Air endurance control method of multi-rotor unmanned aerial vehicle with new energy auxiliary drive

Abstract

This invention discloses an in-flight endurance control method for a multi-rotor UAV driven by new energy sources, relating to the field of multi-rotor UAV control. This invention utilizes the existing rotor state information of the UAV to identify the shearing ripple synchronized with the rotor from the output current or power of the photovoltaic sub-region. Next, an obstruction time slot mask is generated; within the obstruction time slot, relevant samples are removed and weighted down; within the non-obstruction time slot, the true power of the photovoltaic sub-region is reconstructed, and this reconstructed power is used for MPPT updates and photovoltaic sub-region access status determination. This allows the multi-rotor UAV, when performing solar power tracking, to no longer simply execute a "the higher the photovoltaic power, the better" control strategy, but instead establishes a net benefit evaluation of candidate attitudes, reducing the interference of rotor shearing on the photovoltaic power tracking system, reducing erroneous updates, erroneous bypasses, and ineffective solar power tracking attitude adjustments, and improving the actual benefit of the multi-rotor UAV's new energy-assisted endurance.

Description

Technical Field

[0001] This invention relates to the field of multi-rotor unmanned aerial vehicle (UAV) control, and specifically to a method for controlling the in-flight endurance of a multi-rotor UAV driven by a new energy source. Background Technology

[0002] Multi-rotor drones consist of an airframe, multiple rotors, motors, propellers, an electronic speed controller (ESC), a flight controller, and a battery. Each rotor is independently driven by its corresponding motor. Multi-rotor drones offer advantages such as vertical takeoff and landing, hovering, low-speed maneuvering, and flexible control, and are widely used in tasks such as power line inspection, environmental monitoring, emergency communication, agricultural and forestry patrols, and security patrols. Currently, most mainstream electric multi-rotor drones rely on lithium batteries for power. However, lithium batteries have limited energy density, and drones require significant propulsion power during hovering, attitude maintenance, and maneuvering. Therefore, the effective operating time per mission is typically short, and endurance remains a crucial factor limiting the long-duration mission capabilities of multi-rotor drones.

[0003] To extend the endurance of drones, common methods in existing technologies include increasing battery capacity, using hybrid power, fuel cells, solar-assisted power supply, and wind energy harvesting. While increasing battery capacity can improve total energy storage, it also increases takeoff weight, leading to increased hovering and maneuvering power consumption. Hybrid power systems are complex in structure, resulting in higher vibration, noise, and maintenance costs. External wind energy harvesting devices or small wind turbine structures can easily increase drag, and on multi-rotor platforms, they are also affected by rotor downwash and turbulence, where the energy gained from power generation is offset by the energy consumed in propulsion.

[0004] Solar-assisted power generation is a relatively lightweight new energy source for extended flight. Existing solutions typically involve deploying flexible or rigid photovoltaic (PV) modules on the drone's fuselage, arms, or other sun-receiving surfaces, along with a maximum power point tracking (MPPT) controller to ensure the PV modules operate as close to their maximum power point as possible. However, multi-rotor drones differ from fixed-wing solar-powered aircraft in that they experience frequent attitude changes, yaw adjustments, pitch / roll corrections, and periodic rotor sweeps during flight. For PV skins deployed on the fuselage top cover, upper arm surfaces, or near the rotor area, the rotor blades periodically block the PV sub-regions during rotation, creating a "rotor shearing" phenomenon.

[0005] Rotor shearing causes periodic drops in the output current or power of the photovoltaic sub-region, which are related to the rotor speed and the number of blades. These power drops are not due to ordinary cloud shadows or a fault in the photovoltaic sub-region itself, but rather to periodic shading caused by the structure and movement of the multi-rotor drone. If a traditional MPPT controller directly updates its data using voltage, current, or power data containing samples of this shading, it may mistakenly interpret instantaneous power drops as changes in the maximum power point, leading to operating point oscillations or incorrect updates. Furthermore, if the power management unit determines the inefficiency of a photovoltaic sub-region based on the window average power, rotor shearing may misclassify that sub-region as abnormal or inefficient, resulting in unnecessary bypassing, deweighting, or reconfiguration, ultimately reducing the effective output of the photovoltaic array.

[0006] Existing photovoltaic (PV) anti-shading technologies mainly target building shadows, cloud shadows, partial shading of modules, or static shading of PV arrays. Common methods include bypass diodes, multiple MPPTs, global scanning, and array reconfiguration. These methods typically do not utilize the existing correspondence between the rotor speed, blade passing frequency, rotor phase, and the geometric position of the PV sub-region of the multi-rotor UAV itself, nor do they consider the PV current ripple caused by rotor light cutting as an identifiable information source.

[0007] Furthermore, when drones perform photovoltaic (PV) power replenishment, their flight attitude has a significant impact on the replenishment effect. Intuitively, aligning the photovoltaic skin as directly as possible with the sun can increase PV output power. However, for multi-rotor drones, attitude adjustment itself alters rotor thrust distribution, airframe frontal area, and position-keeping power consumption. Especially in the presence of crosswinds, gusts, or relative airflow, the load on each rotor changes under different attitudes. It is possible that a certain attitude may increase PV power but simultaneously increase propulsion system power consumption more significantly.

[0008] The solar power replenishment attitude should not only aim at maximizing instantaneous solar power, but should also establish a net benefit evaluation model for the entire system, taking into account changes in solar power gain, propulsion power consumption, and wind-induced rotor load. By rationally selecting the attitude, it is possible to reduce the thrust demand of the rotor to some extent by utilizing favorable airflow, or to avoid entering attitudes with high wind resistance and severe rotor load imbalance, so that the effects of solar power replenishment and wind energy can jointly participate in the UAV's in-flight endurance control.

[0009] Therefore, a new energy-assisted endurance control method is needed for multi-rotor photovoltaic UAVs: on the one hand, by utilizing the rotor shear ripple in the output current or power of the photovoltaic sub-region, the rotor shading time slot is identified, shading samples are eliminated or downweighted, the real photovoltaic power in the unshaded state is reconstructed, and MPPT updates and sub-region access / bypass control are performed based on the real power; on the other hand, when selecting the refueling attitude, the maximum photovoltaic power is not simply pursued, but the changes in propulsion power consumption and wind-induced rotor load are combined to determine whether the candidate attitude truly improves the net benefit of the whole aircraft, thereby improving the actual effect of new energy-assisted endurance. Summary of the Invention

[0010] To address the aforementioned shortcomings in the existing technology, the in-flight endurance control method for multi-rotor UAVs driven by new energy provided by this invention solves the problem of distorted photovoltaic power evaluation caused by rotor periodic light cutting, which in turn leads to incorrect MPPT updates or low power evaluation.

[0011] To achieve the above-mentioned objectives, the technical solution adopted by this invention is as follows: A method for controlling the in-flight endurance of a multi-rotor unmanned aerial vehicle (UAV) powered by a new energy source is provided, wherein the solar-receiving surface of the UAV is provided with a conformal photovoltaic skin composed of several photovoltaic sub-regions. The method includes the following steps: Collect temperature and current signals from each photovoltaic sub-region; obtain the phase and blade passing frequency of each rotor; The current signal of the photovoltaic sub-region is filtered and synchronized to extract the periodic shear ripple corresponding to the blade passing frequency of the rotor, and the characteristics of the periodic shear ripple are obtained. Based on the characteristics of rotor phase, blade passing frequency, periodic shear ripple, shading duty cycle, and geometric relationship between rotor and photovoltaic sub-region, a shading time slot mask for each photovoltaic sub-region is generated. The credibility of photovoltaic sub-region samples at each sampling time is determined based on the occlusion time slot mask, and the photovoltaic sub-region samples are weighted based on the credibility of the photovoltaic sub-region samples. Based on the shading time slot mask and the sample weight of the photovoltaic sub-region, the true power of the photovoltaic sub-region in the unshading state is reconstructed. Based on the actual power, shading duty cycle, temperature, impact on the bus, and MPPT reliability obtained from the reconstruction of each photovoltaic sub-region under the unshaded state, the corresponding photovoltaic sub-region is controlled to remain connected or bypassed in order to reduce false bypass caused by rotor periodic light cutting. MPPT updates are controlled based on the occlusion time slot mask and the actual power in the unoccluded state obtained from reconstruction. Based on the actual power in the unshaded state obtained from the reconstruction of each photovoltaic sub-region, the candidate attitudes of the multi-rotor UAV are evaluated, and the candidate attitude that satisfies the flight safety constraints and has the largest net benefit is selected as the execution attitude.

[0012] The beneficial effects of this invention are as follows: 1. This invention addresses the unique rotor shedding phenomenon in multi-rotor UAVs by utilizing the synchronous ripple of the output current of the photovoltaic sub-region to identify the shading state, thus achieving self-sensing of the photovoltaic skin. It generates a shading time slot mask based on rotor speed, blade passing frequency, or rotor phase, preventing the MPPT controller from directly updating with abnormal power samples within the shading time slot, thereby reducing MPPT operating point oscillation and output power fluctuations. Furthermore, it reconstructs the true power of the photovoltaic sub-region using non-shading samples, avoiding misjudging short-term power drops caused by rotor periodic shedding as long-term inefficiency of the sub-region, thereby reducing false bypassing, false weighting, or false reconstruction.

[0013] 2. This invention introduces a net benefit evaluation of the entire system into the photovoltaic (PV) power replenishment attitude selection process. Instead of simply using the maximum instantaneous PV power as the attitude adjustment target, it simultaneously considers the incremental PV power, the incremental propulsion power consumption, and the degree of wind-induced rotor load imbalance. This allows the system to maintain its original attitude when facing direct sunlight would significantly increase propulsion energy consumption, and then perform a beam-tracking attitude adjustment only when the PV gain can offset the propulsion cost, making PV power replenishment more consistent with the overall net benefit of the system.

[0014] 3. This invention mainly utilizes the existing ESC, flight control, and photovoltaic sampling information of the UAV. It utilizes wind energy factors, mainly the influence of crosswinds, gusts, and relative airflow on rotor load and propulsion power consumption, and enters the attitude control process through net benefit evaluation. Therefore, it can improve the rationality of energy replenishment attitude selection without significantly increasing hardware complexity. Attached Figure Description

[0015] Figure 1 This is a flowchart illustrating the method. Figure 2 A schematic diagram of the photovoltaic skin arrangement for a drone; Figure 3 This is a schematic diagram of rotor beam cutting. Figure 4 Flowchart for generating the masking time slot; Figure 5 This is a schematic diagram of MPPT time-slotted updates; Figure 6 This is a schematic diagram comparing incorrect bypasses in the photovoltaic sub-region. Detailed Implementation

[0016] The specific embodiments of the present invention are described below to enable those skilled in the art to understand the present invention. However, it should be understood that the present invention is not limited to the scope of the specific embodiments. For those skilled in the art, various changes are obvious as long as they are within the spirit and scope of the present invention as defined and determined by the appended claims. All inventions utilizing the concept of the present invention are protected.

[0017] like Figure 1 As shown, the in-flight endurance control method for this new energy-assisted multi-rotor UAV includes the following steps: S1. Collect temperature and current signals of each photovoltaic sub-region; obtain the phase and blade passing frequency of each rotor; S2. Filter and synchronize the current signal of the photovoltaic sub-region, extract the periodic shear ripple corresponding to the blade passing frequency of the rotor, and obtain the characteristics of the periodic shear ripple. S3. Based on the characteristics of rotor phase, blade passing frequency, periodic shear ripple, shading duty cycle, and geometric relationship between rotor and photovoltaic sub-region, generate shading time slot mask for each photovoltaic sub-region. S4. Confirm the credibility of photovoltaic sub-region samples at each sampling time based on the occlusion time slot mask, and assign weights to the photovoltaic sub-region samples based on the credibility of the photovoltaic sub-region samples. S5. Based on the shading time slot mask and the sample weight of the photovoltaic sub-region, reconstruct the true power of the photovoltaic sub-region in the unshading state; S6. Based on the actual power, shading duty cycle, temperature, impact on the bus, and MPPT reliability of each photovoltaic sub-region under the unshaded state obtained from the reconstruction, control the corresponding photovoltaic sub-region to maintain access or bypass in order to reduce false bypass caused by rotor periodic light cutting. S7. Control MPPT updates based on the occlusion time slot mask and the actual power in the unoccluded state obtained from reconstruction; S8. Based on the actual power in the unshaded state obtained from the reconstruction of each photovoltaic sub-region, evaluate the candidate attitudes of the multi-rotor UAV and select the candidate attitude that satisfies the flight safety constraints and has the largest net benefit as the execution attitude.

[0018] In the specific implementation process, such as Figure 2 As shown, conformal photovoltaic (PV) skin is arranged on the top cover of the UAV fuselage, the upper surface of the arms, the upper surface of the landing gear, or other light-receiving surfaces. To reduce the impact of local shading on the overall output, the PV skin is divided into N PV sub-regions, where N ≥ 2. Each PV sub-region is equipped with an independent voltage and current sampling channel or partially shared with other PV sub-regions, and the PV power supply bus is managed through a DC / DC converter, an MPPT controller, and an access switch. i Each photovoltaic sub-region is in The output power at time t can be expressed as:

[0019] in, For the first i The output voltage of each photovoltaic sub-region For the first i The output current of each photovoltaic sub-region.

[0020] For the i The output current of each photovoltaic sub-region can be approximately decomposed as follows:

[0021] in, This represents the low-frequency or quasi-steady-state components related to actual irradiation and operating point. This represents the synchronization ripple component caused by the periodic shearing of the rotor. Indicates noise.

[0022] like Figure 3 As shown, when a photovoltaic sub-region exhibits significant ripple at the frequency of a rotor blade, and the ripple has a stable correspondence with the frequency of the rotor, it can be determined that the photovoltaic sub-region is affected by the periodic shearing effect of the rotor.

[0023] To identify the periodic shear ripples caused by rotor shear, the photovoltaic sampling frequency is preferably more than 5 times the rotor blade passage frequency, and more preferably more than 10 times. The sampled data can be sent to the flight control center for real-time processing.

[0024] This method targets a flight controller or ESC in a new energy-assisted multi-rotor UAV that can acquire at least one of the following information: rotor speed, commutation frequency, electrical angle, mechanical phase, or blade passage frequency. When the rotor angular velocity is expressed in rad / s, the blade passage frequency can be expressed as:

[0025] in Indicates the first j The frequency of each rotor blade; For the first j The number of blades in each rotor; For the first j The angular velocity of each rotor; In radians, that is, 180°; Indicates time.

[0026] In some embodiments, such as Figure 4 As shown, the specific methods for filtering and synchronizing the current signal of the photovoltaic sub-region to extract the periodic shear ripple corresponding to the blade passage frequency include: Perform filtering: For the first Each photovoltaic sub-region is in Output current signal at time Filtering is performed to estimate the low-frequency or quasi-steady-state components that are related to the actual irradiance and operating point variations. The estimated low-frequency or quasi-steady-state component is subtracted from the output current signal to obtain the residual signal. : ; From residual signal In the middle, with the first j The blades of the first rotor undergo frequency selection processing to extract signals within a preset frequency band, thus obtaining signals consistent with the first rotor blade. j The signal corresponding to each spin is used as a candidate synchronous ripple component. The frequency selection process includes bandpass filtering centered on the blade passage frequency or frequency domain filtering targeting the corresponding frequency band. As the blade passage frequency changes with the rotor speed, the preset frequency band is updated accordingly. The bandwidth of the preset frequency band is determined based on the rotor speed fluctuation range and the photovoltaic sampling frequency.

[0027] Perform synchronous analysis: construct and Time of the first j Each rotor blade is synchronized by a frequency-corresponding reference signal. And calculate the correlation between the output current of the photovoltaic sub-region and the synchronization reference signal:

[0028] in, express Time of the first i The output current of the photovoltaic sub-region is related to the first j The correlation of the rotor blades with the corresponding frequency of the synchronization reference signal; For the corr function; when Greater than the preset correlation threshold and candidate synchronization ripple components When the amplitude is greater than the preset amplitude threshold, the first... i The photovoltaic sub-region exists due to the first j Periodic shearing caused by individual rotors; candidate synchronous ripple components It is identified as a periodic shear wave; the characteristics of the periodic shear wave include amplitude, frequency and phase.

[0029] For example, when it is possible to obtain the first One rotor Mechanical phase at time At that time, it can be based on the number of blades Determine the phase through which the blade passes:

[0030] When only the first can be obtained The rotor blades pass through a frequency When the blade passes through, the phase can be estimated using the following formula:

[0031] in, Indicates the initial blade passing phase; synchronization reference signal It can be phased with the blade Synchronous periodic functions.

[0032] It should be noted that after identifying the periodic light cut-off, the above process does not directly filter out the ripple as ordinary noise, but uses it as a basis for judging the sample credibility. For power drop samples that are highly synchronized with the rotor blade passing frequency, they are marked as low-credibility samples affected by rotor obstruction; for samples not within the obstruction phase interval, they are marked as non-obstruction credible samples. Subsequent MPPT updates and sub-region access judgments no longer directly rely on the full-window average power, but instead rely on the valid samples after obstruction state differentiation.

[0033] In some embodiments, the expression for generating the shading time slot mask of the photovoltaic sub-region is:

[0034] in, express Time of the first Shading time slot mask for each photovoltaic sub-region; Indicates the first Each photovoltaic sub-region forms a periodically shaded set of associated rotors; Indicates the first One rotor The blade at time traverses the phase and satisfies ; Indicates the first The number of blades in each rotor; Indicates the first One rotor Mechanical phase at any given moment; Indicates the first The rotor pairs with the first The shading center phase when a photovoltaic sub-region forms a shading; express The rotor pairs with the first Shading duty cycle of each photovoltaic sub-region; This indicates that the phase difference is constrained to a preset principal value range; The indicator function takes the value 1 when the condition in parentheses is true, and 0 otherwise. The shading time slot mask is represented in binary form. When the shading time slot mask is 1, it means that the sampling point of the corresponding photovoltaic sub-region is in the rotor shading time slot. When the shading time slot mask is 0, it means that the sampling point of the corresponding photovoltaic sub-region is in the non-shading time slot.

[0035] In some embodiments, the reliability of photovoltaic sub-region samples at each sampling time is determined based on the shading time slot mask, and the specific method for assigning weights to photovoltaic sub-regions based on the reliability of photovoltaic sub-region samples includes: sampling time Distance between the nearest occlusion slot boundary As Time of the first i Confidence of individual photovoltaic sub-region samples; based on Transition width with preset occlusion boundary The size relationship is used to assign weights to the photovoltaic sub-regions, and the expression is as follows:

[0036] in for Time of the first i Each photovoltaic sub-region sample weight.

[0037] It should be noted that, and The same time unit or phase unit is used. When the sampling time is in an unobstructed time slot but close to the boundary of an obstructed time slot, the sample weight increases with the distance between the sampling time and the boundary of the obstructed time slot, taking into account the possible deviation in the judgment of the obstruction state caused by rotor phase error, sampling delay or filtering delay; when the distance is not less than the preset obstruction boundary transition width, the sample weight is 1.

[0038] In some embodiments, the expression for calculating the actual power of the photovoltaic sub-region under unshaded conditions is as follows:

[0039] in, Indicates the first i The actual power of a photovoltaic sub-region within a sampling window W under unshaded conditions; Indicates the first i Each photovoltaic sub-region is in The occlusion time slot mask at any given time; the occlusion time slot mask is represented in binary form. When the occlusion time slot mask is 0, it means that the sampling point of the corresponding photovoltaic sub-region is in an unoccluded time slot. For the first i Each photovoltaic sub-region is in The output power at each time step; a sampling window W contains several sampling times.

[0040] It should be noted that for photovoltaic sub-region samples (referring to the output data of the photovoltaic sub-region, including one or more of voltage, current, and power) within the shading time slot, they can be directly discarded or their weight can be reduced; for (power) samples within the unshading time slot, they are used as the basis for the true (power) evaluation. If the number of unshading (power) samples within the sampling window is sufficient, the mean of the unshading (power) samples is used as the true (power) estimate of the photovoltaic sub-region; if the number of unshading (power) samples is small, estimation can be performed by interpolation of adjacent unshading (power) samples, historical unshading (power) sample values, etc., to obtain the (power) estimate of the unshading state of the photovoltaic sub-region.

[0041] By employing the above method, this invention replaces the traditional "average power of the sample window including shading" with "reconstructed power of the unshading reliable sample". This reconstructed power is closer to the actual power generation capacity of the photovoltaic sub-region under the current attitude and illumination conditions, and therefore can serve as a unified power basis for MPPT updates, bypass judgments, and energy replenishment attitude evaluations. This avoids misinterpreting short-term power drops caused by rotor light cutting as long-term inefficiency of the photovoltaic sub-region, and also reduces false bypasses caused by an excessively low average power of the window.

[0042] In this embodiment, the MPPT controller controls the maximum power point tracking (MPPT) of each photovoltaic sub-region based on the shading time slot mask and the reconstructed true power of the unshading time slot. Unlike traditional MPPT which directly uses all power samples within the sampling window, this invention first determines whether the power sample is in a rotor shading time slot before updating the MPPT. Power drop samples within shading time slots are not directly used as the basis for updating other MPPT algorithms; reliable power samples within unshading time slots are used to evaluate the impact of operating point changes on the true output power. It should be noted that the above operation can also be performed based on voltage or current samples; power is not the only choice.

[0043] It should be noted that MPPT reliability is used to evaluate whether reliable photovoltaic sampling data within the current MPPT update sampling window is sufficient. When the MPPT update reliability meets the preset conditions, MPPT updates are allowed, and sub-region status evaluation is allowed based on the unshaded state power estimate of the current sampling window; when it is lower than the preset conditions, the current MPPT operating point is maintained, and the photovoltaic sub-region access status is not changed according to the current sampling window.

[0044] Specifically, when controlling MPPT updates, if an occlusion slot is detected in the current sampling window, at least one of the following operations is performed: (1) Suspend the use of voltage, current or power samples to update MPPT during the blocking time slot; (2) Keep the current MPPT duty cycle or operating voltage unchanged; (3) Reduce the weight of shaded photovoltaic sub-region samples in MPPT updates; where photovoltaic sub-region samples refer to the output data of photovoltaic sub-regions; (4) Delay MPPT updates to non-occlusion time slots; (5) Use the real power obtained from the reconstruction under the unshaded state to update the MPPT.

[0045] The above operations can reduce the interference of instantaneous power drop caused by rotor cut-off on MPPT control, and reduce operating voltage oscillation and low-frequency power fluctuations. Figure 5 This describes one implementation of a time-slotted update method in the above-described MPPT control approach. During unshaded time slots, the MPPT controller performs MPPT updates using reliable photovoltaic sampling data; during rotor shaded time slots, MPPT updates are paused while maintaining the current MPPT duty cycle or operating voltage; when the shaded time slot ends and the system re-enters an unshaded time slot, MPPT updates resume based on newly acquired reliable photovoltaic sampling data.

[0046] In some embodiments, specific methods for controlling the corresponding photovoltaic sub-region to remain connected or bypassed include: When the average power of a photovoltaic sub-region calculated based on all power samples within the evaluation window is lower than the first average power threshold, the proportion of low-power samples within the rotor shading time slot to the total number of low-power samples within the evaluation window exceeds a preset proportion threshold, and the power estimate of the unshaded state of the photovoltaic sub-region is within a preset normal power range, it is determined that the power decrease of the photovoltaic sub-region is caused by rotor periodic light cutting, and the photovoltaic sub-region is kept connected; where low-power samples are power samples within the evaluation window that are lower than the first power sample threshold; When the actual power of a photovoltaic sub-region under unshaded conditions obtained by reconstruction is lower than the second preset power threshold within a set time period, and is accompanied by abnormal temperature rise, open circuit / short circuit characteristics, or the impact on the bus reaches the preset impact threshold, the photovoltaic sub-region is controlled to bypass, or the weight of the photovoltaic sub-region in the power evaluation is reduced.

[0047] The impact of the photovoltaic sub-region on the photovoltaic supplementary power bus can be determined based on the differences or changes in the bus operating parameters when the photovoltaic sub-region is in the connected and bypass states. Bus operating parameters include at least one of the following: bus voltage, bus voltage fluctuation amplitude, bus net input power, and reverse current. When at least one of the following—the decrease in bus voltage, the increase in bus voltage fluctuation amplitude, the decrease in bus net input power, or the reverse current—caused by the connection of the photovoltaic sub-region reaches the corresponding preset impact threshold, the impact of the photovoltaic sub-region on the photovoltaic supplementary power bus is determined to have reached the preset impact threshold.

[0048] The above operation changes the sub-region bypass judgment from the traditional "whether the average power of the window is too low" to "whether the actual power of the unshaded area is too low for a long time", thereby reducing false bypass caused by rotor periodic light cutting and improving the effective access time and total output power of the photovoltaic array.

[0049] It should be noted that when a multi-rotor UAV uses photovoltaic skin for power replenishment, the flight controller can evaluate the small yaw, pitch, or roll attitude based on the reconstructed actual power in the unobstructed state. This attitude evaluation is not simply about keeping the UAV always facing the sun, but rather about determining whether a candidate attitude can bring positive benefits at the overall UAV level.

[0050] Specifically, a candidate attitude may improve the incident angle of the photovoltaic skin, thereby increasing photovoltaic output power; however, it may also alter the airframe's frontal area, rotor inflow conditions, and thrust distribution among the rotors, increasing propulsion system power consumption. Especially in the presence of crosswinds, gusts, or relative airflow, different attitudes will cause varying degrees of wind-induced rotor load changes. If the attitude is chosen appropriately, some of the incoming airflow can reduce the thrust requirements of certain rotors or decrease ineffective attitude correction power consumption while maintaining flight stability; if the attitude is chosen inappropriately, it may increase wind resistance or cause unbalanced rotor load, increasing propulsion power consumption.

[0051] Therefore, in this embodiment, the specific method for evaluating the candidate attitudes of a multi-rotor UAV based on the actual power in the unshaded state reconstructed from each photovoltaic sub-region includes the following steps: S8-1. In response to the photovoltaic energy replenishment attitude adjustment command, calculate the photovoltaic power increment brought about by the candidate attitude based on the real power in the unshaded state obtained by the reconstruction of each photovoltaic sub-region; at the same time, calculate the power consumption increment of the propulsion system and the degree of imbalance of the power change of each rotor caused by the attitude adjustment. S8-2. Calculate the net benefit evaluation value of the corresponding candidate attitude based on the unevenness of the photovoltaic power increment, the propulsion system power consumption increment, and the power change of each rotor. S8-3. Determine if there is a candidate attitude with a positive net benefit. If so, select the candidate attitude that satisfies the flight safety constraints and has the largest net benefit as the execution attitude; otherwise, maintain the current attitude.

[0052] The formula for calculating the net income evaluation value is as follows:

[0053] in, Indicate candidate pose Net income valuation; The photovoltaic power increment represents the actual power obtained under the unshaded state after reconstruction; This represents the increase in propulsion system power consumption caused by attitude adjustment; Indicates the first j The power change of each rotor relative to the current attitude under the candidate attitude; It indicates the degree of imbalance in the power variation of each rotor, and is used to reflect the load imbalance caused by crosswinds, gusts, or relative airflow. These are the weighting coefficients.

[0054] It should be noted that in this candidate attitude evaluation process, the current attitude is used as a fixed comparison benchmark, and the photovoltaic power increment, propulsion system power consumption increment, and rotor power change corresponding to each candidate attitude are all calculated relative to the current attitude.

[0055] In practical control, small yaw, pitch, or roll attitudes near the current attitude can be used to form a candidate attitude set. The flight controller calculates the values ​​for each candidate attitude. and select those that meet flight safety constraints and The maximum attitude is taken as the energy replenishment attitude, that is:

[0056] Flight safety constraints include attitude error not exceeding permissible limits, motor current not exceeding limits, bus voltage within safe ranges, and no significant long-term spikes in rotor power. When all candidate attitudes... When all values ​​are less than the preset threshold, the drone maintains its current attitude.

[0057] By employing the above method, this invention does not simply pursue maximum photovoltaic power, but rather incorporates the incremental photovoltaic power after reconstruction, changes in propulsion power consumption, and wind-induced rotor load changes into the judgment. When a candidate attitude can increase the reconstructed photovoltaic power without significantly increasing propulsion power consumption or causing rotor load imbalance, that attitude is executed; when a candidate attitude increases photovoltaic power but causes a significant increase in wind resistance or rotor load imbalance, the current attitude is maintained or another attitude is selected. This reduces the situation where "photovoltaic power increases but propulsion power consumption increases even more," making the UAV's in-flight endurance control more aligned with the overall energy efficiency of the aircraft.

[0058] To verify the effectiveness of this method, this embodiment establishes an energy balance estimation model between the revenue from photovoltaic power generation and the cost of increased weight of the photovoltaic system. This model is used to determine whether the equivalent average discharge power of the UAV battery decreases after adding photovoltaic skin, and whether the equivalent flight time after adding the photovoltaic system is greater than the flight time without it.

[0059] Using publicly available industrial-grade multirotor drone parameters as a reference, the DJI Matrice 350 RTK-class industrial multirotor drone was selected as the benchmark. This type of drone has unfolded dimensions of approximately 810 mm × 670 mm × 430 mm and weighs approximately [weight missing] with two TB65 batteries. It weighs approximately 6.47 kg and has a maximum flight time of approximately 55 minutes. Each TB65 battery has an energy of 263.2 Wh, therefore the total energy of both batteries is: ; Assume the normal driving range without a photovoltaic system is: ; The average power consumption without a photovoltaic system is: ; Substituting the data, we get: ; This value represents the average power required by the battery to maintain normal flight when the drone is not equipped with a photovoltaic system.

[0060] Considering conformal mounting areas such as the fuselage top cover and upper arm surface, as well as non-conformal mounting areas such as edges, connectors, and rotor mounting areas, the effective photovoltaic coverage area is taken as: ; This embodiment preferably uses a lightweight, flexible gallium arsenide photovoltaic skin. Lightweight, flexible gallium arsenide photovoltaics have high conversion efficiency and low areal density, making them suitable for conformal mounting on the surface of multi-rotor drone fuselages or arms. The photovoltaic conversion efficiency is taken as... Take the surface density of photovoltaic materials as .

[0061] Considering the mass of additional components such as adhesive layers, wires, and small DC / DC or MPPT power management circuits, the additional mass is taken as... .

[0062] The added mass of the photovoltaic system is: ; Substituting the data, we get: ; Solar irradiance under standard photovoltaic testing conditions is typically taken as G =1000W / m².

[0063] Considering the effects of attitude angle changes, photovoltaic temperature rise, DC / DC conversion losses, rotor periodic light shearing and non-perfect solar orientation during actual drone flight, the actual reduction factor is taken as: k = 0.65.

[0064] The average effective output power of the photovoltaic skin is: ; Substituting the data, we get: .

[0065] This power rating represents the average effective electrical power that the photovoltaic skin can provide to the drone's power system under the aforementioned lighting and installation conditions.

[0066] On the other hand, adding photovoltaic skin increases the weight of the drone, requiring the rotor to provide greater lift. In hovering or low-speed flight, the rotor-induced power is approximately a 3 / 2 power of the aircraft's weight. Therefore, the additional propulsion power caused by the increased weight of the photovoltaic system can be estimated as follows:

[0067] Where m0 represents the mass of the drone without the photovoltaic system installed. Add quality to photovoltaic systems.

[0068] but: .

[0069] Considering the additional aerodynamic drag and power consumption of the power management circuit caused by the bonded photovoltaic skin, we conservatively take:

[0070] in, This indicates the additional aerodynamic drag power consumption. This indicates the power consumption of the photovoltaic power management circuit.

[0071] Since the electrical energy output from the photovoltaic skin can directly supply the onboard electrical load or be fed back to the power battery side through the module, the average effective output power of the photovoltaic system can be considered as a offsetting factor against the battery discharge power in the equivalent range estimation. After installing the photovoltaic system, the equivalent average discharge power that the battery side needs to bear is: ; Substituting the data yields .

[0072] Therefore, the equivalent driving range after adding photovoltaic skin is: ; Substituting the data, we get:

[0073] ; Therefore, the improvement in range after adding lightweight, flexible gallium arsenide photovoltaic skin is as follows:

[0074] ; Under the aforementioned reference conditions, the drone's flight time without a photovoltaic system is approximately 55 minutes. With the addition of a lightweight, flexible gallium arsenide photovoltaic skin, the equivalent flight time is approximately 59.8 minutes, representing an increase of about 4.8 minutes. This result demonstrates that the benefits of photovoltaic power generation can outweigh the flight time losses caused by the increased weight, added drag, and power consumption of the photovoltaic system.

[0075] However, in multi-rotor UAVs, the actual effective output of photovoltaic skin does not solely depend on material efficiency and coverage area. Because rotor blades periodically block the photovoltaic sub-regions located on the fuselage or arm surfaces, the output current or power of these sub-regions experiences periodic drops synchronized with rotor speed and blade pass-through frequency. If traditional MPPT or window-averaged power evaluation methods are still used, the short-term power drop caused by rotor shearing may be mistaken for a change in the maximum power point or long-term inefficiency of the sub-region. This can lead to incorrect MPPT updates, operating point oscillations, or incorrect bypassing of the photovoltaic sub-region, preventing the theoretical power generation benefits of the aforementioned photovoltaic skin from being fully converted into actual range gains.

[0076] Therefore, this invention does not merely add photovoltaic materials to the drone, but rather, based on the confirmation that lightweight photovoltaic skin has positive endurance benefits, further addresses the unique rotor periodic light cutting problem of multi-rotor drones. It utilizes rotor speed, blade passage frequency, or rotor phase information to identify shading time slots and generate shading time slot masks; it then removes or downweights shading photovoltaic sub-region samples, reconstructs the true power of the photovoltaic sub-region using unshading reliable photovoltaic sub-region samples, and uses the reconstructed power for MPPT updates and photovoltaic sub-region access status determination. Figure 6 As shown, the above control method can reduce MPPT erroneous updates and sub-region erroneous bypasses caused by rotor light cutting, making the power generation revenue of lightweight photovoltaic skin more stable and effectively participate in the drone's in-flight endurance.

[0077] To verify the effectiveness of this method, a simulation model of rotor shearing in the photovoltaic sub-region of a multi-rotor UAV was further established. In the simulation, the photovoltaic sub-region was affected by periodic rotor shading, with the shading duty cycle set to 15% and the shading depth set to 55%. Rotor shearing caused periodic drops in the output power of the photovoltaic sub-region.

[0078] In MPPT window contamination simulation, the traditional method uses the average power of the window including occluded samples for MPPT updates. Simulation results show that the proposed method reduces low-frequency output power fluctuations by approximately 7.14% and operating voltage fluctuations by approximately 20.52%, indicating that the proposed method can reduce the impact of rotor cut-off samples on MPPT update stability.

[0079] In the simulation of erroneous bypass in photovoltaic sub-regions, traditional methods use the average power of the window containing shading samples to evaluate the state of the photovoltaic sub-region. Because the rotor periodic light cutting causes the average power of the window to remain below the bypass threshold for an extended period, traditional methods misclassify this sub-region as inefficient and bypass it, leading to a decrease in the total output power of the photovoltaic array. This method removes shading samples and reconstructs the true power of the sub-region using non-shading samples, thus maintaining the sub-region's accessibility.

[0080] Simulation results show that under this condition, the erroneous bypass time ratio of the traditional method is 100%, while that of the present method is 0%. The average total output power of the array is 259.91W under the traditional method, while that of the present method is 338.97W, representing an increase of approximately 30.42% in average total output power under this erroneous bypass condition.

[0081] The simulation results above demonstrate that the rotor shear ripple is not merely noise; it can serve as a basis for determining occlusion time slots. By eliminating occlusion samples and reconstructing the true power of non-occlusion systems, erroneous MPPT updates and sub-region bypass errors can be reduced, thereby improving the stability and effective output of the multi-rotor UAV photovoltaic power replenishment system.

[0082] In summary, this invention addresses the sampling distortion problem caused by the periodic shading of photovoltaic sub-regions by rotor blades in multi-rotor UAVs. It utilizes existing rotor state information of the UAV, such as blade passage frequency or rotor phase information, to identify the shearing ripple synchronized with the rotor from the output current or power of the photovoltaic sub-region. Next, a shading time slot mask is generated; relevant samples are removed and downweighted within the shading time slots, and the true power of the photovoltaic sub-region is reconstructed within the non-shading time slots. This reconstructed power is then used for MPPT updates and photovoltaic sub-region access status determination. Consequently, when performing solar power supplementation, multi-rotor UAVs no longer simply execute a "the higher the photovoltaic power, the better" control strategy. Instead, they establish a net benefit evaluation of candidate attitudes, reducing the interference of rotor shearing on the photovoltaic power supplementation system, minimizing erroneous updates, bypassing, and ineffective solar power supplementation attitude adjustments, and improving the actual benefits of new energy-assisted endurance for multi-rotor UAVs.

Claims

1. A method for controlling the in-flight endurance of a multi-rotor unmanned aerial vehicle (UAV) powered by a new energy source, characterized in that, The light-receiving surface of a multi-rotor drone is provided with a conformal photovoltaic skin composed of several photovoltaic sub-regions. The method includes the following steps: Collect temperature and current signals from each photovoltaic sub-region; obtain the phase and blade passing frequency of each rotor; The current signal of the photovoltaic sub-region is filtered and synchronized to extract the periodic shear ripple corresponding to the blade passing frequency of the rotor, and the characteristics of the periodic shear ripple are obtained. Based on the characteristics of rotor phase, blade passing frequency, periodic shear ripple, shading duty cycle, and geometric relationship between rotor and photovoltaic sub-region, a shading time slot mask for each photovoltaic sub-region is generated. The credibility of photovoltaic sub-region samples at each sampling time is determined based on the occlusion time slot mask, and the photovoltaic sub-region samples are weighted based on the credibility of the photovoltaic sub-region samples. Based on the shading time slot mask and the sample weight of the photovoltaic sub-region, the true power of the photovoltaic sub-region in the unshading state is reconstructed. Based on the actual power, shading duty cycle, temperature, impact on the bus, and MPPT reliability obtained from the reconstruction of each photovoltaic sub-region under the unshaded state, the corresponding photovoltaic sub-region is controlled to remain connected or bypassed in order to reduce false bypass caused by rotor periodic light cutting. MPPT updates are controlled based on the occlusion time slot mask and the actual power in the unoccluded state obtained from reconstruction. Based on the actual power in the unshaded state obtained from the reconstruction of each photovoltaic sub-region, the candidate attitudes of the multi-rotor UAV are evaluated, and the candidate attitude that satisfies the flight safety constraints and has the largest net benefit is selected as the execution attitude.

2. The in-flight endurance control method for a multi-rotor UAV with new energy-assisted drive according to claim 1, characterized in that, Each photovoltaic sub-area is equipped with an independent voltage and current sampling channel or a partial shared channel with other photovoltaic sub-areas, and the photovoltaic power supply bus is managed through a DC / DC converter, MPPT controller, and access switch.

3. The in-flight endurance control method for a multi-rotor UAV with new energy-assisted drive according to claim 1, characterized in that, The expression for calculating the blade passage frequency is: in Indicates the first j The frequency of each rotor blade; For the first j The number of blades in each rotor; For the first j The angular velocity of each rotor; It is 180°; Indicates time.

4. The in-flight endurance control method for a multi-rotor UAV with new energy-assisted drive according to claim 1, characterized in that, The specific methods for filtering and synchronizing the current signal of the photovoltaic sub-region to extract the periodic shear ripple corresponding to the blade passage frequency of the rotor include: Perform filtering: For the first Each photovoltaic sub-region is in Output current signal at time Filtering is performed to estimate the low-frequency or quasi-steady-state components that are related to the actual irradiance and operating point variations. The estimated low-frequency or quasi-steady-state component is subtracted from the output current signal to obtain the residual signal. : ； From residual signal Extracting from the first j The signal corresponding to each spin is used as a candidate synchronous ripple component. ; Perform synchronous analysis: construct and Time of the first j Each rotor blade is synchronized by a frequency-corresponding reference signal. And calculate the correlation between the output current of the photovoltaic sub-region and the synchronization reference signal: in, express Time of the first i The output current of the photovoltaic sub-region is related to the first j The correlation of the rotor blades with the corresponding frequency of the synchronization reference signal; For the corr function; when Greater than the preset correlation threshold and candidate synchronization ripple components When the amplitude is greater than the preset amplitude threshold, the first... i The photovoltaic sub-region exists due to the first j Periodic shearing caused by individual rotors; candidate synchronous ripple components It is identified as a periodic shear wave; the characteristics of the periodic shear wave include amplitude, frequency and phase.

5. The in-flight endurance control method for a multi-rotor UAV with new energy-assisted drive according to claim 1, characterized in that, The expression for generating the shading time slot mask for the photovoltaic sub-region is: in, express Time of the first Shading time slot mask for each photovoltaic sub-region; Indicates the first Each photovoltaic sub-region forms a periodically shaded set of associated rotors; Indicates the first One rotor The blade at time traverses the phase and satisfies ; Indicates the first The number of blades in each rotor; Indicates the first One rotor Mechanical phase at any given moment; Indicates the first The rotor pairs with the first The shading center phase when a photovoltaic sub-region forms a shading; express The rotor pairs with the first Shading duty cycle of each photovoltaic sub-region; In radians; This indicates that the phase difference is constrained to a preset principal value range; The indicator function takes the value 1 when the condition in parentheses is true, and 0 otherwise. The shading time slot mask is represented in binary form. When the shading time slot mask is 1, it means that the sampling point of the corresponding photovoltaic sub-region is in the rotor shading time slot. When the shading time slot mask is 0, it means that the sampling point of the corresponding photovoltaic sub-region is in the non-shading time slot.

6. The in-flight endurance control method for a multi-rotor UAV with new energy-assisted drive according to claim 1, characterized in that, The reliability of photovoltaic sub-region samples at each sampling time is determined based on the occlusion time slot mask. The specific method for assigning weights to photovoltaic sub-regions based on the reliability of photovoltaic sub-region samples includes: sampling time Distance between the nearest occlusion slot boundary As Time of the first i Confidence of individual photovoltaic sub-region samples; based on Transition width with preset occlusion boundary The size relationship is used to assign weights to the photovoltaic sub-regions, and the expression is as follows: in for Time of the first i Each photovoltaic sub-region sample weight.

7. The in-flight endurance control method for a multi-rotor UAV with new energy-assisted drive according to claim 6, characterized in that, The reconstructed expression for the actual power of the photovoltaic sub-region under unshaded conditions is: in, Indicates the first i The actual power of a photovoltaic sub-region within a sampling window W under unshaded conditions; Indicates the first i Each photovoltaic sub-region is in The occlusion time slot mask at any given time; the occlusion time slot mask is represented in binary form. When the occlusion time slot mask is 0, it means that the sampling point of the corresponding photovoltaic sub-region is in an unoccluded time slot. for Time of the first i Each photovoltaic sub-region sample weight; For the first i Each photovoltaic sub-region is in Output power at any given moment.

8. The in-flight endurance control method for a multi-rotor UAV with new energy-assisted drive according to claim 1, characterized in that, When controlling MPPT updates, if an occlusion slot is detected in the current sampling window, perform at least one of the following operations: (1) Suspend the use of voltage, current or power samples to update MPPT during the blocking time slot; (2) Keep the current MPPT duty cycle or operating voltage unchanged; (3) Reduce the weight of shaded photovoltaic sub-region samples in MPPT updates; The photovoltaic sub-region sample refers to the output data of the photovoltaic sub-region; (4) Delay MPPT updates to non-occlusion time slots; (5) Use the real power obtained from the reconstruction under the unshaded state to update the MPPT.

9. The in-flight endurance control method for a multi-rotor UAV with new energy-assisted drive according to claim 1, characterized in that, Specific methods for controlling the corresponding photovoltaic sub-region to maintain access or bypass include: When the average power of a photovoltaic sub-region calculated based on all power samples within the evaluation window is lower than the first average power threshold, the proportion of low-power samples within the rotor shading time slot to the total number of low-power samples within the evaluation window exceeds a preset proportion threshold, and the power estimate of the unshaded state of the photovoltaic sub-region is within a preset normal power range, it is determined that the power decrease of the photovoltaic sub-region is caused by rotor periodic light cutting, and the photovoltaic sub-region is kept connected; where low-power samples are power samples within the evaluation window that are lower than the first power sample threshold; When the actual power of a photovoltaic sub-region under unshaded conditions obtained by reconstruction is lower than the second preset power threshold within a set time period, and is accompanied by abnormal temperature rise, open circuit / short circuit characteristics, or the impact on the bus reaches the preset impact threshold, the photovoltaic sub-region is controlled to bypass, or the weight of the photovoltaic sub-region in the power evaluation is reduced.

10. The in-flight endurance control method for a multi-rotor UAV with new energy-assisted drive according to claim 1, characterized in that, The specific method for evaluating the candidate attitudes of a multi-rotor UAV based on the actual power under unshaded conditions obtained from the reconstruction of each photovoltaic sub-region includes the following steps: In response to the photovoltaic energy replenishment attitude adjustment command, the photovoltaic power increment brought about by the candidate attitude is calculated based on the real power in the unshaded state obtained by reconstructing each photovoltaic sub-region; at the same time, the power consumption increment of the propulsion system and the degree of imbalance of the power change of each rotor caused by the attitude adjustment are calculated. The net benefit evaluation value of the corresponding candidate attitude is calculated based on the unevenness of the photovoltaic power increment, the propulsion system power consumption increment, and the power change of each rotor. Determine if there is a candidate attitude with a positive net benefit evaluation value. If so, select the candidate attitude that satisfies flight safety constraints and has the largest net benefit as the execution attitude; otherwise, maintain the current attitude. The formula for calculating the net income evaluation value is as follows: in, Indicate candidate pose Net income valuation; The photovoltaic power increment represents the actual power obtained under the unshaded state after reconstruction; This represents the increase in propulsion system power consumption caused by attitude adjustment; Indicates the first j The power change of each rotor relative to the current attitude under the candidate attitude; It indicates the degree of imbalance in the power variation of each rotor, and is used to reflect the load imbalance caused by crosswinds, gusts, or relative airflow. These are the weighting coefficients.

Images