Energy-power conversion system of unmanned aerial vehicle

Abstract

The invention discloses an unmanned aerial vehicle energy-power conversion system, and relates to the technical field of unmanned aerial vehicle energy, and the system comprises a battery module which is formed by combining a plurality of single batteries in a series connection or parallel connection manner, and is responsible for providing electric energy required by the flight of an unmanned aerial vehicle. And the energy management module performs real-time monitoring, optimal distribution and intelligent regulation and control on energy use of the unmanned aerial vehicle through an integrated sensor, algorithm and control technology, and comprises a data acquisition unit, a charging management unit and a power conversion unit. The charging management unit is used for managing an unmanned aerial vehicle battery charging process, externally connecting a charging power supply and controlling an unmanned aerial vehicle battery charging strategy according to the battery power reserve level; and the power conversion unit adopts a dynamic power distribution strategy, implements closed-loop control based on the real-time power spectrum characteristics of the driving module, and realizes on-demand power output adjustment. And the configuration of longer endurance time of the same power supply weight or fewer batteries under the same endurance time is realized.

Description

Technical Field

[0001] The present invention relates to the technical field of unmanned aerial vehicle energy, and particularly to an energy-power conversion system for unmanned aerial vehicles. Background Art

[0002] With the rapid development of the low-altitude economy and the field of intelligent equipment, the energy system of unmanned aerial vehicles (UAVs) has become the core bottleneck restricting their performance breakthrough. Traditional drones are limited by the battery energy density and power conversion efficiency, and generally have problems such as short endurance, weak payload capacity, and poor environmental adaptability. In this context, the technological innovation of the energy-power conversion system has become the focus of the industry. Its core goal is to achieve a qualitative leap in the energy utilization efficiency of drones through efficient energy management and dynamic power distribution.

[0003] Currently, most mainstream drones use lithium polymer batteries (Li-Po) as the power source, with an energy density of about 250-300 Wh / kg, but the actual endurance is still difficult to break through the limitations of 30 minutes (consumer grade) to several hours (industrial grade).

[0004] Therefore, the present invention provides an energy-power conversion system for unmanned aerial vehicles. Summary of the Invention

[0005] (I) Technical Problems to be Solved

[0006] In view of the deficiencies of the prior art, the present invention provides an energy-power conversion system for unmanned aerial vehicles. The present invention adopts a dynamic power distribution strategy and implements closed-loop control based on the real-time power spectrum characteristics of the drive module to achieve on-demand power output regulation. The dynamic power distribution strategy breaks through the limitations of traditional static configurations and flexibly adjusts the power output by real-time monitoring of channel conditions, user requirements, and interference situations. The closed-loop control based on the real-time power spectrum characteristics of the drive module extracts key characteristics such as the peak frequency and band energy of the signal through Fourier transform or FFT algorithm to achieve precise perception of the system state. The closed-loop control system feeds back the output quantity (such as power, voltage) to the input end, compares it with the set value, and automatically corrects the deviation. For example, in an industrial power supply system, the closed-loop control monitors the output voltage fluctuation, adjusts the power supply working state in real time, ensures that the output is stable at the set value, suppresses harmonic interference, and improves the power supply quality. The on-demand power output regulation realizes the refined adjustment of the power output through technologies such as PID control and frequency conversion control, achieving a longer endurance time with the same power supply weight or a configuration with fewer batteries under the same endurance time, thereby solving the technical problems recorded in the background art.

[0007] (II) Technical Solutions

[0008] To achieve the above objectives, the present invention is realized through the following technical solutions: An energy-power conversion system for unmanned aerial vehicles, comprising:

[0009] The battery module is composed of multiple battery cells combined in series or parallel, and is responsible for providing the electrical energy required for the drone to fly.

[0010] The energy management module, through integrating sensors, algorithms and control technologies, conducts real-time monitoring, optimized allocation and intelligent regulation of the energy usage of the drone, including a data acquisition unit, a charging management unit and a power conversion unit.

[0011] The data acquisition unit uses high-precision Hall effect sensors or shunts to capture transient current changes, synchronously measures the voltages of each power supply module through a differential voltage sampling circuit, monitors the battery temperature changes, and monitors the attitude, speed and position of the drone in real time through an IMU inertial measurement unit, and records power data through the flight control system FCU;

[0012] The charging management unit is used to manage the charging process of the drone battery, connect to an external charging power supply and control the charging strategy of the drone battery according to the battery charge reserve level;

[0013] The power conversion unit adopts a dynamic power distribution strategy, implements closed-loop control based on the real-time power spectrum characteristics of the drive module, and realizes the adjustment of on-demand power output.

[0014] Furthermore, high-precision Hall effect sensors or shunts are used, with a typical accuracy of ±0.5% FS, a bandwidth covering 0 - 10 kHz, and a sampling rate ≥100 Hz to capture transient current changes, and differential input, shielded cables and digital filtering algorithms are used to suppress electromagnetic interference.

[0015] Furthermore, the voltages of each power supply module are synchronously measured through a differential voltage sampling circuit, with a sampling frequency of 1 kHz. NTC thermistors are surface-mounted on the battery pack to monitor temperature changes, and the attitude, speed and position of the drone are monitored in real time through an IMU inertial measurement unit.

[0016] Power data is recorded through the flight control system FCU, including information such as timestamps, instantaneous power values, flight modes, etc. The flight modes include takeoff, cruise and landing.

[0017] Furthermore, the charging of the drone battery is specifically as follows:

[0018] During the fast charging stage with SOC 0 - 80%, the constant current CC mode is adopted to charge with the maximum allowable current, and at the same time, the input power is dynamically adjusted to avoid power supply overload.

[0019] During the slow charging stage with SOC 80% - 100%, the constant voltage CV mode is adopted, the voltage is constant, the current gradually decreases, and pulse charging can be selected for intermittent charging to reduce heat generation.

[0020] When the SOC ≥ 80% or the monomer voltage reaches the upper limit (such as 4.2V), switch the mode and gradually reduce the current through the PID algorithm to avoid voltage step.

[0021] Furthermore, the input power adapter interface supports wide voltage input and is compatible with fast charging protocols. Rectification and filtering convert AC to DC and filter it, outputting a stable DC voltage. The bidirectional Buck-Boost topology is adopted to adapt to different battery voltages. High-frequency MOSFETs and low-loss magnetic components are selected to support high conversion efficiency. It supports the switching between fast charging and slow charging modes.

[0022] When used for the first time, calibrate the initial value through full charge with SOC = 100% or full discharge with SOC = 0%, and automatically reset the cumulative power after each full charge / full discharge to eliminate long-term errors.

[0023] Furthermore, the SOC estimation adopts coulomb meter + open circuit voltage OCV calibration, specifically as follows:

[0024] Automatically detect through a hardware comparator or software state machine to trigger the OCV measurement process. When the current continuously < 0.05C for more than 30 minutes, it is determined to be in a static state, where C is the battery capacity.

[0025] In an incubator, charge and discharge the battery at a rate of 0.1C, record the static OCV values at different SOC points, and establish an OCV-SOC-T three-dimensional look-up table model. For example, for every 10°C increase, the SOC estimation error needs to be corrected by ≤ 1%.

[0026] In the static state, sample the battery terminal voltage with a high-precision ADC, take the average of ≥ 10 sampling times, obtain the SOC value corresponding to the current OCV through the OCV-SOC-T three-dimensional look-up table model, perform weighted fusion with the coulomb meter result, calibrate the coulomb meter result, and adjust the weights of the coulomb meter and OCV according to parameters such as current magnitude and temperature change rate. For example: when the high load is I > 1C, the coulomb meter weight = 90%, the OCV weight = 10%; when static, the coulomb meter weight = 30%, the OCV weight = 70%.

[0027] Furthermore, receive the PWM signal or digital signal (such as S.Bus protocol) sent by the remote control, convert the physical position of the throttle lever amount into a percentage value of 0 - 100%. The middle position of the remote control joystick corresponds to 50% throttle, and full push to the top is 100%.

[0028] Furthermore, statistically analyze the average power corresponding to different throttle positions through historical flight data, set an independent mapping table for special modes, and automatically update the curve after each flight to adapt to the efficiency changes caused by battery aging or propeller loss.

[0029] Furthermore, the pitch / roll / yaw angular velocity is acquired at a frequency of 200 Hz, and high-frequency vibration noise is filtered out. If the pitch angular velocity is greater than 100° / s and the altimeter shows a continuous increase in altitude, it is judged as an emergency climb, and a pre-charge instruction is sent to the supercapacitor group in advance to ensure that 1500 W of power can be released within 0.2 seconds and non-essential loads are limited.

[0030] If the roll angular velocity is greater than 80° / s and the yaw angular velocity is greater than 60° / s, it is predicted to be an obstacle avoidance maneuver, and an additional 300W of power is allocated to the capacitor bank to compensate for the efficiency loss caused by centrifugal force.

[0031] Furthermore, UAV flight missions can usually be divided into multiple phases, and the power requirements of each phase are significantly different:

[0032] During the takeoff phase, the motor runs at full power to overcome gravity, and the power demand is the highest. During the climb phase, a certain thrust is maintained to increase the altitude, and the power demand is second. During the cruise phase, the aircraft flies horizontally at an economic speed, and the power demand is stable. During the mission phase, the aircraft performs tasks such as shooting and delivery, and the power demand depends on the load operation. During the return phase, the power may be limited due to insufficient battery power, or the aircraft may need to accelerate the return due to an emergency. During the landing phase, the power demand gradually decreases to a minimum.

[0033] Furthermore, the power data are aligned with the phase markers in the flight log, which include GPS coordinates and mission trigger signals. When the power exceeds a certain threshold for more than T seconds, it is determined to be a new phase, and K-means clustering is performed on the power curve to automatically identify the phase boundary.

[0034] Furthermore, for each segmented power curve, the following characteristic parameters are extracted: average power, peak power, energy consumption, power variance, and stage duration.

[0035] Average power, the average value of power within a stage, reflects the overall energy consumption level. Peak power, the maximum instantaneous power within a stage, is used to assess the risk of system overload. Energy consumption, the integral of power over time within a stage, estimates battery power consumption. Power variance, reflects the degree of power fluctuation and is related to flight stability. Stage duration, the time span from the beginning to the end, is used for mission planning optimization.

[0036] Further, time series analysis is performed based on the ARIMA model:

[0037] Trends and seasonality are eliminated through differential processing, the optimal (p, d, q) parameters are selected using the AIC / BIC criteria, the model is fitted through maximum likelihood estimation, and the effectiveness is tested using residual analysis, and the power forecast value and confidence interval for the next N steps are output.

[0038] Furthermore, the LSTM model is used to capture long-term dependencies in time series:

[0039] Normalize the power data and construct sliding window samples (e.g., use the data of the past 10 seconds to predict the next 1 second). Design an LSTM layer (e.g., 64 units), a fully connected layer, and an output layer. Train the model using historical data and prevent overfitting through cross-validation.

[0040] Establish a statistical relationship between the flight phase and the power demand. Query the reference power from the mapping table according to the current phase, and then adjust the deviation through a time series model.

[0041] (III) Beneficial effects

[0042] The present invention provides an energy-power conversion system for an unmanned aerial vehicle, having the following beneficial effects:

[0043] 1. Manage the charging process of the UAV battery. Connect an external charging power supply and control the charging strategy of the UAV battery according to the battery power reserve level. By dynamically adjusting the charging mode, the system can automatically switch the charging stage: when the battery reaches 80%, it switches to trickle charging to avoid battery swelling or capacity attenuation caused by overcharging. At the same time, the intelligent charging cabinet can preferentially charge the low-power batteries, support multiple modes such as fast charging, standard charging, and trickle charging, and prevent overheating through temperature monitoring to ensure charging safety. For example, in agricultural plant protection or logistics distribution, when multiple UAVs cooperate, the system can automatically identify the battery status, give priority to restoring key mission equipment, and reduce waiting time.

[0044] 2. Adopt a dynamic power distribution strategy to implement closed-loop control based on the real-time power spectrum characteristics of the drive module, and achieve on-demand power output regulation. The dynamic power distribution strategy breaks through the limitations of traditional static configuration. By real-time monitoring of the channel conditions, user requirements, and interference situations, it flexibly adjusts the power output. The closed-loop control based on the real-time power spectrum characteristics of the drive module extracts key features such as the signal peak frequency and band energy through Fourier transform or FFT algorithm to achieve precise perception of the system state. The closed-loop control system feeds back the output quantity (such as power, voltage) to the input end, compares it with the set value, and automatically corrects the deviation. For example, in an industrial power supply system, the closed-loop control monitors the output voltage fluctuation, adjusts the power supply working state in real time, ensures that the output is stable at the set value, suppresses harmonic interference at the same time, and improves the power quality. The on-demand power output regulation realizes the refined adjustment of the power output through technologies such as PID control and frequency conversion control. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 It is a schematic structural diagram of an energy-power conversion system for an unmanned aerial vehicle according to the present invention;

[0046] Figure 2 It is a schematic diagram of the charging process for managing the energy feedback to the power conversion module according to the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0047] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0048] Please refer to Figure 1 , the present invention provides an energy-power conversion system for an unmanned aerial vehicle, including:

[0049] A battery module, which is composed of multiple battery cells combined in series or parallel, and is responsible for providing the electric energy required for the drone to fly.

[0050] An energy management module, through integrating sensors, algorithms and control technologies, monitors, optimally allocates and intelligently regulates the energy use of the drone in real time, including a data acquisition unit, a charging management unit and a power conversion unit.

[0051] The data acquisition unit uses high-precision Hall effect sensors or shunts, with a typical accuracy of ±0.5% FS, a bandwidth covering 0 - 10 kHz, and a sampling rate ≥100 Hz to capture transient current changes, and adopts differential input, shielded cables and digital filtering algorithms to suppress electromagnetic interference.

[0052] The voltages of each power supply module are synchronously measured through a differential voltage sampling circuit, with a sampling frequency of 1 kHz. NTC thermistors are surface-mounted on the battery pack to monitor temperature changes, and the attitude, speed and position of the drone are real-time monitored through an IMU inertial measurement unit.

[0053] The power data, including information such as timestamps, instantaneous power values, flight modes, etc., are recorded through the flight control system FCU. The flight modes include takeoff, cruise and landing.

[0054] The charging management unit is used to manage the charging process of the drone battery, connect an external charging power supply and control the charging strategy of the drone battery according to the battery power reserve level;

[0055] Specifically, the charging of the drone battery is as follows:

[0056] In the fast charging stage with SOC 0 - 80%, the constant current CC mode is adopted to charge with the maximum allowable current, and at the same time, the input power is dynamically adjusted to avoid power overload.

[0057] In the slow charging stage with SOC 80% - 100%, the constant voltage CV mode is adopted, the voltage is constant, and the current gradually decreases. Pulse charging can be selected for intermittent charging to reduce heat generation.

[0058] When the SOC ≥ 80% or the monomer voltage reaches the upper limit (such as 4.2V), switch the mode and gradually reduce the current through the PID algorithm to avoid voltage step.

[0059] The input power adapter interface supports wide voltage input and is compatible with fast charging protocols. Rectification and filtering convert alternating current to direct current and filter it to output a stable direct current voltage. The bidirectional Buck-Boost topology is adopted to adapt to different battery voltages. High-frequency MOSFETs and low-loss magnetic components are selected to support high conversion efficiency. It supports the switching between fast charging and slow charging modes.

[0060] When used for the first time, calibrate the initial value through full charge SOC = 100% or full discharge SOC = 0%, and automatically reset the cumulative power after each full charge / full discharge to eliminate long-term errors.

[0061] The SOC estimation adopts coulomb meter + open circuit voltage OCV calibration, specifically:

[0062] Automatically detect through a hardware comparator or software state machine to trigger the OCV measurement process. When the current continuously < 0.05C for more than 30 minutes, it is determined to be in a static state, where C is the battery capacity.

[0063] In a thermostat, charge and discharge the battery at a rate of 0.1C, record the static OCV values at different SOC points, and establish a three-dimensional look-up table model of OCV-SOC-T. For example, when the temperature rises by 10°C, the SOC estimation error needs to be corrected by ≤ 1%.

[0064] In the static state, sample the battery terminal voltage with a high-precision ADC. The sampling times are ≥ 10 times and take the average. Obtain the SOC value corresponding to the current OCV through the three-dimensional look-up table model of OCV-SOC-T, perform weighted fusion with the coulomb meter result, calibrate the coulomb meter result, and adjust the weights of the coulomb meter and OCV according to parameters such as current magnitude and temperature change rate. For example, when the high load I > 1C, the coulomb meter weight = 90%, the OCV weight = 10%; when static, the coulomb meter weight = 30%, the OCV weight = 70%.

[0065] Manage the charging process of the UAV battery, connect an external charging power supply and control the charging strategy of the UAV battery according to the battery power reserve level. By dynamically adjusting the charging mode, the system can automatically switch the charging stage: when the battery reaches 80% of its capacity, switch to trickle charging to avoid battery swelling or capacity attenuation caused by overcharging. At the same time, the intelligent charging cabinet can preferentially charge the low-power batteries, support multiple modes such as fast, standard, and trickle charging, and prevent overheating through temperature monitoring to ensure charging safety. For example, in agricultural plant protection or logistics distribution, when multiple UAVs are operating in coordination, the system can automatically identify the battery status, give priority to restoring key mission equipment, and reduce waiting time.

[0066] The power conversion unit adopts a dynamic power allocation strategy and implements closed-loop control based on the real-time power spectrum characteristics of the drive module to achieve on-demand power output adjustment.

[0067] Receive the PWM signal or digital signal (such as S.Bus protocol) sent by the remote control, convert the physical position of the throttle stick into a percentage value of 0-100%. The middle position of the remote control stick corresponds to 50% throttle, and fully pushed to the top is 100%.

[0068] The average power corresponding to different throttle positions is calculated through historical flight data, and an independent mapping table is set for special modes. The curve is automatically updated after each flight to adapt to efficiency changes caused by battery aging or propeller loss.

[0069] The pitch / roll / yaw angular velocity is acquired at a frequency of 200Hz, and high-frequency vibration noise is filtered out. If the pitch angular velocity is greater than 100° / s and the altimeter shows a continuous increase in altitude, it is judged as an emergency climb, and a pre-charge instruction is sent to the supercapacitor group in advance to ensure that 1500W of power can be released within 0.2 seconds and non-essential loads are limited.

[0070] If the roll angular velocity is greater than 80° / s and the yaw angular velocity is greater than 60° / s, it is predicted to be an obstacle avoidance maneuver, and an additional 300W of power is allocated to the capacitor bank to compensate for the efficiency loss caused by centrifugal force.

[0071] A drone flight mission can usually be divided into several phases, and the power requirements of each phase are significantly different:

[0072] During the takeoff phase, the motor runs at full power to overcome gravity, and the power demand is the highest. During the climb phase, a certain thrust is maintained to increase the altitude, and the power demand is second. During the cruise phase, the aircraft flies horizontally at an economic speed, and the power demand is stable. During the mission phase, the aircraft performs tasks such as shooting and delivery, and the power demand depends on the load operation. During the return phase, the power may be limited due to insufficient battery power, or the aircraft may need to accelerate the return due to an emergency. During the landing phase, the power demand gradually decreases to a minimum.

[0073] The power data is aligned with the phase markers in the flight log. The phase markers include GPS coordinates and mission trigger signals. When the power exceeds a certain threshold for more than T seconds, it is determined to be a new phase. K-means clustering is performed on the power curve to automatically identify the phase boundary.

[0074] For each segmented power curve, the following characteristic parameters are extracted: average power, peak power, energy consumption, power variance and stage duration.

[0075] Average power, the average value of power within a stage, reflects the overall energy consumption level. Peak power, the maximum instantaneous power within a stage, is used to evaluate the system overload risk. Energy consumption, the integral of power over time within a stage, estimates the battery power consumption. Power variance, reflecting the degree of power fluctuation, is related to flight stability. Stage duration, the time span from start to end, is used for mission planning optimization.

[0076] Perform time series analysis based on the ARIMA model:

[0077] Eliminate trends and seasonality through differencing, select the optimal (p, d, q) parameters using the AIC / BIC criterion, fit the model through maximum likelihood estimation, and use residual analysis to test the effectiveness, outputting the power prediction values and confidence intervals for the next N steps.

[0078] Capture long-term dependencies in time series based on the LSTM model:

[0079] Normalize the power data and construct sliding window samples (e.g., using the data of the past 10 seconds to predict the next 1 second). Design the LSTM layer (e.g., 64 units), fully connected layer, and output layer. Train the model using historical data and prevent overfitting through cross-validation.

[0080] Establish a statistical relationship between flight stages and power requirements, query the reference power from the mapping table according to the current stage, and then adjust the deviation through a time series model.

[0081] Adopt a dynamic power allocation strategy, implement closed-loop control based on the real-time power spectrum characteristics of the drive module, and achieve on-demand power output regulation. The dynamic power allocation strategy breaks through the limitations of traditional static configurations. By real-time monitoring of channel conditions, user requirements, and interference situations, it flexibly adjusts the power output. The closed-loop control based on the real-time power spectrum characteristics of the drive module extracts key features such as signal peak frequency and band energy through Fourier transform or FFT algorithm to achieve precise perception of the system state. The closed-loop control system feeds back the output quantity (such as power, voltage) to the input end, automatically corrects the deviation after comparing with the set value. For example, in an industrial power supply system, the closed-loop control monitors the output voltage fluctuation, adjusts the power supply working state in real time to ensure that the output is stable at the set value, while suppressing harmonic interference and improving the power supply quality. The on-demand power output regulation realizes the refined adjustment of power output through technologies such as PID control and variable frequency control.

[0082] The above embodiments can be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented using software, the above embodiments can be implemented in whole or in part in the form of a computer program product. Those of ordinary skill in the art will appreciate that the units and algorithm steps of the examples described in connection with the embodiments disclosed herein can be implemented in electronic hardware, or in a combination of computer software and electronic hardware. Whether these functions are executed in hardware or software depends on the specific application and design constraints of the technical solution.

[0083] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0084] As described above, the specific implementation manners of the present application are only described, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present application can easily think of changes or substitutions, which should all be covered within the protection scope of the present application.

Claims

1. An energy-power conversion system for an unmanned aerial vehicle, characterized in that: Including: A battery module, which is composed of multiple battery cells combined in series or parallel, and is responsible for providing the electrical energy required for the drone to fly; An energy management module, which integrates sensors, algorithms and control technologies to monitor, optimize the allocation and intelligently control the energy usage of the drone in real time, including a data acquisition unit, a charging management unit and a power conversion unit; The data acquisition unit uses high-precision Hall effect sensors or shunts to capture transient current changes, synchronously measures the voltages of each power supply module through a differential voltage sampling circuit, monitors the battery temperature change, monitors the attitude, speed and position of the drone in real time through an IMU inertial measurement unit, and records power data through a flight control system FCU; The charging management unit is used to manage the charging process of the drone battery, connect an external charging power supply and control the charging strategy of the drone battery according to the battery charge reserve level; The power conversion unit adopts a dynamic power distribution strategy, implements closed-loop control based on the real-time power spectrum characteristics of the drive module, and realizes the regulation of on-demand power output.

2. The energy-power conversion system of an unmanned aerial vehicle according to claim 1, wherein: The charging of the drone battery is specifically as follows: In the fast charging stage with SOC of 0-80%, the constant current CC mode is adopted to charge with the maximum allowable current, and at the same time, the input power is dynamically adjusted to avoid power supply overload; In the slow charging stage with SOC of 80%-100%, the constant voltage CV mode is adopted, the voltage is constant, and the current gradually decreases. Pulse charging can be selected, and intermittent charging is used to reduce heat generation; When SOC≥80% or the single cell voltage reaches the upper limit, the mode is switched, and the current is gradually reduced through the PID algorithm to avoid voltage step.

3. The energy-power conversion system of an unmanned aerial vehicle according to claim 1, wherein: The input power adapter interface supports wide voltage input and is compatible with fast charging protocols; rectification and filtering convert alternating current into direct current and filter it to output a stable direct current voltage; a bidirectional Buck-Boost topology is adopted to adapt to different battery voltages; High-frequency MOSFETs and low-loss magnetic components are selected to support high conversion efficiency; it supports the switching between fast charging and slow charging modes.

4. The energy-power conversion system of an unmanned aerial vehicle according to claim 1, wherein: When used for the first time, the initial value is calibrated by full charge SOC = 100% or full discharge SOC = 0%, and the cumulative power is automatically reset after each full charge / full discharge to eliminate long-term errors.

5. The energy-power conversion system of an unmanned aerial vehicle according to claim 1, wherein: The SOC estimation adopts coulomb meter + open circuit voltage OCV calibration, specifically: Automatically detected through a hardware comparator or a software state machine, triggering the OCV measurement process. When the current continuously <0.05C exceeds 30 minutes, it is determined to be in a static state, where C is the battery capacity; In a thermostat, the battery is charged and discharged at a rate of 0.1C, and the static OCV values at different SOC points are recorded to establish a three-dimensional look-up table model of OCV-SOC-T; In the static state, the battery terminal voltage is sampled by a high-precision ADC. The sampling is repeated ≥10 times and averaged. The SOC value corresponding to the current OCV is obtained through the OCV-SOC-T three-dimensional look-up table model, and is weighted and fused with the coulomb meter result to calibrate the coulomb meter result. The weights of the coulomb meter and OCV are adjusted according to parameters such as the current magnitude and the temperature change rate.

6. The energy-power conversion system of an unmanned aerial vehicle according to claim 1, characterized in that: It receives the PWM signal or digital signal sent by the remote controller, converts the physical position of the throttle stick amount into a percentage value of 0-100%. The middle position of the remote controller joystick corresponds to 50% throttle, and full push to the top is 100%. The average power corresponding to different throttle positions is statistically analyzed through historical flight data. An independent mapping table is set for special modes, and the curve is automatically updated after each flight to adapt to the efficiency changes caused by battery aging or propeller loss. The pitch / roll / yaw angular velocities are obtained at a frequency of 200 Hz, and high-frequency vibration noise is filtered out. If the pitch angular velocity > 100° / s and the altimeter shows that the altitude is continuously rising, it is determined as an emergency climb. A pre-charge command is sent to the supercapacitor bank in advance to ensure that 1500 W of power can be released within 0.2 seconds, and unnecessary loads are restricted. If the roll angular velocity > 80° / s and the yaw angular velocity > 60° / s, it is predicted as an obstacle avoidance maneuver, and an additional 300 W of power is allocated to the capacitor bank to compensate for the efficiency loss caused by centrifugal force.

7. The energy-power conversion system of an unmanned aerial vehicle according to claim 1, characterized in that: The power data is aligned with the phase markers in the flight log. The phase markers include GPS coordinates and mission trigger signals. When the power continuously exceeds a certain threshold for more than T seconds, it is determined as a new phase, and K-means clustering is performed on the power curve to automatically identify the phase boundary.

8. The energy-power conversion system of an unmanned aerial vehicle according to claim 1, characterized in that: Time series analysis is performed based on the ARIMA model: The trend and seasonality are eliminated through differencing processing. The optimal parameters are selected using the AIC / BIC criterion. The model is fitted through maximum likelihood estimation, and the effectiveness is tested through residual analysis. The power prediction values and confidence intervals for the next N steps are output.

9. The energy-power conversion system of an unmanned aerial vehicle according to claim 1, characterized in that: The long-term dependencies in the time series are captured based on the LSTM model: The power data is normalized to construct sliding window samples; an LSTM layer, a fully connected layer, and an output layer are designed; the model is trained using historical data, and overfitting is prevented through cross-validation.

Images