Air-ground combined intelligent cleaning system and cleaning method

Abstract

This invention relates to the field of photovoltaic panel cleaning technology, specifically to an air-to-ground combined intelligent cleaning system and method. The system comprises a ground-mounted photovoltaic panel power generation detection system and an air-to-ground combined intelligent cleaning robot system, which interact via a communication module. The ground-mounted photovoltaic panel power generation detection system collects photovoltaic panel power generation parameters and environmental data in real time, accurately identifies polluted photovoltaic panels and extracts their three-dimensional coordinates, and verifies the cleaning effect by retesting power generation after cleaning. The air-to-ground combined intelligent cleaning robot integrates a drone and a tracked mobile mechanism, possessing dual-mode movement capabilities of aerial flight and ground movement. It can plan the optimal path using the SOM-TSP algorithm, achieve precise landing by combining PID control and lidar, and complete efficient cleaning through high-pressure atomized spraying and elastic wipers. This invention constructs a closed-loop operation system of "detection-cleaning-verification," breaking through the limitations of traditional cleaning technology scenarios, improving cleaning efficiency and quality, and reducing operation and maintenance costs and safety risks.

Description

Technical Field

[0001] This invention relates to the field of photovoltaic panel cleaning technology, specifically to an air-land combined intelligent cleaning system and cleaning method. Background Technology

[0002] With the deepening of the global clean energy strategy, photovoltaic power plants, as the core carrier of renewable energy utilization, have seen their large-scale deployment and efficient operation and maintenance become crucial for industry development. Photovoltaic panels, as the core power generation components of photovoltaic power plants, are exposed to the natural environment for extended periods, easily accumulating pollutants such as dust, gravel, and bird droppings, significantly negatively impacting power generation efficiency. Relevant data shows that when the dust content on the surface of photovoltaic modules reaches 5g / ㎡, its power generation efficiency is only 93% of that in the dust-free state; when the dust accumulation reaches 15g / ㎡, the power generation efficiency drops to 79%. More seriously, the "hot spot effect" caused by pollutants can lead to a sudden increase in localized temperature of the photovoltaic panels, accelerating module aging and degradation, significantly shortening their lifespan, and causing huge economic losses to the power plant. Therefore, regular cleaning of photovoltaic panels has become a necessary step to ensure the stable operation of the power plant.

[0003] However, current mainstream photovoltaic panel cleaning technologies still face many technical bottlenecks, making it difficult to meet the high-efficiency, safe, and universally applicable cleaning requirements of large-scale distributed photovoltaic power plants. 1. Manual cleaning relies on workers using hand tools, which is not only labor-intensive and inefficient, but also requires significant manpower for large-scale photovoltaic power plants, resulting in high cleaning and maintenance costs. Furthermore, manual cleaning poses extremely high safety risks and further limits efficiency for distributed photovoltaic panels installed in complex terrains such as deserts, coastlines, and mountains, as well as on rooftops and other high-altitude locations, making it unsuitable for cleaning needs in these challenging environments.

[0004] 2. Among traditional automated cleaning equipment, ground-mounted tracked cleaning equipment is greatly affected by the flatness of the ground, and the cleaning end is prone to shaking, resulting in low cleaning efficiency and easy scratching and damage to the surface of photovoltaic panels; wall-mounted cleaning devices have strict limitations on the arrangement, installation height and spacing of photovoltaic panels, and can only be used for photovoltaic arrays of specific specifications, with poor versatility and difficulty in adapting to the diverse layout scenarios of distributed photovoltaic power stations.

[0005] 3. Although single drone cleaning equipment has the flexibility of aerial movement and can cover photovoltaic panels in different locations, it is significantly affected by airflow disturbances, has insufficient flight stability, makes it difficult to achieve precise cleaning operations close to the surface of photovoltaic panels, and has limited effectiveness in removing stubborn stains, failing to meet the ideal cleaning quality requirements. Summary of the Invention

[0006] The purpose of this invention is to provide an air-to-land combined intelligent cleaning system and cleaning method, which realizes a closed-loop operation of identifying and locating the photovoltaic panels to be cleaned, planning the optimal path for the area to be cleaned, and verifying the cleaning effect through stable and efficient cleaning. This improves cleaning efficiency and quality, reduces costs and safety risks, and is adaptable to various layouts of photovoltaic panel arrays.

[0007] To achieve the above objectives, the technical solution of this application is: an air-land combined intelligent cleaning system, comprising: The ground-mounted photovoltaic panel power generation detection system identifies photovoltaic panels with declining power generation by detecting the power generation status of each panel. Based on a preset photovoltaic array coordinate system, it extracts the three-dimensional coordinates of a corner of the photovoltaic panel within the overall photovoltaic array and transmits the coordinate information to an air-land combined intelligent cleaning robot via a communication interface. After the cleaning operation is completed, the power generation of the photovoltaic panel is re-measured to quantitatively verify the cleaning effect. The air-to-ground combined intelligent cleaning robot receives the coordinate information of all abnormal photovoltaic panels transmitted by the ground-based photovoltaic panel power generation detection system, calls the som-tsp path planning algorithm to obtain the optimal operation path, and flies to each photovoltaic panel area to be cleaned according to the planned path to complete the cleaning operation.

[0008] In another implementation of the present invention, the air-land combined intelligent cleaning robot includes a drone, the drone having: The casing houses a microprocessor. The camera, connected to the microprocessor, is located at the front of the device's casing; The motor, connected to the microprocessor, is connected to the corner of the fuselage shell via a propeller connecting rod. The propeller is connected to the output shaft of the motor; The GPS positioning module, connected to the microprocessor, is vertically positioned and close to the fuselage. It is detachably mounted on one of the fuselage propeller connecting rods via a ring buckle. The water tank is located below the outer casing of the machine body, and its side is fixedly connected to the connecting rod of the machine body; A lidar unit, installed at the bottom of the water tank, measures the distance to the photovoltaic panel surface; The water spray pipe is located at the front end of the water tank and is vertically connected to the bottom of the water inlet of the water tank. A water inlet cover is provided on the water inlet. A high-pressure water pump, connected to a microprocessor, is installed on the water spray pipe; The nozzle is installed at the bottom of the water spray duct and is positioned towards the cleaning surface of the photovoltaic panel.

[0009] In another embodiment of the present invention, the water tank is generally trapezoidal in shape, and its top is provided with a square groove for battery installation. The battery is embedded in the groove and fixed to provide power to the microprocessor, camera, motor, GPS positioning module and lidar.

[0010] In another implementation of the present invention, the air-land combined intelligent cleaning robot further includes a tracked movement mechanism, which is connected to both sides of the bottom of the fuselage shell via a fuselage link; a drone base rod is provided between the bottom of the fuselage link on the same side.

[0011] In another implementation of the present invention, the tracked movement mechanism has: The drone track wheel connecting column adopts a hollow structure and is correspondingly sleeved on the outside of the drone's base rod; The bottom plate of the tracked mobile unit is connected to the outer side of the connecting column of the UAV track wheel; Track wheels are respectively embedded in the four corners of the bottom plate of the track moving unit; The rubber track is nested around the track wheel to form a ground-based travel transmission structure. The track wheel drive motor is connected to the microprocessor and is fixedly installed inside the bottom plate of the track moving unit. Its output shaft is connected to the corresponding track wheel through a coupling to provide power for the track rotation.

[0012] In another implementation of the invention, the drone also has a wiper blade, which is embedded inside the bottom plate of the two tracked moving units and connected to the front end of the drone's base rod.

[0013] This invention also provides an air-land combined intelligent cleaning method, comprising the following steps: The ground-mounted photovoltaic panel power generation detection system uses the location information of photovoltaic panels with abnormal power generation as the area to be cleaned, and transmits it to the air-land combined intelligent cleaning robot. The air-land combined intelligent cleaning robot uses the som-tsp path planning algorithm to perform global path optimization on the coordinates of all areas to be cleaned, plan the optimal operation path covering all areas to be cleaned, and mark the start and end points of the operation. The air-land combined intelligent cleaning robot takes off from the ground starting point and flies along the planned path to the initial coordinate point of the first area to be cleaned; it adjusts the motor speed through PID control algorithm and, together with the real-time ranging data of lidar, realizes the landing on the area to be cleaned; After landing, the air-to-land combined intelligent cleaning robot switches its motion control mode from flight control mode to tracked wheel ground movement mode. At the same time, it starts a high-pressure water pump and sprays cleaning medium onto the surface of the photovoltaic panel through nozzles, and completes the cleaning operation of the photovoltaic panel in conjunction with the movement of the track. After the current area to be cleaned is cleaned, the tracked wheel drives the vehicle to move to the next area to be cleaned on the same board or adjacent to it along the shortest path planned by the som-tsp algorithm. During the journey, the camera collects environmental data between the photovoltaic panels in real time and obtains the actual distance between the panels based on the collected data. If the distance between the panels is less than a preset threshold, the tracked wheel travel mode is maintained to complete the crossing operation. If the distance between the panels is greater than the preset threshold, the drone flight mode is switched to cross the panels. After landing on the next photovoltaic panel, the tracked travel mode is resumed. After completing the cleaning of the last point to be cleaned in the path planned by the som-tsp algorithm, the robot starts the flight mode and returns to the ground starting point along the optimal return path; The ground-mounted photovoltaic panel power generation testing system retests the power generation of all cleaned photovoltaic panels. By comparing the power generation before and after cleaning with the standard power generation data, it quantitatively determines whether the cleaning effect is qualified.

[0014] In another implementation of the present invention, if the power generation of the photovoltaic panel after cleaning does not reach the preset ratio of the standard power generation, the ground photovoltaic panel power generation detection system sends a secondary cleaning instruction to the air-land combined intelligent cleaning robot until the cleaning effect is qualified.

[0015] In another implementation of the present invention, the som-tsp path planning algorithm is implemented as follows: for the coordinates of the current area to be cleaned... (indicating the first) (One location to be cleaned), find the winning neuron in the circular neural network, and its index Determined by the following formula: in, Indicates the first The weight vector (i.e., the coordinate position of the neuron) of each neuron in the two-dimensional working plane. Indicates Euclidean distance; A Gaussian neighborhood function is used to determine the strength of the influence of the winning neuron on other neurons in its neighborhood: in, Represents neurons With winning neurons Index distance in a ring network topology For time The neighborhood radius parameter of decay; Based on the influence of the winning neuron and its neighborhood, the weight vectors of all neurons in the ring network are iteratively updated: in, For time The decaying learning rate parameter is used to control the step size of weight updates.

[0016] In another implementation of the present invention, the control law of the PID control algorithm is: in, This is the motor speed adjustment output of the PID control. for The robot's height deviation value at any given time, where The target altitude during the robot's descent. for Real-time robot height This is the proportionality coefficient. The integral coefficient is... These are the differential coefficients; According to the adjustment amount Real-time correction of the motor's target speed, so that... The robot quickly converges to zero, enabling a smooth landing relative to the photovoltaic panel and avoiding body jolting or scraping of the photovoltaic panel caused by sudden changes in motor speed.

[0017] By adopting the above technical solution, the present invention can achieve the following technical effects: 1. Adopting an innovative air-to-ground combination design, it combines the flexibility of drones moving in the air with the stability of tracked vehicles on the ground. Through intelligent switching between flight and tracked movement modes, it can smoothly and efficiently cross different spacing between panels. It can not only get close to the surface of photovoltaic panels for precise cleaning, but also adapt to photovoltaic panels of various layouts such as plain power plants, rooftop arrays, and mountain power plants, completely breaking through the scene limitations of traditional cleaning equipment.

[0018] 2. Relying on the ground-mounted photovoltaic panel power generation detection system, the system can identify and locate polluted photovoltaic panels, avoiding blind cleaning; the robot is equipped with the som-tsp path planning algorithm, which can plan the optimal path covering the entire area to be cleaned, abandoning the full-coverage cleaning mode, greatly reducing the mileage of ineffective operations, significantly improving cleaning efficiency, and reducing resource waste.

[0019] 3. Construct a closed-loop operation system for the entire process of "detection-cleaning-verification". Before cleaning, identify the contaminated area. During cleaning, use dual-mode operation to ensure the cleaning effect. After cleaning, use power generation retesting for quantitative verification. This allows for a direct assessment of the recovery of photovoltaic panel power generation efficiency, effectively avoiding incomplete cleaning and ensuring that cleaning quality meets standards.

[0020] 4. The fully automated operation requires no human intervention, completely eliminating safety hazards in high-altitude and complex terrain operations and significantly saving labor costs. At the same time, optimal path planning optimizes flight and travel energy consumption, further reducing the overall cost of power plant clean operation and maintenance, demonstrating significant economic practicality. Attached Figure Description

[0021] To more clearly illustrate the technical solutions in this invention 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 some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0022] Figure 1 Flowchart of an air-land combined intelligent cleaning method; Figure 2 Flowchart for visually perceiving specific locations to be cleaned; Figure 3 A flowchart for cleaning a single photovoltaic panel; Figure 4 Schematic diagram for selecting the motion mode of the moving photovoltaic panel; Figure 5 Front view of the air-land combined intelligent cleaning robot; Figure 6 Side view of the air-land combined intelligent cleaning robot; Figure 7 A structural diagram of the tracked movement mechanism of an air-land combined intelligent cleaning robot; Figure 8 This is a structural diagram of the quadcopter flight unit of an air-land combined intelligent cleaning robot.

[0023] The numbers in the diagram are as follows: 1. Drone; 2. Tracked movement mechanism; 3. Fuselage shell; 4. Camera; 5. GPS positioning module; 6. Fuselage propeller connecting rod; 7. Motor; 8. Propeller; 9. Fuselage connecting rod; 10. Battery; 11. Water tank; 12. Water inlet cap; 13. Water inlet; 14. Water spray guide tube; 15. Nozzle; 16. Sweeper; 17. Base rod; 18. Drone track wheel connecting column; 19. Track wheel; 20. Tracked movement unit base plate; 21. Rubber track; 22. Track wheel drive motor. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention 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. The following description of at least one exemplary embodiment is merely illustrative and is in no way intended to limit the present invention or its application or use. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0025] It should be noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit the exemplary embodiments of the present invention. As used herein, the singular form is intended to include the plural form as well, unless the context clearly indicates otherwise. Furthermore, it should be understood that when the terms "comprising" and / or "including" are used in this specification, they indicate the presence of features, steps, operations, devices, components, and / or combinations thereof.

[0026] Unless otherwise specifically stated, the relative arrangement, numerical expressions, and values ​​of the components and steps set forth in these embodiments do not limit the scope of the invention. It should also be understood that, for ease of description, the dimensions of the various parts shown in the accompanying drawings are not drawn to actual scale. Techniques, methods, and devices known to those skilled in the art may not be discussed in detail, but where appropriate, such techniques, methods, and devices should be considered part of the specification. In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not as limitations. Therefore, other examples of exemplary embodiments may have different values.

[0027] Example 1 This embodiment discloses an air-to-ground combined intelligent cleaning robot system, which includes a ground-based photovoltaic panel power generation detection system and an air-to-ground combined intelligent cleaning robot. The two interact with each other through a communication module and work together to complete the photovoltaic panel cleaning operation.

[0028] The ground-mounted photovoltaic panel power generation detection system acts as the "sensing center," responsible for identifying the pollution status of photovoltaic panels, locating their positions, transmitting data, and verifying the cleaning effect. Specifically, it includes the combiner box of the photovoltaic array, the inverter, and an intelligent monitoring unit for each photovoltaic panel. This unit is responsible for real-time collection of power generation parameters such as output voltage, current, and power of each individual photovoltaic panel, with a sampling frequency of up to once per minute. The system has a set standard power generation capacity and is equipped with a calculation unit. After the collected real-time power generation parameters of each photovoltaic panel are converted into power generation capacity by the calculation unit, the difference between this value and the standard power generation capacity is calculated to determine whether the photovoltaic panel needs cleaning.

[0029] For photovoltaic panels identified as needing cleaning, the three-dimensional coordinates of a specified angle are extracted based on a pre-constructed photovoltaic array coordinate system (X-axis for horizontal position, Y-axis for vertical position, and Z-axis for height). This coordinate system establishes a one-to-one correspondence between the coordinates and the photovoltaic panel number, with the starting point of the air-to-land combined intelligent cleaning robot as the origin, ensuring that the cleaning robot accurately locates the target photovoltaic panel.

[0030] The communication module establishes a stable connection with the air-land combined intelligent cleaning robot through a standardized communication interface, transmitting the coordinate data of the photovoltaic panels to be cleaned to the robot's microprocessor. Simultaneously, it receives the end signal from the robot after its operation, activates the power generation monitoring mode, collects the power generation parameters of the photovoltaic panels after cleaning, compares them with the parameters before cleaning and the standard parameters, and determines whether the cleaning is qualified based on the percentage of power generation after cleaning that reaches the standard parameters. If it is not qualified, a second cleaning command is sent, forming a closed-loop control.

[0031] The air-to-ground combined intelligent cleaning robot system serves as the "executor" for cleaning operations, integrating dual-mode movement capabilities of aerial flight and ground travel. Its structure comprises a drone and a tracked mobile mechanism. The drone integrates core components such as flight power, cleaning execution, energy supply, and perception and control, while the tracked mobile mechanism provides stable support for ground cleaning. The two are firmly fixed together by a high-strength connection structure, ensuring no relative displacement during flight and ground travel, thus balancing flexibility and stability. The drone includes: The fuselage shell is made of ABS material and has an overall octahedral structure. The top two sides are slightly lower than the middle to form a streamlined profile, which reduces air resistance during flight. The shell is 3mm thick and has good impact resistance and waterproof and dustproof rating, making it suitable for complex outdoor environments.

[0032] Flight propulsion system: It adopts a quadcopter flight structure. Four high-power density brushless DC motors are fixed to the four corners of the fuselage shell through four aluminum alloy fuselage propeller connecting rods. The propellers are made of ABS material and are fixed to the motor output shaft in the center through a coupling. Vertical take-off and landing, hovering, turning and other actions are achieved by adjusting the speed difference of each motor.

[0033] Vision and positioning components: A high-definition depth camera is located at the center of the front of the fuselage to collect real-time images of the photovoltaic panel surface and information such as the spacing between panels. Image recognition algorithms are used to achieve boundary recognition, stain detection, and spacing measurement. The GPS positioning module has an accuracy of ±2cm and is vertically mounted close to the fuselage. It is detachably embedded in the fuselage propeller connecting rod via a ring buckle. The lidar has a range of 50m and a ranging accuracy of ±3cm, collecting distance data in real time to provide a basis for precise landing and obstacle avoidance.

[0034] Lower Support and Functional Components: The fuselage linkage and drone base rod are located below the fuselage shell. The drone base rod consists of two parallel aluminum alloy rods, which are fixed to the bottom of the fuselage shell by four fuselage linkages to form a support frame. The cleaning execution components (water tank, water inlet cap, water inlet, water spray guide, nozzle, and wiper) are fixed to the four fuselage linkages. The water tank has a trapezoidal structure (10L capacity, made of polyethylene) with a square groove on top for placing the battery. The water inlet is located at the upper front of the water tank. The water spray guide is made of stainless steel and connects vertically to the bottom of the water inlet. The nozzle is a high-pressure atomizing nozzle. The wiper is made of elastic natural rubber and is embedded in the inner side of the bottom plate of the two tracked moving units and fixed to the front of the drone base rod. The battery is a high-capacity lithium polymer battery, embedded in the groove above the water tank, and passively cooled by the water inside the tank.

[0035] Core microprocessor: Based on the ARM Cortex-A9 architecture airborne microprocessor with a main frequency of 1GHz, it has multi-tasking capabilities, can simultaneously access and process 16 channels of sensor data in real time, and has a built-in embedded operating system to support algorithm program upgrades and function expansion.

[0036] The tracked movement mechanism includes: Connection and support: It is directly connected to the body through the fuselage linkage. Two hollow cylindrical track wheel connecting columns are nested on the outside of the bottom rods on both sides of the UAV and fixed by locking bolts. The lower end is welded and fixed to the middle of the track moving unit base plate (5mm thick aluminum alloy plate, size 60cm×20cm).

[0037] The travel drive assembly has one track wheel at each of the four corners of the tracked moving unit base plate. The rubber track (high-strength wear-resistant material with diamond-shaped anti-slip texture on the surface) is nested on the outside of the four track wheels to form a closed travel structure. Two adjustable speed DC track wheel drive motors are fixed inside the base plate. The output shaft is connected to the corresponding track wheel shaft through a coupling. The forward, backward, and turning actions are achieved by adjusting the speed difference of the motors.

[0038] Example 2: Air-Land Combined Intelligent Cleaning Method Based on the system described in Embodiment 1, this embodiment provides a fully automated photovoltaic panel cleaning method, including 9 core steps, to achieve a closed-loop operation of "equipment calibration - task reception - path planning - flight landing - photovoltaic panel cleaning - area switching - mode switching - return trip - effect verification". The specific steps are as follows: 2.1 Equipment Calibration Phase The air-land combined intelligent cleaning robot performs calibration at a preset starting point on the ground. After calibration, it sends a "ready to standby" signal to the ground detection system and waits for cleaning task instructions.

[0039] 2.2 Task Reception Phase After completing a comprehensive inspection of the photovoltaic array, the ground-mounted photovoltaic panel power generation detection system packages and transmits the data information of all contaminated photovoltaic panels to the robot. After receiving the data, the robot's microprocessor generates a simplified map of the photovoltaic array through image recognition algorithms, marks the location of the area to be cleaned, and sends a "data reception completed" feedback signal after completing the task.

[0040] 2.3 Path Planning Stage The microprocessor calls the som-tsp path planning algorithm to process the coordinates of the area to be cleaned, plan the optimal operation path covering all areas to be cleaned, and mark the start and end points.

[0041] 2.4 Flight and Landing Phase The robot takes off from the ground, flies along the planned path to the initial coordinates of the first area to be cleaned, and completes a smooth landing through PID control and lidar ranging.

[0042] 2.5 Floor Cleaning Stage After landing, the microprocessor automatically switches the main control mode from flight control mode to tracked movement mode, moves to the specific location to be cleaned, starts the high-pressure water pump, and sprays water through the water spray pipe to clean, while simultaneously scraping away the wet stains and residual moisture.

[0043] 2.6 Area Switching Phase After the current area to be cleaned is finished, the track drive motor starts, and the robot moves to the next area to be cleaned on the same board along the shortest path obtained by the som-tsp algorithm.

[0044] 2.7 Sports Mode Switching Phase Determine if the next point on the optimal path is an adjacent photovoltaic panel: if not adjacent, activate flight mode; if adjacent, determine the spacing between panels using visual data collected by the camera. If the spacing exceeds a threshold, activate flight mode; if the spacing is less than the threshold, activate tracked wheel travel mode to reach the adjacent panel.

[0045] 2.8 Return trip phase After completing the cleaning of the last point to be cleaned in the som-tsp algorithm, the robot sends a "cleaning complete" signal to the ground photovoltaic power generation detection system, activates the flight obstacle avoidance module, and selects a straight flight to return to the ground starting point.

[0046] 2.9 Cleaning effect verification stage After receiving the "cleaning operation completed" signal, the ground detection system immediately retests the power generation of all cleaned photovoltaic panels. If the power generation meets the standard, it marks "cleaning qualified"; if it does not meet the standard, it marks "cleaning unqualified" and triggers a secondary cleaning process.

[0047] Application Example 1: Cleaning operations of large-scale centralized photovoltaic power plants in plains environments This application example is a large-scale centralized photovoltaic power station in a plain area. The power station adopts a matrix-style compact layout with uniform and standardized spacing between panels (30cm~80cm). There are no tall obstacles in the surrounding area, and the lighting conditions are stable. However, in certain seasons, it is susceptible to wind and sand, which can cause dust accumulation on the surface of the photovoltaic panels, requiring regular cleaning. The operation configuration includes the air-land combined intelligent cleaning system and method of this invention, along with a ground-mounted photovoltaic panel power generation detection system and multiple evenly distributed ground supply stations. The supply stations have automatic charging and water replenishment functions, supporting continuous robot operation. Pre-operation preparations: Technicians completed the debugging of the ground-mounted photovoltaic panel power generation testing system, constructed a photovoltaic array coordinate system, and set the three-dimensional coordinate axes corresponding to the horizontal and vertical positions and installation height of the photovoltaic panels, with the designated refueling station as the origin. The coordinates of the characteristic angles of each photovoltaic panel were calibrated using positioning technology, establishing a unique correspondence between the coordinates and the photovoltaic panel number. The standard power generation was determined based on the rated power of the photovoltaic panels, and pollution judgment thresholds and cleanliness qualification thresholds were set. This photovoltaic power station covers an area of ​​100 acres and has 10,000 photovoltaic panels measuring 1600mm × 900mm installed.

[0048] Operation implementation process: The ground-mounted photovoltaic panel power generation detection system uses an inverter to collect power generation parameters such as output voltage, current, and power of each photovoltaic panel in real time. After being converted into power generation by the calculation unit, the data is compared with the standard power generation. It was found that the power generation of 200 photovoltaic panels was less than 80% of the standard value, and they were identified as polluted photovoltaic panels. The system also extracts the three-dimensional coordinates of the upper left corner of the panels and transmits them to the robot system through the communication module.

[0049] After receiving the coordinate data, the robot system calls the som-tsp path planning algorithm to optimize the path of 200 pollution points within 0.5 seconds, planning the optimal operation path with a total distance of 5km, which is 40% shorter than the disordered path, and marking the starting point (replenishment station) and the end point.

[0050] The robot takes off from the supply station, ascends vertically along a planned path to a safe altitude, and then transitions to horizontal flight. During flight, it uses GPS positioning and visual navigation to correct its course, precisely reaching the area above the first contaminated photovoltaic panel. It then hovers, with lidar collecting real-time distance data to the photovoltaic panel surface. This data, combined with a PID control algorithm, adjusts the motor speed to achieve a smooth landing.

[0051] After landing, the robot automatically switches the main control mode to tracked movement mode and moves along an S-shaped path at a speed of 0.3 m / s. At the same time, the high-pressure water pump is activated, and cleaning water is delivered to the nozzle through the water spray pipe and sprayed onto the surface of the photovoltaic panel. The squeegee sticks to the surface of the panel and scrapes away the wet stains and residual water. The cleaning time for a single photovoltaic panel is 30 seconds.

[0052] After the current photovoltaic panel is cleaned, the robot moves along the planned path to the next area to be cleaned. When it reaches the boundary between the panels, the camera captures images of the panels, calculates the actual distance, and compares it with a preset threshold of 50cm: if some panels are detected to be 70cm apart (greater than the threshold), the robot automatically switches to flight mode to cross them; for areas with a distance of 30cm, it maintains tracked mode to move directly, ensuring the continuity of the operation.

[0053] After cleaning 200 photovoltaic panels, the robot activates flight mode, returns to the supply station along a straight path, automatically connects to the charging and water replenishment interfaces, and sends a "cleaning complete" signal to the ground detection system.

[0054] After receiving the signal, the ground-based detection system immediately retested the power generation of all cleaned photovoltaic panels. The results showed that 98% of the photovoltaic panels had recovered to more than 95% of the standard value, and were marked as "cleaning qualified," demonstrating excellent cleaning results. A single robot completed the operation in just 5 hours, increasing efficiency by 60 times compared to traditional manual cleaning.

[0055] Application Example 2: Cleaning operations for agricultural-solar hybrid photovoltaic power stations and rooftop distributed photovoltaic arrays This application example includes two types of application scenarios, both of which utilize the air-land combined intelligent cleaning system and cleaning method of this invention, as detailed below: Scenario 1: Cleaning operations at an agricultural-solar hybrid photovoltaic power station In this scenario, photovoltaic panels are installed above farmland with large spacing between them. The field contains obstacles such as crops that have grown taller and irrigation pipes. The area is also rainy, and the surface of the photovoltaic panels easily accumulates pollutants such as mud and fallen leaves. Operations must avoid damaging crops. The cleaning robot of this invention is used in the operation, equipped with a ground detection system and a field supply station. The robot's camera is equipped with crop recognition capabilities to ensure that the operation does not crush or contaminate crops.

[0056] Operation Implementation Process Before the operation, the ground detection system completes a comprehensive inspection of the photovoltaic array. By collecting the power generation parameters of individual photovoltaic panels and comparing them with standard values, it identifies contaminated photovoltaic panels, extracts their three-dimensional coordinates, and transmits them to the robot system.

[0057] The robot's microprocessor calls a path planning algorithm to plan the optimal working path based on the distribution of obstacles in the field. During the flight phase, it activates an automatic obstacle avoidance function to avoid obstacles such as crops on the path.

[0058] Once the device reaches the target photovoltaic panel, it checks the position of the photovoltaic panel support and controls the landing point to maintain a safe distance from the support to avoid collision damage.

[0059] After switching to tracked travel mode, it can identify ground obstacles such as irrigation pipes in real time and automatically adjust its travel path to avoid them; at the same time, it can activate high-pressure spraying and water scraping cleaning functions to efficiently remove pollutants such as mud and fallen leaves from the plate surface.

[0060] After completing the task, the robot returned to the field supply station. The ground detection system retested and showed that the photovoltaic panel power generation efficiency had returned to the expected level, and the crops were undamaged, achieving synergistic protection of cleanliness and agricultural production.

[0061] Scenario 2: Cleaning operation of rooftop distributed photovoltaic array In this scenario, a factory roof is equipped with 500 photovoltaic panels with a spacing of 20cm to 50cm between them. Some areas have obstacles such as ventilation ducts and air conditioning units, and manual cleaning requires working at height, which poses a high safety risk.

[0062] Operation Implementation Process The ground-based detection system collects power generation parameters from rooftop photovoltaic panels via wireless communication. After analysis, it identifies 30 polluted photovoltaic panels and transmits their coordinate data to the robotic system.

[0063] When the robot's microprocessor plans its work path, it automatically avoids obstacles on the roof. Then it takes off from the ground, flies to the area above the roof photovoltaic panels, and uses a GPS positioning module and lidar to achieve a precise landing, avoiding collisions with the roof facilities.

[0064] After landing, the robot switches to tracked mode and starts the cleaning function. During the cleaning process, the camera detects stubborn bird droppings on the surface of a photovoltaic panel. The robot automatically reduces its speed and increases the spraying pressure, successfully removing the stains.

[0065] During the process of crossing the panels, since the spacing between the panels is between 20cm and 50cm (less than the preset threshold), the robot maintains the track mode and moves continuously to complete the cleaning of all contaminated photovoltaic panels.

[0066] After the operation was completed, the robot flew back to the starting point on the ground along the original path. The ground detection system retested the power generation of the cleaned photovoltaic panels. The power generation of all photovoltaic panels recovered to more than 96% of the standard value, and they were marked as "cleaning qualified". The entire process did not require manual climbing onto the roof, completely eliminating the safety hazards at height.

[0067] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. An air-land combined intelligent cleaning system, characterized in that, include: The ground-mounted photovoltaic panel power generation detection system identifies photovoltaic panels with declining power generation by detecting the power generation status of each panel. Based on the preset photovoltaic array coordinate system, the three-dimensional coordinates of one corner of the photovoltaic panel in the overall photovoltaic array are extracted, and the coordinate information is transmitted to the air-land combined intelligent cleaning robot through the communication interface. After the cleaning operation is completed, the power generation of the photovoltaic panel will be retested to quantitatively verify the cleaning effect. The air-to-ground combined intelligent cleaning robot receives the coordinate information of all abnormal photovoltaic panels transmitted by the ground-based photovoltaic panel power generation detection system, calls the som-tsp path planning algorithm to obtain the optimal operation path, and flies to each photovoltaic panel area to be cleaned according to the planned path to complete the cleaning operation.

2. The air-land combined intelligent cleaning system according to claim 1, characterized in that, The air-land combined intelligent cleaning robot includes a drone (1), which has: The casing (3) contains a microprocessor; The camera (4), connected to the microprocessor, is located at the front of the casing (3); The motor (7) is connected to the microprocessor and is connected to the corner of the fuselage shell (3) via the fuselage propeller connecting rod (6); The propeller (8) is connected to the output shaft of the motor; The GPS positioning module (5) is connected to the microprocessor, is set vertically upward and close to the fuselage shell (3), and is detachably embedded in one of the fuselage propeller connecting rods (6) by a ring buckle; The water tank (11) is located below the outer casing (3) of the machine body, and its side is fixedly connected to the connecting rod (9) of the machine body; A lidar is installed at the bottom of the water tank (11) to measure the distance to the photovoltaic panel. A water spray pipe (14) is installed at the front end of the water tank (11) and is vertically connected to the bottom of the water inlet (13) of the water tank (11). A water inlet cover (12) is provided on the water inlet (13). A high-pressure water pump, connected to a microprocessor, is installed on the water spray pipe (14); The nozzle (15) is installed at the bottom of the water spray conduit (14) and is positioned toward the cleaning surface of the photovoltaic panel.

3. The air-land combined intelligent cleaning system according to claim 2, characterized in that, The water tank (11) has a trapezoidal structure and a square groove on its top for battery installation. The battery is embedded in the groove and fixed to provide power to the microprocessor, camera, motor, GPS positioning module and lidar.

4. The air-land combined intelligent cleaning system according to claim 2, characterized in that, The air-land combined intelligent cleaning robot also includes a tracked movement mechanism (2), which is connected to the bottom sides of the fuselage shell (3) via a fuselage link (9); a drone base rod (17) is provided between the bottoms of the fuselage link (9) on the same side.

5. The air-land combined intelligent cleaning system according to claim 4, characterized in that, The tracked moving mechanism (2) has: The UAV track wheel connecting column (18) adopts a hollow structure and is correspondingly sleeved on the outside of the UAV base rod (17); The tracked mobile unit base plate (20) is connected to the outside of the UAV track wheel connecting column (18); Track wheels (19) are respectively embedded in the four corners of the bottom plate (20) of the track moving unit; The rubber track (21) is nested inside the track wheel (19) to form a ground travel transmission structure; The track wheel drive motor (22) is connected to the microprocessor and is fixedly installed on the inner side of the track moving unit base plate (20). Its output shaft is connected to the corresponding track wheel (19) through a coupling to provide power for track rotation.

6. The air-land combined intelligent cleaning system according to claim 5, characterized in that, The drone (1) also has a wiper blade (16) which is embedded inside the bottom plate (20) of the two tracked moving units and connected to the front end of the drone's bottom rod (17).

7. An air-land combined intelligent cleaning method, characterized in that, Includes the following steps: The ground-mounted photovoltaic panel power generation detection system uses the location information of photovoltaic panels with abnormal power generation as the area to be cleaned, and transmits it to the air-land combined intelligent cleaning robot. The air-land combined intelligent cleaning robot uses the som-tsp path planning algorithm to perform global path optimization on the coordinates of all areas to be cleaned, plan the optimal operation path covering all areas to be cleaned, and mark the start and end points of the operation. The air-land combined intelligent cleaning robot takes off from the ground starting point and flies along the planned path to the initial coordinate point of the first area to be cleaned; it adjusts the motor speed through PID control algorithm and, together with the real-time ranging data of lidar, realizes the landing on the area to be cleaned; After landing, the air-to-land combined intelligent cleaning robot switches its motion control mode from flight control mode to tracked wheel ground movement mode. At the same time, it starts a high-pressure water pump and sprays cleaning medium onto the surface of the photovoltaic panel through nozzles, and completes the cleaning operation of the photovoltaic panel in conjunction with the movement of the track. After the current area to be cleaned is cleaned, the tracked wheel drives the vehicle to move to the next area to be cleaned on the same board or adjacent to it along the shortest path planned by the som-tsp algorithm. During the journey, the camera collects environmental data between the photovoltaic panels in real time and obtains the actual distance between the panels based on the collected data. If the distance between the panels is less than a preset threshold, the tracked wheel travel mode is maintained to complete the crossing operation. If the distance between the panels is greater than the preset threshold, the drone flight mode is switched to cross the panels. After landing on the next photovoltaic panel, the tracked travel mode is resumed. After completing the cleaning of the last point to be cleaned in the path planned by the som-tsp algorithm, the robot starts the flight mode and returns to the ground starting point along the optimal return path; The ground-mounted photovoltaic panel power generation testing system retests the power generation of all cleaned photovoltaic panels. By comparing the power generation before and after cleaning with the standard power generation data, it quantitatively determines whether the cleaning effect is qualified.

8. The air-land combined intelligent cleaning method according to claim 7, characterized in that, If the power generation of the photovoltaic panels does not reach the preset proportion of the standard power generation after cleaning, the ground photovoltaic panel power generation detection system sends a secondary cleaning instruction to the air-land combined intelligent cleaning robot until the cleaning effect is qualified.

9. The air-land combined intelligent cleaning method according to claim 7, characterized in that, The SOM-TSP path planning algorithm is implemented as follows: for the coordinates of the current area to be cleaned... Find the winning neuron in a circular neural network, its index Determined by the following formula: in, Indicates the first The weight vector of each neuron in the two-dimensional working plane Indicates Euclidean distance; A Gaussian neighborhood function is used to determine the strength of the influence of the winning neuron on other neurons in its neighborhood: in, Represents neurons With winning neurons Index distance in a ring network topology For time The neighborhood radius parameter of decay; Based on the influence of the winning neuron and its neighborhood, the weight vectors of all neurons in the ring network are iteratively updated: in, For time The decaying learning rate parameter is used to control the step size of weight updates.

10. The air-land combined intelligent cleaning method according to claim 7, characterized in that, The control law of the PID control algorithm is: in, This is the motor speed adjustment output of the PID control. for The robot's height deviation value at any given time, where The target altitude for the robot during its descent. for Real-time robot height This is the proportionality coefficient. The integral coefficient is... These are the differential coefficients; According to the adjustment amount Real-time correction of the motor's target speed, so that... The robot rapidly converges to zero, enabling a smooth descent relative to the photovoltaic panel.

Images