Unmanned aerial vehicle cleaning task execution method, device and equipment for photovoltaic power station and medium

Abstract

The invention discloses an unmanned aerial vehicle cleaning task execution method and device for a photovoltaic power station, equipment and a medium, and relates to the technical field of photovoltaics, and the method comprises the steps: dividing a total cleaning route of the photovoltaic power station into cleaning sub-routes according to the unit operation length of an unmanned aerial vehicle; determining a fixed or non-fixed target position optimization mode of a vehicle-mounted nest parking point; establishing a photovoltaic power station cleaning task according to the target position optimization mode, and constructing a target time function corresponding to the photovoltaic power station cleaning task by using the starting point and the ending point of the cleaning sub-route; the dependent variable of the target time function is the time for completing the cleaning task of the photovoltaic power station; constructing a position optimization model taking the minimization of the target time function as a target, and iteratively searching a target optimal parking point position in the allowed parking area of the vehicle-mounted machine nest by using the position optimization model; and controlling the unmanned aerial vehicle to take water from the vehicle-mounted machine nest at the target optimal parking point position so as to execute a photovoltaic power station cleaning task. And the unmanned aerial vehicle cleaning task execution efficiency of the photovoltaic power station is improved.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic technology, and in particular to a method, apparatus, equipment and medium for performing drone cleaning tasks in photovoltaic power plants. Background Technology

[0002] The accumulation of contaminants on the surface of photovoltaic (PV) modules reduces their light transmittance and triggers hot spot effects, significantly impacting the power generation performance of PV power plants. Therefore, cleaning is necessary to maintain the power generation efficiency of PV power plants. In existing technologies, manual cleaning of modules is difficult to guarantee in scenarios with high module supports and undulating terrain, such as mountainous PV, floating power plants, and carport PV systems, due to challenges in cleaning efficiency, effectiveness, and personnel safety. PV cleaning robots consume a lot of electricity, have high initial investment and operating costs, are difficult to move on uneven PV arrays, hindering widespread adoption, and their own weight and vibrations during operation may damage the module surface and shorten the lifespan of the PV modules.

[0003] Cleaning drones, as an emerging technology, can replace high-risk manual operations and have high cleaning efficiency. However, at present, drone cleaning tasks cannot be coordinated with the location of the vehicle-mounted drone nest. Using only the photovoltaic power station water replenishment and charging point as a single take-off and landing point will result in excessive non-operation distance, which will seriously affect the efficiency of cleaning operations.

[0004] In summary, improving the efficiency of drone cleaning tasks in photovoltaic power plants is a problem that needs to be solved in this field. Summary of the Invention

[0005] In view of this, the purpose of this invention is to provide a method, apparatus, equipment, and medium for performing drone cleaning tasks in photovoltaic power plants, thereby improving the efficiency of drone cleaning tasks in photovoltaic power plants. The specific solution is as follows: Firstly, this application discloses a method for performing drone cleaning tasks in photovoltaic power plants, including: The total cleaning route of the photovoltaic power station is divided into various cleaning sub-routes based on the unit operating length of the drone. Determine the target location optimization mode for the vehicle-mounted unit's parking point; wherein, the target location optimization mode is either a first optimization mode where the parking point location is not fixed or a second optimization mode where the parking point location is fixed; A photovoltaic power station cleaning task is established based on the target location optimization mode, and a target time function corresponding to the photovoltaic power station cleaning task is constructed using the start and end points of the cleaning sub-route; wherein, the dependent variable of the target time function is the time used to complete the photovoltaic power station cleaning task. A location optimization model is constructed with the objective of minimizing the target time function, and the location optimization model is used to iteratively search for the target optimal parking point location in the allowed parking area of ​​the vehicle-mounted vehicle nest; The drone is controlled to collect water from the vehicle-mounted air pod at the optimal parking point of the target in order to perform the cleaning task of the photovoltaic power station.

[0006] Optionally, the step of dividing the total cleaning route of the photovoltaic power station into various cleaning sub-routes based on the unit operating length of the UAV includes: Plan the overall cleaning route for the photovoltaic power station; The ratio of the drone's onboard water tank volume to the unit cleaning water volume is determined as the drone's unit operating time, and the product of the drone's unit operating time and the drone's unit operating speed is determined as the drone's unit operating length. The total cleaning route is divided into sub-cleaning routes based on the unit operating length of the UAV.

[0007] Optionally, establishing the photovoltaic power station cleaning task based on the target location optimization mode includes: The cleaning sub-routes are divided into cleaning route groups based on the water tank capacity of the vehicle-mounted air pod and the airborne water tank capacity of the UAV. Construct a single cleaning task corresponding to each of the aforementioned cleaning route groups; If the target location optimization mode is the first optimization mode, then each of the individual cleaning tasks will be sequentially determined as a photovoltaic power station cleaning task. If the target location optimization mode is the second optimization mode, then the sum of the tasks of each individual cleaning task is determined as the photovoltaic power station cleaning task.

[0008] Optionally, the step of constructing a target time function corresponding to the photovoltaic power station cleaning task using the start and end points of the cleaning sub-path includes: The drone operation time and drone water refill time corresponding to the photovoltaic power station cleaning task are determined, and the round-trip time function and drone battery swapping delay time function corresponding to the photovoltaic power station cleaning task are constructed using the start and end points of the cleaning sub-flight route; wherein, the drone operation time is the sum of the unit operation time of the drones corresponding to the photovoltaic power station cleaning task; A first time function corresponding to the photovoltaic power station cleaning task is constructed based on the drone operation time, the drone water refill time, the drone round-trip time function, and the drone battery swapping delay time function. A second time function is constructed by using the time function of the vehicle-mounted air cell's round trip to the water replenishment point and the water replenishment time of the vehicle-mounted air cell at the water replenishment point; Construct a target time function that includes the first time function and the second time function.

[0009] Optionally, the drone battery swapping delay time function corresponding to the photovoltaic power station cleaning task is constructed using the start and end points of the cleaning sub-flight route, including: A function for the number of times the drone can be swapped is constructed using the start and end points of the cleaning sub-route, the drone's empty flight range, the drone's full-load flight range, and the drone's flight power consumption. The product of the number of times the drone battery swaps and the battery swap delay time is determined as the drone battery swap delay time function corresponding to the photovoltaic power station cleaning task; wherein, the battery swap delay time is the delay time caused by the drone's onboard water tank swapping batteries while adding water.

[0010] Optionally, the step of iteratively searching for the optimal target parking point location within the permitted parking area of ​​the vehicle-mounted vehicle nest using the location optimization model includes: The current parking point location is determined from the permitted parking area of ​​the vehicle-mounted unit nest, and the current optimal parking point location is substituted into the target time function to obtain the first time corresponding to the current optimal parking point location; The current parking point location is generated using a global search mechanism or a local generation mechanism to determine the parking point location to be compared. If the location of the parking point to be compared is within the permitted parking area of ​​the vehicle-mounted unit nest, then the location of the parking point to be compared is substituted into the target time function to obtain the second time corresponding to the location of the parking point to be compared; If the first time is not less than the second time and the preset stopping condition is met, then the parking point location to be compared is determined as the target optimal parking point location; If the first time is not less than the second time and the preset stopping condition is not met, a new current parking point location is determined from the allowed parking area of ​​the vehicle-mounted unit nest, and the parking point location to be compared is determined as the new current optimal parking point location. Then, the process jumps back to the step of generating the parking point location to be compared of the current parking point location through a global search mechanism or a local generation mechanism.

[0011] Optionally, generating the parking point location to be compared from the current parking point location through a global search mechanism or a local generation mechanism includes: The target exploration probability corresponding to the current iteration number is determined by using the current iteration number, the preset initial exploration probability, the preset final exploration probability, and the preset maximum iteration number. Based on the target exploration probability, the target mechanism corresponding to the current iteration number is determined from the global search mechanism and the local generation mechanism; If the target mechanism is the global search mechanism, then a reference parking point location is randomly selected from the allowed parking area of ​​the vehicle-mounted unit nest, and the reference parking point location and the current parking point location are used to generate a parking point location to be compared with the current parking point location; If the target mechanism is the local generation mechanism, then the current optimal parking point position, the preset local search radius, and the normalized random number are used to generate the parking point position to be compared with the current parking point position.

[0012] Secondly, this application discloses a drone cleaning task execution device for a photovoltaic power station, comprising: The route division module is used to divide the total cleaning route of the photovoltaic power station into various cleaning sub-routes based on the unit operation length of the drone. The mode determination module is used to determine the target location optimization mode of the vehicle-mounted unit's parking point; wherein, the target location optimization mode is either a first optimization mode where the parking point location is not fixed or a second optimization mode where the parking point location is fixed. The function construction module is used to establish a photovoltaic power station cleaning task based on the target location optimization mode, and to construct a target time function corresponding to the photovoltaic power station cleaning task using the start and end points of the cleaning sub-path; wherein, the dependent variable of the target time function is the time used to complete the photovoltaic power station cleaning task; The location optimization module is used to construct a location optimization model with the goal of minimizing the target time function, and to use the location optimization model to iteratively search for the target optimal parking point location in the allowed parking area of ​​the vehicle-mounted vehicle nest; The task execution module is used to control the UAV to collect water from the vehicle-mounted air pod at the optimal parking point of the target in order to perform the photovoltaic power station cleaning task.

[0013] Thirdly, this application discloses an electronic device, including: Memory, used to store computer programs; A processor is used to execute the computer program to implement the steps of the aforementioned disclosed method for performing drone cleaning tasks in photovoltaic power plants.

[0014] Fourthly, this application discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, it implements the steps of the aforementioned disclosed method for executing drone cleaning tasks in a photovoltaic power station.

[0015] The beneficial effects of this application are as follows: This application divides the total cleaning route of the photovoltaic power station into various cleaning sub-routes based on the unit operation length of the UAV; determines the target location optimization mode for the vehicle-mounted UAV nest parking point; wherein, the target location optimization mode is a first optimization mode where the parking point location is not fixed or a second optimization mode where the parking point location is fixed; establishes a photovoltaic power station cleaning task based on the target location optimization mode, and constructs a target time function corresponding to the photovoltaic power station cleaning task using the start and end points of the cleaning sub-routes; wherein, the dependent variable of the target time function is the time used to complete the photovoltaic power station cleaning task; constructs a location optimization model with the goal of minimizing the target time function, and uses the location optimization model to iteratively search for the target optimal parking point location in the allowed parking area of ​​the vehicle-mounted UAV nest; controls the UAV to take water from the vehicle-mounted UAV nest at the target optimal parking point location to perform the photovoltaic power station cleaning task. Therefore, this application, by dividing the total cleaning route of the photovoltaic power station into cleaning sub-routes according to the unit operation length of the UAV, enables each sub-route to be adapted to the cleaning capacity of the UAV with a full water tank in a single operation, avoiding frequent ineffective round trips due to insufficient water and ensuring the continuity of the operation; it determines two target location optimization modes, fixed or non-fixed parking point locations, which can adapt to different photovoltaic power station environments, improving flexibility and applicability; it establishes cleaning tasks based on the target location optimization mode and constructs a target time function using the start and end points of the sub-routes, constructs a location optimization model that minimizes the target time function, and iteratively searches for the optimal parking point within the allowed parking area, ensuring that the parking point location is legal, i.e., not located in the module installation area, and reduces the total time consumption through iterative optimization, controlling the UAV to collect water from the target optimal parking point to execute the task, which can reduce ineffective movement based on the optimal parking point location, improve operation efficiency, and ensure that the operation process does not interfere with the photovoltaic modules, ensuring safety. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0017] Figure 1 This application discloses a flowchart of a method for performing drone cleaning tasks in a photovoltaic power station. Figure 2 This application discloses a specific flowchart of a drone cleaning management system; Figure 3 This is a schematic diagram illustrating the location optimization of a specific vehicle-mounted aircraft nest parking point disclosed in this application; Figure 4This is a schematic diagram of the structure of a drone cleaning task execution device for a photovoltaic power station disclosed in this application; Figure 5 This is a structural diagram of an electronic device disclosed in this application. Detailed Implementation

[0018] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0019] The accumulation of contaminants on the surface of photovoltaic (PV) modules reduces their light transmittance and triggers hot spot effects, significantly impacting the power generation performance of PV power plants. Therefore, cleaning is necessary to maintain the power generation efficiency of PV power plants. In existing technologies, manual cleaning of modules is difficult to guarantee in scenarios with high module supports and undulating terrain, such as mountainous PV, floating power plants, and carport PV systems, due to challenges in cleaning efficiency, effectiveness, and personnel safety. PV cleaning robots consume a lot of electricity, have high initial investment and operating costs, are difficult to move on uneven PV arrays, hindering widespread adoption, and their own weight and vibrations during operation may damage the module surface and shorten the lifespan of the PV modules.

[0020] Cleaning drones, as an emerging technology, can replace high-risk manual operations and have high cleaning efficiency. However, at present, drone cleaning tasks cannot be coordinated with the location of the vehicle-mounted drone nest. Using only the photovoltaic power station water replenishment and charging point as a single take-off and landing point will result in excessive non-operation distance, which will seriously affect the efficiency of cleaning operations.

[0021] Therefore, this application provides a drone cleaning task execution scheme for photovoltaic power plants, which improves the efficiency of drone cleaning tasks in photovoltaic power plants.

[0022] See Figure 1 As shown in the figure, this application discloses a method for performing drone cleaning tasks in a photovoltaic power station, including: Step S11: Divide the total cleaning route of the photovoltaic power station into various cleaning sub-routes according to the unit operation length of the drone.

[0023] In the process of using drones to clean photovoltaic power plants, preparation work for the cleaning task also needs to be completed. This preparation work is done by different modules, among which the equipment management module is responsible for the following tasks: 1) Main equipment for drone cleaning: including cleaning drones, vehicle-mounted drone nests and photovoltaic power station auxiliary equipment: (1) Cleaning drones include drones, handheld mobile devices, airborne cleaning spray booms and water pumps, drone airborne water tanks, and cameras. (2) Vehicle-mounted drone nests include vehicles, drone nests, vehicle water tanks, drone backup batteries, and drone battery vehicle-mounted supercharging stations. The vehicles used for vehicle-mounted drone nests must have high-precision navigation capabilities, such as being equipped with a vehicle navigation system or a vehicle-mounted RTK (Real-time kinematic) receiver. (3) Photovoltaic power station auxiliary equipment includes component dust detectors, water meters, and electricity meters. The drones, vehicle navigation systems, or vehicle-mounted RTK receivers used all use a unified coordinate system setting such as WGS84 or CGCS2000, and all enable RTK real-time differential mode based on the photovoltaic power station RTK reference station or network RTK.

[0024] 2) Equipment management functions: (1) Basic equipment information: equipment number, maintenance status, management personnel. (2) Equipment usage log: current power station in use, planned usage time, and user personnel. (3) Equipment call application: equipment call application and approval, transportation and receipt status.

[0025] The real-time acquisition module obtains various types of data in real time through monitoring equipment, including image data, measurement data, location information data, equipment status information data, and data accessed from the photovoltaic operation and maintenance platform. The specific tasks handled by the real-time acquisition module are as follows: 1) Image Data: Photos and videos taken by the cleaning drone's onboard camera, and real-time footage from the monitoring camera in the vehicle-mounted drone housing. 2) Measurement Data: Readings from added sensors or measuring devices, including water levels in the drone's onboard water tank, the vehicle's water tank, total water consumption monitored by the water meter, and the pollution ratio and dust thickness monitored in real-time by the component dust detector. 3) Location Information Data: Acquired via IoT protocols and cloud interfaces, providing real-time location data for the cleaning drone's trajectory, latitude and longitude, flight altitude, flight speed, and vehicle-mounted drone housing. 4) Equipment Status Data: Equipment status information acquired via IoT protocols and cloud interfaces, such as drone battery level, drone flight time, onboard cleaning boom retraction / extension status, onboard cleaning pump start / stop, vehicle-mounted drone housing switch status, feedback that the drone in the housing has landed and returned to center, backup battery status for drones charging in the vehicle-mounted drone housing, vehicle overcharge status, and equipment alarm information. 5) Photovoltaic Operation and Maintenance Platform Access Data: This includes power plant operation data monitored by the photovoltaic operation and maintenance platform, such as daily power generation efficiency, hourly current, daily actual power generation, power loss due to faults / power curtailment / external factors (if any), and daily instantaneous irradiance data. This data is used for cleaning decisions and evaluation of cleaning effectiveness. After acquiring the raw information collected in real time, abnormal and duplicate data points are removed, and missing values ​​are supplemented using interpolation and mean imputation methods. The units and transmission time formats of the data are standardized, and the data is output in a structured format and stored in the database. Simultaneously, an index is created to optimize query performance.

[0026] The cleaning decision module is responsible for the following tasks: 1) Annual cleaning plan: If the photovoltaic power station has a fixed annual cleaning plan, the cleaning time will be determined according to the plan, and the power station will be switched to the cleaning-ready state periodically. 2) Measurement data decision: Based on the pollution ratio and dust thickness measured in real time by the component dust detector obtained by the real-time acquisition module, if the pollutant accumulation limit for cleaning is reached, the power station will be switched to the cleaning-ready state. 3) Monitoring data decision: If the photovoltaic operation and maintenance platform monitors a power generation efficiency reduction of more than 10%-15% or a photovoltaic power station current reduction of more than 8%, and after ruling out other equipment failures and weather factors, it is necessary to perform cleaning, then the power station needs to be switched to the cleaning-ready state.

[0027] The power station management module is responsible for the following tasks: 1) Power station list: Displaying a list of power stations to be cleaned, and allowing modification and deletion of existing power station information. 2) Power station addition: Creating a power station by filling in basic information such as the name and capacity of the power station to be cleaned, importing local weather data, and uploading aerial base map files. The power station base map is usually obtained by using image data captured by camera equipment carried by inspection drones during power station inspections, which is then encrypted using aerial triangulation and used to construct dense point clouds. The file formats include: oblique photogrammetry OSGB (Open Scene Gragh Binary) files, vector data GEOJSON files, Digital Surface Model (DSM) TIF (Tagged Image File Format) files, terrain data TERRAIN files, and point cloud data LAS (Lidar Aerial Survey) files.

[0028] The cleaning task module is responsible for the following tasks: 1) Selecting a cleaning power station: Select a cleaning power station from the power station list and enter the current power station's base map. 2) Preparing cleaning equipment: Import equipment that has been requested and accepted from the equipment management module, such as cleaning drones, handheld mobile devices, cleaning mounts (water pumps, spray booms), vehicle-mounted air pods, cameras, component dust detectors, water meters, and electricity meters. Install water meters and electricity meters at the water and power supply ends of the cleaning task at the power station's fixed replenishment points. Verify equipment components and serial numbers, assemble the equipment, and check that all equipment functions correctly to confirm that the cleaning equipment preparation is complete.

[0029] Furthermore, the cleaning task module is also responsible for task route planning. During the planning process, it is necessary to determine the optimal parking point of the vehicle-mounted drone nest, thereby completing the planning of the complete flight route of the drone in performing the cleaning task, and then executing the cleaning task. In the process of determining the optimal parking point of the vehicle-mounted drone nest, it is necessary to determine each cleaning sub-route, the optimization mode, and also to construct a time function, thereby completing the optimization of the parking point location.

[0030] In this embodiment, dividing the total cleaning route of the photovoltaic power station into various cleaning sub-routes based on the unit operating length of the UAV includes: planning the total cleaning route of the photovoltaic power station; determining the ratio of the UAV's onboard water tank volume to the unit cleaning water volume as the unit operating time of the UAV, and determining the product of the UAV's unit operating time and the UAV's unit operating speed as the unit operating length of the UAV; and dividing the total cleaning route into various cleaning sub-routes based on the unit operating length of the UAV.

[0031] First, plan the overall cleaning flight path for the photovoltaic power station. It's important to note that this overall cleaning flight path refers to the total flight path of the drone during the photovoltaic power station cleaning process, excluding the drone's flight path when not cleaning. In other words, the overall cleaning flight path is the total path taken by the drone during operation, excluding its path when not in operation. There are three methods for planning the overall cleaning flight path: selecting existing paths used in previous cleaning tasks, manually creating a new path, or automatically creating a new path. The manual method requires obtaining the path drawn by the worker using a mouse to select points and connect them sequentially; this is the overall cleaning flight path for the photovoltaic power station, and calculating the initial total length D of the overall cleaning flight path. The automatic method involves selecting the polygon vertices of each photovoltaic module installation area in a clockwise or counterclockwise order. The latitude and longitude of the boundary points of the photovoltaic module installation areas are denoted as... The toolbox of MATLAB R2021a is called. Using the uavCoveragePlanner, the angle between the path direction and the horizontal axis (0° for the horizontal direction and 90° for the vertical direction) and the width of the flight path, i.e. the cleaning width W of the cleaning drone, are input. The toolbox automatically generates the waypoint coordinates of the initial full-coverage bow-shaped flight path and automatically calculates the initial total length D of the cleaning flight path.

[0032] It is important to note that whether creating a flight path manually or automatically, you will need to enter the flight path name, fixed flight altitude or terrain-following flight altitude, and the drone's onboard water tank volume. Vehicle water tank capacity Cleaning drone unit operation speed Cleaning drone takeoff / return speed Unit cleaning water volume (L / min), estimated vehicle speed Water filling time for drone onboard water tank Vehicle water tank filling time The extended time caused by simultaneously refilling the drone's onboard water tank and swapping the battery during a single cleaning operation. Personnel's working time per session The latitude and longitude coordinates of the fixed replenishment point of the power station (used for adding water to the vehicle's water tank) are set as P. In this way, the route division and the location optimization of the vehicle-mounted aircraft nest parking point can be carried out based on these parameters.

[0033] Next, flight paths will be defined. The ratio of the drone's onboard water tank volume to the unit cleaning water volume will be determined as the drone's unit operating time. ,Right now The unit operating length of the drone is determined by multiplying the unit operating time of the drone by the unit operating speed of the drone. ,Right now The total cleaning route is divided into sub-routes based on the unit operating length of the drone. Since the drone's onboard water tank needs to return to the vehicle's water tank to refill after its water supply is depleted, the unit operating length... This represents the unit operating length supported by the drone after its onboard water tank is filled with water. (Based on the unit operating length) Divide the initial total length D of the route into n segments of length D. Sub-routes, namely Extract the start and end points of each sub-route. The latitude and longitude coordinates of the start point of each sub-route are marked as follows: The latitude and longitude coordinates of the endpoint of each sub-route are as follows: .

[0034] Understandably, if the location S of the vehicle-mounted aircraft carrier parking point is determined, then the non-operational distance, including the distance from location S to the starting point of each sub-route, is denoted as... And the distance from the end of each sub-route to the parking point S of the vehicle-mounted aircraft nest is denoted as Each operational route The expected flight lengths are respectively .

[0035] Step S12: Determine the target location optimization mode for the vehicle-mounted unit's parking point; wherein, the target location optimization mode is either a first optimization mode where the parking point location is not fixed or a second optimization mode where the parking point location is fixed.

[0036] In this embodiment, the target location optimization mode for the vehicle-mounted cleaning unit's docking station includes two modes: a first optimization mode where the docking station location is not fixed, and a second optimization mode where the docking station location is fixed. In the first optimization mode where the docking station location is not fixed, the docking station location is fixed throughout the entire photovoltaic power station cleaning task. That is, after each water filling, the vehicle-mounted cleaning unit returns to its original docking station location. Since network RTK is highly dependent on 4G signal quality, if the photovoltaic power station being cleaned is located in a remote mountainous area with poor 4G signal quality and lacks an RTK reference station, therefore... This mode is suitable for scenarios with poor information transmission signal quality. Conversely, in the second optimization mode with a fixed parking point, the parking point of the vehicle-mounted drone nest is not fixed during the entire photovoltaic power station cleaning task. For example, when the drone is cleaning power stations 1, 2, and 3, the location of the vehicle-mounted drone nest is S1, while when the drone is cleaning power stations 4, 5, and 6, the location of the vehicle-mounted drone nest is S2. That is, the parking point of the vehicle-mounted drone nest may change depending on the location of the power station being cleaned. When the signal transmission quality is good or the cleaning time is short, the second optimization mode can be selected.

[0037] Step S13: Establish a photovoltaic power station cleaning task according to the target location optimization mode, and construct a target time function corresponding to the photovoltaic power station cleaning task using the start and end points of the cleaning sub-route; wherein, the dependent variable of the target time function is the time used to complete the photovoltaic power station cleaning task.

[0038] In this embodiment, establishing a photovoltaic power station cleaning task based on the target location optimization mode includes: dividing each cleaning sub-route into cleaning route groups based on the water tank volume of the vehicle-mounted drone and the airborne water tank volume of the UAV; constructing a single cleaning task corresponding to each cleaning route group; if the target location optimization mode is the first optimization mode, then each single cleaning task is sequentially determined as a photovoltaic power station cleaning task; if the target location optimization mode is the second optimization mode, then the sum of the tasks of each single cleaning task is determined as a photovoltaic power station cleaning task.

[0039] First, based on the water tank capacity of the vehicle-mounted water tank and the airborne water tank capacity of the drone, each cleaning sub-route is divided into cleaning route groups. In other words, the number of times the drone can collect water (F) after the vehicle's water tank is fully filled is determined. Q1 is the volume of the UAV's onboard water tank, and Q2 is the volume of the vehicle's water tank. Next, calculate the number of times the vehicle's water tank needs to be refilled, G. Therefore, the number of times the vehicle's water tank needs to be refilled, G, is equal to the number of cleaning route groups. Each cleaning route group consists of F cleaning sub-routes. In other words, since a full vehicle water tank can only support F refills of the drone's onboard water tank at a time, and the vehicle's water tank needs to be refilled at a fixed power station point after it runs out of water, the F cleaning sub-routes... Divide into one route group, for example, the first to third cleaning sub-routes. Divided into the first route group, and the fourth to sixth cleaning sub-routes. Divided into the second route group, and so on, correspondingly, since the drones also need to fly from the vehicle-mounted pod to the starting point of each cleaning sub-route (distance b) n ) and flying from the end of the cleaning sub-course to the vehicle-mounted nest (distance e) n ), and d n= b n +L0+e n Therefore, the routes executed by each route group are as follows: That is, to execute routes d1, d2, d3...d n The routes are divided into G groups, each consisting of F routes. Each group has F routes to be executed. The first group of routes is denoted as d. 11 d 21 d 31 …d F1The G-group route is denoted as d. 1G d 2G …d FG Next, we will construct individual cleaning tasks corresponding to each cleaning route group. In other words, an individual cleaning task refers to the cleaning task completed by the UAV for a single cleaning route group, including tasks related to the UAV and tasks related to the vehicle-mounted air pod.

[0040] If the target location optimization mode is the first optimization mode, then each individual cleaning task is sequentially assigned as a photovoltaic power station cleaning task. In other words, if it's the first optimization mode, meaning the parking location of the vehicle-mounted drone's nest is not fixed, then after the drone completes the cleaning task corresponding to the current flight path group, the vehicle-mounted drone's nest changes its parking location and moves to the next vehicle-mounted drone's nest parking location. In this way, different cleaning flight path groups correspond to different vehicle-mounted drone's nest parking locations, and different photovoltaic power station cleaning tasks correspond to different vehicle-mounted drone's nest parking locations. That is, the vehicle-mounted drone's nest parking location corresponding to photovoltaic power station cleaning task R1 is S1, ..., photovoltaic power station cleaning task R... G The corresponding vehicle-mounted unit parking point is S. G If the target location optimization mode is the second optimization mode, then the sum of the tasks of each individual cleaning task is determined as the photovoltaic power station cleaning task. In other words, if it is the second optimization mode, then the sum of the tasks of all cleaning tasks is the photovoltaic power station cleaning task. That is, under the complete power station cleaning task, there is only one vehicle-mounted rack parking point. In other words, under this mode, there is only 1 photovoltaic power station cleaning task R and only 1 corresponding vehicle-mounted rack parking point S.

[0041] In this embodiment, constructing a target time function corresponding to the photovoltaic power station cleaning task using the start and end points of the cleaning sub-flight route includes: determining the drone operation time and drone water refill time corresponding to the photovoltaic power station cleaning task, and constructing a drone round-trip time function and a drone battery swapping delay time function corresponding to the photovoltaic power station cleaning task using the start and end points of the cleaning sub-flight route; wherein, the drone operation time is the sum of the unit operation time of the drones corresponding to the photovoltaic power station cleaning task; constructing a first time function corresponding to the photovoltaic power station cleaning task based on the drone operation time, the drone water refill time, the drone round-trip time function, and the drone battery swapping delay time function; constructing a second time function using the time function of the vehicle-mounted drone's round-trip to the water replenishment point and the water refill time of the vehicle-mounted drone's nest at the water replenishment point; and constructing a target time function that includes the first time function and the second time function.

[0042] Constructing a target time function corresponding to a photovoltaic power station cleaning task can be understood as follows: In the first optimization mode, the target time function only corresponds to the current single cleaning task, that is, the total time required to complete the current single cleaning task. In this mode, the dependent variable of the target time function is the time required to complete the current single cleaning task. In the second optimization mode, the target time function corresponds to the complete cleaning task, the total time required to complete all single cleaning tasks. Specifically, for example, if there are a total of 10 flight path groups, then there are 10 single cleaning tasks. In the first optimization mode, the 1st, 2nd, ..., 10th single cleaning tasks are sequentially treated as photovoltaic power station cleaning tasks, thus constructing a target time function for each single cleaning task. In the second optimization mode, although there are 10 flight path groups, the sum of the 10 single cleaning tasks is considered as one photovoltaic power station cleaning task, and only this one target time function needs to be constructed.

[0043] The drone operation time and drone water refill time for the photovoltaic power station cleaning task are determined. The drone operation time is the total unit operation time of the drones for the photovoltaic power station cleaning task. In other words, under the first optimization mode, the drone operation time refers to the time spent cleaning the power station within one flight path group, and the drone water refill time refers to the total time required for water refilling within one flight path group. Therefore, under the first optimization mode, the drone operation time is... The time required for the drone to add water is In the second optimization mode, the drone operation time refers to the total time for cleaning the power station across all flight paths, and the drone water refill time refers to the total time required for water refilling across all flight paths. Therefore, in the second optimization mode, the drone operation time is... The time required for the drone to add water is .

[0044] The round-trip time function of the UAV for the photovoltaic power station cleaning task is constructed using the start and end points of the cleaning sub-flight route. It is understandable that, given the start and end points of the cleaning sub-flight route, if the location of the vehicle-mounted UAV's docking point is determined, then the round-trip distance (b, e) is also determined. Therefore, in the first optimization mode, the UAV round-trip time function is: .

[0045] That is, the drone makes F round trips to the vehicle-mounted nest in a single cleaning task.

[0046] If the second optimization mode is used, then the round-trip time function of the drone is: That is, the total number of round trips between the UAV and the vehicle-mounted nest under all cleaning tasks, n, where n is the number of cleaning sub-routes.

[0047] The drone battery swapping delay time function for the photovoltaic power station cleaning task is constructed using the start and end points of the cleaning sub-flight route. It is understandable that, given the start and end points of the cleaning sub-flight route, if the location of the vehicle-mounted drone's docking station is determined, the round-trip distance (b, e) can be determined. Therefore, in the first optimization mode, the drone battery swapping delay time function is: ,in, This refers to the battery swapping delay time for a single battery swap of a drone, while This represents the number of battery swaps in the first optimization mode. In the second optimization mode, the drone battery swap delay time function is... ,in, This refers to the battery swapping delay time for a single battery swap of a drone, while This refers to the number of battery swaps performed in the second optimization mode.

[0048] Next, based on the drone operation time, drone water refill time, drone round-trip time function, and drone battery swapping delay time function, a first time function corresponding to the photovoltaic power station cleaning task is constructed. The percentage of drone power consumption when the drone's onboard water tank is depleted is P0. This is obtained by measuring the percentage of drone power consumed when the drone is fully charged and its onboard water tank is full, cleaning with a unit cleaning water volume Q0 until the onboard water tank is depleted, which completes one unit operation length L0. In the first optimization mode, the first time function is specifically as follows: If it is the second optimization mode, the first time function is specifically as follows: A second time function is constructed using the time function of the onboard unit's round trip to the water replenishment point and the water replenishment time of the onboard unit at the water replenishment point. Under the first optimization mode, the second time function is: ,in, The time for a single water refill at the water replenishment point for the vehicle-mounted air purifier. This refers to the distance between the parking point and the water replenishment point of the vehicle-mounted air purifier. The speed of the vehicle-mounted air pod traveling to and from the water replenishment point and the parking point, in the second optimization mode, is the second time function as follows: , The time for a single water refill at the water replenishment point for the vehicle-mounted air purifier. This refers to the distance between the parking point and the water replenishment point of the vehicle-mounted air purifier. To determine the speed of the vehicle-mounted air pod traveling between the water replenishment point and the parking point, a target time function is constructed, comprising a first time function and a second time function. Specifically, the target time function in the first optimization mode is... for: .

[0049] in, .

[0050] Objective time function in the second optimization mode for: ; in, .

[0051] In this embodiment, the UAV battery swapping delay time function corresponding to the photovoltaic power station cleaning task is constructed using the start and end points of the cleaning sub-route. This includes: constructing a UAV battery swapping count function using the start and end points of the cleaning sub-route, the UAV's empty-load range, the UAV's fully loaded range, and the UAV's flight power consumption; and determining the UAV battery swapping delay time function corresponding to the photovoltaic power station cleaning task by multiplying the UAV battery swapping count function by the battery swapping delay time. The battery swapping delay time is the delay caused by the UAV's onboard water tank swapping batteries while adding water.

[0052] The following describes the specific process of constructing the UAV battery swapping delay time function under the first optimization mode: Since the vehicle's water tank can only be filled F times per trip, and the vehicle needs to be refilled at a fixed power station point after the water tank is depleted, the routes d1, d2, d3, ..., d... will be executed. n The routes are divided into G groups, with each group consisting of F routes. Each group has F routes in operation.

[0053] The dynamic adjustment system reselects the parking location after each water refill. When the cleaning drone performs the k-th route group flight mission, it uses the selected S... k For takeoff and landing positions, the parking position of the vehicle after the kth water refill is only related to the start and end points of the sub-routes included in the kth route group and the location P of the power station fixed replenishment point (used for refilling the vehicle's water tank). The selected parking position of the vehicle's onboard nest for the kth time... The distance to location P, the fixed replenishment point at the power station (used for refilling vehicle water tanks), is denoted as . Location of the vehicle's onboard unit after the kth water refill. The distance to the starting point of each sub-route is denoted as b. 1,k b 2,k b 3,k …b F,k Location of the vehicle's onboard air pod after the kth water refill, from the end of each sub-route. The distance is denoted as e1,k e 2,k e 3,k …e F,k .

[0054] Before the first execution of each new flight path group, the cleaning drone's onboard water tank is refilled and its battery is swapped. Subsequently, each time the drone returns to the vehicle-mounted pod, its onboard water tank is refilled via the vehicle's water storage tank, and the drone's battery level is verified. This process continues until a specific flight path d of the k-th flight path group is completed. r,k If the remaining battery percentage of the cleaning drone meets the following formula, then the battery swapping of the cleaning drone will be performed simultaneously with the water addition: ; .

[0055] in, Indicates the first One operational route, This represents the percentage of total power consumed by the drone while flying at full load, from the "distance S from the vehicle-mounted drone docking point to the starting point of each cleaning sub-route." Similarly, This represents the percentage of total power consumed by the drone during its flight from the vehicle-mounted drone docking point S to the end of each cleaning sub-route when it is unloaded. The percentage of power consumed during flight when the drone's onboard water tank is depleted is P0. This percentage is obtained by measuring the power consumed when the drone is fully charged and its onboard water tank is full, cleaning with a unit water volume Q0 until the onboard water tank is depleted, which completes one unit of work length L0. Indicates that the drone is used up The percentage of the drone's power consumption during flight when the onboard water tank is full, and the percentage of the cleaning drone's remaining power. To complete a certain execution route d of the k-th route group r,k The remaining battery percentage of the cleaning drone is calculated as the difference between the absolute value of the battery consumption percentage of the executed route and the floor value of the battery consumption percentage of the executed route.

[0056] In summary, the specific function for the drone battery swapping delay time in the first optimization mode is as follows: .

[0057] The following describes the specific process of constructing the UAV battery swapping delay time function under the second optimization mode: Each time the drone's onboard water tank is used up, the percentage of the drone's power consumption during flight is P0. This percentage is obtained by measuring the percentage of power consumed by the drone after it is fully charged and its onboard water tank is full, and then cleaning with a unit cleaning volume of Q0 until the onboard water tank is used up, which completes one unit of operation length L0.

[0058] Cleaning drone idle range The data is obtained by measuring the takeoff / return speed v of the cleaning drone when it is fully charged and its onboard water tank is empty. r The cumulative flight distance until the drone's battery is depleted.

[0059] Cleaning drone full-load flight range The method of obtaining the data is by measuring the takeoff / return speed v of the cleaning drone after it is fully charged and its onboard water tank is full. r The cumulative flight distance until the drone's battery is depleted.

[0060] The distances from the vehicle-mounted air purifier parking point S to the starting points of each cleaning sub-line are b1, b2, b3…b n This cleaning process involves the drone being fully loaded (onboard water tank full); completing steps b1, b2, b3…b n The percentage of drone battery power consumed in each flight segment is as follows: The distances from the end point of each cleaning sub-route to the vehicle-mounted aircraft parking point S are e1, e2, e3…e n During this cleaning process, the drone is in an unloaded state (the water in the onboard tank has been depleted); completing steps e1, e2, e3…e n The percentage of drone battery power consumed in each flight segment is as follows: .

[0061] In summary, the formula for calculating the expected number of battery swaps K for the cleaning drone is as follows: ; Therefore, the extended time caused by simultaneously cleaning and refilling the drone's onboard water tank and swapping the battery is the total time spent. : ; The specific function for the drone battery swapping delay time in the second optimization mode is as follows: ; Each time the drone returns to its vehicle-mounted pod, its onboard water tank is refilled via the vehicle's water storage tank, and the drone's battery level is checked during the cleaning process. This process is repeated after completing a specific flight path. rIf the remaining battery percentage of the cleaning drone meets the following formula, then the battery swapping of the cleaning drone will be performed simultaneously with the water addition: ; ; Where r represents the r-th executed route, This represents the percentage of total power consumed by the drone while flying at full load, from the "distance S from the vehicle-mounted drone docking point to the starting point of each cleaning sub-route." Similarly, This represents the percentage of total power consumed by the drone during its flight from the vehicle-mounted drone docking point S to the end of each cleaning sub-route when it is unloaded. Each time the drone's onboard water tank is depleted, the percentage of power consumed during flight is P0. This percentage is obtained by measuring the percentage of power consumed by the drone after it is fully charged and its onboard water tank is full, cleaning with a unit cleaning volume Q0 until the onboard water tank is depleted, which completes one unit of work length L0. To complete a certain execution route d r The remaining battery percentage of the cleaning drone is calculated as the difference between the absolute value of the battery consumption percentage of the executed route and the floor value of the battery consumption percentage of the executed route.

[0062] Step S14: Construct a location optimization model with the objective of minimizing the target time function, and use the location optimization model to iteratively search for the target optimal parking point location in the allowed parking area of ​​the vehicle-mounted vehicle nest.

[0063] After constructing the target time function, it is also necessary to determine the permitted parking area for the vehicle-mounted drone nest, i.e., the area where the vehicle-mounted drone nest can be parked. This allows for the search for the optimal target parking point within the permitted parking area. It is understood that when the vehicle-mounted drone nest is parked at the optimal target parking point, the time required for the drone to complete the photovoltaic power station cleaning task is minimized. The permitted parking area for the vehicle-mounted drone nest is the entire power station area excluding the component installation area. The component installation area refers to the area where the power station components are installed; these areas already have equipment installed, so the vehicle-mounted drone nest cannot be parked in this area. The latitude and longitude of the vertices of the polygon representing the component installation area are denoted as follows: Therefore, the formula for the permitted parking area for vehicle-mounted aircraft housings is as follows: ; in, This indicates the coordinates of the permitted parking points for the vehicle-mounted unit. These represent the minimum and maximum dimensions of the entire power plant area, respectively. The expressions representing the minimum and maximum longitude values ​​for the entire power plant area, respectively, and those not located within the component installation area, are: ,in, This is the negation sign.

[0064] In this embodiment, the step of iteratively searching for the target optimal parking point location within the permitted parking area of ​​the vehicle-mounted vehicle's housing using the location optimization model includes: determining the current parking point location from the permitted parking area of ​​the vehicle-mounted vehicle's housing, and substituting the current optimal parking point location into the target time function to obtain the first time corresponding to the current optimal parking point location; generating a parking point location to be compared with the current parking point location through a global search mechanism or a local generation mechanism; if the parking point location to be compared is within the permitted parking area of ​​the vehicle-mounted vehicle's housing, then substituting the parking point location to be compared into the target time function to obtain... The second time corresponding to the parking point location to be compared; if the first time is not less than the second time and the current preset stopping condition is met, then the parking point location to be compared is determined as the target optimal parking point location; if the first time is not less than the second time and the current preset stopping condition is not met, then a new current parking point location is determined from the vehicle-mounted unit's allowed parking area, and the parking point location to be compared is determined as the new current optimal parking point location, and then the process jumps back to the step of generating the parking point location to be compared for the current parking point location through a global search mechanism or a local generation mechanism.

[0065] The process of iteratively searching for the optimal parking point within the permitted parking area of ​​the vehicle-mounted device nest using the spider-bee algorithm is to continuously update the current optimal parking point position. The permitted parking area of ​​the vehicle-mounted device nest is the search space, and parameter initialization is also required. Specifically, the search space U is divided into multiple non-overlapping sub-regions using a partitioning and filling method. Within each sub-region, populations are randomly generated according to area ratios, and a population size of [missing information]. initial population .

[0066] The iterative optimization process in the first optimization mode is described below: Input F known latitude and longitude coordinates of the starting points of the cleaning sub-routes contained in the k-th route group, B 1,k B 2,k B 3,k …B F,k Recorded as F known k groups of execution routes contain the latitude and longitude coordinates of the endpoints of the cleaning sub-routes, E 1,k E 2,k E 3,k …E F,k Recorded as The latitude and longitude coordinates of the power station's fixed replenishment point are as follows: Set the maximum number of iterations T to 1000, and the preset stopping condition to reach the maximum number of iterations T or the change in the objective function is less than a threshold (10). -6This completes the initialization process.

[0067] Next, we will consider each parking spot within the permitted parking area of ​​the vehicle-mounted aircraft nest as a spider-like structure. Taking latitude optimization as an example, we will sequentially analyze each spider-like structure... Once the current parking spot location is determined, then the optimal parking spot location is determined. Substituting the target time function, we can obtain the first time corresponding to the current optimal parking point position, that is, the time required for the drone to complete the photovoltaic power station cleaning task when the vehicle-mounted drone nest is parked at the current optimal parking point position; and generate the parking point positions to be compared at the current parking point position through a global search mechanism or a local generation mechanism. If the parking point to be compared is within the permitted parking area of ​​the vehicle-mounted vehicle's nest, then the parking point to be compared is substituted into the target time function to obtain the second time corresponding to the parking point to be compared. If the first time is not less than the second time and the current preset stopping condition is met, then the parking point to be compared is determined as the target optimal parking point. If the first time is not less than the second time and the current preset stopping condition is not met, then a new current parking point is determined from the permitted parking area of ​​the vehicle-mounted vehicle's nest, and the parking point to be compared is determined as the new current optimal parking point. Then, the process jumps back to the step of generating the parking point to be compared for the current parking point through a global search mechanism or a local generation mechanism. It can be understood that if the first time is less than the second time and the current preset stopping condition is not met, then the current optimal parking point remains unchanged, and a new current parking point is determined from the permitted parking area of ​​the vehicle-mounted vehicle's nest. Then, the process jumps back to the step of generating the parking point to be compared for the current parking point through a global search mechanism or a local generation mechanism. Similarly, the longitude optimization process is the same as described above. In summary, the update formula for the current optimal parking point location is as follows: ; ; In this way, the global optimal solution is updated. The output makes Minimum location of vehicle-mounted vehicle housing coordinate ,and minimum value Obtain the time taken for each route group. The smallest F dynamically adjusted vehicle-mounted vehicle housing parking point positions After obtaining the coordinates, each execution route and vehicle water replenishment route can be generated and displayed on the power station base map.

[0068] The iterative optimization process in the second optimization mode is described below: Input n known latitude and longitude coordinates of the starting points of the cleaning sub-routes, B1, B2, B3…B nRecorded as Given n known latitude and longitude coordinates of the endpoints of the cleaning sub-routes, E1, E2, E3…E n Recorded as The latitude and longitude coordinates of the power station's fixed replenishment point are as follows: Set the maximum number of iterations T to 1000, and the preset stopping condition to reach the maximum number of iterations T or the change in the objective function is less than a threshold (10). -6 ), complete initialization.

[0069] Next, we will consider each parking spot within the permitted parking area of ​​the vehicle-mounted aircraft nest as a spider-like structure. Taking latitude optimization as an example, we will sequentially analyze each spider-like structure... Once the current parking spot location is determined, then the optimal parking spot location is determined. Substituting the target time function, we can obtain the first time corresponding to the current optimal parking point position, that is, the time required for the drone to complete the photovoltaic power station cleaning task when the vehicle-mounted drone nest is parked at the current optimal parking point position; and generate the parking point positions to be compared at the current parking point position through a global search mechanism or a local generation mechanism. If the parking point to be compared is within the permitted parking area of ​​the vehicle-mounted vehicle's nest, then the parking point to be compared is substituted into the target time function to obtain the second time corresponding to the parking point to be compared. If the first time is not less than the second time and the current preset stopping condition is met, then the parking point to be compared is determined as the target optimal parking point. If the first time is not less than the second time and the current preset stopping condition is not met, then a new current parking point is determined from the permitted parking area of ​​the vehicle-mounted vehicle's nest, and the parking point to be compared is determined as the new current optimal parking point. Then, the process jumps back to the step of generating the parking point to be compared for the current parking point through a global search mechanism or a local generation mechanism. It can be understood that if the first time is less than the second time and the current preset stopping condition is not met, then the current optimal parking point remains unchanged, and a new current parking point is determined from the permitted parking area of ​​the vehicle-mounted vehicle's nest. Then, the process jumps back to the step of generating the parking point to be compared for the current parking point through a global search mechanism or a local generation mechanism. Similarly, the longitude optimization process is the same as described above. In summary, the update formula for the current optimal parking point location is as follows: ; ; In this way, the global optimal solution is continuously updated. That is, continuously update the current optimal parking point position until the output is satisfactory. Minimum location of vehicle-mounted vehicle housing coordinate ,and minimum value , obtain Minimum location of vehicle-mounted vehicle housing After obtaining the coordinates (i.e., the optimal parking point location), each execution route and vehicle water replenishment route can be generated and displayed on the power station base map.

[0070] In this embodiment, generating the comparison parking point position of the current parking point position through a global search mechanism or a local generation mechanism includes: determining the target exploration probability corresponding to the current iteration number using the current iteration number, a preset initial exploration probability, a preset final exploration probability, and a preset maximum iteration number; determining the target mechanism corresponding to the current iteration number from the global search mechanism and the local generation mechanism based on the target exploration probability; if the target mechanism is the global search mechanism, then randomly selecting a reference parking point position from the allowed parking area of ​​the vehicle-mounted vehicle nest, and generating the comparison parking point position of the current parking point position using the reference parking point position and the current parking point position; if the target mechanism is the local generation mechanism, then generating the comparison parking point position of the current parking point position using the current optimal parking point position, a preset local search radius, and a normalized random number.

[0071] First, explore probability. Using adaptive probability and dynamically balancing search and development, the formula for obtaining the target exploration probability is as follows: ; In the formula, To preset the initial exploration probability, it can be set to 0.9. The preset final exploration probability can be set to 0.2, T is the preset maximum number of iterations, and t is the current iteration number.

[0072] The following describes the location of the parking point to be compared in the first optimization mode. Determination process: On the one hand, if the global search mechanism determines the target mechanism corresponding to the current iteration number based on the target exploration probability, then for the current parking point position, i.e., each spider wasp Randomly select reference parking points, i.e., spider wasps Moving closer to it, the initial formula for obtaining the latitude of the parking point location to be compared is as follows: ; In the formula: To control the step size for approaching the optimal solution, it can be set to 0.5. This is the initial parking point position to be compared with the current parking point position.

[0073] On the other hand, if the target mechanism corresponding to the current iteration number is determined by the local generation mechanism based on the target exploration probability, then in the current optimal solution... The nearby random search, i.e., randomly searching around the current optimal parking point location, uses the following formula to initially obtain the latitude of the parking point locations to be compared: ; In the formula: This is the initial, current optimal parking point position. To preset the local search radius, set it to 10% of the search space. To represent normally distributed random numbers (perturbations). .

[0074] The formula for truncating out-of-bounds coordinates is as follows: ; in, This is the dimension of the final parking point location to be compared.

[0075] and Similarly, we can obtain : ; in, This refers to the longitude of the final parking point to be compared.

[0076] The following describes the location of the parking point to be compared in the second optimization mode. Determination process: On the one hand, if the global search mechanism determines the target mechanism corresponding to the current iteration number based on the target exploration probability, then for the current parking point position, i.e., each spider wasp Randomly select reference parking points, i.e., spider wasps Moving closer to it, the initial formula for obtaining the latitude of the parking point location to be compared is as follows: ; In the formula, To control the step size for approaching the optimal solution, it can be set to 0.5. This is the initial parking point position to be compared with the current parking point position.

[0077] On the other hand, if the target mechanism corresponding to the current iteration number is determined by the local generation mechanism based on the target exploration probability, then in the current optimal solution... The nearby random search, i.e., randomly searching around the current optimal parking point location, uses the following formula to initially obtain the latitude of the parking point locations to be compared: ; In the formula, This is the initial, current optimal parking point position. To preset the local search radius, set it to 10% of the search space. To represent normally distributed random numbers (perturbations). .

[0078] The formula for truncating out-of-bounds coordinates is as follows: ; in, This is the dimension of the final parking point location to be compared.

[0079] and Similarly, we can obtain : ; in, This refers to the longitude of the final parking point to be compared.

[0080] Step S15: Control the UAV to collect water from the vehicle-mounted air pod at the optimal parking point of the target in order to perform the photovoltaic power station cleaning task.

[0081] After determining the optimal parking point, the vehicle-mounted drone is positioned at the corresponding optimal parking point while the drone is performing the photovoltaic power station cleaning task. In this way, the drone can collect water from the vehicle-mounted drone at the optimal parking point to perform the photovoltaic power station cleaning task.

[0082] For example Figure 2 The flowchart shown is a specific UAV cleaning management system. The system begins by executing a photovoltaic power station cleaning task, specifically including: (1) selecting a right-angle turn or a U-turn during flight; (2) selecting the nose direction as nose lock or nose follow; (3) setting the radar obstacle avoidance distance; and (4) setting the starting points B1, B2, B3…B of each sub-flight route. n Then the vehicle-mounted air pod closes, the spray boom lowers, and the water pump starts; the drone flies to the end point of each sub-route E1, E2, E3…E n Then the spray bar retracts, the water pump shuts off, and the vehicle-mounted drone nest opens. (5) Turn on RTK and turn on the camera; (6) Set up fault alarms and select hover or return to home when a fault occurs; (7) Turn on the one-click return to home function on the platform and the manual takeover function of the on-site operator using a handheld mobile terminal device to facilitate equipment fault handling; (8) The platform transmits the route planning and parameter setting instructions to the cleaning drone, triggering its autonomous take-off. The cleaning drone flies along the planned route according to the input parameters and begins to perform the cleaning task.

[0083] Furthermore, when performing photovoltaic power station cleaning tasks, the photovoltaic power station cleaning tasks can be monitored in real time. During the cleaning process, the cleaning drone displays information on the front end of the platform for real-time monitoring, including image data, measurement data, location information and equipment status data, and personnel operation confirmation information. Specifically, it includes: (1) Image data: photos and videos taken by the cleaning drone through the high-definition camera or infrared camera, and real-time images of the vehicle-mounted drone's monitoring camera; (2) Measurement data: readings of the added sensors or measuring equipment, such as the water volume of the drone's onboard water tank, the water volume of the vehicle's water tank, the total water consumption monitored by the water meter, and the pollution ratio and dust thickness monitored in real time by the component dust detector; (3) Location information: the flight path, latitude and longitude, flight altitude, flight speed of the cleaning drone, and the real-time positioning of the vehicle-mounted drone; (4) Equipment status data: the status information of the cleaning drone and the vehicle-mounted drone itself, such as battery power, flight time, spray boom retraction status, water pump start and stop, drone switch, feedback that the drone in the drone has landed and returned to center, the status of the vehicle-mounted drone's backup battery, the vehicle-mounted overcharge status, and equipment alarm information. (5) Personnel operation confirmation information: Water and power supply support personnel at the power station's fixed replenishment points confirm the start and completion of water filling for the UAV's onboard water tank and the start and completion of UAV battery swapping via handheld mobile devices. Vehicle-mounted UAV support personnel confirm the start and completion of water filling for the vehicle's storage tank via handheld mobile devices.

[0084] When performing a photovoltaic power station cleaning task, a cleaning task progress can also be generated. The task progress includes cleaning task preparation, cleaning progress, and report generation. Specifically, it includes: (1) confirming each cleaning task preparation item, that is, confirming that the vehicle has arrived at the vehicle-mounted drone nest parking point S; confirming that the power station fixed replenishment point P has been prepared for water and power supply; confirming that each equipment is installed in place and the communication connection status is normal; checking that there are no irrelevant personnel or temporary obstacles in the cleaning area; notifying personnel of the photovoltaic power station who are not related to the flight task to be prohibited from approaching the work area during the cleaning drone's mission. After each cleaning task preparation item is completed, the operator will send a confirmation signal to the platform via a handheld mobile terminal device; (2) cleaning progress, the platform and the handheld mobile terminal device will synchronize the task route group and the flight task completion status of the execution route; (3) generating a cleaning report. The cleaning report includes: 1) the basis for the cleaning decision; 2) the name of each route, fixed flight altitude or terrain-following flight altitude, drone onboard water tank volume Q1′, vehicle water tank volume Q2′, cleaning drone cleaning speed v0′, and cleaning drone take-off / return speed v r Unit cleaning water volume Q0′ (L / min), vehicle speed v c3) The time for adding water to the vehicle's water tank is t1′, the time for adding water to the drone's onboard water tank is t2′, the time for battery swapping during the addition of water to the drone's onboard water tank is t3′, and the time for each operation by personnel is t4′; 4) The images show the flight paths of each route in the actual cleaning task and the parking locations of the vehicle-mounted drone's nest and the fixed replenishment points of the power station are S′ and P′; 5) The total time for completing the cleaning task, the total water consumption for the cleaning task monitored by the water meter at the fixed replenishment point of the power station, and the total electricity consumption for the cleaning task monitored by the electricity meter at the fixed replenishment point of the power station; 6) The cleaning effect is evaluated, and the cleaning effect improvement rate Q is calculated: the daily actual power generation of the photovoltaic power station for x days before the start of cleaning is collected from the power station operation data monitored by the photovoltaic operation and maintenance platform. (If there is power loss due to faults, power outages, or external factors, the corresponding power loss value should be added.) Daily instantaneous irradiance data. And calculate the daily cumulative irradiance. Collect the daily actual power generation of the photovoltaic power station x days after the cleaning is completed. (If there is power loss due to faults, power outages, or external factors, the corresponding power loss value should be added.) Daily instantaneous irradiance data. And calculate the daily cumulative irradiance. The formula for calculating the daily cumulative irradiance C is as follows: ; In the formula, The selected sunrise time for that day. The selected sunset time for the day. This represents the time interval for collecting instantaneous irradiance data.

[0085] The specific formula for calculating the cleaning effect improvement rate Q is as follows: .

[0086] When performing photovoltaic power plant cleaning tasks, the following tasks can also be completed through the cleaning task recording module: 1) Cleaning record saving. The saved record includes: (1) Task number, task name, cleaning task date and time, related equipment number, and route name; (2) Flight log (flight trajectory), clicking on the flight trajectory will open the map to display the flight path; (3) Task materials (pictures, videos), clicking on the materials (pictures, videos) will allow you to view the corresponding pictures and videos; (4) Alarm log; (5) Cleaning report.

[0087] 2) Cleaning record filtering and querying. Cleaning records can be searched and queried by task number or by task name keywords.

[0088] The beneficial effects of this application are as follows: This application divides the total cleaning route of the photovoltaic power station into various cleaning sub-routes based on the unit operation length of the UAV; determines the target location optimization mode for the vehicle-mounted UAV nest parking point; wherein, the target location optimization mode is a first optimization mode where the parking point location is not fixed or a second optimization mode where the parking point location is fixed; establishes a photovoltaic power station cleaning task based on the target location optimization mode, and constructs a target time function corresponding to the photovoltaic power station cleaning task using the start and end points of the cleaning sub-routes; wherein, the dependent variable of the target time function is the time used to complete the photovoltaic power station cleaning task; constructs a location optimization model with the goal of minimizing the target time function, and uses the location optimization model to iteratively search for the target optimal parking point location in the allowed parking area of ​​the vehicle-mounted UAV nest; controls the UAV to take water from the vehicle-mounted UAV nest at the target optimal parking point location to perform the photovoltaic power station cleaning task. Therefore, this application, by dividing the total cleaning route of the photovoltaic power station into cleaning sub-routes according to the unit operation length of the UAV, enables each sub-route to be adapted to the cleaning capacity of the UAV with a full water tank in a single operation, avoiding frequent ineffective round trips due to insufficient water and ensuring the continuity of the operation; it determines two target location optimization modes, fixed or non-fixed parking point locations, which can adapt to different photovoltaic power station environments, improving flexibility and applicability; it establishes cleaning tasks based on the target location optimization mode and constructs a target time function using the start and end points of the sub-routes, constructs a location optimization model that minimizes the target time function, and iteratively searches for the optimal parking point within the allowed parking area, ensuring that the parking point location is legal, i.e., not located in the module installation area, and reduces the total time consumption through iterative optimization, controlling the UAV to collect water from the target optimal parking point to execute the task, which can reduce ineffective movement based on the optimal parking point location, improve operation efficiency, and ensure that the operation process does not interfere with the photovoltaic modules, ensuring safety.

[0089] The following is based on Figure 3 Taking the specific location optimization diagram of the vehicle-mounted unit's parking point as an example, the optimization process in the first optimization mode and the second optimization mode will be explained respectively.

[0090] 1) The optimization process under the first optimization mode is as follows: Since the vehicle's water tank can only be filled F times per trip, and the vehicle needs to be refilled at a fixed power station point after the water tank is depleted, the routes d1, d2, d3, ..., d... will be executed. n The system is divided into G route groups, with each group consisting of F routes. Each route group has F execution routes. The first optimization mode involves reselecting the parking location after each water refill. When the cleaning drone performs the k-th route group flight mission, it uses the selected S... kFor takeoff and landing positions, the parking position of the vehicle after the kth water refill is only related to the start and end points of the sub-routes included in the kth route group and the location P of the power station fixed replenishment point (used for refilling the vehicle's water tank). The selected parking position of the vehicle's onboard nest for the kth time... The distance to location P, the fixed replenishment point at the power station (used for refilling vehicle water tanks), is denoted as . Location of the vehicle's onboard unit after the kth water refill. The distance to the starting point of each sub-route is denoted as b. 1,k b 2,k b 3,k …b F,k Location of the vehicle's onboard air pod after the kth water refill, from the end of each sub-route. The distance is denoted as e 1,k e 2,k e 3,k …e F,k .

[0091] Before the first execution of each new flight path group, the cleaning drone's onboard water tank is refilled and its battery is swapped. Subsequently, each time the drone returns to the vehicle-mounted pod, its onboard water tank is refilled via the vehicle's water storage tank, and the drone's battery level is verified. This process continues until a specific flight path d of the k-th flight path group is completed. r,k If the remaining battery percentage of the cleaning drone meets the following formula, then the battery swapping of the cleaning drone will be performed simultaneously with the water addition: ; ; Next, the objective function is to minimize the estimated cleaning time for each route group. The total time for the k-th route group is then calculated. Specifically, this includes the time function corresponding to the estimated flight time of the executed route of the k-th route group, the time function corresponding to the estimated water replenishment and round-trip time of the vehicle's water tank, and the time function corresponding to the estimated total battery swapping time of the cleaning drone. The target time function under the first optimization mode is as follows: ; In the formula, .

[0092] Furthermore, the permitted area is determined. The permitted area for the vehicle-mounted unit's parking location is the entire power station area minus the component installation area. The latitude and longitude of the polygon vertices of the component installation area are denoted as follows: The permitted areas for parking vehicle-mounted aircraft shelters are as follows: ; Before iterative optimization, parameters are initialized. The search space U is divided into multiple non-overlapping sub-regions using a partitioning method. Within each sub-region, populations are randomly generated based on area proportions, with a population size of [missing value]. initial population Input F known latitude and longitude coordinates of the starting points of the sub-routes contained in the k-th route group, B 1,k B 2,k B 3,k …B F,k Recorded as F known k groups of executed routes contain the latitude and longitude coordinates of the destinations of the sub-routes, E 1,k E 2,k E 3,k …E F,k Recorded as The latitude and longitude coordinates of the power station's fixed replenishment point are as follows: Set the maximum number of iterations T to 1000. The preset stopping condition is set to reaching the maximum number of iterations T or the change in the objective function being less than a threshold (10). -6 ).

[0093] The probability of exploration is Adaptive probability is employed. Dynamic equilibrium search and development: ; In the formula, The initial probability is set to 0.9; Let the final probability be 0.2; This represents the current iteration number.

[0094] Begin iterative optimization for each spider wasp. Randomly select another spider wasp Approach it: ; In the formula, To control the step size for approaching the optimal solution, we set it to 0.5.

[0095] In the current optimal solution Random search nearby: ; In the formula, This is the current globally optimal solution. The local search radius is set to 10% of the search space. These are random numbers (perturbations) distributed according to a standard normal distribution. .

[0096] The formula for truncating out-of-bounds coordinates is shown below: ; and Similarly, we can obtain : ; Calculate the total time for cleaning the flight path at the new location. : ; ; Update the global optimal solution The output makes Minimum location of vehicle-mounted vehicle housing coordinate ,and minimum value Obtain the time taken for each route group. The smallest F dynamically adjusted vehicle-mounted vehicle housing parking point positions After obtaining the coordinates, each execution route and vehicle water replenishment route can be generated and displayed on the power station base map.

[0097] 2) The optimization process under the second optimization mode is as follows: Total estimated length of the route for: ; Estimated flight time for the route for: ; Calculate the estimated water replenishment and round-trip time for the vehicle's water tank, i.e., establish a second time function (including the time function for the vehicle-mounted drone's round-trip to the water replenishment point and the water replenishment time for the vehicle-mounted drone at the water replenishment point). Specifically, to complete the initial total flight path length D, the number of water collection attempts required by the drone is equal to the number of sub-flight paths n, therefore... The vehicle's water tank supports F number of times the drone can collect water after each full tank is filled. In the formula, Q1 is the volume of the UAV's onboard water tank, Q2 is the volume of the vehicle's water tank, and G is the number of times the vehicle's water tank needs to be refilled. Since the vehicle's water tank can only be filled F times per trip, and the vehicle needs to be refilled at a fixed power station point after the water tank is depleted, the routes d1, d2, d3...d will be executed. n The routes are divided into G groups, each consisting of F routes. Each group has F routes to be executed. The first group is denoted as d. 11 d 21 d 31 …d F1 The G group is denoted as d. 1G d 2G d 3G …d FG The distance from the vehicle-mounted water tank parking point S to the fixed replenishment point P at the power station (used for refilling the vehicle's water tank) is denoted as... Estimated time for refilling the vehicle's water tank and round trip. The expression is: ; Calculate the estimated total battery swapping time for the cleaning drone. The percentage of power consumed during flight when the drone's onboard water tank is depleted is denoted as P0. This percentage is determined by measuring the percentage of power consumed when the cleaning drone is fully charged and its onboard water tank is full, cleaning with a unit water volume Q0 until the onboard water tank is depleted, completing one unit of work length L0. The unloaded flight range of the cleaning drone is also calculated. The data is obtained by measuring the takeoff / return speed v of the cleaning drone when it is fully charged and its onboard water tank is empty. r The total flight distance until the drone's battery is depleted. (This refers to the total range of the cleaning drone at full load.) The method of obtaining the data is by measuring the takeoff / return speed v of the cleaning drone after it is fully charged and its onboard water tank is full. r The total flight distance until the drone's battery is depleted. The distance from the vehicle-mounted drone's parking location S to the starting point of each cleaning sub-route is b1, b2, b3…b n This cleaning process involves the drone being fully loaded (onboard water tank full); completing steps b1, b2, b3…b n The percentage of drone battery power consumed in each flight segment is as follows: The distances from the end point of each cleaning sub-route to the vehicle-mounted aircraft parking point S are e1, e2, e3…e n During this cleaning process, the drone is in an unloaded state (the water in the onboard tank has been depleted); completing steps e1, e2, e3…e n The percentage of drone battery power consumed in each flight segment is as follows: In summary, the formula for the expected number of battery swaps K for the cleaning drone is: ; Total time extended due to simultaneous water replenishment and battery swapping for cleaning the drone's onboard water tank: ; Each time the drone returns to its vehicle-mounted pod, its onboard water tank is refilled via the vehicle's water storage tank, and the drone's battery level is checked during the cleaning process. This process is repeated after completing a specific flight path. r If the remaining battery percentage of the cleaning drone meets the following formula, then the battery swapping of the cleaning drone will be performed simultaneously with the water addition: ; ; When a certain execution route d is completed r Cleaning drone remaining battery percentage It is the difference between the absolute value of the power consumption of the executed route and the floor value of the power consumption of the executed route.

[0098] Estimated total time for airline cleaning The corresponding expression is the sum of the time functions for calculating the estimated flight time of the executed route, the estimated time for refilling the vehicle's water tank and the round-trip time, and the estimated total battery swapping time for the cleaning drone: The calculation formula is as follows: ; ; Obtain the permitted area. The permitted area for the vehicle-mounted unit's parking location is the entire power station area minus the component installation area. The latitude and longitude of the vertices of the polygon representing the component installation area are denoted as follows: The permitted areas for parking vehicle-mounted aircraft shelters are as follows: ; Before iterative optimization, the parameters are initialized. The search space U is divided into multiple non-overlapping sub-regions using a partitioning method. Within each sub-region, populations are randomly generated based on area proportions, with a population size of [missing value]. initial population Input the latitude and longitude coordinates of n known sub-flight origins, B1, B2, B3…B n Recorded as Given n known latitude and longitude coordinates of the destinations of sub-routes, E1, E2, E3…E n Recorded as The latitude and longitude coordinates of the power station's fixed replenishment point are as follows: Set the maximum number of iterations T to 1000. The termination condition is set to reaching the maximum number of iterations T or the change in the objective function being less than a threshold (10). -6 ).

[0099] Explore probability Adaptive probability is employed. Dynamic equilibrium search and development: ; In the formula, The initial probability is set to 0.9; Let be the final probability, set to 0.2; t be the current iteration number.

[0100] Next, iterative optimization will be performed. For each spider wasp... Randomly select another spider wasp Approach it: ; In the formula: To control the step size for approaching the optimal solution, we set it to 0.5.

[0101] In the current optimal solution Random search nearby: ; In the formula: This is the current globally optimal solution. The local search radius is set to 10% of the search space. These are random numbers (perturbations) distributed according to a standard normal distribution. .

[0102] Truncate out-of-bounds coordinates: ; and Similarly, we can obtain : ; Calculate the total time for cleaning the flight path at the new location. : ; ; Update the global optimal solution The output makes Minimum location of vehicle-mounted vehicle housing coordinate ,and minimum value To obtain Minimum location of vehicle-mounted vehicle housing After obtaining the coordinates, each execution route and vehicle water replenishment route can be generated and displayed on the power station base map.

[0103] See Figure 4 As shown in the figure, this application discloses a drone cleaning task execution device for a photovoltaic power station, comprising: The route division module 11 is used to divide the total cleaning route of the photovoltaic power station into various cleaning sub-routes according to the unit operation length of the UAV. The mode determination module 12 is used to determine the target location optimization mode of the vehicle-mounted unit's parking point; wherein, the target location optimization mode is a first optimization mode in which the parking point location is not fixed or a second optimization mode in which the parking point location is fixed. The function construction module 13 is used to establish a photovoltaic power station cleaning task according to the target location optimization mode, and to construct a target time function corresponding to the photovoltaic power station cleaning task using the start and end points of the cleaning sub-route; wherein, the dependent variable of the target time function is the time used to complete the photovoltaic power station cleaning task. The location optimization module 14 is used to construct a location optimization model with the goal of minimizing the target time function, and to use the location optimization model to iteratively search for the target optimal parking point location in the allowed parking area of ​​the vehicle-mounted vehicle nest; The task execution module 15 is used to control the UAV to collect water from the vehicle-mounted air pod at the optimal parking point of the target in order to perform the photovoltaic power station cleaning task.

[0104] The beneficial effects of this application are as follows: This application divides the total cleaning route of the photovoltaic power station into various cleaning sub-routes based on the unit operation length of the UAV; determines the target location optimization mode for the vehicle-mounted UAV nest parking point; wherein, the target location optimization mode is a first optimization mode where the parking point location is not fixed or a second optimization mode where the parking point location is fixed; establishes a photovoltaic power station cleaning task based on the target location optimization mode, and constructs a target time function corresponding to the photovoltaic power station cleaning task using the start and end points of the cleaning sub-routes; wherein, the dependent variable of the target time function is the time used to complete the photovoltaic power station cleaning task; constructs a location optimization model with the goal of minimizing the target time function, and uses the location optimization model to iteratively search for the target optimal parking point location in the allowed parking area of ​​the vehicle-mounted UAV nest; controls the UAV to take water from the vehicle-mounted UAV nest at the target optimal parking point location to perform the photovoltaic power station cleaning task. Therefore, this application, by dividing the total cleaning route of the photovoltaic power station into cleaning sub-routes according to the unit operation length of the UAV, enables each sub-route to be adapted to the cleaning capacity of the UAV with a full water tank in a single operation, avoiding frequent ineffective round trips due to insufficient water and ensuring the continuity of the operation; it determines two target location optimization modes, fixed or non-fixed parking point locations, which can adapt to different photovoltaic power station environments, improving flexibility and applicability; it establishes cleaning tasks based on the target location optimization mode and constructs a target time function using the start and end points of the sub-routes, constructs a location optimization model that minimizes the target time function, and iteratively searches for the optimal parking point within the allowed parking area, ensuring that the parking point location is legal, i.e., not located in the module installation area, and reduces the total time consumption through iterative optimization, controlling the UAV to collect water from the target optimal parking point to execute the task, which can reduce ineffective movement based on the optimal parking point location, improve operation efficiency, and ensure that the operation process does not interfere with the photovoltaic modules, ensuring safety.

[0105] Furthermore, embodiments of this application also provide an electronic device. Figure 5 This is a structural diagram of an electronic device 20 according to an exemplary embodiment. The content of the diagram should not be construed as limiting the scope of this application.

[0106] Figure 5This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Specifically, it may include: at least one processor 21, at least one memory 22, a power supply 23, a communication interface 24, an input / output interface 25, and a communication bus 26. The memory 22 stores a computer program, which is loaded and executed by the processor 21 to implement the relevant steps in the method for executing drone cleaning tasks in a photovoltaic power station, as disclosed in any of the foregoing embodiments.

[0107] In this embodiment, the power supply 23 is used to provide operating voltage for various hardware devices on the electronic device; the communication interface 24 can create a data transmission channel between the electronic device and external devices, and the communication protocol it follows can be any communication protocol applicable to the technical solution of this application, and is not specifically limited here; the input / output interface 25 is used to acquire external input data or output data to the outside world, and its specific interface type can be selected according to specific application needs, and is not specifically limited here.

[0108] The processor 21 may include one or more processing cores, such as a quad-core processor or an octa-core processor. The processor 21 may be implemented using at least one hardware form selected from DSP (Digital Signal Processing), FPGA (Field-Programmable Gate Array), and PLA (Programmable Logic Array). The processor 21 may also include a main processor and a coprocessor. The main processor, also known as a CPU (Central Processing Unit), is used to process data in the wake-up state; the coprocessor is a low-power processor used to process data in the standby state. In some embodiments, the processor 21 may integrate a GPU (Graphics Processing Unit), which is responsible for rendering and drawing the content to be displayed on the screen. In some embodiments, the processor 21 may also include an AI (Artificial Intelligence) processor, which is used to handle computational operations related to machine learning.

[0109] In addition, the memory 22, as a carrier for resource storage, can be a read-only memory, random access memory, disk or optical disk, etc. The resources stored on it include operating system 221, computer program 222 and data 223, etc., and the storage method can be temporary storage or permanent storage.

[0110] The operating system 221 manages and controls the various hardware devices and computer programs 222 on the electronic device to enable the processor 21 to perform calculations and processing on the massive amounts of data 223 in the memory 22. It can be Windows, Unix, Linux, etc. The computer program 222, in addition to including a computer program capable of performing the drone cleaning task execution method for a photovoltaic power station disclosed in any of the foregoing embodiments, may further include computer programs capable of performing other specific tasks. The data 223 may include data received by the electronic device from external devices, as well as data collected by its own input / output interface 25.

[0111] Furthermore, this application also discloses a computer-readable storage medium for storing a computer program; wherein, when the computer program is executed by a processor, it implements the aforementioned method for executing a drone cleaning task in a photovoltaic power station. Specific steps of this method can be found in the corresponding content disclosed in the foregoing embodiments, and will not be repeated here.

[0112] The various embodiments in this specification are described in a progressive manner, with each embodiment focusing on its differences from other embodiments. Similar or identical parts between embodiments can be referred to interchangeably. For the apparatus disclosed in the embodiments, since it corresponds to the method disclosed in the embodiments, the description is relatively simple; relevant parts can be referred to in the method section.

[0113] Those skilled in the art will further recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the components and steps of the various examples have been generally described in terms of functionality in the foregoing description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application. The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein can be implemented directly in hardware, software modules executed by a processor, or a combination of both. The software module may be located in random access memory (RAM), memory, read-only memory (ROM), electrically programmable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), register, hard disk, removable disk, CD-ROM (Compact Disc Read-Only Memory), or any other form of storage medium known in the art.

[0114] Finally, it should be noted that in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0115] The present invention provides a detailed description of a method, apparatus, equipment, and medium for performing drone cleaning tasks in a photovoltaic power station. Specific examples have been used to illustrate the principles and implementation methods of the present invention. The descriptions of the above embodiments are only for the purpose of helping to understand the method and core ideas of the present invention. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of the present invention. Therefore, the content of this specification should not be construed as a limitation of the present invention.

Claims

1. A method for performing a UAV cleaning task of a photovoltaic power station, characterized in that, The application relates to a method for optimizing a parking position of a vehicle-mounted UAV nest. The method comprises the following steps: dividing a total cleaning route of a photovoltaic power station into cleaning sub-routes according to a unit operation length of a UAV; determining a target position optimization mode of a parking point of the vehicle-mounted UAV nest; wherein the target position optimization mode is a first optimization mode in which a parking point position is not fixed or a second optimization mode in which a parking point position is fixed; establishing a photovoltaic power station cleaning task according to the target position optimization mode, and constructing a target time function corresponding to the photovoltaic power station cleaning task by using starting points and ending points of the cleaning sub-routes; wherein a dependent variable of the target time function is a time for completing the photovoltaic power station cleaning task; constructing a position optimization model for minimizing the target time function, and iteratively searching for a target optimal parking point position in a parking area allowed by the vehicle-mounted UAV nest by using the position optimization model; 2.The method of claim 1, wherein, controlling the UAV to take water from the vehicle-mounted UAV nest at the target optimal parking point position to execute the photovoltaic power station cleaning task. The method for dividing the total cleaning route of the photovoltaic power station into the cleaning sub-routes according to the unit operation length of the UAV comprises the following steps: planning a total cleaning route of a photovoltaic power station; determining a unit operation time of the UAV by taking a ratio of a volume of an on-board water tank of the UAV to a unit cleaning water volume, and determining a unit operation length of the UAV by multiplying the unit operation time of the UAV by a unit operation speed of the UAV; 3.The method of claim 1, wherein, dividing the total cleaning route into the cleaning sub-routes according to the unit operation length of the UAV. The method for establishing the photovoltaic power station cleaning task according to the target position optimization mode comprises the following steps: dividing each of the cleaning sub-routes into a cleaning route group according to a volume of a water storage tank of the vehicle-mounted UAV nest and the volume of the on-board water tank of the UAV; constructing a single cleaning task corresponding to each of the cleaning route groups; if the target position optimization mode is the first optimization mode, sequentially determining each of the single cleaning tasks as a photovoltaic power station cleaning task; 4.The method of claim 1, wherein, if the target position optimization mode is the second optimization mode, determining a task sum of each of the single cleaning tasks as a photovoltaic power station cleaning task. The method for constructing the target time function corresponding to the photovoltaic power station cleaning task by using the starting points and the ending points of the cleaning sub-routes comprises the following steps: determining a UAV operation time corresponding to the photovoltaic power station cleaning task, a UAV water filling time, and constructing a UAV round trip time function and a UAV battery replacement delay time function corresponding to the photovoltaic power station cleaning task by using the starting points and the ending points of the cleaning sub-routes; wherein the UAV operation time is a unit operation time sum of the UAV corresponding to the photovoltaic power station cleaning task; constructing a first time function corresponding to the photovoltaic power station cleaning task based on the UAV operation time, the UAV water filling time, the UAV round trip time function and the UAV battery replacement delay time function; constructing a second time function by using a time function of a UAV nest round trip water supplement point and a UAV nest water filling time at the water supplement point; 5.The method of claim 4, wherein, constructing the target time function comprising the first time function and the second time function. The method for constructing the UAV battery replacement delay time function corresponding to the photovoltaic power station cleaning task by using the starting points and the ending points of the cleaning sub-routes comprises the following steps: The starting point and the ending point of the cleaning sub-route, the unmanned aerial vehicle empty load endurance mileage, the unmanned aerial vehicle full load endurance mileage, and the unmanned aerial vehicle flight power consumption are used to construct an unmanned aerial vehicle battery replacement frequency function; The product of the unmanned aerial vehicle battery replacement frequency function and the battery replacement delay time is determined as an unmanned aerial vehicle battery replacement delay time function corresponding to the photovoltaic power station cleaning task, wherein the battery replacement delay time is a delay time caused by the on-board water tank of the unmanned aerial vehicle when the water tank is filled with water and the battery is replaced at the same time.

6. The method of claim 1 to 5, wherein, The target optimal parking point position is iteratively searched in the vehicle-mounted machine nest allowed parking area by using the position optimization model, and includes: A current parking point position is determined in the vehicle-mounted machine nest allowed parking area, and the current optimal parking point position is substituted into the target time function to obtain a first time corresponding to the current optimal parking point position; A to-be-compared parking point position of the current parking point position is generated by a global search mechanism or a local generation mechanism; If the to-be-compared parking point position is in the vehicle-mounted machine nest allowed parking area, the to-be-compared parking point position is substituted into the target time function to obtain a second time corresponding to the to-be-compared parking point position; If the first time is not less than the second time and the preset stop condition is met, the to-be-compared parking point position is determined as the target optimal parking point position; If the first time is not less than the second time and the preset stop condition is not met, a new current parking point position is determined in the vehicle-mounted machine nest allowed parking area, and the to-be-compared parking point position is determined as a new current optimal parking point position, and then the step of generating the to-be-compared parking point position of the current parking point position by the global search mechanism or the local generation mechanism is re-executed. 7.The method of claim 6, wherein, The to-be-compared parking point position of the current parking point position is generated by the global search mechanism or the local generation mechanism, and includes: A target exploration probability corresponding to the current iteration number is determined by using the current iteration number, a preset initial exploration probability, a preset final exploration probability, and a preset maximum iteration number; A target mechanism corresponding to the current iteration number is determined from the global search mechanism and the local generation mechanism based on the target exploration probability; If the target mechanism is the global search mechanism, a reference parking point position is randomly selected from the vehicle-mounted machine nest allowed parking area, and the to-be-compared parking point position of the current parking point position is generated by using the reference parking point position and the current parking point position; If the target mechanism is the local generation mechanism, the to-be-compared parking point position of the current parking point position is generated by using the current optimal parking point position, a preset local search radius, and a standard normal distribution random number.

8. An unmanned aerial vehicle cleaning task execution device for a photovoltaic power station, characterized in that, It includes: The route division module is configured to divide a cleaning total route of the photovoltaic power station into each cleaning sub-route according to a unit operation length of the unmanned aerial vehicle; The mode determination module is configured to determine a target position optimization mode of the vehicle-mounted machine nest parking point, wherein the target position optimization mode is a first optimization mode with a non-fixed parking point position or a second optimization mode with a fixed parking point position. a function construction module, configured to establish a photovoltaic power station cleaning task according to the target position optimization mode, and to construct a target time function corresponding to the photovoltaic power station cleaning task by using the start point and the end point of the cleaning sub-route; wherein the dependent variable of the target time function is the time for completing the photovoltaic power station cleaning task; a position optimization module, configured to construct a position optimization model with the objective of minimizing the target time function, and to iteratively search for a target optimal parking point position in a vehicle-mounted UAV nest allowed parking area by using the position optimization model; a task execution module, configured to control the UAV to take water from the vehicle-mounted UAV nest at the target optimal parking point position to execute the photovoltaic power station cleaning task.

9. An electronic device, comprising: comprise: a memory, configured to save a computer program; a processor, configured to execute the computer program to implement the steps of the photovoltaic power station unmanned aerial vehicle cleaning task execution method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, for storing a computer program; wherein the computer program is executed by a processor to implement the steps of the photovoltaic power station unmanned aerial vehicle cleaning task execution method according to any one of claims 1 to 7.

Images