiment of the present application. DETAILED DESCRIPTION

[0084] The following will be combined with the drawings in the embodiments of this application to clearly and completely describe the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0085] In order to make the above-mentioned purposes, features and advantages of the present application more obvious and easy to understand, the present application is further described in detail below with reference to the accompanying drawings and specific implementation methods.

[0086] The present application provides a method for limiting the motor power of a fuel cell aircraft. The method accurately predicts the dynamic change trends of the hydrogen SOC and the power battery SOC, and monitors in real time whether they reach the preset critical thresholds, and then gradually adjusts the motor output power, thereby effectively ensuring the stable flight performance and endurance of the fuel cell aircraft.

[0087] In an exemplary embodiment, Figure 1 As shown, a method for limiting the motor power of a fuel cell aircraft is provided, comprising the following steps:

[0088] Step S1: In the hydrogen storage system, parameters such as hydrogen pressure and flow are collected and the acquired data is transmitted to the HMS. At the same time, parameters such as current, voltage and temperature are collected in the power battery and the acquired data is transmitted to the BMS. The HMS and BMS calculate and update the hydrogen SOC and power battery SOC in real time based on these data. On this basis, the HMS and BMS respectively transmit the latest status information of the hydrogen SOC and power battery SOC to the aircraft control unit.

[0089] Step S2: Based on the historical data of the hydrogen storage system and power battery, a dual-channel multi-task Transformer model is established. Combined with the real-time flight conditions, the changing trends of hydrogen SOC and power battery SOC in the next 5-10 minutes are predicted.

[0090] As an optional implementation, step S2 specifically includes:

[0091] Step S21: If the hydrogen SOC is within the normal range and the power battery SOC is close to the critical threshold (the power battery SOC is less than 25%), the fuel cell is used to power the motor first, and the power battery output is gradually reduced. At the same time, if the flight stability phase allows, the power battery can be charged through the fuel cell;

[0092] Step S22: If the power battery SOC is within the normal range and the hydrogen SOC is close to the critical threshold (hydrogen SOC is lower than 25%), the power battery is used to power the motor first, and the fuel cell output is gradually reduced.

[0093] Step S3: The aircraft control unit instructs the MCU to start implementing progressive power limitation on the motor.

[0094] Step S4: Based on the flight conditions of the fuel cell aircraft, a comprehensive analysis is performed to determine whether to prioritize reducing the motor torque or reducing the speed, so as to achieve a dynamic adjustment strategy with optimal energy efficiency.

[0095] As an optional implementation, step S4 specifically includes:

[0096] Step S41: If the flight speed needs to be kept stable but the power demand is low, the motor torque is reduced first; the hydrogen SOC consumption rate of the hydrogen storage system is obtained. and the consumption rate of the power battery SOC Finally, the motor torque is dynamically adjusted based on the dual SOC consumption rate:

[0097]

[0098] in, It is the basic torque value calculated based on the current load and flight conditions of the fuel cell aircraft; γ is the adjustment coefficient used to control the speed and smoothness of torque changes; and are the real-time consumption rates of hydrogen and power battery respectively; H max rate and E max rate are the maximum allowable consumption rates of hydrogen and power battery power, respectively;

[0099] When the hydrogen SOC or the power battery SOC approaches the critical threshold, a time smoothing factor is introduced to gradually reduce the torque to ensure that the fuel cell aircraft smoothly transitions to the low-energy consumption mode:

[0100]

[0101] Among them, α T To adjust the time smoothing factor of the torque to ensure that the torque adjustment is not too drastic and improve the stability of the fuel cell aircraft flight;

[0102] Step S42: If the overall thrust requirement is low and the impact on flight stability is small, the motor speed is reduced first; the hydrogen SOC consumption rate of the hydrogen storage system is obtained. and the consumption rate of the power battery SOC Finally, the motor speed is dynamically adjusted based on the dual SOC consumption rate:

[0103]

[0104] in, The motor target speed is set based on the current flight conditions and the driver's needs; δ is the adjustment coefficient used to control the change in the target speed;

[0105] When the hydrogen SOC or power battery SOC approaches the critical threshold, a time smoothing factor is introduced to gradually reduce the speed to ensure that the fuel cell aircraft smoothly transitions to low-energy consumption mode:

[0106]

[0107] Among them, α ω To adjust the time smoothing factor of the speed, ensure that the speed adjustment will not be too drastic, and improve the stability of fuel cell aircraft driving.

[0108] Step S5: After the MCU implements progressive power limiting on the motor, the motor's real-time output power is fed back to the aircraft control unit through the MCU. After receiving the motor's real-time output power, the aircraft control unit analyzes and evaluates the aircraft's power demand and energy balance based on the current hydrogen SOC and power battery SOC, as well as the specific power requirements of the flight phase. Based on this real-time data, the aircraft control unit dynamically adjusts the motor's output power, thereby ensuring stable flight of the fuel cell aircraft while maximizing its endurance.

[0109] As an optional implementation, Figure 2 As shown, in step S2, a dual-channel multi-task Transformer model is constructed that considers the charging behavior of the fuel cell to the power battery. The model is used to accurately predict whether the hydrogen SOC and the power battery SOC have reached a critical threshold (the hydrogen SOC or the power battery SOC is less than 25%), specifically including:

[0110] In the first step, a dual-channel design is proposed, which is used to predict hydrogen SOC and power battery SOC respectively, and the charging behavior of the fuel cell to the power battery is introduced as a dynamic adjustment factor.

[0111] The input of the model consists of two parts, the hydrogen SOC input channel and the power battery SOC input channel; the hydrogen SOC input channel is used to process the characteristic data related to the hydrogen storage system, such as the hydrogen pressure P of the hydrogen storage system. H2 , hydrogen flow Q H2 and hydrogen temperature T H2 , the output power P of the fuel cell FC and charging current I FC , fuel cell aircraft speed v aircraft The power battery SOC input channel is used to process characteristic data related to the power battery system, such as the power battery temperature T bat , voltage V bat , current I bat , charging time t charge and charging efficiency η charge , fuel cell aircraft load L aircraft .

[0112] In practical applications, there is charging behavior feedback between the two channels. When the fuel cell charges the power battery, the power battery SOC channel will feed back charging-related information (battery voltage, battery current, battery temperature, charging efficiency, charging time, and power battery SOC) to the hydrogen SOC channel. This feedback indicates that the hydrogen SOC will decrease as the fuel cell charges the power battery. The information fed back to the hydrogen SOC channel during charging can be expressed as the amount of hydrogen SOC reduction, which can be expressed as the following formula:

[0113]

[0114] Among them, ΔSOC H2 is the change in hydrogen SOC, k is the proportional coefficient used to convert the charging power of the power battery into the consumption of hydrogen SOC, V bat is the voltage of the power battery, I bat is the current of the power battery, η charge is the charging efficiency of the power battery, t charge For charging time.

[0115] At the same time, when the hydrogen SOC drops to 30%, the hydrogen SOC channel will feed this information back to the power battery SOC channel to limit the charging behavior of the fuel cell and avoid further consumption of hydrogen. The following formula is used to limit the charging behavior:

[0116] When SOC H2 ≥SOC H2，threshold hour:

[0117]

[0118] When SOC H2 <SOC H2，threshold hour:

[0119]

[0120] in, is the fuel cell charging current after limitation; I FC charging current for the original fuel cell; is the fuel cell charging power after limitation; P FC The original fuel cell charging power; SOC H2 is the current hydrogen SOC; SOC H2，threshold To limit the hydrogen SOC threshold when the fuel cell charges the power battery, it is usually set to 30%; when SOC H2 When it is above the threshold, The charging behavior of the fuel cell should be kept unrestricted, allowing the charging current and power to be maintained at their original values; when SOC H2 When it is less than the threshold, The charging current and power should be gradually reduced to avoid further consumption of hydrogen.

[0121] In the second step, the data output by the hydrogen SOC input channel and the power battery SOC input channel are input into the charging behavior feature processing modules in their respective channels.

[0122] This module extracts the key features of timing changes during the charging process through steps such as feature normalization, time series feature extraction, feature interaction fusion and dynamic weight adjustment, and effectively integrates the relationship between hydrogen consumption and power battery status.

[0123] In the third step, at the output stage of the charging behavior feature processing module, the position encoding vector is generated using the following calculation formula:

[0124]

[0125] Where pos represents the position of the current feature vector in the sequence (for example, the posth time step); i represents the dimension index in the feature vector; d represents the total dimension of the feature vector; and f(λ) is a mapping function used to adjust the amplitude and frequency of the position encoding according to the characteristic parameters in the hydrogen SOC input channel or the power battery SOC input channel.

[0126] The position encoding vector is added or concatenated with the charging behavior feature vector to incorporate the position information into the charging features, ensuring that the model in the encoder layer can understand the sequential relationship between the charging features at different time steps and capture the temporal dynamics of the charging behavior.

[0127] In the fourth step, the data output by the charging behavior feature processing module is position-encoded, integrated with the timing information, and then input into the multi-head attention mechanism of the encoder layer.

[0128] As an optional implementation, the multi-head attention mechanism specifically includes:

[0129] (1) The multi-head attention mechanism of each channel performs a linear transformation on the position-encoded feature data. In the hydrogen SOC channel, in order to focus on the changes in hydrogen SOC-related features, a hydrogen query vector Q is generated. H2 , hydrogen bond vector K H2 and hydrogen value vector V H2 , the features related to hydrogen SOC are weighted separately, thereby enhancing the model's attention to changes in hydrogen SOC. The calculation formula is as follows:

[0130]

[0131] Among them, X H2 is the position-encoded input feature vector of the hydrogen SOC channel; and They are the query, key, and value weight matrices specifically for the hydrogen SOC channel, ensuring that the generation of query, key, and value vectors can reflect the weights of hydrogen SOC-related features.

[0132] In the power battery SOC channel, in order to focus on the changes in the power battery SOC related characteristics, a power battery query vector Q is generated. bat , power battery key vector K bat and the power battery value vector V bat , and weight the features related to the power battery SOC separately. The calculation formula is as follows:

[0133]

[0134] Among them, X bat It is the position-encoded input feature vector of the power battery SOC channel; and They are the weight matrices of query, key, and value specifically for the power battery SOC channel, ensuring that the generation of query, key, and value vectors can reflect the weights of the power battery SOC-related features.

[0135] (2) For each channel, the attention score is calculated using the respective query vector and key vector, and the similarity between each time step is obtained by dot product. The calculation formulas for the attention calculation scores of the hydrogen SOC channel and the power battery SOC channel are:

[0136]

[0137] Among them, d k is the dimension of the key vector, which is used to scale to prevent the value from being too large. The softmax function is used to normalize the score to a probability distribution.

[0138] (3) The multi-head attention mechanism divides the input features of hydrogen SOC and power battery SOC into multiple subspaces by parallel calculation of multiple heads, and processes them with different attention angles.

[0139] Each head independently calculates the attention output of the query vector, key vector, and value vector, and then concatenates the outputs of multiple heads to form the feature aggregation vector of each channel, allowing the model to capture changes in hydrogen SOC or power battery SOC from multiple angles. The calculation formula for each channel is as follows:

[0140]

[0141] in, and It is the output weight matrix, ensuring that each channel in the multi-head attention output focuses on the feature changes of its own SOC.

[0142] In the fifth step, after each channel's encoder layer is processed by its own multi-head attention mechanism, the original features output by the charging behavior feature processing module are added to the output features of the multi-head attention mechanism to form a residual connection to better preserve the original hydrogen SOC features. The output of each layer is then normalized to make the data distribution more even.

[0143] The features extracted by the multi-head attention mechanism are further transformed nonlinearly using a feedforward fully connected layer, enabling the model to learn more complex feature relationships between hydrogen SOC and power battery SOC, selectively retaining and suppressing features. The feedforward fully connected layer is also combined with residual connections and layer normalization to enhance feature expression and maintain model training stability.

[0144] N encoder layers are stacked, with each encoder receiving and processing the output of the previous layer. Through operations such as a multi-head attention mechanism, feedforward fully connected layers, residual connections, and layer normalization, the feature information for predicting hydrogen SOC or power battery SOC is gradually extracted and enriched. In hydrogen SOC and power battery SOC prediction, the value of N can be adjusted based on the complexity of the task and the amount of data. During the layer-by-layer stacking process, the output of each encoder layer becomes the input of the next encoder layer. After the final encoder layer completes processing, a final feature representation is obtained. This final feature representation contains the deep features of the hydrogen SOC or power battery SOC refined by the N encoder layers.

[0145] In the sixth step, the final feature vectors output by the encoder layer in the hydrogen SOC channel and the power battery SOC channel are input into the shared layer to achieve information sharing and feature interaction between the hydrogen SOC channel and the power battery SOC channel.

[0146] The characteristic representation of the hydrogen SOC channel can be written as:

[0147]

[0148] The characteristic representation of the power battery SOC channel can be written as:

[0149] H bat =[T bat ,V bat ,I bat ,t charge ,η charge ,L aircraft ]

[0150] The feature vectors of hydrogen SOC and power battery SOC are combined using a linear weighted fusion method to form a shared feature representation containing the hydrogen SOC channel and the power battery SOC channel:

[0151]

[0152] Among them, α and β are the weight coefficients of each channel, which are learnable parameters and satisfy α+β=1.

[0153] In the seventh step, the features output by the shared layer are a mixture of common information of the two tasks of hydrogen SOC prediction and power battery SOC prediction. This information is helpful for both tasks, but cannot meet the specific needs of each task. The task-specific layer further extracts features that are highly relevant to the specific task based on the shared features to meet the prediction needs of each task.

[0154] In the hydrogen SOC channel, the channel's task-specific layer strengthens its focus on features related to hydrogen SOC. In the power battery SOC channel, the channel's task-specific layer strengthens its focus on features related to power battery SOC. Each channel's task-specific layer consists of fully connected layers, feedforward neural networks, and convolutional layers. These layers deeply process, filter, and transform input features to generate more targeted feature representations, thereby improving prediction accuracy.

[0155] In the eighth step, the task-specific layer outputs the accurate prediction values ​​of hydrogen SOC and power battery SOC respectively.

[0156] As an optional implementation, Figure 3 As shown in the figure, the motor coordinated control system considering hydrogen SOC and power battery SOC includes:

[0157] This coordinated control system involves core components such as the hydrogen storage system, HMS, FCU, fuel cell, unidirectional DC / DC converter, BMS, power battery, bidirectional DC / DC converter, and motor. Through the coordinated scheduling of these controllers and the aircraft control unit, precise control of motor power output is achieved. First, under the control of the HMS, the hydrogen storage system continuously supplies hydrogen to the fuel cell. The fuel cell output voltage is regulated by the unidirectional DC / DC converter to adapt to the needs of the motor or power battery. The HMS monitors key parameters of the hydrogen storage system in real time, including hydrogen SOC, hydrogen flow rate, and hydrogen temperature, to ensure fuel cell operation stability. When the hydrogen SOC approaches a set threshold, the HMS transmits this information to the aircraft control unit via the CAN bus. Upon receiving this information, the aircraft control unit sends control commands to the FCU to appropriately adjust the hydrogen flow rate and fuel cell reaction efficiency, thereby limiting the fuel cell output power. This limiting reduces hydrogen consumption to a certain extent, thereby extending flight range, while also preventing system overload caused by excessive fuel cell output power. The aircraft control unit, based on the information transmitted by the HMS and the current state of the fuel cell, instructs the MCU to gradually limit the power of the motors. This progressive power limiting mechanism not only helps manage energy consumption but also gradually reduces the motor power output without significantly affecting the flight experience, thus ensuring basic flight needs when the hydrogen SOC is low.

[0158] The power battery's voltage, current, temperature and other parameters are monitored in real time through the BMS. The BMS controls the operating mode of the bidirectional DC / DC converter based on the power battery SOC status. When the hydrogen SOC is sufficient and the fuel cell output power is higher than the current motor drive demand, the bidirectional DC / DC allows excess current to be transferred from the fuel cell to the power battery for charging; when the hydrogen SOC is low and the fuel cell power supply is insufficient, the bidirectional DC / DC reversely supplies power to supplement the motor's power demand from the power battery. During this process, the BMS is not only responsible for maintaining the health of the power battery, ensuring its safety and stability during discharge and charging, but also provides the necessary data support to cooperate with the power limitation strategy of the aircraft control unit.

[0159] This collaborative control system dynamically adjusts motor power output through information exchange among control units such as the HMS, BMS, FCU, and MCU, and centralized control by the aircraft control unit. Based on real-time changes in the hydrogen SOC and power battery SOC, this system flexibly manages the power output strategies of the fuel cell and power battery, ensuring flight stability and optimizing endurance performance of the fuel cell aircraft under various operating conditions.

[0160] Compared with the traditional motor output power limitation method, the fuel cell aircraft motor power limitation method of the present application can consider whether the hydrogen SOC and the power battery SOC have reached the preset critical threshold value, and then gradually limit the motor output power to ensure the stable flight and endurance of the fuel cell aircraft.

[0161] The fuel cell aircraft motor power limitation method of the present application takes into account the charging behavior of the fuel cell to the power battery when predicting the hydrogen SOC and the power battery SOC, and can more accurately predict the dynamic change trends of the hydrogen SOC and the power battery SOC, thereby providing an accurate reference for the motor output power limitation.

[0162] The fuel cell aircraft motor power limiting method of the present application can analyze and prioritize whether to reduce the motor torque or reduce the speed based on the flight conditions of the fuel cell aircraft, thereby ensuring a dynamic adjustment strategy for the motor to achieve optimal energy efficiency.

[0163] In an exemplary embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor implements the above-mentioned fuel cell aircraft motor power limitation method when executing the computer program.

[0164] In an exemplary embodiment, a computer-readable storage medium is provided, storing a computer program. When the computer program is executed by a processor, the method for limiting the motor power of a fuel cell aircraft is implemented.

[0165] In an exemplary embodiment, a computer program product is provided, comprising a computer program. When the computer program is executed by a processor, the computer program implements the above-mentioned fuel cell aircraft motor power limitation method.

[0166] In an exemplary embodiment, a computer device is provided. The computer device may be a server or a terminal. The internal structure diagram thereof may be as follows: Figure 4As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, abbreviated as I / O) and a communication interface. The processor, memory and input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a fuel cell aircraft motor power limiting method is implemented.

[0167] Those skilled in the art will understand that Figure 4 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0168] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0169] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, database or other media used in the embodiments provided in this application may include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory may include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM may be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0170] The databases involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, distributed databases based on blockchains. The processors involved in the various embodiments provided herein may include, but are not limited to, general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic units, data processing logic units based on quantum computing, and the like.

[0171] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0172] This document uses specific examples to illustrate the principles and implementation methods of this application. The description of the above examples is only intended to help understand the method and core concept of this application. At the same time, for those skilled in the art, based on the concept of this application, there may be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting this application.

Claims

1. A method for limiting the motor power of a fuel cell aircraft, characterized in that: include: In the hydrogen storage system, parameters such as hydrogen pressure and flow are collected and transmitted to the hydrogen management system (HMS). Simultaneously, parameters such as current, voltage, and temperature are collected in the power battery and transmitted to the battery management system (BMS). The HMS and BMS calculate and update the hydrogen SOC and power battery SOC in real time based on this data. Based on this information, the HMS and BMS transmit the latest status information of the hydrogen SOC and power battery SOC to the aircraft control unit respectively. A dual-channel, multi-task Transformer model is established based on historical data of the hydrogen storage system and power battery. Combined with real-time flight conditions, the model predicts the changing trends of hydrogen SOC and power battery SOC within the next 5-10 minutes. If the hydrogen SOC is within the normal range and the power battery SOC is close to the critical threshold, the fuel cell will be used to power the motor first, and the power battery output will be gradually reduced. At the same time, if the flight stability phase allows, the power battery can be charged through the fuel cell; If the power battery SOC is within the normal range and the hydrogen SOC is close to the critical threshold, the power battery will be used to power the motor first, and the fuel cell output will be gradually reduced; The aircraft control unit instructs the motor control unit (MCU) to implement progressive power limits on the motors; Based on the flight conditions of the fuel cell aircraft, a comprehensive analysis is conducted to determine whether to prioritize reducing motor torque or speed, including: If you need to maintain a stable flight speed but the power demand is low, prioritize reducing the motor torque; If the overall thrust requirement is low and the impact on flight stability is small, prioritize reducing the motor speed.

2. The fuel cell aircraft motor power limitation method according to claim 1, characterized in that: Also includes: If it is determined that the motor torque should be reduced first, then the hydrogen SOC consumption rate of the hydrogen storage system is obtained. and the consumption rate of the power battery SOC Finally, the motor torque is dynamically adjusted based on the dual SOC consumption rate: in, It is the basic torque value calculated based on the current load and flight conditions of the fuel cell aircraft; γ is the adjustment coefficient used to control the speed and smoothness of torque changes; and are the real-time consumption rates of hydrogen and power battery respectively; H maxrate and E maxrate are the maximum allowable consumption rates of hydrogen and power battery power, respectively; When the hydrogen SOC or the power battery SOC approaches the critical threshold, a time smoothing factor is introduced to gradually reduce the torque to ensure that the fuel cell aircraft smoothly transitions to the low-energy consumption mode: Among them, α T To adjust the time smoothing factor of the torque to ensure that the torque adjustment is not too drastic and improve the stability of the fuel cell aircraft flight; If it is determined that the motor speed should be reduced first, then the hydrogen SOC consumption rate of the hydrogen storage system is obtained. and the consumption rate of the power battery SOC Finally, the motor speed is dynamically adjusted based on the dual SOC consumption rate: in, The motor target speed is set based on the current flight conditions and the driver's needs; δ is the adjustment coefficient used to control the change in the target speed; When the hydrogen SOC or power battery SOC approaches the critical threshold, a time smoothing factor is introduced to gradually reduce the speed to ensure that the fuel cell aircraft smoothly transitions to low-energy consumption mode: Among them, α ω To adjust the time smoothing factor of the speed, ensure that the speed adjustment will not be too drastic, and improve the stability of fuel cell aircraft driving.

3. The fuel cell aircraft motor power limitation method according to claim 1, characterized in that: Also includes: After the MCU implements progressive power limiting on the motor, the motor's real-time output power is fed back to the aircraft control unit through the MCU. After receiving the motor's real-time output power, the aircraft control unit analyzes and evaluates the aircraft's power demand and energy balance based on the current hydrogen SOC and power battery SOC, as well as the specific power requirements of the flight phase. Based on this real-time data, the aircraft control unit dynamically adjusts the motor's output power.

4. The fuel cell aircraft motor power limitation method according to claim 1, characterized in that: Construct a dual-channel multi-task Transformer model that considers the fuel cell charging behavior of the power battery, including: It has a dual-channel design, which is used to predict hydrogen SOC and power battery SOC respectively, and introduces the charging behavior of the fuel cell to the power battery as a dynamic adjustment factor; The model input consists of two parts: the hydrogen SOC input channel and the power battery SOC input channel. The hydrogen SOC input channel is used to process characteristic data related to the hydrogen storage system, and the power battery SOC input channel is used to process characteristic data related to the power battery system. The data output from the hydrogen SOC input channel and the power battery SOC input channel are input into the charging behavior feature processing module in their respective channels. This module extracts the key features of the time series changes during the charging process through steps such as feature normalization, time series feature extraction, feature interaction fusion, and dynamic weight adjustment, and effectively integrates the relationship between hydrogen consumption and power battery status. At the output stage of the charging behavior feature processing module, the position encoding vector is added or concatenated with the charging behavior feature vector to incorporate position information into the charging features. This ensures that the model in the encoder layer can understand the sequential relationship between charging features at different time steps and capture the temporal dynamics of charging behavior. The data output by the charging behavior feature processing module is position-encoded, integrated with the timing information, and then input into the encoder layer; The final feature vectors output by the encoder layer in the hydrogen SOC channel and the power battery SOC channel are input into the shared layer to achieve information sharing and feature interaction between the hydrogen SOC channel and the power battery SOC channel. The feature vectors of the hydrogen SOC and the power battery SOC are combined using a linear weighted fusion method to form a shared feature representation containing the hydrogen SOC channel and the power battery SOC channel: Among them, α and β are the weight coefficients of each channel, which are learnable parameters and satisfy α + β = 1; The task-specific layer further extracts features highly relevant to the specific task based on the shared features to meet the prediction requirements of each task. In the hydrogen SOC channel, the task-specific layer of this channel will strengthen the focus on features related to hydrogen SOC, while in the power battery SOC channel, the task-specific layer of this channel will strengthen the focus on features related to power battery SOC. The task-specific layer of each channel consists of a fully connected layer, a feedforward neural network, and a convolutional layer. The task-specific layer of each channel outputs accurate prediction values ​​of hydrogen SOC and power battery SOC respectively.

5. The fuel cell aircraft motor power limitation method according to claim 4, characterized in that: The dual-channel multi-task Transformer model introduces the charging behavior of the fuel cell to the power battery as a dynamic adjustment factor, specifically including: When the fuel cell charges the power battery, the power battery SOC channel will feed back charging-related information (battery voltage, battery current, battery temperature, charging efficiency, charging time, and power battery SOC) to the hydrogen SOC channel. This feedback indicates that the hydrogen SOC will decrease due to the fuel cell charging the power battery. The information fed back by the power battery to the hydrogen SOC channel during charging can represent the reduction in hydrogen SOC, which can be expressed by the following formula: Among them, ΔSOC H2 is the change in hydrogen SOC, k is the proportional coefficient used to convert the charging power of the power battery into the consumption of hydrogen SOC, V bat is the voltage of the power battery, I bat is the current of the power battery, η charge is the charging efficiency of the power battery, t charge is the charging time; When the hydrogen SOC drops to 30%, the hydrogen SOC channel will feed this information back to the power battery SOC channel to limit the charging behavior of the fuel cell and avoid further consumption of hydrogen. The following formula is used to limit the charging behavior: When SOC H2 ≥SOC H2，threshold hour: When SOC H2 <SOC H2，threshold hour: in, is the fuel cell charging current after limitation; I FC charging current for the original fuel cell; is the fuel cell charging power after limitation; P FC The original fuel cell charging power; SOC H2 is the current hydrogen SOC; SOC H2，threshold To limit the hydrogen SOC threshold when the fuel cell charges the power battery, it is usually set to 30%; when SOC H2 When it is above the threshold, The charging behavior of the fuel cell should be kept unrestricted, allowing the charging current and power to be maintained at their original values; when SOC H2 When it is less than the threshold, The charging current and power should be gradually reduced to avoid further consumption of hydrogen.

6. The fuel cell aircraft motor power limitation method according to claim 4, characterized in that: The encoder layer specifically includes: After each channel's encoder layer captures changes in hydrogen SOC or power battery SOC through its respective multi-head attention mechanism, it then adds the original features output by the charging behavior feature processing module to the output features of the multi-head attention mechanism to form a residual connection to better preserve the original hydrogen SOC features. The output of each layer is then normalized to ensure a more even data distribution. The features extracted by the multi-head attention mechanism are further transformed nonlinearly using a feedforward fully connected layer, enabling the model to learn more complex feature relationships between hydrogen SOC and power battery SOC, selectively retaining and suppressing features. The feedforward fully connected layer is also combined with residual connections and layer normalization to enhance feature expression and maintain model training stability. N encoder layers are stacked, and each encoder layer receives and processes the output of the previous layer. Through operations such as multi-head attention mechanism, feedforward fully connected layers, residual connections and layer normalization, the feature information for predicting hydrogen SOC or power battery SOC is gradually extracted and enriched. In the process of stacking layer by layer, the output of each encoder layer will become the input of the encoder of the next layer. After the last encoder layer completes processing, a final feature representation will be obtained. This final feature representation contains the deep features of hydrogen SOC or power battery SOC refined by N layers of encoders.

7. The fuel cell aircraft motor power limitation method according to claim 6, characterized in that: The multi-head attention mechanism specifically includes: The multi-head attention mechanism of each channel performs a linear transformation on the position-encoded feature data; In the hydrogen SOC channel, in order to focus on the changes in hydrogen SOC-related features, a hydrogen query vector, a hydrogen bond vector, and a hydrogen value vector are generated, and the features related to hydrogen SOC are weighted separately; In the power battery SOC channel, in order to focus on the changes in the power battery SOC-related features, a power battery query vector, a power battery key vector, and a power battery value vector are generated, and the power battery SOC-related features are individually weighted. For each channel, the attention score is calculated using the respective query vector and key vector, and the similarity between each time step is obtained by dot product; The multi-head attention mechanism divides the input features of hydrogen SOC and power battery SOC into multiple subspaces by parallel computing multiple heads, and processes them from different perspectives. Each head independently calculates the attention output of the query vector, key vector, and value vector, and then splices the outputs of multiple heads together to form the feature aggregation vector of each channel.

8. The fuel cell aircraft motor power limitation method according to claim 1, characterized in that: The specific coordinated control system includes: hydrogen storage system, HMS, fuel cell control unit (FCU), fuel cell, unidirectional DC / DC, BMS, power battery, bidirectional DC / DC, inverter, motor and aircraft control unit; The collaborative control system dynamically adjusts the power output of the motor according to the real-time status changes of hydrogen SOC and power battery SOC through information interaction among control units such as HMS, BMS, FCU, MCU and centralized control of the aircraft control unit.

9. A computer device comprising: A memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the fuel cell aircraft motor power limitation method according to any one of claims 1 to 8.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the fuel cell aircraft motor power limiting method according to any one of claims 1 to 8 is implemented.

11. A computer program product comprising a computer program, characterized in that When the computer program is executed by a processor, the fuel cell aircraft motor power limiting method according to any one of claims 1 to 8 is implemented.

Images