Hydrogen energy unmanned aerial vehicle dynamic hybrid control system based on fuzzy PID and endurance optimization method

Abstract

The invention relates to the technical field of fuzzy PID (Proportion Integration Differentiation) control, in particular to a hydrogen energy unmanned aerial vehicle dynamic hybrid control system based on fuzzy PID and an endurance optimization method. Change values are paired with proportion, integral and differential coefficient initial values by calculating change rates of a pitch angle, a roll angle and a vertical speed; comparing the power variation amplitude with a set threshold value by adopting a PID control algorithm to obtain a trend mark, selecting proportion, integral and differential coefficient adjustment amounts according to a trend to form an adjustment coefficient set, realizing parameter presetting in a pre-control stage, and combining an attitude error, an error variation amount and a power difference, so as to obtain a control result; combining a multi-target particle swarm optimization algorithm to carry out condition check on the upper limit and the lower limit of a combination result, a fuel cell power value and a lithium battery electric quantity value, rejecting non-conforming combinations and reserving effective records, calculating a comprehensive value of attitude errors and power differences for compliant combinations, and screening high-quality groups according to an ascending order; and extracting proportion, integral and differential coefficients to form a control instruction set.

Description

Technical Field

[0001] This invention relates to the field of fuzzy PID control technology, and in particular to a dynamic hybrid control system for hydrogen-powered unmanned aerial vehicles based on fuzzy PID and a method for optimizing their range. Background Technology

[0002] Fuzzy PID control technology targets the nonlinear, multivariable, and time-varying characteristics of the controlled object. By defining membership functions and fuzzy rule bases, it achieves online adaptive adjustment of PID parameters. It is often used for system control with operating condition disturbances, model uncertainties, or dynamic response lags. It can maintain system stability and robustness even when the parameter change rate is higher than the traditional PID tuning time constant.

[0003] A dynamic hybrid control system for hydrogen-powered unmanned aerial vehicles (UAVs) based on fuzzy PID is designed to address the power distribution imbalance problem caused by load changes, operating condition disturbances, and energy state fluctuations during flight of UAVs powered by a combination of hydrogen fuel cells and batteries. By using fuzzy logic reasoning to adjust the proportional coefficient Kp, integral coefficient Ki, and derivative coefficient Kd in real time, the system dynamically optimizes the power supply ratio of the two energy sources in the hybrid power system, improves power response speed, enhances energy utilization efficiency, and maintains the stability of the UAV's thrust output, thereby improving endurance and stabilizing flight attitude control.

[0004] In the dynamic hybrid control of hydrogen-powered drones based on fuzzy PID, existing technologies mostly rely on real-time attitude and power change feedback to adjust PID parameters, lacking advance prediction of power demand. This results in the inability to pre-adjust parameters before sudden load changes, leading to response delays, insufficient energy state constraints, the risk of generating out-of-limit combinations, and increased execution instability. In the power allocation phase, dynamic sorting optimization based on the power ratio of fuel cells and lithium batteries is rarely used, potentially causing efficiency fluctuations during energy switching. The lack of synchronous mapping between the load response phase and power allocation prevents precise correlation between operational data and execution actions, limiting the reliability of subsequent dynamic adjustments. Energy operation data storage primarily relies on segmented recording, lacking continuous task recording throughout the entire process, which affects the data's utilization value in task review and scheduling optimization. Summary of the Invention

[0005] The purpose of this invention is to overcome the shortcomings of existing technologies and propose a dynamic hybrid control system for hydrogen-powered unmanned aerial vehicles based on fuzzy PID and a method for optimizing endurance.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a dynamic hybrid control system for hydrogen-powered unmanned aerial vehicles based on fuzzy PID control includes: Flight power demand prediction module: Based on the power of hydrogen fuel cell and the capacity of lithium battery, calculate the attitude change rate and pair it with the initial values ​​of Kp, Ki, and Kd. Use PID control algorithm to compare the power change amplitude with the threshold to judge the trend. Select ΔKp, ΔKi, and ΔKd according to the results to generate a set of predicted power adjustment coefficients. Control parameter adaptive generation module: Based on the predicted power adjustment coefficient set, the particle swarm optimization algorithm is used to match the coefficients with the attitude error and the change, and the available combinations are screened in combination with power and energy limits. The selection is sorted by error and power difference to generate a fusion control coefficient table. Power allocation execution module: Based on the fusion control coefficient table, the control coefficients are matched with the power parameters of the fuel cell and lithium battery, out-of-limit combinations are eliminated, and the remaining combinations are sorted and numbered in descending order of power ratio to generate an effective energy allocation sequence; Flight load response module: Based on the effective energy distribution sequence, load the power values ​​of fuel cells and lithium batteries and record the vibration frequency during the execution cycle, synchronously acquire thrust and attitude changes, calculate the thrust difference range, and generate flight thrust response records; Energy Operation Management Module: Based on the flight thrust response records, thrust differences, power values, vibration frequencies, and attitude changes are sorted and stored according to mission number and flight segment number to establish a complete energy scheduling archive.

[0007] As a further embodiment of the present invention, the predicted power adjustment coefficient set includes proportional coefficient adjustment, integral coefficient adjustment, and differential coefficient adjustment; the fused control coefficient table includes proportional coefficient, integral coefficient, and differential coefficient; the effective energy allocation sequence includes fuel cell power value, lithium battery power value, and energy output ratio; the flight thrust response record includes thrust difference value, vibration frequency value, and attitude change amplitude; and the energy scheduling full-process archive includes mission number, flight segment number, full-process power record, and attitude change data.

[0008] As a further aspect of the present invention, the flight power demand prediction module includes: Power and attitude calculation submodule: Based on the power of hydrogen fuel cell and the capacity of lithium battery, calculate the adjacent time difference of pitch angle, roll angle and vertical velocity and divide it by the time interval to obtain the rate of change. The prediction unit using PID control algorithm stores each set of rate of change in pairs with the initial values ​​of proportional coefficient, integral coefficient and derivative coefficient according to the index position to generate attitude power paired dataset. Trend State Determination Submodule: Based on the attitude power pairing dataset, compare the power change rate with the preset threshold record to determine whether the record is positive or negative, and determine whether the state is rising or falling. The absolute value of the change rate is compared with the interval range to determine the level code, and is paired with the state mark and written into the table to generate a power trend state table. The coefficient adjustment generation submodule: Based on the power trend status table, it retrieves the proportional coefficient adjustment amount, integral coefficient adjustment amount, and derivative coefficient adjustment amount corresponding to the combination of status marker and level code, arranges the selected adjustment amounts into a data sequence according to the original index order, and generates a set of predicted power adjustment coefficients.

[0009] As a further aspect of the present invention, the PID control algorithm, upon receiving input data such as attitude angle change rate, fuel cell power, and lithium battery capacity, first reads the current proportional coefficient, integral coefficient, and derivative coefficient. It then multiplies the attitude error value by the proportional coefficient to obtain the proportional term output value. The algorithm accumulates the error sequence collected over time and multiplies it by the integral coefficient to obtain the integral term output value. It subtracts the previous cycle error value from the current error value and divides it by the corresponding time interval to obtain the error change rate, which is then multiplied by the derivative coefficient to obtain the derivative term output value. Finally, it adds the proportional, integral, and derivative terms to form the comprehensive output result for the current cycle. This comprehensive output value is used to adjust control commands and records the corresponding time index. Simultaneously, it updates the accumulated error and the previous cycle error value to the latest values.

[0010] As a further aspect of the present invention, the adaptive control parameter generation module includes: Coefficient Error Matching Submodule: Based on the predicted power adjustment coefficient set, the particle swarm optimization algorithm is used to combine the proportional coefficient adjustment, integral coefficient adjustment, and differential coefficient adjustment with the attitude error value, error change value, and power difference value one by one within the adjustable parameter range. Each combination is recorded in the matching table and the performance index of the corresponding individual particle is retained to generate the error matching result table. Restriction Condition Filtering Submodule: Based on the error matching result table, compare the fuel cell power and lithium battery capacity with their respective upper and lower limits, mark and remove records that do not meet all conditions, retain records that meet the conditions and update the index, and generate an effective error power combination table; The parameter optimization output submodule calculates the sum of the attitude error value and the power difference as a comprehensive value based on the effective error power combination table. The comprehensive value is sorted in ascending order, and the proportional coefficient, integral coefficient, and derivative coefficient are extracted from the first record to form a control parameter group, generating a fusion control coefficient table.

[0011] As a further aspect of the present invention, the particle swarm optimization algorithm generates a combined vector for each particle during the initialization phase, containing proportional coefficient adjustment, integral coefficient adjustment, differential coefficient adjustment, attitude error value, error change value, and power difference value. Simultaneously, each particle is assigned a random velocity vector, and its current position is recorded as its individual adaptation position. Positions with fitness values ​​that match the overall particle fitness are marked as global adaptation positions. During iteration, the fitness value of each particle is calculated sequentially and compared with the fitness of its individual adaptation position. When the current fitness value is better, the individual adaptation position is updated. The current fitness value is then compared with the global adaptation position fitness; when the current fitness value is better, the global adaptation position is updated synchronously. Based on the current velocity, current position, individual adaptation position, and global adaptation position, a velocity increment is calculated according to a set weight and added to the current velocity. The updated velocity is then added to the current position to obtain a new search position. The proportional coefficient adjustment, integral coefficient adjustment, differential coefficient adjustment, and attitude parameter combination in the position are written into a matching table. After reaching a preset number of iterations, the fitness-fitted particle position is output as the result of this round of search.

[0012] As a further aspect of the present invention, the power allocation execution module includes: Power matching screening submodule: Based on the fusion control coefficient table, the proportional coefficient, integral coefficient, and derivative coefficient are compared with the fuel cell power and lithium battery power in the same row and column by column to determine whether each item is greater than or less than the corresponding safety threshold. Any combination that does not meet the conditions is eliminated, and all the remaining combinations are stored in the list in the original order to generate a list of effective power combinations. Energy ratio sorting submodule: Based on the effective power combination list, the ratio of fuel cell power to lithium battery power is obtained for each combination data. The ratio value and combination number are recorded in pairs and sorted from high to low according to the ratio value. The sorting results are assigned consecutive numbers and the numbers and ratios are integrated into the record sequence to generate an effective energy allocation sequence.

[0013] As a further aspect of the present invention, the flight load response module includes: Power loading and binding submodule: Based on the effective energy allocation sequence, the fuel cell power and lithium battery power under the same time index are loaded to the converter power port, the vibration frequency values ​​at the same time are written into the combined record, and the power and frequency data are bound in the record. The command sequence is stored in sequence to generate a power frequency binding sequence. Thrust difference calculation submodule: Based on the power frequency binding sequence, the thrust value and attitude change amplitude of the propulsion unit are collected at each time index. The absolute difference of the thrust value and the attitude amplitude are combined into a data pair. The maximum and minimum value ranges of all data pairs are defined, and the definition results are written into a table to generate a flight thrust response record.

[0014] As a further aspect of the present invention, the energy operation management module includes: Task segment data integration submodule: Based on the flight thrust response record, the thrust difference value, fuel cell power, lithium battery power, vibration frequency, and attitude change amplitude are grouped according to the task number and flight segment number. Records under the same number are merged into a single-row data table to generate the energy task segment record table. Full-process archive generation submodule: Based on the energy task segment record table, the records in each row are connected in chronological order to form a continuous link table, and all data in the link table is written to a long-term storage medium to generate a full-process energy scheduling archive.

[0015] A method for optimizing the endurance of a hydrogen-powered drone based on fuzzy PID dynamic hybrid control, wherein the method is executed based on the aforementioned fuzzy PID-based dynamic hybrid control system for hydrogen-powered drones, and includes the following steps: Step 1: Based on the power of the hydrogen fuel cell and the capacity of the lithium battery, calculate the pitch angle, roll angle and vertical velocity change rate in time sequence. Use the feedforward-fuzzy PID hybrid control algorithm to pair the change rate with the initial values ​​of the proportional, integral and derivative coefficients, compare the power change amplitude with the threshold marking trend, select the adjustment amount of the proportional, integral and derivative coefficients according to the trend, and generate a set of predicted power adjustment coefficients. Step 2: Based on the predicted power adjustment coefficient set, the multi-objective particle swarm optimization algorithm is used to combine the coefficients with attitude error, error change, and power difference in sequence. The upper and lower limits of fuel cell power and lithium battery capacity are compared to screen available combinations. The attitude error and power difference are calculated and sorted. The proportional, integral, and derivative coefficients of the top combinations are selected to form a control parameter group and generate a fusion control coefficient table. Step 3: Based on the fusion control coefficient table, match the control coefficients with the power of the fuel cell and lithium battery, eliminate combinations that exceed limits, calculate the power ratio, sort and number them, and generate an effective energy allocation sequence; Step 4: Based on the effective energy distribution sequence, load power to the distribution converter according to the time index and record the vibration frequency, collect thrust and attitude changes, calculate the thrust difference interval, and generate flight thrust response record; Step 5: Based on the flight thrust response record, sort and connect the thrust difference, power, vibration frequency, and attitude change according to the mission and flight segment number to establish a complete energy scheduling archive.

[0016] Compared with the prior art, the advantages and positive effects of the present invention are as follows: In this invention, by calculating the pitch angle, roll angle and vertical velocity change rate, the change values ​​are paired with the initial values ​​of proportional, integral and derivative coefficients. The PID control algorithm is used to compare the power change amplitude with the set threshold to obtain trend markers. The proportional, integral and derivative coefficients are selected according to the trend to form an adjustment coefficient set, thereby realizing parameter preset in the control stage. In this invention, attitude error, error change and power difference are combined, and the combination result is checked against the upper and lower limits of fuel cell power value and lithium battery charge value by combining multi-objective particle swarm optimization algorithm. Unqualified combinations are eliminated and valid records are retained. For compliant combinations, the comprehensive value of attitude error and power difference is calculated and high-quality groups are selected in ascending order. The proportional, integral and differential coefficients are extracted to form a control instruction set. In this invention, an energy distribution sequence is constructed by eliminating out-of-bounds records and arranging them in descending order of power ratio. This sequence is used as a driving force to load fuel cell power and lithium battery power to the distribution conversion port during the execution cycle. The corresponding vibration frequency is recorded and thrust and attitude changes are collected synchronously. The flight load characteristics are defined by the maximum and minimum difference of the thrust value. Attached Figure Description

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

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

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

[0020] Example 1 Please see Figure 1 This invention provides a technical solution: a dynamic hybrid control system for hydrogen-powered unmanned aerial vehicles based on fuzzy PID control, comprising: Flight power demand prediction module: Based on the power of hydrogen fuel cell and the capacity of lithium battery, calculate the attitude change rate and pair it with the initial values ​​of Kp, Ki, and Kd. Use PID control algorithm to compare the power change amplitude with the threshold to judge the trend. Select ΔKp, ΔKi, and ΔKd according to the results to generate a set of predicted power adjustment coefficients. The adaptive generation module for control parameters: Based on the predicted power adjustment coefficient set, the particle swarm optimization algorithm is used to match the coefficients with the attitude error and the change, and the available combinations are screened in combination with power and energy limits. The selection is sorted by error and power difference to generate a fusion control coefficient table. Power allocation execution module: Based on the fusion control coefficient table, the control coefficients are matched with the power parameters of fuel cells and lithium batteries, out-of-limit combinations are eliminated, and the remaining combinations are sorted and numbered in descending order of power ratio to generate an effective energy allocation sequence; Flight load response module: Based on the effective energy distribution sequence, it loads the power values ​​of fuel cells and lithium batteries and records the vibration frequency during the execution cycle, simultaneously acquires thrust and attitude changes, calculates the thrust difference range, and generates flight thrust response records; Energy Operation Management Module: Based on flight thrust response records, thrust differences, power values, vibration frequencies, and attitude changes are sorted and stored according to mission number and flight segment number to establish a complete energy dispatch archive.

[0021] The predicted power adjustment coefficient set includes proportional coefficient adjustment, integral coefficient adjustment, and derivative coefficient adjustment; the fused control coefficient table includes proportional coefficient, integral coefficient, and derivative coefficient; the effective energy allocation sequence includes fuel cell power value, lithium battery power value, and energy output ratio; the flight thrust response record includes thrust difference value, vibration frequency value, and attitude change amplitude; and the energy scheduling full-process archive includes mission number, flight segment number, full-process power record, and attitude change data.

[0022] The flight power demand forecasting module includes: Power and attitude calculation submodule: Based on the power of hydrogen fuel cell and the capacity of lithium battery, calculate the adjacent time difference of pitch angle, roll angle and vertical velocity and divide it by the time interval to obtain the rate of change. The prediction unit using PID control algorithm stores each set of rate of change in pairs with the initial values ​​of proportional coefficient, integral coefficient and derivative coefficient according to the index position to generate attitude power paired dataset. Trend State Determination Submodule: Based on the attitude power pairing dataset, compare the power change rate with the preset threshold record to mark it as positive or negative, determine the state as rising or falling, compare the absolute value of the change rate according to the interval range to determine the level code, and pair it with the state mark to write it into the table to generate a power trend state table. The coefficient adjustment generation submodule: Based on the power trend status table, it retrieves the proportional coefficient adjustment, integral coefficient adjustment, and derivative coefficient adjustment corresponding to the combination of status marker and level code, arranges the selected adjustment amounts into a data sequence according to the original index order, and generates a set of predicted power adjustment coefficients. Power and attitude calculation submodule: Based on the power of hydrogen fuel cell and the capacity of lithium battery, the prediction unit adopts the feedforward-fuzzy PID hybrid control algorithm. It arranges the collected pitch angle, roll angle and vertical velocity in time index order, calculates the angle difference and velocity difference between two adjacent records and divides them by the corresponding time interval to obtain the rate of change. Each set of rate of change is paired with the initial values ​​of proportional coefficient, integral coefficient and derivative coefficient according to the same index position and stored in the structured parameter table. The paired data is recorded to the attitude power database according to the index order to generate the attitude power paired dataset. Trend State Determination Submodule: Based on the attitude power pairing dataset, the trend analysis unit adopts a feedforward-fuzzy PID hybrid control algorithm to compare each power change rate with a preset threshold one by one. If the value is greater than the threshold, it is marked as positive, and if it is less than the threshold, it is marked as negative. The marking results are recorded in the state mark column, and the absolute value of the change rate is calculated. The results are compared with the lower limit and upper limit values ​​in the graded interval table to determine the grade code. The grade code and the corresponding state mark are stored in the same column in the table to generate a power trend parameter list and a power trend state table. The coefficient adjustment generation submodule, based on the power trend state table, employs a parameter fusion unit with a feedforward-fuzzy PID hybrid control algorithm. It retrieves the position of the state marker and level code for each record in the corresponding parameter mapping matrix, reads the proportional coefficient adjustment, integral coefficient adjustment, and derivative coefficient adjustment, and writes these three adjustments sequentially into the output sequence according to their original index positions. The output sequence is then stored in the coefficient adjustment table, maintaining a one-to-one correspondence with the corresponding time index, thus generating a set of predicted power adjustment coefficients. The PID control algorithm, upon receiving input data such as attitude angle change rate, fuel cell power, and lithium battery capacity, first reads the current proportional coefficient, integral coefficient, and derivative coefficient. It then multiplies the attitude error value by the proportional coefficient to obtain the proportional term output value. The algorithm accumulates the error sequence over time and multiplies it by the integral coefficient to obtain the integral term output value. Finally, it subtracts the previous cycle's error value from the current error value and divides it by the corresponding time interval to obtain the error change rate, which is then multiplied by the derivative coefficient to obtain the derivative term output value. The proportional, integral, and derivative terms are then added together to form the comprehensive output result for the current cycle. This comprehensive output value is used to adjust control commands and records the corresponding time index. Simultaneously, the accumulated error and the error value from the previous cycle are updated to the latest values.

[0023] The adaptive generation module for control parameters includes: Coefficient Error Matching Submodule: Based on the predicted power adjustment coefficient set, the particle swarm optimization algorithm is used to combine the proportional coefficient adjustment, integral coefficient adjustment, and differential coefficient adjustment with the attitude error value, error change value, and power difference value one by one within the adjustable parameter range. Each combination is recorded in the matching table and the performance index of the corresponding individual particle is retained to generate the error matching result table. The constraint filtering submodule compares the fuel cell power and lithium battery capacity with their respective upper and lower limits based on the error matching result table. It marks and removes records that do not meet all the conditions, retains records that meet the conditions, updates the index, and generates an effective error power combination table. The parameter optimization output submodule calculates the sum of attitude error value and power difference as a comprehensive value based on the effective error power combination table. The comprehensive value is sorted in ascending order, and the proportional coefficient, integral coefficient and derivative coefficient are extracted from the first record to form a control parameter group, generating a fusion control coefficient table. The coefficient error matching submodule: Based on the predicted power adjustment coefficient set, the multi-objective particle swarm optimization algorithm is used to generate the initial position and velocity vector of particles within the preset range of proportional coefficient adjustment, integral coefficient adjustment, and differential coefficient adjustment. The proportional coefficient adjustment is paired with the attitude error value, the integral coefficient adjustment is paired with the error change value, and the differential coefficient adjustment is paired with the power difference value to form a complete parameter combination vector. The combined numerical sequence is written into the matching table and the particle number is recorded. At the same time, the fitness value corresponding to each combination is calculated and stored in the performance index column to generate the error matching result table. The constraint filtering submodule, based on the error matching result table, uses the constraint verification steps of the multi-objective particle swarm optimization algorithm to compare the fuel cell power value in each record with the set upper and lower power limits, and compare the lithium battery charge value with the corresponding upper and lower charge limits. If any comparison result exceeds the range, the record is marked as invalid and removed from the matching table. The combination records marked as valid are retained and reordered according to the original particle number. The updated data is written into the valid combination table to generate the valid error power combination table. The parameter optimization output submodule: Based on the effective error power combination table, the multi-objective particle swarm optimization algorithm reads the attitude error value and power difference of each record in the optimization stage and performs addition to obtain the comprehensive value. All comprehensive values ​​are sorted from smallest to largest. The proportional coefficient, integral coefficient and differential coefficient of the combination with the highest sort position are extracted to form a new parameter set. The parameters are written into the output parameter table in order and stored in association with the corresponding particle number to generate the fusion control coefficient table.

[0024] The particle swarm optimization algorithm generates a combined vector for each particle during the initialization phase, containing proportional coefficient adjustment, integral coefficient adjustment, differential coefficient adjustment, attitude error value, error change value, and power difference value. Simultaneously, each particle is assigned a random velocity vector, and its current position is recorded as its individual fit position. Positions with the best fitness values ​​among all particles are marked as global fit positions. During iteration, the fitness value of each particle is calculated sequentially and compared with the fitness of its individual fit position. When the current fitness value is better, the individual fit position is updated. The current fitness value is then compared with the global fit position fitness; if better, the global fit position is updated synchronously. Based on the current velocity, current position, individual fit position, and global fit position, a velocity increment is calculated according to set weights and added to the current velocity. The updated velocity is then added to the current position to obtain a new search position. The proportional coefficient adjustment, integral coefficient adjustment, differential coefficient adjustment, and attitude parameter combination in the position are written into a matching table. After reaching a preset number of iterations, the fitness-fitted particle position is output as the result of this round of search.

[0025] The power allocation execution module includes: Power matching screening submodule: Based on the fusion control coefficient table, the proportional coefficient, integral coefficient, and derivative coefficient are compared with the fuel cell power and lithium battery power in the same row and column by column to determine whether each item is greater than or less than the corresponding safety threshold. Any combination that does not meet the conditions is eliminated, and all the remaining combinations are stored in the list in the original order to generate a list of effective power combinations. Energy Ratio Sorting Submodule: Based on the list of effective power combinations, the ratio of fuel cell power to lithium battery power is obtained for each combination data. The ratio value and combination number are recorded in pairs. The combination is sorted from high to low according to the ratio value. The sorting results are assigned consecutive numbers. The numbers and ratios are integrated into the record sequence to generate an effective energy allocation sequence. Power matching and filtering submodule: Based on the fused control coefficient table, a threshold matching and filtering algorithm is used to read the proportional coefficient, integral coefficient, derivative coefficient and fuel cell power and lithium battery power column by column in the same row number. The parameter value of each column is compared with the corresponding safety threshold. If the value is greater than the upper threshold or less than the lower threshold, the combination is marked as invalid and the reason for rejection is recorded. The unmarked combinations are written into the data list according to the original index order, and each combination is ensured to include the proportional coefficient, integral coefficient and derivative coefficient and the corresponding fuel cell power and lithium battery power values, generating a list of valid power combinations; Energy Ratio Sorting Submodule: Based on the effective power combination list, the power ratio sorting algorithm reads the fuel cell power value and lithium battery power value in each combination record in turn, performs division to obtain the ratio result, generates a pair record with each ratio and the combination number, sorts all ratio data from high to low according to the numerical value, assigns consecutive numbers to all records according to the sorting position, and writes the number, ratio and combination parameters into a new record sequence to generate an effective energy allocation sequence.

[0026] The flight load response module includes: Power loading and binding submodule: Based on the effective energy allocation sequence, the fuel cell power and lithium battery power under the same time index are loaded to the converter power port, the vibration frequency values ​​at the same time are written into the combined record, and the power and frequency data are bound in the record. The command sequence is stored in sequence to generate the power frequency binding sequence. Thrust difference calculation submodule: Based on the power frequency binding sequence, the thrust value and attitude change amplitude of the propulsion unit are collected at each time index. The absolute difference of the thrust value and the attitude amplitude are combined into a data pair. The maximum and minimum value ranges of all data pairs are defined. The definition results are written into a table to generate flight thrust response records. Power loading and binding submodule: Based on the effective energy allocation sequence, the power frequency binding algorithm reads the fuel cell power value and lithium battery power value under the same time index, loads the fuel cell power value to the first power port corresponding to the converter, loads the lithium battery power value to the second power port corresponding to the converter, extracts the vibration frequency value under the same time index and writes it into the independent frequency field of the combined record, binds the power value and frequency value in the record in a key-value pair manner, and writes the bound records into the continuous storage buffer according to the time index order to form an instruction sequence, generating a power frequency binding sequence; The thrust difference calculation submodule, based on the power-frequency binding sequence, uses a thrust difference calculation algorithm to call the thrust value and attitude change amplitude value of the propulsion unit under each time index. It subtracts two adjacent thrust values ​​and takes the absolute value, and merges the absolute value and attitude change amplitude into a binary data pair. It stores the data pairs corresponding to all time indices into a list, compares the absolute thrust values ​​in the list one by one to identify the maximum and minimum values, defines the maximum and minimum values ​​as the difference range, matches them with the corresponding time index records, and writes them into the result table to generate a flight thrust response record.

[0027] The energy operation management module includes: Task segment data integration submodule: Based on flight thrust response records, thrust difference values, fuel cell power, lithium battery power, vibration frequency, and attitude change amplitude are grouped by mission number and flight segment number, and records under the same number are merged into a single-row data table to generate an energy mission segment record table; Full-process archive generation submodule: Based on the energy task segment record table, the records of each row are connected in chronological order to form a continuous link table, and all data in the link table is written to the long-term storage medium to generate a full-process energy scheduling archive. The task segment data integration submodule: Based on the flight thrust response record, the task number grouping integration algorithm is used to read the thrust difference value, fuel cell power, lithium battery power, vibration frequency, and attitude change amplitude line by line. The task number and flight segment number are used as a joint key value for matching. Multiple records with the same key value are sorted in ascending order by time index and the fields are merged. The merged parameter values ​​are written into a single-row format data table, and the task number and flight segment number columns are retained to generate an energy task segment record table. The full-process archive generation submodule: Based on the energy task segment record table, the time index field of each record is extracted using the time link construction algorithm. All records are sorted in ascending order by time index. The sorted records are then connected in row order to form a continuous link table. All fields in the link table, such as task number, flight segment number, thrust difference value, fuel cell power, lithium battery power, vibration frequency, and attitude change amplitude, are written into the long-term storage medium in the original column structure to generate a full-process energy dispatch archive.

[0028] A method for optimizing the range of a hydrogen-powered drone based on fuzzy PID dynamic hybrid control is provided. This method, implemented based on the aforementioned fuzzy PID-based dynamic hybrid control system for hydrogen-powered drones, includes the following steps: Step 1: Based on the power of the hydrogen fuel cell and the capacity of the lithium battery, calculate the pitch angle, roll angle and vertical velocity change rate in time sequence. Use the feedforward-fuzzy PID hybrid control algorithm to pair the change rate with the initial values ​​of the proportional, integral and derivative coefficients, compare the power change amplitude with the threshold marking trend, select the adjustment amount of the proportional, integral and derivative coefficients according to the trend, and generate a set of predicted power adjustment coefficients. Step 2: Based on the predicted power adjustment coefficient set, the multi-objective particle swarm optimization algorithm is used to combine the coefficients with attitude error, error change and power difference in sequence. The upper and lower limits of fuel cell power and lithium battery capacity are compared to screen available combinations. The attitude error and power difference are calculated and sorted. The proportional, integral and derivative coefficients of the top combinations are selected to form a control parameter group and generate a fusion control coefficient table. Step 3: Based on the fusion control coefficient table, match the control coefficients with the power of fuel cells and lithium batteries, eliminate combinations that exceed limits, calculate the power ratio, sort and number them, and generate an effective energy allocation sequence; Step 4: Based on the effective energy distribution sequence, load power to the distribution converter according to the time index and record the vibration frequency, collect thrust and attitude changes, calculate the thrust difference interval, and generate flight thrust response records; Step 5: Based on the flight thrust response record, sort and connect the thrust difference, power, vibration frequency, and attitude change according to the mission and flight segment number to establish a complete energy scheduling archive.

[0029] The above are merely preferred embodiments of the present invention and are not intended to limit the present invention in any other way. Any person skilled in the art may make changes or modifications to the above-disclosed technical content to create equivalent embodiments that can be applied to other fields. However, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A dynamic hybrid control system for hydrogen-powered unmanned aerial vehicles based on fuzzy PID control, characterized in that, The system includes: Flight power demand prediction module: Based on the power of hydrogen fuel cell and the capacity of lithium battery, calculate the attitude change rate and pair it with the initial values ​​of Kp, Ki, and Kd. Use PID control algorithm to compare the power change amplitude with the threshold to judge the trend. Select ΔKp, ΔKi, and ΔKd according to the results to generate a set of predicted power adjustment coefficients. Control parameter adaptive generation module: Based on the predicted power adjustment coefficient set, the particle swarm optimization algorithm is used to match the coefficients with the attitude error and the change, and the available combinations are screened in combination with power and energy limits. The selection is sorted by error and power difference to generate a fusion control coefficient table. Power allocation execution module: Based on the fusion control coefficient table, the control coefficients are matched with the power parameters of the fuel cell and lithium battery, out-of-limit combinations are eliminated, and the remaining combinations are sorted and numbered in descending order of power ratio to generate an effective energy allocation sequence; Flight load response module: Based on the effective energy distribution sequence, load the power values ​​of fuel cells and lithium batteries and record the vibration frequency during the execution cycle, synchronously acquire thrust and attitude changes, calculate the thrust difference range, and generate flight thrust response records; Energy Operation Management Module: Based on the flight thrust response records, thrust differences, power values, vibration frequencies, and attitude changes are sorted and stored according to mission number and flight segment number to establish a complete energy scheduling archive.

2. The dynamic hybrid control system for hydrogen-powered unmanned aerial vehicles based on fuzzy PID according to claim 1, characterized in that, The predicted power adjustment coefficient set includes proportional coefficient adjustment, integral coefficient adjustment, and derivative coefficient adjustment; the fused control coefficient table includes proportional coefficient, integral coefficient, and derivative coefficient; the effective energy allocation sequence includes fuel cell power value, lithium battery power value, and energy output ratio; the flight thrust response record includes thrust difference value, vibration frequency value, and attitude change amplitude; and the energy scheduling full-process archive includes mission number, flight segment number, full-process power record, and attitude change data.

3. The dynamic hybrid control system for hydrogen-powered unmanned aerial vehicles based on fuzzy PID according to claim 1, characterized in that, The flight power demand prediction module includes: Power and attitude calculation submodule: Based on the power of hydrogen fuel cell and the capacity of lithium battery, calculate the adjacent time difference of pitch angle, roll angle and vertical velocity and divide it by the time interval to obtain the rate of change. The prediction unit using PID control algorithm stores each set of rate of change in pairs with the initial values ​​of proportional coefficient, integral coefficient and derivative coefficient according to the index position to generate attitude power paired dataset. Trend State Determination Submodule: Based on the attitude power pairing dataset, compare the power change rate with the preset threshold record to determine whether the record is positive or negative, and determine whether the state is rising or falling. The absolute value of the change rate is compared with the interval range to determine the level code, and is paired with the state mark and written into the table to generate a power trend state table. The coefficient adjustment generation submodule: Based on the power trend status table, it retrieves the proportional coefficient adjustment amount, integral coefficient adjustment amount, and derivative coefficient adjustment amount corresponding to the combination of status marker and level code, arranges the selected adjustment amounts into a data sequence according to the original index order, and generates a set of predicted power adjustment coefficients.

4. The dynamic hybrid control system for hydrogen-powered unmanned aerial vehicles based on fuzzy PID according to claim 3, characterized in that, The PID control algorithm, upon receiving input data such as attitude angle change rate, fuel cell power, and lithium battery capacity, first reads the current proportional coefficient, integral coefficient, and derivative coefficient. It then multiplies the attitude error value by the proportional coefficient to obtain the proportional term output value. The algorithm accumulates the error sequence over time and multiplies it by the integral coefficient to obtain the integral term output value. It subtracts the previous cycle's error value from the current error value and divides it by the corresponding time interval to obtain the error change rate, which is then multiplied by the derivative coefficient to obtain the derivative term output value. Finally, it adds the proportional, integral, and derivative terms to form the comprehensive output result for the current cycle. This comprehensive output value is used to adjust control commands and records the corresponding time index. Simultaneously, it updates the accumulated error and the previous cycle's error value to the latest values.

5. The dynamic hybrid control system for hydrogen-powered unmanned aerial vehicles based on fuzzy PID according to claim 1, characterized in that, The adaptive control parameter generation module includes: Coefficient Error Matching Submodule: Based on the predicted power adjustment coefficient set, the particle swarm optimization algorithm is used to combine the proportional coefficient adjustment, integral coefficient adjustment, and differential coefficient adjustment with the attitude error value, error change value, and power difference value one by one within the adjustable parameter range. Each combination is recorded in the matching table and the performance index of the corresponding individual particle is retained to generate the error matching result table. Restriction Condition Filtering Submodule: Based on the error matching result table, compare the fuel cell power and lithium battery capacity with their respective upper and lower limits, mark and remove records that do not meet all conditions, retain records that meet the conditions and update the index, and generate an effective error power combination table; The parameter optimization output submodule calculates the sum of the attitude error value and the power difference as a comprehensive value based on the effective error power combination table. The comprehensive value is sorted in ascending order, and the proportional coefficient, integral coefficient, and derivative coefficient are extracted from the first record to form a control parameter group, generating a fusion control coefficient table.

6. The dynamic hybrid control system for hydrogen-powered unmanned aerial vehicles based on fuzzy PID according to claim 6, characterized in that, The particle swarm optimization algorithm generates a combined vector for each particle during the initialization phase, containing proportional coefficient adjustment, integral coefficient adjustment, differential coefficient adjustment, attitude error value, error change value, and power difference value. Simultaneously, each particle is assigned a random velocity vector, and its current position is recorded as its individual adaptation position. Positions with the best fitness values ​​among all particles are marked as global adaptation positions. During iteration, the fitness value of each particle is calculated sequentially and compared with the fitness of its individual adaptation position. When the current fitness value is better, the individual adaptation position is updated. The current fitness value is then compared with the global adaptation position fitness; if better, the global adaptation position is updated synchronously. Based on the current velocity, current position, individual adaptation position, and global adaptation position, a velocity increment is calculated according to a set weight and added to the current velocity. The updated velocity is then added to the current position to obtain a new search position. The proportional coefficient adjustment, integral coefficient adjustment, differential coefficient adjustment, and attitude parameter combination in the position are written into a matching table. After reaching a preset number of iterations, the fitness-adapted particle position is output as the result of this round of search.

7. The dynamic hybrid control system for hydrogen-powered unmanned aerial vehicles based on fuzzy PID according to claim 1, characterized in that, The power allocation execution module includes: Power matching screening submodule: Based on the fusion control coefficient table, the proportional coefficient, integral coefficient, and derivative coefficient are compared with the fuel cell power and lithium battery power in the same row and column by column to determine whether each item is greater than or less than the corresponding safety threshold. Any combination that does not meet the conditions is eliminated, and all the remaining combinations are stored in the list in the original order to generate a list of effective power combinations. Energy ratio sorting submodule: Based on the effective power combination list, the ratio of fuel cell power to lithium battery power is obtained for each combination data. The ratio value and combination number are recorded in pairs and sorted from high to low according to the ratio value. The sorting results are assigned consecutive numbers and the numbers and ratios are integrated into the record sequence to generate an effective energy allocation sequence.

8. The dynamic hybrid control system for hydrogen-powered unmanned aerial vehicles based on fuzzy PID according to claim 1, characterized in that, The flight load response module includes: Power loading and binding submodule: Based on the effective energy allocation sequence, the fuel cell power and lithium battery power under the same time index are loaded to the converter power port, the vibration frequency values ​​at the same time are written into the combined record, and the power and frequency data are bound in the record. The command sequence is stored in sequence to generate a power frequency binding sequence. Thrust difference calculation submodule: Based on the power frequency binding sequence, the thrust value and attitude change amplitude of the propulsion unit are collected at each time index. The absolute difference of the thrust value and the attitude amplitude are combined into a data pair. The maximum and minimum value ranges of all data pairs are defined, and the definition results are written into a table to generate a flight thrust response record.

9. The dynamic hybrid control system for hydrogen-powered unmanned aerial vehicles based on fuzzy PID according to claim 1, characterized in that, The energy operation management module includes: Task segment data integration submodule: Based on the flight thrust response record, the thrust difference value, fuel cell power, lithium battery power, vibration frequency, and attitude change amplitude are grouped according to the task number and flight segment number. Records under the same number are merged into a single-row data table to generate the energy task segment record table. Full-process archive generation submodule: Based on the energy task segment record table, the records in each row are connected in chronological order to form a continuous link table, and all data in the link table is written to a long-term storage medium to generate a full-process energy scheduling archive.

10. A method for optimizing the endurance of a hydrogen-powered unmanned aerial vehicle (UAV) based on fuzzy PID control, characterized in that, The execution of the dynamic hybrid control system for hydrogen-powered unmanned aerial vehicles based on fuzzy PID according to any one of claims 1-9 includes the following steps: Step 1: Based on the power of the hydrogen fuel cell and the capacity of the lithium battery, calculate the pitch angle, roll angle and vertical velocity change rate in time sequence. Use the feedforward-fuzzy PID hybrid control algorithm to pair the change rate with the initial values ​​of the proportional, integral and derivative coefficients, compare the power change amplitude with the threshold marking trend, select the adjustment amount of the proportional, integral and derivative coefficients according to the trend, and generate a set of predicted power adjustment coefficients. Step 2: Based on the predicted power adjustment coefficient set, the multi-objective particle swarm optimization algorithm is used to combine the coefficients with attitude error, error change, and power difference in sequence. The upper and lower limits of fuel cell power and lithium battery capacity are compared to screen available combinations. The attitude error and power difference are calculated and sorted. The proportional, integral, and derivative coefficients of the top combinations are selected to form a control parameter group and generate a fusion control coefficient table. Step 3: Based on the fusion control coefficient table, match the control coefficients with the power of the fuel cell and lithium battery, eliminate combinations that exceed limits, calculate the power ratio, sort and number them, and generate an effective energy allocation sequence; Step 4: Based on the effective energy distribution sequence, load power to the distribution converter according to the time index and record the vibration frequency, collect thrust and attitude changes, calculate the thrust difference interval, and generate flight thrust response record; Step 5: Based on the flight thrust response record, sort and connect the thrust difference, power, vibration frequency, and attitude change according to the mission and flight segment number to establish a complete energy scheduling archive.

Images