Low-altitude unmanned aerial vehicle hydrogen-electricity hybrid energy management system and control device

Abstract

The invention relates to the technical field of unmanned aerial vehicle power systems, in particular to a low-altitude unmanned aerial vehicle hydrogen-electricity hybrid energy management system and control device.According to the low-altitude unmanned aerial vehicle hydrogen-electricity hybrid energy management system and control device, the residual hydrogen amount can be accurately quantified through combined collection and fusion calculation of hydrogen flow signals and air pressure signals; and the opening state of the proportional valve and the air inlet valve is adjusted in real time under data driving, so that the hydrogen input is matched with the actual power requirement, the voltage, the current and the temperature are synchronously monitored, and fluctuation feature extraction and life parameter calibration are performed on a power output sequence. The operation health degree and the life loss trend of the fuel cell can be dynamically obtained before power distribution, an accurate reference basis is provided for follow-up power allocation, the internal resistance change trend is measured and calculated through voltage and current sampling, and dynamic adjustment of the charge-discharge depth and charge-discharge process parameters is carried out in combination with the health degree. And the usability and the load capacity of the lithium battery are accurately represented.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) power system technology, and in particular to a hydrogen-electric hybrid energy management system and control device for low-altitude UAVs. Background Technology

[0002] In the field of unmanned aerial vehicle (UAV) propulsion system technology, UAV propulsion configuration design, energy system integration methods, energy conversion and distribution control methods, and control strategies for dynamically adjusting energy output at different stages of flight missions are developed to improve flight time under unit load conditions, reduce working stress of energy components, improve energy utilization efficiency, and extend the cycle life of propulsion components.

[0003] The purpose of a low-altitude unmanned aerial vehicle (UAV) hydrogen-electric hybrid energy management system is to achieve optimal energy distribution at different flight stages by coordinating and controlling the power output of hydrogen fuel cells and lithium batteries. This aims to improve the UAV's endurance under rated load and set flight altitude conditions, reduce instantaneous power surges and thermal stresses during the cyclic operation of hydrogen fuel cells and lithium batteries, thereby extending the cycle life of power components, improving energy utilization efficiency, and maintaining stable power output.

[0004] In existing systems, hydrogen consumption and inventory estimation rely heavily on single sensor signals, which are highly susceptible to interference from environmental parameters and pressure fluctuations, affecting the accuracy of fuel supply. During fuel cell operation and management, conventional methods rely solely on power, voltage, or single temperature monitoring to determine the operating status, lacking synchronous quantitative analysis of load fluctuation characteristics and lifespan status. This results in the inability to detect performance degradation signs in a timely manner during long-term operation. Furthermore, lithium battery operation monitoring employs periodic health assessment methods, with assessment results lagging behind the operating status, making real-time adjustments to charge and discharge depth impossible. This can lead to excessive discharge or thermal stress accumulation from repeated high-power charging during high-load phases. Existing models use fixed ratios or simple rules for power sharing between fuel cells and lithium batteries, which are difficult to adapt to the rapid changes in load demands during flight missions. In power demand assessment during the flight phase, only a single variable, such as speed or altitude, is typically considered, failing to establish a dynamic correlation between flight attitude parameters, thrust commands, and energy health status, resulting in discrepancies between predicted results and actual demands. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of the existing technology and to propose a hydrogen-electric hybrid energy management system and control device for low-altitude unmanned aerial vehicles.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a low-altitude unmanned aerial vehicle (UAV) hydrogen-electric hybrid energy management system, comprising: Hydrogen supply module: Based on the hydrogen flow transmitter signal, it obtains the hydrogen flow rate, calculates the remaining hydrogen amount by combining the data from the pressure transmitter, adjusts the status of the proportional valve and the inlet valve, regulates the hydrogen flow rate, and obtains the remaining hydrogen amount data. Fuel cell system module: Based on the remaining hydrogen data, adjust the fuel cell power output, monitor the battery voltage, current and temperature, extract load fluctuation characteristics, obtain battery life data, perform power regulation, and obtain fuel cell power output and life loss data; Lithium battery system module: Based on the power output and life loss data of the fuel cell, acquire the lithium battery current and voltage, calculate the internal resistance change, evaluate the battery health status, adjust the lithium battery charge and discharge status, and generate lithium battery health status data. Energy scheduling module: Based on the lithium battery health status data, it uses support vector machine to evaluate the battery load, determine whether the load requirements are met, select an appropriate battery load allocation method, and generate an appropriate energy scheduling scheme. Power allocation module: Based on the adaptive energy scheduling scheme, it adopts multivariate predictive control to analyze the power demand during flight, determine the battery health status and power limitations, adjust the battery power output, optimize power allocation, execute power regulation, and generate real-time power allocation commands.

[0007] As a further aspect of the present invention, the hydrogen remaining quantity data includes real-time flow rate, gas pressure data, and information on the remaining hydrogen quantity in the hydrogen cylinder; the fuel cell power output and lifespan loss data includes battery voltage, current, load fluctuations, and temperature changes; the lithium battery health status data includes battery internal resistance, cycle count, and battery voltage and current fluctuations; the adaptive energy scheduling scheme includes the power allocation ratio between the hydrogen fuel cell and the lithium battery, load allocation strategy, and scheduling timing; and the real-time power allocation command includes the power output setpoint, power adjustment range, control time, and phased targets for each battery.

[0008] As a further aspect of the present invention, the hydrogen supply module includes: Hydrogen flow monitoring submodule: Based on the hydrogen flow transmitter signal, it acquires hydrogen flow data, combines it with the pressure data output by the pressure transmitter, calculates the remaining hydrogen in the hydrogen cylinder, and tracks the hydrogen flow change in real time through the monitor to generate hydrogen remaining data. Hydrogen flow regulation submodule: Based on the remaining hydrogen data, it adjusts the on / off state of the proportional valve, controls the hydrogen flow through the intake valve, adjusts the hydrogen intake in real time, performs flow control, and outputs the adjusted hydrogen flow data to obtain hydrogen flow regulation data.

[0009] As a further aspect of the present invention, the fuel cell system module includes: Power output regulation submodule: Based on the hydrogen flow regulation data, the power output of the fuel cell is adjusted in real time. By monitoring the changes in battery voltage and current, the real-time power of the battery is obtained, and the output power is adjusted according to the battery status to obtain the fuel cell power output data. Lifetime loss monitoring submodule: Based on the fuel cell power output data, it monitors the battery temperature in real time, extracts battery load fluctuation characteristics, analyzes the battery health status, obtains battery life loss information, and generates fuel cell life loss data.

[0010] As a further aspect of the present invention, the lithium battery system module includes: Battery status monitoring submodule: Based on the power output and life loss data of the fuel cell, acquire lithium battery current and voltage signals, measure the change of internal resistance of the battery in real time, monitor battery voltage fluctuations and charging and discharging processes, evaluate battery health status, and generate lithium battery health status data. Charge and discharge management submodule: Based on the lithium battery health status data, analyze the battery health status, determine whether to adjust the charge and discharge depth, monitor the current and voltage changes during the charging process in real time, adjust the charge and discharge state, and generate lithium battery charge and discharge control data.

[0011] As a further aspect of the present invention, the energy scheduling module includes: Load assessment submodule: Based on the lithium battery health status data, the module acquires the battery's voltage and current signals, uses a support vector machine to measure the battery's internal resistance and charging status in real time, analyzes the battery load, calculates whether the battery can meet the current power demand, and generates a battery load assessment result. Battery load judgment submodule: Based on the battery load assessment results, judge the battery health status, confirm whether the battery can stably provide the required power output, judge the load satisfaction according to the battery health status, and generate the load satisfaction judgment result; Energy allocation optimization submodule: Based on the load satisfaction judgment result, select an appropriate battery load allocation scheme and generate an appropriate energy scheduling scheme by adjusting the power output ratio of hydrogen fuel cell and lithium battery.

[0012] As a further aspect of the present invention, the support vector machine first converts the collected battery voltage and current signals into time series data at a fixed sampling rate, constructing a feature vector consisting of the average voltage, average current, instantaneous voltage drop amplitude, equivalent internal resistance calculation value, and charging state value. Then, based on a preset training sample set, which contains multiple sets of feature vectors under different battery health states and corresponding load capacity labels, the radial basis function kernel function is used to calculate the inner product between samples in the feature space. Then, the real-time feature vector of the battery to be evaluated is input into the support vector machine model, and kernel function mapping and decision function calculation are performed to output the discrimination result. A result value of −1 indicates insufficient load capacity, and a result value of +1 indicates that the load capacity meets the requirements. Finally, the discrimination result is combined with the internal resistance measurement value and the charging state value to form the battery load evaluation result.

[0013] As a further aspect of the present invention, the power distribution module includes: Flight Phase Analysis Submodule: Based on the aforementioned adaptive energy scheduling scheme, multivariate predictive control is used to analyze the power demand of the current flight phase, determine the power requirements of different flight phases, evaluate the battery's ability to meet the power demand, and generate flight phase power demand analysis results. Battery health status control submodule: Based on the power demand analysis results of the flight phase, multivariate predictive control is used to determine whether the battery health status meets the power output requirements, adjust the peak value of the battery output power, control the battery power allocation ratio, and generate battery power output control data. Power regulation submodule: Based on the battery power output control data, it performs power regulation, adjusts the battery power output and optimizes power distribution, ensures that the power output of each battery is within the appropriate range by adjusting the output value, and generates real-time power distribution instructions.

[0014] As a further aspect of the present invention, the multivariate predictive control first constructs a state vector based on multiple state variables collected during the current flight phase of the aircraft, such as speed, altitude, pitch angle, thrust command, battery voltage, battery current, and fuel cell output power, and updates it according to a fixed sampling period. Then, a state-space model is used to describe the dynamic relationship between the power demand during the flight phase and each state variable. The state matrix, input matrix, and output matrix are all obtained based on historical test data and system identification calculations. Within each sampling period, a rolling time-domain optimization method is used to solve the optimization objective function based on quadratic programming. The objective function consists of the sum of squares of the difference between the predicted power demand and the available battery output power, and the sum of squares of the control input increments. After obtaining the optimal control increment sequence, only the first step of the control quantity is executed, and the prediction and optimization process is repeated in the next sampling period. Finally, the output control quantity adjusts the battery power allocation ratio and peak limit.

[0015] A hydrogen-electric hybrid energy management device for low-altitude unmanned aerial vehicles (UAVs) includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the aforementioned hydrogen-electric hybrid energy management system for low-altitude UAVs.

[0016] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, by jointly acquiring and fusing hydrogen flow and pressure signals, the remaining amount of hydrogen can be accurately quantified, and the opening of the proportional valve and the intake valve can be adjusted in real time under data-driven conditions to keep the hydrogen input in line with the actual power demand. In this invention, by synchronously monitoring voltage, current and temperature, and extracting fluctuation characteristics and calibrating lifetime parameters of the power output sequence, the operating health and life loss trend of the fuel cell can be dynamically obtained before power allocation, providing an accurate reference for subsequent power allocation. In this invention, the internal resistance change trend is calculated by sampling voltage and current, and the charging and discharging depth and charging and discharging process parameters are dynamically adjusted in combination with health status, so as to accurately characterize the usability and load capacity of lithium batteries. In this invention, an input vector containing features such as real-time voltage, current, internal resistance, and charging / discharging status is constructed, and a classification algorithm is used to determine the load satisfaction. The determination result is then used as a constraint to perform global power allocation optimization. In this invention, a multivariable predictive control model is introduced, which uses multiple parameters such as flight speed, altitude, vertical climb rate, thrust command and battery output status to make rolling predictions of future time-domain power demand. The predicted values ​​are then solved in conjunction with the health status and peak constraints to generate dynamic power allocation and output control commands. This allows the energy output process to adapt to the operating state, reduces transient impacts on devices, and maintains long-term stable operation. Attached Figure Description

[0017] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0018] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0019] In the description of this invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," and "outer," etc., indicating orientation or positional relationships, are based on the orientation or positional relationships shown in the accompanying drawings and are only for the convenience of describing the invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation of the invention. Furthermore, in the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0020] Example 1 Please see Figure 1 This invention provides a technical solution: a hydrogen-electric hybrid energy management system for low-altitude unmanned aerial vehicles, comprising: Hydrogen supply module: Based on the hydrogen flow transmitter signal, it obtains the hydrogen flow rate, calculates the remaining hydrogen amount by combining the data from the pressure transmitter, adjusts the status of the proportional valve and the inlet valve, regulates the hydrogen flow rate, and obtains the remaining hydrogen amount data. Fuel cell system module: Based on hydrogen remaining data, adjust fuel cell power output, monitor battery voltage, current and temperature, extract load fluctuation characteristics, obtain battery life data, perform power regulation, and obtain fuel cell power output and life loss data; Lithium battery system module: Based on fuel cell power output and life loss data, acquire lithium battery current and voltage, calculate internal resistance changes, assess battery health status, adjust lithium battery charge and discharge state, and generate lithium battery health status data. Energy scheduling module: Based on lithium battery health status data, it uses support vector machine to evaluate battery load, determine whether the load demand is met, select an appropriate battery load allocation method, and generate an appropriate energy scheduling scheme. Power allocation module: Based on the adaptive energy scheduling scheme, it adopts multivariate predictive control to analyze the power demand during flight, determine the battery health status and power limitations, adjust the battery power output, optimize power allocation, execute power regulation and generate real-time power allocation commands.

[0021] The hydrogen remaining data includes real-time flow rate, pressure data, and information on the amount of hydrogen remaining in the hydrogen cylinder. The fuel cell power output and lifespan loss data includes battery voltage, current, load fluctuations, and temperature changes. The lithium battery health status data includes battery internal resistance, cycle count, and battery voltage and current fluctuations. The adaptive energy dispatch scheme includes the power allocation ratio between hydrogen fuel cells and lithium batteries, load allocation strategies, and dispatch timing. The real-time power allocation instructions include the power output setpoints, power adjustment ranges, control times, and phased targets for each battery.

[0022] The hydrogen supply module includes: Hydrogen flow monitoring submodule: Based on the hydrogen flow transmitter signal, it acquires hydrogen flow data, combines it with the pressure data output by the pressure transmitter, calculates the remaining hydrogen in the hydrogen cylinder, and tracks the hydrogen flow change in real time through the monitor to generate hydrogen remaining data. Hydrogen flow regulation submodule: Based on the remaining hydrogen amount data, it adjusts the on / off state of the proportional valve, controls the hydrogen flow through the intake valve, adjusts the hydrogen intake in real time, performs flow control, and outputs the adjusted hydrogen flow data to obtain hydrogen flow regulation data. Hydrogen Flow Monitoring Submodule: Based on the hydrogen flow transmitter signal, the raw hydrogen flow sampling data is processed using a Kalman filter algorithm. The initial state vector is initialized with the first term representing the flow rate in liters per minute (L / min) and the second term representing the flow rate change in liters per minute (L / s). The diagonal elements of the initial state covariance matrix are all set to 0.01. The state transition matrix has the first row and first column set to 1, the first row and second column set to the sampling time interval of 0.1 seconds, the second row and first column set to 0, and the second row and second column set to 1. The observation matrix has the first row and first column set to 1 and the first row and second column set to 0. The process noise covariance matrix has its diagonal elements set to 0.00001. The measurement noise covariance matrix has the first row and first column set to 0.00001. One column is set to 0.0001. At each sampling time, the prediction step is first executed to update the state vector and state covariance matrix using the state transition matrix. Then, the Kalman gain value is calculated and the flow measurement value and the predicted value are merged proportionally to update the state vector and state covariance. After that, the smoothed flow value obtained by Kalman filtering is combined with the pressure data provided by the pressure transmitter. The amount of substance in the gas cylinder is calculated using the pressure unit megapascal, the gas cylinder volume unit cubic meter, the universal gas constant value of 8.314, and the temperature unit Kelvin in the ideal gas equation of state. Then, the remaining hydrogen mass value is obtained by converting it according to the molar mass unit of hydrogen per mole, and the remaining hydrogen amount data is recorded and generated. Hydrogen Flow Regulation Submodule: Based on the remaining hydrogen quantity data, a PID control algorithm is used to adjust the opening of the proportional valve. The proportional coefficient Kp is set to 0.85, the integral coefficient Ki to 0.15, the derivative coefficient Kd to 0.05, and the sampling period is 0.1 seconds. The control deviation is defined as the set flow rate value minus the currently measured flow rate value. At each sampling moment, the proportional term is calculated based on the difference between the current control deviation and the control deviation at the previous sampling moment. The integral term is accumulated based on the product of the control deviation value and the sampling period. The differential term is calculated based on the current deviation and the previous two deviations. The proportional term, integral term, and differential term are summed according to the set proportional coefficient to obtain the control output value of the current cycle. The output value is mapped as a percentage to the maximum output value to obtain the opening percentage of the proportional valve. The maximum opening percentage is set to 100%. The corresponding control current value in milliamperes is sent as a drive signal to the proportional valve drive module, which drives the intake valve to open and close, thereby adjusting the hydrogen intake flow rate. The adjusted flow rate value is measured and recorded, and hydrogen flow regulation data is output and generated.

[0023] The fuel cell system module includes: Power output regulation submodule: Based on hydrogen flow regulation data, the power output of the fuel cell is adjusted in real time. By monitoring changes in battery voltage and current, the real-time power of the battery is obtained, and the output power is adjusted according to the battery status to obtain the fuel cell power output data. Lifetime loss monitoring submodule: Based on fuel cell power output data, it monitors battery temperature in real time, extracts battery load fluctuation characteristics, analyzes battery health status, obtains battery life loss data, and generates fuel cell life loss data. Power Output Regulation Submodule: Based on hydrogen flow regulation data, a fuzzy PID control algorithm is used to regulate the fuel cell power output in real time. The proportional coefficient of the PID controller is initialized to 0.9, the integral coefficient to 0.12, the derivative coefficient to 0.04, and the sampling period to 0.1 seconds. The input is set as the difference between the target power value (kW) and the measured power value (kW) of the fuel cell as the control deviation. The fuzzy logic module input is the control deviation and the rate of change of the control deviation. The domain of discourse for the control deviation is defined as -10 to 10, and the domain of discourse for the rate of change is defined as -5 to 5. The domain of discourse is divided into seven fuzzy subsets: negative large, negative medium, negative small, zero, positive small, positive medium, and positive large, using a triangular membership function. The Mamdani inference method is used to dynamically adjust the proportional coefficient, integral coefficient, and derivative coefficient of the PID according to a preset rule table, with the values ​​varying within ±30%. In each sampling period, the output control quantity (Amperes) is calculated based on the adjusted PID parameters and sent to the fuel cell power control module. The real-time power value is obtained by monitoring the battery voltage (Volts) and current (Amperes) according to a multiplicative relationship. The adjustment results are used to generate fuel cell power output data. The lifespan loss monitoring submodule extracts the load fluctuation characteristics of the fuel cell based on the fuel cell power output data using a sliding window FFT frequency domain analysis method. The sliding window length is set to 1024 sampling points, the window sliding step size to 128 sampling points, and the sampling frequency to 100 Hz. The power output value sequence within each window is input into a Fast Fourier Transform function to calculate the spectral amplitude array. A frequency resolution of 0.097 Hz is defined, and amplitude data within the range of 0 to 5 Hz are used to analyze low-frequency load fluctuations. The mean and peak values ​​of the spectral amplitude, as well as the energy distribution ratio, are calculated. The battery temperature value in the time domain is recorded in degrees Celsius. The frequency domain characteristic data and temperature data are matched one-to-one using timestamps. The module indexes the parameters using a lifespan decay parameter table under different temperature and frequency conditions from the fuel cell factory lifespan curve database, reads the lifespan loss percentage, and generates fuel cell lifespan loss data.

[0024] The lithium battery system module includes: Battery status monitoring submodule: Based on fuel cell power output and life loss data, acquire lithium battery current and voltage signals, measure battery internal resistance changes in real time, monitor battery voltage fluctuations and charging and discharging processes, assess battery health status, and generate lithium battery health status data. Charge and discharge management submodule: Based on lithium battery health status data, analyze the battery health status, determine whether to adjust the depth of charge and discharge, monitor the current and voltage changes during the charging process in real time, adjust the charge and discharge status, and generate lithium battery charge and discharge control data. Battery Status Monitoring Submodule: Based on fuel cell power output and lifespan loss data, the module uses AC impedance spectroscopy analysis to acquire lithium battery current and voltage signals. It sets the AC excitation signal amplitude to 5 mV, the frequency scanning range to 0.1 Hz to 1 kHz, and 50 sampling points with a logarithmic frequency step. By applying AC excitation and simultaneously acquiring current and voltage signals, the module calculates the amplitude ratio and phase difference. The amplitude ratio and phase difference data are recorded in the form of complex plane impedance. The battery internal resistance value in milliohms is obtained by fitting the series ohmic resistance, parallel charge transfer resistance, and electrochemical double-layer capacitance in the equivalent circuit model. Simultaneously, the module records the battery voltage and current change curves during charging and discharging at a sampling frequency of 0.05 seconds within the acquisition period. It statistically analyzes the voltage rise and fall amplitudes during the discharge and charging phases. The internal resistance change, charge / discharge fluctuation amplitude, and charge / discharge curve characteristics are input into the status assessment calculation process to generate lithium battery health status data. The charge / discharge management submodule, based on lithium battery health status data, employs an adaptive charge / discharge depth control algorithm to adjust parameters during the charge / discharge process. The initial charge / discharge depth threshold is set to 80% of the rated capacity. Input quantities are defined as the remaining capacity percentage, voltage fluctuation amplitude, and internal resistance change rate in the healthy state. The remaining capacity percentage is mapped to the charging cut-off voltage and discharging cut-off voltage adjustment ranges. The charging cut-off voltage adjustment range per volt is set to 4.15 to 4.25, and the discharging cut-off voltage adjustment range is set to 3.0 to 3.2. During charging, the module monitors changes in current (amperes) and voltage (volts) with a 0.1-second sampling period. When the capacity percentage exceeds the threshold or the internal resistance increase rate in the health status display exceeds 1%, the charging cut-off voltage is lowered to the lower limit of the adjustment range. During discharging, the voltage drop rate is monitored, and when the rate exceeds the set value of 0.02 volts per second, the discharging cut-off voltage is raised to the upper limit of the adjustment range. Control parameters are output to the BMS execution unit in real time, generating lithium battery charge / discharge control data.

[0025] The energy dispatch module includes: Load assessment submodule: Based on lithium battery health status data, it acquires battery voltage and current signals, uses support vector machine to measure battery internal resistance and charging status in real time, analyzes battery load, calculates whether the battery can meet the current power demand, and generates battery load assessment results. Battery load judgment submodule: Based on the battery load assessment results, judge the battery health status, confirm whether the battery can stably provide the required power output, judge the load satisfaction according to the battery health status, and generate the load satisfaction judgment result; Energy allocation optimization submodule: Based on the load satisfaction judgment result, select an appropriate battery load allocation scheme and generate an appropriate energy dispatch scheme by adjusting the power output ratio of hydrogen fuel cells and lithium batteries; Load assessment submodule: Based on lithium battery health status data, it acquires battery voltage and current signals, and uses a support vector machine (SVM) algorithm to classify the acquired signals. It constructs a feature vector consisting of the average voltage in volts, the average current in amperes, the discharge voltage drop rate in volts per second, the equivalent internal resistance in milliohms, and the state of charge percentage. It loads an SVM model trained from historical sample data. The model parameters include a radial basis function kernel width of 0.5 and a penalty factor of 100. The input is a five-dimensional feature vector in the training sample space. The classification label is defined as 1 indicating that the power requirement can be met and 0 indicating that the power requirement cannot be met. The real-time acquired feature vector is input into the model for inference. The similarity between the feature vector and the support vector is calculated by calling the kernel function, and the classification result is obtained by weighting according to the weight coefficient of the support vector and the training bias. At the same time, the internal resistance test value in milliohms and the state of charge percentage are recorded, and the battery load assessment result is generated. Battery load judgment submodule: Based on the battery load assessment results, a threshold comparison judgment method is used to determine the battery health status. The internal resistance threshold is set to 5 milliohms, the charging status threshold is 30%, and the load demand label is 1. When the measured internal resistance value is not greater than the set threshold, the charging status percentage is not lower than the set threshold, and the classification label is 1, the satisfying flag value 1 is output; otherwise, the satisfying flag value 0 is output. The internal resistance detection value, charging status percentage, and satisfying flag value are encoded and stored as a combined data packet to obtain the load satisfying judgment result. The energy allocation optimization submodule: Based on the load satisfaction judgment result, a linear programming allocation algorithm is used to select an appropriate battery load allocation scheme. The decision variables are defined as the output power of the hydrogen fuel cell (kW) and the output power of the lithium battery (kW). The objective function is to minimize the sum of the squares of the output power of the hydrogen fuel cell and the output power of the lithium battery. The constraints include that the sum of the output power equals the target power demand, the output power of the hydrogen fuel cell is not lower than its minimum technical power limit, not higher than its rated power, and the output power of the lithium battery does not exceed the upper limit of the available discharge power. The constraints and objective function are loaded as inputs into the allocation calculation function. The simplex method solver is called to search for the optimal combination of hydrogen fuel cell output power and lithium battery output power in the feasible region according to the power demand. The obtained power output ratio is used as the allocation parameter to generate an appropriate energy scheduling scheme.

[0026] The Support Vector Machine (SVM) first converts the collected battery voltage and current signals into time-series data at a fixed sampling rate, constructing a feature vector consisting of the mean voltage, mean current, instantaneous voltage drop, calculated equivalent internal resistance, and state of charge. Then, based on a pre-set training sample set containing multiple sets of feature vectors under different battery health states and corresponding load capacity labels, the inner product between samples is calculated in the feature space using a radial basis function kernel. The real-time feature vector of the battery to be evaluated is then input into the SVM model, performing kernel function mapping and decision function calculation, and outputting a judgment result. A result value of -1 indicates insufficient load capacity, while a result value of +1 indicates that the load capacity meets the requirements. Finally, the judgment result is combined with the internal resistance measurement and state of charge values ​​to form the battery load evaluation result.

[0027] The power distribution module includes: Flight Phase Analysis Submodule: Based on the adaptive energy scheduling scheme, multivariate predictive control is used to analyze the power demand of the current flight phase, determine the power requirements of different flight phases, evaluate the battery's ability to meet the power demand, and generate flight phase power demand analysis results. Battery health status control submodule: Based on the power demand analysis results during flight, multivariate predictive control is used to determine whether the battery health status meets the power output requirements, adjust the peak value of the battery output power, control the battery power distribution ratio, and generate battery power output control data. Power regulation submodule: Based on battery power output control data, it performs power regulation, adjusts battery power output and optimizes power distribution, ensures that the power output of each battery is within the appropriate range by adjusting the output value, and generates real-time power distribution instructions; Flight Phase Analysis Submodule: Based on the adaptive energy scheduling scheme, a multivariate predictive control algorithm is used to analyze the power demand during the flight phase. The input data is defined as the flight speed (m / s), flight altitude (m), vertical climb rate (m / s), pitch angle (degrees), thrust control command percentage, battery voltage (volts), battery current (amperes), and fuel cell output power (kilowatts) for the current flight phase. The sampling period of the prediction model is set to 0.1 seconds, the prediction time domain length is set to 20 steps, and the control time domain length is set to 5 steps. Based on historical test data and system identification results, a state matrix, input matrix, and output matrix are constructed. Constraints are set, including the available battery power range, the fuel cell power change rate, and the upper limit of battery current. The current state is read and input into the predictive control calculation function in each sampling period. The rolling time domain optimization process is called to generate the future prediction curve. The power demand value at each future moment is calculated. The power demand curve is compared with the available battery output power curve point by point to generate the power demand analysis results for the flight phase. Battery health status control submodule: Based on the power demand analysis results during flight, a multivariate predictive control algorithm is used to determine the battery health status and adjust the output. The input quantities are set as the future power demand value sequence in the analysis results, the battery internal resistance in milliohms, the remaining capacity percentage, the current battery temperature in degrees Celsius, and the maximum allowable current in amperes. In the predictive control model, power output peak constraints, current change rate constraints, and internal resistance change rate constraints are set. The control objective is to make the battery output power meet the demand curve and limit the peak value. In each sampling period, the prediction process is executed to calculate the optimal peak allocation table in the future time domain. The optimized output power adjustment ratio and peak limit value are recorded as control commands and encapsulated as battery power output control data. Power Regulation Submodule: Based on battery power output control data, it performs power regulation using a direct power allocation calculation method. The input parameters are defined as the target peak power (kW), output proportional parameter (%), real-time battery voltage (volts), and current (amperes) in the control data. In each sampling period, it calculates the current power output value and subtracts it from the target value. The difference is mapped proportionally to the power allocation adjustment amount between the fuel cell and the battery. The power allocation execution function is called to update the battery output value (kW) and the fuel cell output value (kW) is updated according to the allocation adjustment amount. The real-time output power of the fuel cell and lithium battery is recorded and a real-time power allocation command is generated.

[0028] Multivariable predictive control (MVC) first constructs a state vector based on multiple state variables collected during the current flight phase of the aircraft, including speed, altitude, pitch angle, thrust command, battery voltage, battery current, and fuel cell output power. This vector is updated at a fixed sampling period. Next, a state-space model is used to describe the dynamic relationship between power demand and each state variable during the flight phase. The state matrix, input matrix, and output matrix are all obtained based on historical test data and system identification calculations. Within each sampling period, a rolling time-domain optimization method is used to solve the objective function based on quadratic programming. The objective function consists of the sum of squares of the difference between the predicted power demand and the available battery output power, and the sum of squares of the control input increments. After obtaining the optimal control increment sequence, only the first step of the control input is executed before entering the next sampling period and repeating the prediction and optimization process. The final output control input adjusts the battery power allocation ratio and peak limit.

[0029] A hydrogen-electric hybrid energy management device for low-altitude unmanned aerial vehicles (UAVs) includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the aforementioned hydrogen-electric hybrid energy management system for low-altitude UAVs.

[0030] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A hydrogen-electric hybrid energy management system for low-altitude unmanned aerial vehicles (UAVs), characterized in that, Includes the following steps: Hydrogen supply module: Based on the hydrogen flow transmitter signal, it obtains the hydrogen flow rate, calculates the remaining hydrogen amount by combining the data from the pressure transmitter, adjusts the status of the proportional valve and the inlet valve, regulates the hydrogen flow rate, and obtains the remaining hydrogen amount data. Fuel cell system module: Based on the remaining hydrogen data, adjust the fuel cell power output, monitor the battery voltage, current and temperature, extract load fluctuation characteristics, obtain battery life data, perform power regulation, and obtain fuel cell power output and life loss data; Lithium battery system module: Based on the power output and life loss data of the fuel cell, acquire the lithium battery current and voltage, calculate the internal resistance change, evaluate the battery health status, adjust the lithium battery charge and discharge status, and generate lithium battery health status data. Energy scheduling module: Based on the lithium battery health status data, it uses support vector machine to evaluate the battery load, determine whether the load requirements are met, select an appropriate battery load allocation method, and generate an appropriate energy scheduling scheme. Power allocation module: Based on the adaptive energy scheduling scheme, it adopts multivariate predictive control to analyze the power demand during flight, determine the battery health status and power limitations, adjust the battery power output, optimize power allocation, execute power regulation, and generate real-time power allocation commands.

2. The low-altitude unmanned aerial vehicle hydrogen-electric hybrid energy management system according to claim 1, characterized in that, The hydrogen remaining data includes real-time flow rate, pressure data, and information on the remaining hydrogen in the hydrogen cylinder. The fuel cell power output and lifespan loss data includes battery voltage, current, load fluctuations, and temperature changes. The lithium battery health status data includes battery internal resistance, cycle count, and battery voltage and current fluctuations. The adaptive energy scheduling scheme includes the power allocation ratio between the hydrogen fuel cell and the lithium battery, load allocation strategy, and scheduling timing. The real-time power allocation command includes the power output setpoint, power adjustment range, control time, and phased targets for each battery.

3. The low-altitude unmanned aerial vehicle hydrogen-electric hybrid energy management system according to claim 1, characterized in that, The hydrogen supply module includes: Hydrogen flow monitoring submodule: Based on the hydrogen flow transmitter signal, it acquires hydrogen flow data, combines it with the pressure data output by the pressure transmitter, calculates the remaining hydrogen in the hydrogen cylinder, and tracks the hydrogen flow change in real time through the monitor to generate hydrogen remaining data. Hydrogen flow regulation submodule: Based on the remaining hydrogen amount data, adjust the on / off state of the proportional valve, control the hydrogen flow through the intake valve, adjust the hydrogen intake in real time, perform flow control, and output the adjusted hydrogen flow data to obtain hydrogen flow regulation data.

4. The low-altitude unmanned aerial vehicle hydrogen-electric hybrid energy management system according to claim 1, characterized in that, The fuel cell system module includes: Power output regulation submodule: Based on the hydrogen flow regulation data, the power output of the fuel cell is adjusted in real time. By monitoring the changes in battery voltage and current, the real-time power of the battery is obtained, and the output power is adjusted according to the battery status to obtain the fuel cell power output data. Lifetime loss monitoring submodule: Based on the fuel cell power output data, it monitors the battery temperature in real time, extracts battery load fluctuation characteristics, analyzes the battery health status, obtains battery life loss information, and generates fuel cell life loss data.

5. The low-altitude unmanned aerial vehicle hydrogen-electric hybrid energy management system according to claim 1, characterized in that, The lithium battery system module includes: Battery status monitoring submodule: Based on the power output and life loss data of the fuel cell, acquire lithium battery current and voltage signals, measure the change of internal resistance of the battery in real time, monitor battery voltage fluctuations and charging and discharging processes, evaluate battery health status, and generate lithium battery health status data. Charge and discharge management submodule: Based on the lithium battery health status data, analyze the battery health status, determine whether to adjust the charge and discharge depth, monitor the current and voltage changes during the charging process in real time, adjust the charge and discharge state, and generate lithium battery charge and discharge control data.

6. The low-altitude unmanned aerial vehicle hydrogen-electric hybrid energy management system according to claim 1, characterized in that, The energy scheduling module includes: Load assessment submodule: Based on the lithium battery health status data, the module acquires the battery's voltage and current signals, uses a support vector machine to measure the battery's internal resistance and charging status in real time, analyzes the battery load, calculates whether the battery can meet the current power demand, and generates a battery load assessment result. Battery load judgment submodule: Based on the battery load assessment results, judge the battery health status, confirm whether the battery can stably provide the required power output, judge the load satisfaction according to the battery health status, and generate the load satisfaction judgment result; Energy allocation optimization submodule: Based on the load satisfaction judgment result, select an appropriate battery load allocation scheme and generate an appropriate energy scheduling scheme by adjusting the power output ratio of hydrogen fuel cell and lithium battery.

7. The low-altitude unmanned aerial vehicle hydrogen-electric hybrid energy management system according to claim 6, characterized in that, The support vector machine (SVM) first converts the collected battery voltage and current signals into time-series data at a fixed sampling rate, constructing a feature vector consisting of the average voltage, average current, instantaneous voltage drop, calculated equivalent internal resistance, and state of charge. Then, based on a preset training sample set containing multiple sets of feature vectors under different battery health states and corresponding load capacity labels, the radial basis function (RBF) kernel is used to calculate the inner product between samples in the feature space. The real-time feature vector of the battery to be evaluated is then input into the SVM model, performing kernel function mapping and decision function calculation, and outputting a judgment result. A result value of -1 indicates insufficient load capacity, while a result value of +1 indicates that the load capacity meets the requirements. Finally, the judgment result is combined with the internal resistance measurement and state of charge values ​​to form the battery load evaluation result.

8. The low-altitude unmanned aerial vehicle hydrogen-electric hybrid energy management system according to claim 1, characterized in that, The power distribution module includes: Flight Phase Analysis Submodule: Based on the aforementioned adaptive energy scheduling scheme, multivariate predictive control is used to analyze the power demand of the current flight phase, determine the power requirements of different flight phases, evaluate the battery's ability to meet the power demand, and generate flight phase power demand analysis results. Battery health status control submodule: Based on the power demand analysis results of the flight phase, multivariate predictive control is used to determine whether the battery health status meets the power output requirements, adjust the peak value of the battery output power, control the battery power allocation ratio, and generate battery power output control data. Power regulation submodule: Based on the battery power output control data, it performs power regulation, adjusts the battery power output and optimizes power distribution, ensures that the power output of each battery is within the appropriate range by adjusting the output value, and generates real-time power distribution instructions.

9. The low-altitude unmanned aerial vehicle hydrogen-electric hybrid energy management system according to claim 8, characterized in that, The multivariate predictive control first constructs a state vector based on multiple state variables collected during the current flight phase of the aircraft, such as speed, altitude, pitch angle, thrust command, battery voltage, battery current, and fuel cell output power. This vector is then updated at a fixed sampling period. Next, a state-space model is used to describe the dynamic relationship between power demand and each state variable during the flight phase. The state matrix, input matrix, and output matrix are all obtained based on historical test data and system identification calculations. Within each sampling period, a rolling time-domain optimization method is used to solve the optimization objective function based on quadratic programming. The objective function consists of the sum of squares of the difference between the predicted power demand and the available battery output power, and the sum of squares of the control input increments. After obtaining the optimal control increment sequence, only the first step of the control is executed before entering the next sampling period and repeating the prediction and optimization process. The final output control quantity adjusts the battery power allocation ratio and peak limit.

10. A hydrogen-electric hybrid energy management device for low-altitude unmanned aerial vehicles, comprising a memory and a processor, characterized in that, The memory stores a computer program, and when the processor executes the computer program, it implements the low-altitude unmanned aerial vehicle hydrogen-electric hybrid energy management system according to any one of claims 1 to 9.

Images