Thermal management system for propulsion system of large-scale unmanned rotorcraft
Through the combination of modular heat dissipation structure, multi-sensor data fusion and dynamic PID algorithm, the thermal management problem of large rotor UAV propulsion systems under complex operating conditions is solved, efficient heat dissipation and accurate fault diagnosis are achieved, and the real-time and safety of the system are improved.
Patent Information
- Application Number
- CN202510483043.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-04-17
- Publication Date
- 2025-08-01
AI Technical Summary
In the thermal management of large rotor UAV propulsion systems, the existing technology cannot respond to dynamic changes in the flight stage in real time, lacks the ability to buffer transient thermal shock, a single temperature sensor cannot fully reflect the thermal state of the system, poor communication security, inaccurate fault diagnosis, and safety hazards.
The modular thermal dissipation structure design is adopted, combined with multi-sensor data fusion and LSTM neural network prediction model, and the dynamic PID algorithm realizes adaptive thermal dissipation power adjustment, dual-bus encrypted communication and three-level fault diagnosis mechanism to ensure the real-time and reliability of the system.
It realizes efficient heat dissipation of the propulsion system in high-altitude and high load environments, accurately diagnoses faults, ensures real-time safety and reliability of data transmission, and improves the safe operation capabilities of the drone.
Smart Images
Figure CN120397327A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of thermal management of unmanned aerial vehicles (UAVs), and in particular to a thermal management system for a propulsion system of a large rotary-wing UAV. Background Art
[0002] Large rotary-wing drones, due to their long flight times and high payload capacity, are widely used in logistics, emergency rescue, military reconnaissance, and other fields. Their propulsion systems, including key components like motors and controllers, generate significant heat during operation. This heat load increases dramatically during complex flight conditions, such as takeoff, climb, and high-load maneuvers. If this heat cannot be dissipated promptly, it can lead to component performance degradation, shortened lifespan, and even system failures or safety incidents.
[0003] Currently, existing thermal management solutions for UAV propulsion systems primarily rely on liquid cooling, combined with traditional control and monitoring technologies to achieve thermal regulation. Temperature sensors are used to collect the surface temperature of the motor or controller, and preset proportional (P), integral (I), and differential (D) parameters are used to adjust the water pump speed or fan power to control the coolant flow. For example, when the temperature exceeds a threshold, the water pump speed is increased to improve heat dissipation. However, these parameters remain unchanged during flight, relying solely on a single temperature signal for feedback regulation. Temperature sensors are used to collect the temperature of key components at a single point, coupled with a liquid cooling system consisting of a radiator, water pump, water tank, and liquid pipelines. Coolant circulates through the pipelines, removing heat through heat exchange with the outside air through the radiator. The water tank is typically a conventional liquid storage structure, relying on the coolant's sensible heat absorption to cope with thermal fluctuations, lacking the ability to buffer transient thermal shocks. Communication modules often use a single bus architecture and unencrypted protocols (such as Modbus RTU), transmitting only basic data such as temperature and flow. Fault diagnosis relies on a single sensor threshold, triggering an alarm when the temperature is abnormal, lacking the ability to integrate and analyze multi-dimensional data and accurately locate faults.
[0004] Fixed-parameter PID control cannot respond in real time to dynamic changes in flight phases (takeoff, climb, cruise, etc.), altitude, and motor load rate. The water tank of traditional liquid cooling solutions serves only as a liquid storage container and lacks the ability to dynamically adapt to heat dissipation power. When a drone encounters a transient thermal shock (such as a sudden high load), the coolant's sensible heat absorption capacity is limited, and the temperature fluctuates violently. A single temperature sensor cannot fully reflect the system's thermal state, and key parameters such as pressure and flow are not effectively utilized, resulting in a high risk of missed or misjudgment. The single-bus architecture has weak anti-interference capabilities, and non-encrypted data can be easily intercepted or tampered with. Data transmission delays are long, and there is a lack of a rapid switching mechanism for faulty buses, resulting in the flight control system being unable to obtain thermal management status information in a timely manner, posing a safety hazard.
[0005] In summary, due to the deficiencies of the existing technologies in aspects such as intelligent control, heat dissipation efficiency, fault diagnosis, and communication security, there is an urgent need for a thermal management system that can adapt to multiple working conditions, integrate multi-dimensional monitoring and prediction, and have high reliability to solve the heat dissipation and safe operation problems of the propulsion system of large rotor unmanned aerial vehicles in extreme environments. Summary of the Invention
[0006] Object of the Invention: In order to overcome the deficiencies of the existing technologies, the present invention provides a thermal management system for the propulsion system of large rotor unmanned aerial vehicles. Through the modular heat dissipation structure design, efficient heat dissipation and convenient maintenance in a compact space are achieved. The dynamic perception and early warning capabilities of the heat load under complex working conditions are enhanced by using multi-sensor data fusion and the LSTM neural network prediction model. The adaptive adjustment of the heat dissipation power in all flight stages is realized by combining the dynamic PID algorithm, and the real-time performance, reliability, and security of the system are ensured through dual-bus encrypted communication and a three-level fault diagnosis mechanism, thereby solving the problems of temperature control, fault location, and safe operation of the propulsion system under high load, multiple working conditions, and extreme environments.
[0007] A thermal management system for the propulsion system of large rotor unmanned aerial vehicles constructed by the present invention includes a heat dissipation device, which includes a radiator, a water pump, a water tank, and a liquid pipeline connected through a standardized quick-release interface. The liquid pipeline is provided with a branch with a filter and a check valve; the installation space of the radiator does not exceed 300 mm × 190 mm × 150 mm, the installation space of the water pump does not exceed 120 mm × 70 mm × 100 mm, and the water tank is located at the rear of the radiator; the diameter of the liquid pipeline is 10 mm and the total length does not exceed 1 m. The water tank is internally provided with a paraffin-based phase change material unit with a gradient porosity and a metal foam composite structure. The metal foam is distributed in an orthogonal network shape, and the volume ratio is 15%-25%, and it is evenly distributed in the phase change material unit. The liquid inlet and outlet of the water tank are respectively provided with temperature-controlled shape memory alloy valves;
[0008] A sensor module, on which temperature, pressure, flow, vibration, and infrared thermal imaging sensors are installed for collecting the working condition data of the motor, the water inlet pipe, the controller, the water pump, and the pipeline; the sensors output standard signals and transmit them to the control module; the temperature sensor has a measurement range of -40 °C to 120 °C, the pressure sensor has a measurement range of 0 to 1 MPa, and the flow sensor has a measurement range of 0 to 100 L / min;
[0009] A control module, which is internally provided with a preset heat dissipation power database for flight stages, a dynamic PID algorithm, and a heat load early warning mechanism based on the prediction model of the fusion of the LSTM neural network and fuzzy logic, for adjusting the water pump speed and driving auxiliary cooling; the flight stages include takeoff, climb, cruise, descent, and landing; among them, the heat dissipation power requirements for takeoff, climb, cruise, descent, and landing are 9 kW, 7 kW, 5 kW, 5 kW, and 5 kW;
[0010] A communication module. The communication module adopts a dual RS485 bus architecture, supports SM4 encryption and packet polling, realizes high-speed bidirectional communication with the flight control computer, and has a bus health self-diagnosis mechanism. It is connected to the radiator, water pump, water tank and liquid pipeline through a standardized quick-release interface, supporting flexible layout in a compact space; temperature, pressure and flow sensors monitor the working conditions of the motor, inlet pipe and controller in real time; the dual RS485 bus architecture combines SM4 encryption and TDMA time division multiplexing to ensure the real-time performance (delay < 15ms) and anti-interference ability of data transmission.
[0011] Further, in the dynamic PID algorithm described in this application, the altitude H, motor load rate L, flight acceleration a and attitude angle θ are introduced, and their proportionality coefficient and integral coefficient are dynamically adjusted according to the following formula:
[0012] Kp = Kp0 × [1 + 0.2 × sin(π × H / 5000) + 0.1 × a + 0.05 × θ]
[0013] Ki = Ki0 / [1 + 0.5 × L + 0.2 × |sinθ|]
[0014] Among them, Kp0 is the initial value of the proportionality coefficient under the static condition at sea level, and its value range is 1.0 - 2.0; Kp1 is the initial value of the integral coefficient, and its value range is 0.5 - 1.0. When the temperature change rate > 2°C / s, the water pump speed is immediately increased to the preset threshold. The proportionality coefficient Kp is dynamically adjusted with the altitude H, and the formula introduces a sine function to compensate for the decrease in heat dissipation efficiency caused by the thin air at high altitudes. The integral coefficient Ki is inversely proportional to the motor load rate L to suppress the temperature overshoot at high loads.
[0015] Further, the sensor module in this application performs multi-sensor data fusion and calculates the heat load index (HLI) according to the following formula:
[0016] HLI = 0.6ΔT + 0.3ΔP + 0.1Q + 0.1V
[0017] Among them, ΔT is the real-time change rate of the motor temperature and the inlet pipe temperature (unit: °C / s), ΔP is the difference between the inlet pipe pressure and the rated working pressure (unit: MPa), Q is the coolant volume flow (unit: L / min), and V is the effective value of the vibration signal (vibration acceleration, unit: g); each parameter is weighted and calculated after normalization; when HLI ≥ 80, multi-level fault diagnosis is triggered.
[0018] Further, the multi-level fault diagnosis in the present application includes: the first level, verifying the flowmeter data by using the pipeline cross-sectional area and the coolant density, with an allowable deviation of ±5% FS. Meanwhile, analyzing the vibration sensor data to determine whether there is vibration interference in the flowmeter; the second level, performing FFT analysis on the energy of the pipeline acoustic signal in the frequency band of 100 - 500 Hz. When the cumulative energy value > the baseline value by 15 dB, it is determined that there is an abnormal bubble. Meanwhile, observing the temperature distribution on the pipeline surface by using an infrared thermal imaging sensor; the third level, positioning the overheated part by comparing the temperature differences of the three phases of the motor and combining the Hall sensor, vibration sensor, and infrared thermal imaging data, and generating a diagnostic code.
[0019] Further, the water tank of the heat dissipation device in the present application is internally provided with a paraffin-based phase change material unit. The porosity of this unit linearly transitions from 50% at the bottom to 80% at the top, and the difference in porosity between adjacent layers ≤ 5%. And it is equipped with a shape memory alloy actuator. When the coolant temperature > 50°C, the actuator is powered on (12V DC), and the honeycomb pore diameter is dynamically adjusted at a rate of 0.2 mm / s within the range of 0.5 mm to 2 mm (the adjustment step is 0.1 mm) until the temperature < 45°C and then resets; the porosity of the metal foam is 70% - 90%, the pore diameter is 0.2 mm - 0.5 mm, the phase change temperature of the paraffin-based phase change material unit is 50 - 60°C, and the latent heat ≥ 200 kJ / kg. The latent heat absorption efficiency is optimized through the gradient porosity design, and the metal foam composite structure improves the heat transfer performance.
[0020] Further, the communication module in the present application inserts a 16-byte dynamic check code (encrypted from the first 8 bytes of the data frame by using the SM4 algorithm) and a 4-byte timestamp into the data frame based on the Modbus RTU protocol, and adopts the preemptive fault frame transmission and TDMA time division multiplexing method (the TDMA period is 15 ms, the sensor module is allocated a 5 ms transmission time slot, the control module is allocated 8 ms, and 2 ms redundant time slots are reserved), to achieve a data transmission delay < 15 ms and anti-interference ability. The bus health self-diagnosis mechanism monitors the voltage, current, and signal quality of the bus in real time.
[0021] Further, the control module in the present application outputs according to the sensor data and the LSTM-fuzzy logic prediction model. When the predicted heat load index in the next 30 seconds ≥ 75, the preparatory cooling measure is started 200 ms in advance, the water pump speed is increased, and the auxiliary cooling fan is driven; the fuzzy logic module establishes a two-dimensional correction table: when the atmospheric humidity H ≥ 60% and the wind speed V < 2 m / s, the correction coefficient of the LSTM prediction result is +5%; when H ∈ [40%, 60%) and V ∈ [2 - 6 m / s], the correction coefficient is 0; when V ≥ 8 m / s, the correction coefficient is -5%.
[0022] Furthermore, a fault handling strategy library is also deployed in the control module of the present application. The fault diagnosis results are mapped to local loop isolation, redundant pump switching, and degraded air-cooling emergency operations. The fault handling strategy library selects corresponding handling measures according to the fault code and flight phase, supporting the UAV to continue flying after a fault or triggering an emergency landing instruction.
[0023] From the above technical solutions, it can be seen that the present invention has the following beneficial effects:
[0024] The thermal management system for the propulsion system of a large rotor UAV of the present invention realizes the efficient dynamic regulation of the thermal load of the propulsion system, the accurate diagnosis and positioning of complex faults, and the real-time safety and reliability of data transmission through the organic combination of modular heat dissipation structure design, multi-sensor data fusion technology, dynamic PID control algorithm, LSTM neural network prediction model, and dual-bus high-security communication mechanism. It significantly improves the heat dissipation efficiency, temperature control stability, and fault handling ability of the UAV in extreme environments such as high altitude and high load, providing an all-round technical guarantee for the safe and reliable operation of large rotor UAVs. BRIEF DESCRIPTION OF THE DRAWINGS
[0025] Figure 1 is the architecture diagram of the thermal management system for the propulsion system of a large rotor UAV of the present invention;
[0026] Figure 2 is the module interaction timing diagram (taking "thermal load warning during takeoff phase" as an example) of the thermal management system for the propulsion system of a large rotor UAV of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS
[0027] Embodiment
[0028] The embodiment of the present invention provides a thermal management system for the propulsion system of a large rotor UAV. Its overall architecture is as Figure 1 shown, including a heat dissipation device, a sensor module, a control module, and a communication module. Taking a certain six-rotor cargo UAV (takeoff weight 120 kg) performing a material transportation task at a plateau airport (altitude 3000 meters) as an application scenario, the implementation process of the thermal management system is described.
[0029] 1. Architecture implementation (refer to Figure 1 )
[0030] The thermal management system is integrated with the UAV propulsion system through a standardized quick-release interface:
[0031] Heat dissipation device: A flat radiator made of aluminum alloy (size 280mm×180mm×140mm) is installed on the top of the motor compartment, and copper liquid cooling pipelines (diameter 10mm) are arranged around the motor axis. The water tank (volume 2.5L) is manufactured by a layered injection molding process. The bottom is filled with a paraffin / expanded graphite composite phase change material with a porosity of 50% (phase change temperature 58°C, latent heat value 218kJ / kg), and the top is a lauric acid-based phase change material with a porosity of 80% (phase change temperature 42°C). A nickel-based foamed metal with a porosity of 85% and a pore diameter of 0.3mm is embedded inside.
[0032] Sensor module: PT1000 temperature sensors (range -40°C to 200°C) are arranged at the motor stator windings, MEMS pressure sensors (range 0 to 1.2MPa) are installed on the inlet pipe, turbine flow meters (accuracy ±1.5% FS) are set at the pipeline branch, and three-axis vibration sensors (frequency range 5Hz to 5kHz) are configured on the motor base.
[0033] Control module: An ARM Cortex-M7 processor is adopted, and the heat dissipation power curve of a typical flight profile (takeoff for 5 minutes, climb for 3 minutes, cruise for 25 minutes, descent for 4 minutes) is pre-stored. The input layer of the LSTM neural network contains 8 nodes (temperature, pressure, flow rate, vibration, altitude, attitude angle, motor current, ambient temperature and humidity), there are 3 hidden layers (number of nodes 32-16-8), and the output layer predicts the heat load value in the next 30 seconds.
[0034] Communication module: The dual RS485 bus architecture adopts a daisy chain topology, the baud rate is set to 115200bps, a dynamic SM4 key is inserted into each frame of data (updated every 5 seconds), and the bus health monitoring circuit detects signal attenuation in real time (an alarm is triggered when the threshold is set to a voltage <2.7V or noise >200mV).
[0035] 2. Workflow implementation (refer to Figure 2 )
[0036] Taking the takeoff stage as an example to illustrate the system operation:
[0037] (1) Initialization stage: After the flight control computer sends a takeoff command, the control module reads the preset database to load the basic heat dissipation power parameters (9kW) for the takeoff stage. After confirming the normal two-channel communication through bus health self-diagnosis, the main water pump (initial speed 2500rpm) is started.
[0038] (2) Data acquisition stage: The temperature sensor samples the motor winding temperature at a sampling rate of 100Hz (rising from 25°C to 82°C), the pressure sensor monitors the pressure fluctuation of the inlet pipe (0.35MPa ± 0.02MPa), and the vibration sensor detects abnormal vibration with a fundamental frequency of 120Hz (effective value 0.8g).
[0039] (3) Dynamic adjustment stage: The control module dynamically adjusts the PID parameters according to the altitude of 3000m, pitch angle, and motor load rate of 92% collected in real time, according to the formula:
[0040] Kp = 1.2×[1 + 0.2×sin(π×3000 / 5000) + 0.1×3.5 + 0.05×15] = 1.56
[0041] Ki = 0.8 / [1 + 0.5×0.92 + 0.2×|sin15°|] = 0.49
[0042] When the motor temperature change rate is detected to reach 2.8 °C / s, immediately increase the water pump speed to the threshold of 3800 rpm.
[0043] (4) Fault diagnosis stage: Calculate the heat load index
[0044] HLI = 0.6×2.8 + 0.3×0.12 + 0.1×18 + 0.1×0.8 = 83.6, triggering a three-level diagnosis:
[0045] First level: Verify the flowmeter data (theoretical value 20 L / min, measured value 18.5 L / min, deviation 7.5%), and combine vibration spectrum analysis to confirm that the sensor is interfered by mechanical vibration at 120 Hz, and enable software filtering compensation;
[0046] Second level: The acoustic sensor detects a sudden increase in energy in the 350 Hz frequency band of the pipeline (15 dB higher than the baseline), and the infrared thermal imager shows a 3 °C cold spot on the outer wall of the No. 3 pipeline, determining that there is air bubble accumulation in this branch;
[0047] Third level: Compare the temperature differences of the three phases of the motor U / V / W (82 °C / 79 °C / 85 °C), and combine the current fluctuation detected by the Hall sensor to locate the local overheating of the W-phase winding.
[0048] (5) Disposal stage: According to the fault code F1032 (F1 represents local overheating, 03 represents air bubble accumulation, 2 represents the third branch), the strategy library starts the emergency procedure: close the shape memory alloy valve of the third branch, switch to the redundant water pump, and at the same time increase the fan speed to 6500 rpm to ensure that the system heat dissipation power is maintained above 8.2 kW.
[0049] 3. Implementation details of key components
[0050] (1) Phase change material unit: Under the working condition at an altitude of 3000 m, when the coolant temperature rises to 50 °C, the high-density phase change material at the bottom starts to absorb heat (latent heat value 218 kJ / kg), and the low-density material at the top starts a secondary phase change when the coolant temperature exceeds 42 °C. The shape memory alloy actuator expands the honeycomb aperture from 1.2 mm to 1.8 mm according to the control signal, increasing the heat transfer surface area by 40%.
[0051] (2) Communication protocol implementation: Each frame of data (example: AA 09F1 00 01 02 03 04 05 06
[0052] 07 08 CRC16) inserts a dynamic verification code (generates a 16-byte ciphertext using the SM4 algorithm) and a timestamp (Unix time, accurate to milliseconds). The bus adopts a TDMA time slot allocation mechanism, with each 15 ms as a communication cycle, ensuring that the transmission delay of control instructions ≤ 12 ms.
[0053] (3) Prediction model training: Use historical flight data (100 flights) to train the LSTM network. The input features include 12-dimensional parameters such as motor temperature gradient, ambient pressure, and attitude angle change rate, and the output prediction error is controlled within ±4%. The fuzzy logic module corrects and compensates the prediction result by +5% according to the real-time wind speed (6 m / s) and relative humidity (45%).
[0054] 4. Implementation effect verification
[0055] In the plateau test of a certain type of UAV, this system shows the following advantages compared with the traditional scheme:
[0056] The motor temperature rise rate is reduced by 62% (from 4.2 °C / s to 1.6 °C / s)
[0057] The fault location accuracy is increased to 98.7% (the traditional scheme is 82%)
[0058] The communication interruption rate drops from 1.2% to 0.05%
[0059] The phase change material improves the transient thermal shock buffering ability by 3.8 times
[0060] The above implementation manners are exemplary, and their purpose is to illustrate the technical concept and characteristics of the present invention, so that those skilled in this field can understand the content of the present invention and implement it accordingly, and cannot be used to limit the protection scope of the present invention. All changes or modifications made according to the spirit and essence of the present invention should be covered within the protection scope of the present invention.
Claims
1. A thermal management system for a propulsion system of a large rotor unmanned aerial vehicle, characterized in that, Comprising: A heat dissipation device, including a radiator, a water pump, a water tank, and a liquid pipeline connected through a standardized quick-release interface. The liquid pipeline is provided with a branch with a filter and a check valve; the water tank is internally provided with a composite structure of a paraffin-based phase change material unit with a gradient porosity and metal foam. The metal foam is orthogonally reticulated and has a volume ratio of 15%-25%, and is evenly distributed in the phase change material unit. The inlet and outlet of the water tank are respectively provided with temperature-controlled shape memory alloy valves; A sensor module, which is equipped with temperature, pressure, flow rate, vibration, and infrared thermal imaging sensors, and is used to collect the working condition data of the motor, the inlet pipe, the controller, the water pump, and the pipeline; A control module, which is internally provided with a preset heat dissipation power database for different flight stages, a dynamic PID algorithm, and a heat load warning mechanism based on a prediction model that combines an LSTM neural network and fuzzy logic, and is used to adjust the water pump speed and drive auxiliary cooling; A communication module, which adopts a dual RS485 bus architecture, supports SM4 encryption and packet polling, realizes high-speed two-way communication with the flight control computer, and has a bus health self-diagnosis mechanism.
2. The thermal management system for a propulsion system of a large rotor unmanned aerial vehicle according to claim 1, characterized in that In the dynamic PID algorithm, the altitude H, the motor load rate L, the flight acceleration a, and the attitude angle θ are introduced, and its proportionality coefficient Kp and integral coefficient Ki are dynamically adjusted according to the following formula: Kp = Kp0 × [1 + 0.2 × sin(π × H / 5000) + 0.1 × a + 0.05 × θ] Ki = Ki0 / [1 + 0.5 × L + 0.2 × ∣sinθ∣] Where, Kp0 is the initial value of the proportionality coefficient under the static condition at sea level, and its value range is 1.0-2.0; Ki1 is the initial value of the integral coefficient, and its value range is 0.5-1.
0. When the temperature change rate > 2°C / s, immediately increase the water pump speed to the preset threshold.
3. The thermal management system for a propulsion system of a large rotor unmanned aerial vehicle according to claim 1, wherein, The sensor module performs multi-sensor data fusion and calculates the heat load index (HLI) according to the following formula: Calculate the heat load index HLI = 0.6ΔT + 0.3ΔP + 0.1Q + 0.1V Where ΔT is the real-time change rate of the motor temperature and the inlet pipe temperature, ΔP is the difference between the inlet pipe pressure and the rated working pressure, Q is the volume flow rate of the coolant, and V is the effective value of the vibration signal; each parameter is weighted and calculated after being normalized; when HLI ≥ 80, trigger multi-level fault diagnosis.
4. The thermal management system for a propulsion system of a large rotor unmanned aerial vehicle according to claim 3, characterized in that, The multi-level fault diagnosis includes: the first level, using the cross-sectional area of the pipeline and the density of the coolant to verify the flowmeter data, and at the same time analyzing the vibration sensor data to judge whether there is vibration interference in the flowmeter; the second level, performing FFT analysis on the energy of the pipeline acoustic signal in the frequency band of 100-500Hz to judge bubble abnormalities, and at the same time using the infrared thermal imaging sensor to observe the temperature distribution on the surface of the pipeline; the third level, positioning the overheated part by comparing the temperature difference of the three phases of the motor and combining the Hall sensor, vibration sensor, and infrared thermal imaging data, and generating a diagnostic code.
5. The thermal management system for a large rotor UAV propulsion system according to claim 1, characterized in that, The water tank of the heat dissipation device is internally provided with a paraffin-based phase change material unit. The porosity of this unit transitions from 50% at the bottom to 80% at the top, and is configured with a shape memory alloy actuator to dynamically adjust the honeycomb pore diameter within the range of 0.5 mm to 2 mm according to the real-time heat dissipation power. The porosity of the metal foam is 70%-90%, and the pore diameter is 0.2 mm-0.5 mm.
6. The thermal management system for a propulsion system of a large rotor unmanned aerial vehicle according to claim 1, wherein The communication module inserts a 16-byte dynamic check code and a 4-byte timestamp into the Modbus RTU protocol data frame, and adopts a preemptive fault frame transmission and TDMA time division multiplexing method to achieve a data transmission delay <15 ms and anti-interference ability. The bus health self-diagnosis mechanism monitors the voltage, current and signal quality of the bus in real time.
7. A thermal management system for a propulsion system of a large rotor unmanned aerial vehicle according to claim 1, characterized in that, The control module outputs according to the sensor data and the LSTM-fuzzy logic prediction model. When the predicted heat load index in the next 30 seconds is ≥75, it starts the preparatory cooling measures 200 ms in advance, increases the pump speed and drives the auxiliary heat dissipation fan. The fuzzy logic module corrects the prediction result of LSTM according to the atmospheric humidity and wind speed.
8. A thermal management system for a propulsion system of a large rotor unmanned aerial vehicle according to any one of claims 1-7, characterized in that The control module is also deployed with a fault handling strategy library, which maps the fault diagnosis results to local loop isolation, redundant pump switching and degraded air-cooled emergency operations. The fault handling strategy library selects corresponding handling measures according to the fault code and flight phase.