Liquid hydrogen storage and supply system for unmanned aerial vehicle and closed-loop intelligent control method of liquid hydrogen storage and supply system

Abstract

The invention discloses a liquid hydrogen storage and supply system for an unmanned aerial vehicle and a closed-loop intelligent control method of the liquid hydrogen storage and supply system. The method comprises the following steps: constructing a discrete time state space model for describing a dynamic coupling relationship among flow, pressure and temperature; constructing an optimization problem for minimizing the objective function; receiving a target set value sent by an unmanned aerial vehicle flight control system; adjusting the weight of the tracking error or the weight of the control energy consumption according to the received target set value; based on a discrete time state space model, rolling optimization calculation is carried out on an optimization problem according to the target function after weight adjustment and flow, pressure and temperature data collected in real time; and distributing a calculation result to an execution mechanism. By adopting an advanced multivariable predictive control algorithm, rapid, accurate and stable closed-loop control over the hydrogen flow, pressure and temperature is achieved, and the dynamic requirement of a propelling system within the full power range is met.

Description

Technical Field

[0001] This invention relates to unmanned aerial vehicle (UAV) propulsion system technology in the aerospace field, and more particularly to a liquid hydrogen storage and supply system for UAVs and its closed-loop intelligent control method. Background Technology

[0002] Liquid hydrogen propulsion systems, especially when combined with high-efficiency fuel cells, can significantly extend the flight time of drones and achieve zero carbon emissions, showing broad application prospects.

[0003] However, the application of liquid hydrogen faces severe technical challenges. Liquid hydrogen needs to be stored and transported at an extremely low temperature of -253°C (approximately 20 K), which places extremely stringent requirements on the structural design of storage tanks, efficient insulation, thermal management, supply systems, and overall safety. Any minor thermal leak could lead to rapid evaporation of liquid hydrogen (boil-off gas, BOG), resulting in fuel loss and increased pressure inside the storage tank, endangering flight safety.

[0004] Currently, existing airborne liquid hydrogen storage and supply technologies mainly focus on two extremes: one is large high-altitude long-endurance (HALE) platforms, which are relatively less sensitive to weight and can accommodate more complex and heavier storage and supply systems; the other is small (<25 kg) UAV demonstrator platforms, which have small system scale and low power requirements, and whose technical solutions are difficult to directly extend to medium and large UAVs. For 600 kg-class long-endurance UAV platforms flying at low altitudes below 3 km, there is currently a lack of mature and systematic liquid hydrogen storage and supply solutions on the market. Specifically, existing technologies have not yet systematically solved the following two problems.

[0005] First, achieving real-time, precise closed-loop control of hydrogen flow rate, pressure, and temperature is challenging. During missions, unmanned aerial vehicles (UAVs) exhibit highly variable flight profiles (e.g., takeoff, climb, cruise, maneuvering, hovering, landing), causing the power demands of propulsion systems (such as fuel cells) to change rapidly and dynamically over a wide range (e.g., from near-zero kW in standby mode to over 30 kW at maximum takeoff power). This necessitates that the liquid hydrogen supply system can precisely and in real-time adjust the output hydrogen flow rate, pressure, and temperature to match the dynamic demands of the propulsion system. Existing solutions often employ simple open-loop or PID control, which struggles to address the challenges of multivariable strong coupling, large time lag, and nonlinear control. This can easily lead to unstable hydrogen supply, affecting fuel cell performance and lifespan, and even causing safety issues.

[0006] Secondly, it cannot be deeply integrated with the flight control system. As the "heart" of the UAV, the liquid hydrogen storage and supply system's operational status directly affects flight safety and mission success. Existing propulsion systems often exist as independent "black boxes," with interaction with the UAV's flight control computer (FCU) limited to simple power commands and limited status feedback. This loose coupling fails to achieve real-time coordination between propulsion needs and flight control commands, lacking unified, in-depth status monitoring, fault diagnosis, and fault-tolerant control capabilities. If the storage and supply system malfunctions, the flight control system cannot obtain accurate fault information and adjust flight strategies in a timely manner, potentially leading to catastrophic consequences.

[0007] Therefore, there is an urgent need to invent an innovative technical solution to systematically solve the above problems and provide a liquid hydrogen storage and supply system that integrates lightweight design, precise closed-loop intelligent control and deep coupling with the flight control system to meet the stringent application requirements of low-altitude long-endurance liquid hydrogen-powered UAVs. Summary of the Invention

[0008] To address the aforementioned shortcomings in the existing technology, this invention provides a liquid hydrogen storage and supply system for unmanned aerial vehicles (UAVs) and its closed-loop intelligent control method, which solves the problems of existing liquid hydrogen storage and supply systems being unable to achieve real-time and accurate closed-loop control of hydrogen flow, pressure, and temperature, and being unable to deeply integrate with flight control systems.

[0009] To achieve the aforementioned objectives, the technical solution adopted by this invention is: a closed-loop intelligent control method for a liquid hydrogen storage and supply system for unmanned aerial vehicles (UAVs), comprising: Construct a discrete-time state-space model to describe the dynamic coupling relationship between flow rate, pressure, and temperature in a liquid hydrogen storage and supply system; Construct an optimization problem that minimizes the objective function; Receive the target setpoint sent by the UAV flight control system; adjust the weight of tracking error or control energy consumption according to the received target setpoint; perform rolling optimization calculation on the optimization problem based on the discrete-time state-space model, according to the objective function after weight adjustment and the real-time collected flow, pressure and temperature data of the liquid hydrogen storage and supply system; and distribute the calculation results to the actuators.

[0010] Secondly, the present invention also provides a liquid hydrogen storage and supply system for implementing a closed-loop intelligent control method for a liquid hydrogen storage and supply system for unmanned aerial vehicles, comprising: Vacuum-insulated composite storage tank for storing liquid hydrogen; The liquid hydrogen supply control unit, as the control center of the liquid hydrogen storage and supply system, includes an input module, an MPC control module, an adaptive gain scheduling module, a control allocation module, and an output module. The input module is used to receive target setpoints from the UAV flight control system and flow, pressure and temperature data collected in real time from the sensor array; The MPC control module is used to perform rolling optimization calculations, including a state predictor and an optimization solver. The state predictor is used to construct a discrete-time state-space model describing the dynamic coupling relationship between flow rate, pressure, and temperature in the liquid hydrogen storage and supply system; construct an optimization problem that minimizes the objective function; use the discrete-time state-space model to predict the behavior of the liquid hydrogen storage and supply system within a finite time window in the future, and obtain the prediction result; the optimization solver is used to solve the optimization problem that minimizes the objective function based on the prediction result, and generate the optimal control sequence.

[0011] The adaptive gain scheduling module is used to adjust the weight of tracking error or the weight of energy consumption control in the objective function according to the received target setpoint. The control allocation module is used to apply the first control action in the optimal control sequence to the actuator; The output module is used to upload real-time collected hydrogen flow, pressure and temperature data to the UAV flight control system; The actuators, including proportional valves and regulating valves, are used to adjust the fluid state according to relevant instructions issued by the liquid hydrogen supply control unit; The liquid hydrogen evaporation / vaporization unit is used to controllably vaporize the liquid hydrogen output from the vacuum insulated composite storage tank and adjust the temperature of the gaseous hydrogen. A sensor array is used to collect real-time data on hydrogen flow rate, pressure, and temperature in a liquid hydrogen storage and supply system. The safety protection subsystem includes an overpressure safety valve, a cryogenic alarm, and redundant sensor configurations for monitoring pressure and temperature anomalies.

[0012] The present invention also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor performs the steps of a closed-loop intelligent control method for a liquid hydrogen storage and supply system for unmanned aerial vehicles.

[0013] The present invention also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the steps of a closed-loop intelligent control method for a liquid hydrogen storage and supply system for unmanned aerial vehicles.

[0014] The beneficial effects of this invention are as follows: By employing advanced multivariable predictive control algorithms, the controller can pre-calculate the optimal control commands across the entire operating range of the propulsion system, from low-power cruise to high-power maneuvering. This enables rapid, accurate, and stable closed-loop control of the three controlled variables: hydrogen flow rate, pressure, and temperature. This is ensured through the algorithm's rolling optimization and feedback correction mechanisms. Even when there is a slow drift in the external environment or system characteristics, the control quality can be maintained, thus comprehensively meeting the complex and dynamic operational requirements of the propulsion system.

[0015] Through deep integration with the UAV flight control system, unified status monitoring, fault diagnosis, predictive maintenance, and flight-propulsion coordinated control are achieved, significantly improving the system's intelligence level and overall flight safety. During flight, the flight control system can dynamically optimize the flight trajectory and power distribution strategy based on mission requirements and the real-time capabilities of the propulsion system, achieving true flight-propulsion coordinated control. This deep integration fundamentally breaks through the limitations of traditional information silos between systems, significantly improving the intelligence level, mission adaptability, and overall flight safety of the entire UAV platform. Attached Figure Description

[0016] Figure 1 A schematic diagram of a liquid hydrogen storage and supply system for a drone provided for an embodiment; Figure 2 Control logic block diagram of the liquid hydrogen supply closed-loop intelligent control method provided in the embodiment; Figure 3 This is a data interaction architecture diagram illustrating the deep integration of a liquid hydrogen storage and supply system with a drone flight control system, provided in this embodiment. Figure 4 A diagram illustrating the collaborative control architecture between a liquid hydrogen storage and supply system and a drone flight control system, provided for an embodiment. The components are as follows: 101, inner liner of the vacuum-insulated composite storage tank; 102, outer shell of the vacuum-insulated composite storage tank; 103, multilayer vacuum insulation material (MLI); 104, low thermal conductivity composite material support structure; 201, electric heat exchanger; 202, miniature spiral heat pump; 3011, first pressure sensor; 3012, second pressure sensor; 3013, third pressure sensor; 3021, first temperature sensor; 3022, second temperature sensor; 3023, third temperature sensor; 303, mass flow meter; 401, actuator proportional valve (PV); 402, variable capacity pressure regulating valve (VRV); 403, adjustable power electric heater (HE); 500, liquid hydrogen supply control unit (LCU); 601, overpressure safety protection valve. Detailed Implementation

[0017] The specific embodiments of the present invention are described below to enable those skilled in the art to understand the present invention. However, it should be understood that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, various changes are obvious as long as they are within the spirit and scope of the present invention as defined and determined by the appended claims. All inventions utilizing the concept of the present invention are protected.

[0018] In one embodiment of the present invention, a liquid hydrogen storage and supply system for a drone is provided, such as Figure 1As shown, the vacuum-insulated composite storage tank (VACS) adopts a lightweight and high-strength design. Its structure is a double-walled vacuum jacket. The inner liner 101 of the vacuum-insulated composite storage tank is made of high-strength aluminum-lithium alloy (Al-Li) or titanium alloy with low-temperature resistance and good hydrogen compatibility to withstand the pressure and temperature shock of cryogenic liquid hydrogen. The outer shell 102 of the vacuum-insulated composite storage tank is made of carbon fiber reinforced polymer (CFRP) to provide high specific strength and specific stiffness and resist external loads. A high-vacuum environment exists between the two walls, and multi-layer vacuum insulation material (MLI) 103 and a low thermal conductivity composite material support structure 104 are arranged to minimize heat conduction and heat radiation.

[0019] The Liquid Hydrogen Supply Control Unit (LCU) 500, serving as the system's "brain," employs a dual-core heterogeneous architecture comprised of a highly reliable industrial-grade microcontroller (MCU) and a field-programmable gate array (FPGA). The MCU is responsible for running complex control algorithms and communication protocols with the flight controller, while the FPGA handles high-speed data acquisition, actuator driving, and hardware-level safety interlocking logic. The LCU runs the core closed-loop intelligent control algorithm of this invention, such as... Figure 2 As shown, it includes an input module, an MPC control module, an adaptive gain scheduling module, a control allocation module, and an output module; The input module is used to receive target setpoints from the UAV flight control system and flow, pressure and temperature data collected in real time from the sensor array; The MPC control module performs rolling optimization calculations and includes a state predictor and an optimization solver. The state predictor, based on the thermodynamic equations of liquid hydrogen evaporation, the pipeline fluid pressure drop model, and the dynamic response characteristics of proportional and regulating valves, constructs a discrete-time state-space model describing the dynamic coupling relationship between flow rate, pressure, and temperature in the liquid hydrogen storage and supply system through system identification or mechanistic modeling. This discrete-time state-space model is then used to predict the behavior of the liquid hydrogen storage and supply system within a finite future time window, yielding the prediction results. The optimization solver solves the optimization problem of minimizing the objective function based on the prediction results, generating the optimal control sequence (including proportional valve opening, pressure regulating valve position, and heater power).

[0020] Within each control cycle, the rolling optimization calculation steps include: real-time acquisition of hydrogen flow rate, pressure, and temperature data in the liquid hydrogen storage and supply system; based on the real-time acquired flow rate, pressure, and temperature data, using a discrete-time state-space model to predict the behavior of the liquid hydrogen storage and supply system within a future finite time window, and obtaining the prediction results; solving the optimization problem of minimizing the objective function based on the prediction results, generating the optimal control sequence; and applying the first control action in the optimal control sequence to the actuator.

[0021] The discrete-time state-space model typically consists of the following two equations: Equations of state: ; Output equation: ; In the formula, Indicates the index of discrete time, for example Indicates the current moment. This indicates the time of the next control cycle. The state vector is a column vector containing n state variables that completely describe the system at time t. The internal state of the system. The choice of state variables is crucial; they should be the smallest set of variables that can determine the future behavior of the system. In this embodiment, the pressure inside the storage tank, the temperature of the hydrogen at the outlet of the supply pipeline, and the mass flow rate at the hydrogen outlet are selected to construct a three-dimensional state vector. The input vector, or control vector, is a column vector containing r control inputs. These represent externally applied control quantities that can be actively adjusted by the system, such as valve opening, pump speed, and heater power. In this embodiment, the opening of the flow control valve, the power of the vaporizer / heater, and the power or speed of the booster pump are used as the input vectors. The output vector is a column vector containing p output variables, representing system physical quantities that can be directly measured by sensors. In this embodiment, the output vector is consistent with the state vector. This is the state matrix, which describes the interactions between the internal states of the system. The input matrix describes how each control input affects every state variable within the system. The output matrix describes how the system's internal state variables combine to form the final measurable output. A direct transfer matrix or feedforward matrix describes the direct, instantaneous effect of the input vector on the output vector, i.e., without the transfer of state variables.

[0022] The objective function is:

[0023] In the formula, Describe the objective function. Indicates the length of the prediction time domain. Indicates the time index. Indicates that the system is in Real-time output (flow rate, pressure, temperature). This indicates the target value requested by the flight control system. To represent the square of the Euclidean norm, Indicates that the system is in Real-time control inputs (valve opening, heating power). As the weight of the tracking error, The weighting for controlling energy consumption is used to balance tracking accuracy and control motion smoothness. This represents the tracking error term, which measures the difference between the predicted output and the reference trajectory. This represents the energy consumption control term; mathematically, it is... .

[0024] In this embodiment, when constructing the optimization problem, corresponding constraints are set for the liquid hydrogen storage and supply system, including: input variable constraints (e.g., valve opening must be between 0% and 100%, heater power cannot exceed its rated maximum power, and cannot be negative), input rate of change constraints (the response speed of the actuator is limited; for example, the valve cannot instantaneously go from fully closed to fully open), state variable constraints (for safety and process requirements, the internal state of the system must be maintained within a certain range. For example, the tank pressure cannot exceed its design pressure limit, nor can it be lower than the lower limit for maintaining the liquid phase; the hydrogen temperature must be within the allowable range of the material), and output variable constraints (sometimes the output is directly constrained, for example, the outlet hydrogen flow rate is required not to be lower than the minimum requirement of the fuel cell).

[0025] Combining the objective function with the constraints described above constitutes a complete constrained optimization problem that needs to be solved in each control cycle k. Since the objective function is quadratic and the constraints are linear, this type of problem is called a quadratic programming (QP) problem. Many mature numerical computation toolkits can efficiently solve QP problems.

[0026] In this embodiment, pressure and temperature are used as state vectors, heater power is used as input vectors, and the output vector is consistent with the state vectors. Through system identification, the following simplified discrete-time state-space model matrix (sampling time 1 second) is obtained: ; ; ; ; The specific parameter settings for the MPC control module are as follows: prediction time domain Control time domain Output weight matrix Input weight matrix Control increment .

[0027] Constraints include: Heater power: kW; Power change rate: kW / s; Pressure safety: MPa.

[0028] At the present moment The system's initial state was collected from the sensors, and the UAV flight control system issued the target settings. The current state is... The control quantity at the previous moment was The target value is set to .

[0029] The system state at three future time points is predicted using a state-space model; these predicted states will be the future control sequence. The function. Note that, because ,so ,Right now .

[0030] The variable that needs optimization is the control increment sequence: ; ; ; .

[0031] predict The state at any given moment:

[0032]

[0033] .

[0034] predict The state at any given moment: ; Substituting the above equation into... and The expression, after being rearranged, will yield a result about and The linear expression: .

[0035] predict The state at any given moment: .

[0036] Based on the above predictions, we solve the optimization problem of minimizing the objective function. Substituting the predictions into the objective function, the final objective function J can be rearranged into a standard quadratic programming (QP) form:

[0037] Among them, matrix sum vector It is a system matrix Weight Current state and target value The constant obtained from the calculation.

[0038] At the same time, the constraints must also be transformed into conditions concerning Linear inequalities:

[0039]

[0040]

[0041] .

[0042] Given a standard QP problem, input the QP problem into a problem solver (e.g., the cvxpy or quadprog library in Python, or the quadprog function in MATLAB).

[0043] The solver will return a value that makes The minimum optimal control increment sequence that satisfies all constraints. Assume the solver provides the following result:

[0044] This result means: At the present moment The optimal control increment is +2.0 kW.

[0045] In the next moment The optimal control increment is +1.5 kW.

[0046] This solution is the "optimal" decision obtained after comprehensively considering multiple factors, such as reaching the target pressure of 12MPa and the target temperature of 85K as quickly as possible, while avoiding drastic changes in heater power and ensuring that the pressure does not exceed 15MPa.

[0047] Based on the rolling time-domain principle of MPC, only the first action in the optimal sequence is adopted to calculate the actual applied control quantity: ; The control system sends a command to the heater's power controller to set its power to 3.0 kW. The system operates at a heating power of 3.0 kW for one control cycle (1 second).

[0048] exist time: (1) Reacquire data: Measure the new system state, for example... (Note: This actual measurement may differ from previous predictions.) There are deviations because the model is imperfect or contains disturbances, which is precisely the significance of closed-loop control.

[0049] (2) Update target value: The flight control system may send a new target value. .

[0050] (3) Repeat the process: in a new state and new target value Starting from this point, calculate the new optimal control increment sequence. Then the first control action Apply to the system.

[0051] This "predict-optimize-execute-remeasure" cycle continues, enabling the system to converge intelligently toward the target state while adapting to changes in the target and external disturbances.

[0052] The Adaptive Gain Scheduling (AGS) module adjusts the weights of tracking error and control energy consumption in the objective function based on the received target setpoints. Specifically, it determines the current operating condition based on real-time collected hydrogen flow, pressure, and temperature data, according to pre-defined operating condition intervals. It then adaptively adjusts the weights of tracking error and control energy consumption based on the current operating condition. Specifically, when the hydrogen flow rate is higher than a set threshold, the weight of tracking error is increased to ensure rapid response; when the hydrogen flow rate is lower than the set threshold, the weight of control energy consumption is increased to achieve stable energy-saving control. For example, under high flow (high power) demand, the weight of tracking error is increased (…). To ensure rapid response; in low-flow (cruising) conditions, increase the weight given to energy consumption control. This is to achieve stable energy-saving control.

[0053] The power demand of the UAV propulsion system varies greatly, leading to frequent switching of the operating point (condition) of the liquid hydrogen supply system, resulting in significant nonlinearity in its dynamic characteristics. Traditional fixed-parameter MPC controllers struggle to maintain optimal performance under all operating conditions. The adaptive gain scheduling strategy introduced in this invention enables the controller to "intelligently" adapt to changing operating conditions, achieving robust optimal control across the entire flight envelope.

[0054] The control assignment module applies the first control action from the optimal control sequence to the actuator. The LCU applies only the first control action of the calculated control sequence to the actuator. In the next control cycle, the LCU uses the latest system state feedback from the sensors to correct the model and repeats the rolling optimization process. This mechanism enables MPC to effectively handle model mismatch and external disturbances.

[0055] In the next control cycle, hydrogen flow rate, pressure and temperature data in the liquid hydrogen storage and supply system are collected again. Based on the new data, a discrete-time state-space model is used to predict and optimize the problem, so as to continuously correct the decision based on the new information and form a rolling closed loop in the time domain.

[0056] The output module uploads real-time hydrogen flow, pressure, and temperature data to the UAV flight control system. Using the MAVLink-H2 protocol, the LCU reports detailed status information across more than 30 dimensions to the FCU at a high frequency (e.g., 50Hz), including tank pressure, estimated level, hydrogen temperature, flow rate, valve opening, heater power, BOG estimation rate, and system health index. The FCU, in turn, sends precise power demand curves (rather than single power points) to the LCU based on flight mission and attitude control requirements, and can even include power predictions for the next few seconds.

[0057] The liquid hydrogen supply control unit (LCU) incorporates FDIR logic for fault detection of the sensor array. This FDIR logic can detect dozens of potential faults, including sensor failure, actuator jamming, pipeline leaks, and abnormal temperatures. Upon detection, the LCU immediately reports the fault code and severity level to the FCU via MAVLink-H2 alarm messages. Based on the received information, the FCU's fault management system can trigger emergency procedures for the UAV (such as switching to backup power, automatic return-to-home, or emergency landing) and instruct the LCU to enter appropriate degraded operating modes (such as activating redundant sensors or limiting maximum output power), forming a closed-loop fault-tolerant control system of "diagnosis-decision-execution," significantly improving flight safety. The FCU can perform global energy optimization based on mission range, remaining hydrogen (accurately estimated by the LCU), and current flight status. For example, during the return phase, the FCU can instruct the LCU to supply hydrogen in the most energy-efficient mode, thereby maximizing remaining range and ensuring safe recovery.

[0058] like Figure 3As shown, the LCU (Liquid Hydrogen Supply Control Unit) connects to the UAV Flight Control Computer (FCU) via a high-speed, reliable CAN-FD bus (communication rate set to 500 kbps). The communication protocol uses a custom message set based on an extension of the MAVLink protocol, named MAVLink-H2. This message set uses a custom ID range (e.g., 200-210) to define a bidirectional data interaction format including propulsion power requests, system status parameter reporting, detailed fault diagnosis codes, and emergency stop commands, achieving deep coupling between this system and the flight control system.

[0059] The actuator is used to precisely adjust the fluid state according to relevant instructions issued by the liquid hydrogen supply control unit, including: (1) A high-speed response actuator proportional valve (PV) 401 with a flow rate adjustment range of 0~1g / s is installed on the gaseous hydrogen output pipeline to precisely adjust the supply flow rate according to the control command.

[0060] (2) The variable pressure regulating valve (VRV) 402 uses an electric screw valve, which is installed between the evaporator and the proportional valve. It actively adjusts the output pressure by changing the flow volume in the valve, thereby achieving decoupled control of pressure and flow.

[0061] The integrated liquid hydrogen evaporation / vaporization unit design includes a compact electrothermal heat exchanger 201 driven by an adjustable power electric heater (HE) 403 and a miniature spiral heat pump 202 for auxiliary pressurization and circulation. This unit can controllably vaporize 20 K liquid hydrogen from the storage tank and precisely adjust the temperature of the output gaseous hydrogen (GH2) to any setpoint within the range of 30 K to 50 K as needed, to suit the optimal inlet temperature of the fuel cell. The electric heater has a maximum power of 0.5 kW, sufficient to handle the vaporization requirements at maximum flow. The adjustable power electric heater (HE) 403 is integrated with the evaporator and precisely adjusts the heating power via a PWM signal, thereby controlling the hydrogen vaporization rate and outlet temperature.

[0062] The sensor array is used to monitor key system status parameters in real time, acquiring hydrogen flow rate, pressure, and temperature data, including: (1) Three pressure sensors, including a first pressure sensor 3011, a second pressure sensor 3012, and a third pressure sensor 3013, with a measurement range of 0–15 bar and an accuracy of ±0.01 bar. They are respectively arranged inside the storage tank, at the evaporator inlet, and on the final gaseous hydrogen output pipeline to monitor the storage tank pressure, the gasification process pressure, and the supply pressure.

[0063] (2) Three temperature sensors, including a first temperature sensor 3021, a second temperature sensor 3022, and a third temperature sensor 3023, are high-precision low-temperature sensors with a measurement range of 10 K to 300 K and an accuracy better than ±0.1 K. They are respectively arranged on the inner wall of the storage tank, the inlet of the evaporator, and the gaseous hydrogen output pipeline to monitor the liquid hydrogen temperature, vaporization temperature, and supply temperature.

[0064] (3) A mass flow meter 303, which is a thermal micro-flow meter, has a measurement range of 0~1g / s (corresponding to the maximum hydrogen consumption of a fuel cell of about 30kW) and an accuracy of ±0.5%FS. It is installed at the gaseous hydrogen output end to accurately measure the mass flow rate of hydrogen supplied to the fuel cell.

[0065] Safety protection subsystem: Provides multi-level safety protection, including a mechanical overpressure safety valve 601 with a set pressure of 11 bar, a cryogenic alarm with a set temperature below 15 K, redundant configuration of key sensors (such as dual pressure / temperature sensors), and fault detection and intelligent fault-tolerant switching logic (FDIR) implemented in the LCU. When the internal system pressure exceeds the set threshold, it automatically opens to release pressure, preventing the equipment from exploding or being structurally damaged due to excessive pressure. When the temperature falls below the safe range, an alarm is triggered. Multiple sensors simultaneously monitor the same parameter (such as pressure and temperature) to avoid system misjudgment due to the failure of a single sensor.

[0066] This embodiment applies the liquid hydrogen storage and supply system and its closed-loop intelligent control method to a quadcopter drone with a maximum takeoff weight of 600 kg and a peak propulsion power of 30 kW, to illustrate the implementation process of the present invention in detail.

[0067] After the drone is powered on, the liquid hydrogen supply control unit (LCU) 500 first performs system initialization and self-test, checking whether the sensor array and actuators are normal. At the same time, the liquid hydrogen supply control unit (LCU) 500 establishes communication with the drone flight control system (FCU) through the CAN-FD interface, exchanges information, and confirms that the MAVLink-H2 protocol version is compatible.

[0068] Before takeoff, the FCU sends a "standby" command, and the LCU controls the variable pressure regulating valve (VRV) 402 and the actuator proportional valve (PV) 401 to be closed. At the same time, the adjustable power electric heater (HE) 403 in the power regulation unit is activated to slightly preheat the evaporator, stabilizing its outlet temperature at the minimum inlet temperature required by the fuel cell (e.g., 40K). At this time, the liquid hydrogen in the vacuum insulated composite storage tank (VACS) generates a slight positive pressure due to natural evaporation.

[0069] During takeoff and climb (high power demand) phases, the FCU plans the takeoff route and sends a power demand curve to the LCU, showing that the power will linearly climb from 0kW to 30kW within 2 seconds. The AGS module in the LCU detects the high flow demand and immediately increases the weight matrix Q for tracking error in the MPC controller, sacrificing some smoothness for the fastest response speed. Based on the FCU's target flow rate (~1g / s), pressure (~10bar), and temperature (~40K) setpoints, and combined with real-time feedback from the first pressure sensor 3011, the second pressure sensor 3012, and the third pressure sensor 3013, the MPC controller performs a rolling optimization calculation every 10ms. The calculation results are distributed to the actuators. The proportional valve (PV) 401 of the actuators quickly opens to a large opening, the variable capacity pressure regulating valve (VRV) 402 coordinates to stabilize the output pressure, and the adjustable power electric heater (HE) 403 operates at near full power (0.5 kW) to ensure rapid and sufficient vaporization of liquid hydrogen. Throughout the process, the LCU reports actual parameters such as flow rate, pressure, and temperature to the FCU at a frequency of 50Hz in real time. The FCU can then fine-tune the motor speed according to the actual supply situation, achieving close coordination between flight control and propulsion.

[0070] During the cruise phase (low to medium power stable demand), the UAV reaches its cruise altitude, and the power demand transmitted by the FCU stabilizes at around 8 kW. The AGS module detects that the flow rate has stabilized at a low level (~1.3 g / s) and increases the weight matrix R of the control energy consumption in the MPC controller, making control actions (valve adjustment, heating power fluctuations) smoother, thus saving energy and extending the life of the actuators. The MPC controller accurately maintains the supply parameters near the target value with minimal fluctuations, ensuring the stable and efficient operation of the fuel cell. During this phase, the LCU accurately estimates the liquid level and BOG (naturally evaporating gas) production rate in the tank based on changes in tank pressure and temperature, combined with the amount of hydrogen consumed, and reports this key information to the FCU for global range planning and energy management.

[0071] Assuming that during cruise, the third pressure sensor 3013 malfunctions, its reading instantly jumps to 0. The LCU's FDIR logic detects within two cycles (20ms) that the deviation between this reading and the model prediction and redundant sensor readings exceeds a threshold, determining that the sensor has failed. The LCU immediately reports the fault code "P_SENS_FAIL_01" to the FCU via a MAVLink-H2 message (ID 205: Fault Diagnostic Report). Simultaneously, the LCU's control law seamlessly switches to using the reading from the second pressure sensor 3012 at the evaporator inlet (compensated by the pipeline pressure drop model) as pressure feedback to ensure control loop stability. The system enters degraded operation mode. Upon receiving the fault code, the FCU displays a "Hydrogen supply system pressure sensor redundancy" warning to the ground station operator on its cockpit interface and, according to preset rules, decides whether to continue the current mission (because the system has successfully tolerated the fault) or plan an early return.

[0072] Mission accomplished, the drone prepares to land. The power demand sent by the FCU gradually decreases. The LCU smoothly reduces the opening of the proportional valve (PV) 401 and the power of the adjustable power electric heater (HE) 403. After the drone lands safely, the FCU sends a "shutdown" command. The LCU controls all valves to close, stops heating, and sends a final status packet containing total hydrogen consumption and a system health summary to the FCU before entering hibernation mode.

[0073] In this embodiment, the system control of the liquid hydrogen storage and supply system and the UAV flight control system is as follows: Figure 4 As shown, a three-layer structure of decision-coordination-execution is used to achieve deep coupling control between the flight control system and the hydrogen power system.

[0074] The flight mission decision layer is the top-level intelligent decision-making center of the UAV, mainly including mission management and mode switching for takeoff, cruise, maneuvering, glide, landing, and emergency response; the dynamic trajectory planning part performs dynamic planning based on hydrogen energy supply and real-time status such as mission and weather; the energy management strategy makes real-time adjustments to energy utilization efficiency; the fault arbitration decision receives the health status of the power system and the flight control decides whether to change the mode.

[0075] The power coordination control layer is the core of the multi-parameter closed-loop control of this invention. It optimizes the ratio of hydrogen and electric power output based on the MPC model predictive control algorithm, and receives flight control parameters and commands to adjust the supply of liquid hydrogen and hydrogen gas in a timely manner. The thermal management system maintains the temperature of liquid hydrogen storage, gaseous hydrogen combustion and fuel cell stack according to energy demand. The fault diagnosis and alarm system integrates feedback data from various sensors, evaluates the overall health status of the system in real time, and reports to the flight mission decision layer.

[0076] The underlying execution layer contains all the physical actuators, tanks, valves, fuel cells, and sensors of the entire system.

[0077] This invention employs an advanced multivariable predictive control algorithm to achieve rapid, accurate, and stable closed-loop control of hydrogen flow rate, pressure, and temperature, meeting the dynamic requirements of the propulsion system across its entire power range. Through deep integration with the UAV flight control system, it achieves unified status monitoring, fault diagnosis, predictive maintenance, and flight-propulsion coordinated control, significantly improving the system's intelligence level and overall flight safety.

Claims

1. A closed-loop intelligent control method for a liquid hydrogen storage and supply system for unmanned aerial vehicles (UAVs), characterized in that, include: Construct a discrete-time state-space model to describe the dynamic coupling relationship between flow rate, pressure, and temperature in a liquid hydrogen storage and supply system; Construct an optimization problem that minimizes the objective function; Receive target setpoints sent by the UAV flight control system; adjust the weight of tracking error or control energy consumption based on the received target setpoints; Based on the discrete-time state-space model, rolling optimization calculations are performed on the optimization problem according to the objective function with adjusted weights and the real-time collected flow, pressure and temperature data of the liquid hydrogen storage and supply system; the calculation results are then distributed to the actuators.

2. The method according to claim 1, characterized in that, Within each control cycle, the rolling optimization calculation steps include: Real-time acquisition of hydrogen flow rate, pressure, and temperature data in the liquid hydrogen storage and supply system; Based on real-time collected flow, pressure and temperature data, the behavior of the liquid hydrogen storage and supply system within a finite time window in the future is predicted using a discrete-time state-space model, and the prediction results are obtained. Based on the prediction results, an optimization problem that minimizes the objective function is solved to generate the optimal control sequence; Apply the first control action in the optimal control sequence to the actuator.

3. The method according to claim 1, characterized in that, The objective function is: In the formula, Describe the objective function. Indicates the length of the prediction time domain. Indicates the time index. Indicates that the system is in Real-time outputs, including flow rate, pressure, and temperature; This indicates the target value requested by the flight control system. To represent the square of the Euclidean norm, Indicates that the system is in Real-time control inputs include valve opening degree and heating power; As the weight of the tracking error, The weighting of energy consumption is used to balance tracking accuracy and control smoothness. Represents the tracking error term. This indicates the energy consumption control item.

4. The method according to claim 2, characterized in that, The weights of tracking error and control energy consumption in the objective function are adjusted using an adaptive gain scheduling strategy, specifically as follows: Based on real-time collected hydrogen flow, pressure and temperature data, the current working condition is determined according to the pre-set working condition division intervals. The weights of tracking error and energy consumption are adaptively adjusted based on the current operating conditions. Specifically, when the hydrogen flow rate is higher than the set threshold, the weight of tracking error is increased; when the hydrogen flow rate is lower than the set threshold, the weight of energy consumption control is increased.

5. A liquid hydrogen storage and supply system implementing the method of any one of claims 1 to 4, characterized in that, include: Vacuum-insulated composite storage tank for storing liquid hydrogen; The liquid hydrogen supply control unit, as the control center of the liquid hydrogen storage and supply system, includes an input module, an MPC control module, an adaptive gain scheduling module, a control allocation module, and an output module. The input module is used to receive target setpoints from the UAV flight control system and flow, pressure and temperature data collected in real time from the sensor array; The MPC control module is used to perform rolling optimization calculations, including a state predictor and an optimization solver. The state predictor is used to construct a discrete-time state-space model describing the dynamic coupling relationship between flow rate, pressure, and temperature in the liquid hydrogen storage and supply system; construct an optimization problem that minimizes the objective function; use the discrete-time state-space model to predict the behavior of the liquid hydrogen storage and supply system within a finite time window in the future, and obtain the prediction result; the optimization solver is used to solve the optimization problem that minimizes the objective function based on the prediction result, and generate the optimal control sequence. The adaptive gain scheduling module is used to adjust the weight of tracking error or the weight of energy consumption control in the objective function according to the received target setpoint. The control allocation module is used to apply the first control action in the optimal control sequence to the actuator; The output module is used to upload real-time collected hydrogen flow, pressure and temperature data of the liquid hydrogen storage and supply system to the UAV flight control system. The actuators, including proportional valves and regulating valves, are used to adjust the fluid state according to relevant instructions issued by the liquid hydrogen supply control unit; The liquid hydrogen evaporation / vaporization unit is used to controllably vaporize the liquid hydrogen output from the vacuum insulated composite storage tank and adjust the temperature of the gaseous hydrogen. A sensor array is used to collect hydrogen flow, pressure, and temperature data in real time. The safety protection subsystem includes an overpressure safety valve, a cryogenic alarm, and redundant sensor configurations for monitoring pressure and temperature anomalies.

6. The system according to claim 5, characterized in that, The liquid hydrogen supply control unit has built-in FDIR logic for fault detection of the sensor array.

7. A computer device, characterized in that, It includes a memory and a processor. The memory stores a computer program. When the computer program is executed by the processor, it causes the processor to perform the steps of the closed-loop intelligent control method for a liquid hydrogen storage and supply system for an unmanned aerial vehicle as described in any one of claims 1 to 4.

8. A computer-readable storage medium, characterized in that, The system contains a computer program that, when executed by a processor, causes the processor to perform the steps of a closed-loop intelligent control method for a liquid hydrogen storage and supply system for unmanned aerial vehicles as described in any one of claims 1 to 4.

Images