CN115889271A - A method of photovoltaic cleaning combined with drones - Google Patents

Abstract

A photovoltaic cleaning method combined with drones, comprising the following steps: step 1, divide the photovoltaic power plant into several areas, and assign a set of cleaning devices to each area; step 2, use drones to image photovoltaic panels Collect, use the cloud platform to identify stains and set the cleaning track; step 3, use the drone to launch the photovoltaic cleaning robot, and the photovoltaic cleaning robot will clean along the planned path; step 4, the drone will finally transport the photovoltaic cleaning robot to In the energy storage charging compartment. The invention provides a photovoltaic cleaning method combined with an unmanned aerial vehicle, which can clean point-to-point, improves the efficiency of stain cleaning, and reduces the impact of reducing power generation efficiency due to occlusion during cleaning.

Description

A method of photovoltaic cleaning combined with drones

technical field

The invention relates to a solar cell panel cleaning method, in particular to a photovoltaic cleaning method combined with an unmanned aerial vehicle.

Background technique

A photovoltaic panel is a panel array composed of a series of solar panels connected in series and parallel. Due to the occlusion of dust and other stains, it will affect the power generation efficiency of the photovoltaic panel, and even generate photovoltaic hot spots, which will damage the photovoltaic panel components; nowadays, photovoltaic There are various forms of power stations, such as building-integrated photovoltaics, large-scale photovoltaic power stations in desert areas in the northwest, etc.; the solar panels of these photovoltaic power stations are susceptible to dust, bird droppings, scale, oil and other stains. Efficiency drops drastically, causing economic losses, so cleaning of photovoltaic panels is essential.

Nowadays, there are various cleaning methods for photovoltaics. Some use manual cleaning, which is not only costly, but also has low cleaning efficiency, which is time-consuming and laborious; some use manual control of robotic arms (large mechanical equipment) for cleaning. This method is operated It is difficult, and it is necessary to pay attention to the position of the mechanical arm at all times to ensure that the cleaning brush and the photovoltaic panel fit together. , water resources are relatively scarce, and cleaning with high-pressure water guns will lead to waste of water resources, and the scale left behind may form a second shield on the photovoltaic panel; there are also photovoltaic cleaning robots for cleaning, but most cleaning robots have relatively low cleaning paths. Fixed, the cleaning area covers the entire panel, which will not only block the panel when the panel emits light, and affect the power generation efficiency, but also the cleaning efficiency of a large area is not high.

Contents of the invention

The technical problem to be solved by the present invention is to provide a photovoltaic cleaning method combined with an unmanned aerial vehicle, which can realize point-to-point cleaning of photovoltaic panel stains in a way that can be detected first and then cleaned, which improves the efficiency of stain cleaning and reduces cleaning time. The effect of reducing power generation efficiency due to shading.

In order to solve the problems of the technologies described above, the technical solution adopted in the present invention is:

A photovoltaic cleaning method combined with a drone, comprising the following steps:

Step 1. Divide the photovoltaic power station into several regions according to the size of the photovoltaic power station and the input cost of photovoltaic cleaning, and assign a set of cleaning devices to each region;

Step 2. During operation, the cloud platform sends start signals to the energy storage and charging bins and UAVs in each area according to the manually set time period; after receiving the signal, the energy storage and charging bins open the end cover; at the same time, the UAV Power on and go to the scheduled departure point planned by the cloud platform;

Step 3. After arriving at the starting point, the UAV flies along the direction of the photovoltaic panels in the area at a certain speed; when it reaches the predetermined image collection point planned by the cloud platform, the UAV will take an image of the photovoltaic panels through the machine vision camera attached below. Acquisition, the collection results will be uploaded to the cloud platform in real time through the receiver; during this period, the cloud platform will monitor the feedback data of the UAV’s onboard ultrasonic distance sensor in real time to ensure that the flight inspection height of the UAV and the vertical distance of the photovoltaic panel are maintained At a certain height, so as to take into account the flight safety of the UAV and the integrity of the collected pictures;

Step 4. After the cloud platform receives the images from the drone inspection, it will use the open source vision library to identify stains;

Step 5. The cloud platform will establish a coordinate system in a series of pixel points of the image, mark the identified pollutants and record the coordinates of the pollutants in this coordinate system according to the results of stain identification; after the identification process is completed , the cloud platform presents all pollutant marker points in the entire photovoltaic panel area according to their coordinates, and classifies adjacent pollutant marker points as a node, and plans photovoltaic cleaning through the Dijkstra algorithm in graph theory. Robot's initial cleaning path;

Step 6. After the cloud platform plans the cleaning path of the photovoltaic cleaning robot, the cloud platform sends a return command to the drone. When the flight controller of the drone receives the command, it returns to the energy storage and charging compartment at a certain speed, and uses the combined The hook transfers the photovoltaic cleaning robot away from the energy storage and charging bin; after reaching the starting point, the drone drops the photovoltaic cleaning robot onto the photovoltaic panel. After the release is completed, the drone returns to the charging compartment;

Step 7. When the photovoltaic cleaning robot reaches the photovoltaic panel, it will clean along the path planned by the cloud platform until the cleaning is completed;

Step 8. After the photovoltaic cleaning robot finishes cleaning along the path, it stops at the end point and sends a completion signal to the cloud platform. After receiving the signal, the cloud platform orders the drone to start, fly away from the energy storage and charging compartment, and follow the The inspection method and path at the beginning are inspected;

Step 9. The UAV collects the picture of the cleaned photovoltaic panel and uploads it to the cloud platform. The cloud platform uses the open source computer vision library for image recognition, and then uploads the recognition result to the background. The place where there are still stains is marked and displayed on the On the visualization panel in the background, the panel shows which area of the photovoltaic power station the stain belongs to, and which photovoltaic panel it is located in this area; at the same time, the cloud platform will give an alarm prompt and wait for manual further processing; the staff can according to The image result selection will re-call the robot back to clean; or take manual cleaning;

Step 10. After completing all the operations, the photovoltaic cleaning robot will stop at the end position and send a retraction signal to the cloud platform. After receiving the signal, the cloud platform will order the drone to transport the photovoltaic cleaning robot to the energy storage and charging bin .

The cleaning device in the first step includes a drone, a photovoltaic cleaning robot, an energy storage charging bin and a cloud platform, wherein:

The UAV adopts the existing conventional unmanned aerial vehicle, and the unmanned aerial vehicle is additionally equipped with an ultrasonic distance sensor, a machine vision recognition camera, an infrared monitoring probe and a power monitoring module;

The photovoltaic cleaning robot adopts a crawler-type robot body, the robot body is equipped with a control circuit board, and a cleaning mechanism is installed outside the robot body;

The energy storage charging bin includes a bin body and an end cover, wherein a charging plug is provided in the bin body, and the charging plug is plugged into the charging slot of the photovoltaic cleaning robot; the end cover is set on the left and right and opened or closed by the actuator;

The cloud platform communicates with drones, photovoltaic cleaning robots, and energy storage and charging bins for unified management.

The combined hook in the step 6 includes a controllable hook installed on the bottom of the drone and a hanging ring installed on the top of the photovoltaic cleaning robot. The controllable hooks and the hanging rings are multiple groups and correspond one by one.

The controllable hook includes a hook body, the lower end of the hook body is equipped with a hook through a tension sensor, the upper part of the hook body is installed with an integrated control module, and the upper part of the hook body is opposite to the hook. Press and form an opening with the hook. After the photovoltaic cleaning robot is hung, the locking clip moves down through the telescopic mechanism and forms a closed-loop structure with the hook.

The telescopic mechanism includes a push-pull electromagnet, and the driving part of the push-pull electromagnet is installed in the waist-shaped hole of the lever through a pin shaft; one end of the lever is hinged with the bracket, and the other end is hinged with the push rod, and the push rod slides through the chute and is arranged on the One side and the lower end of the bracket are in contact with the locking clip.

The cleaning mechanism includes a disc-shaped cleaning brush, a strip-shaped auxiliary cleaning brush and a dust suction device; wherein the disc-shaped cleaning brushes are arranged at intervals on the front of the robot body and driven by a motor; the strip-shaped auxiliary cleaning brush is arranged on the The rear part of the robot body; the dust suction device is located between the disc-shaped cleaning brush and the strip-shaped auxiliary cleaning brush, and the opening faces the disc-shaped cleaning brush.

The actuator includes a rack fixed on the back of the end cover, the rack is placed up and down in the limit groove of the frame of the bin, the outer side of the rack is engaged with the gear, and the gear is driven by a motor.

There are brush head stain cleaning grooves and track grooves installed in the warehouse body, and the brush head stain cleaning grooves and track grooves correspond to the walking mechanism and cleaning mechanism of the photovoltaic cleaning robot; device, a pressure sensor is arranged in the track groove.

A shrinkable protective sleeve is arranged on the outside of the charging plug, and the shrinkable protective sleeve is fixed on the baffle, and the lower end of the baffle is installed in the shell through a spring.

A photovoltaic cleaning system combined with an unmanned aerial vehicle of the present invention has the following technical effects:

1) Using a photovoltaic cleaning system that combines drones and photovoltaic cleaning robots, the method of detecting first and then cleaning realizes point-to-point cleaning of photovoltaic panel stains, improves the efficiency of stain cleaning, and reduces power generation due to occlusion during cleaning Efficiency impact.

2) Stain recognition based on machine vision, and path planning using Dijkstra algorithm, which realizes fixed-point cleaning of stains and reduces the impact of photovoltaic cleaning robots on photovoltaic panel power generation.

3) The energy storage charging compartment adopts two open end covers to close the charging compartment, which is dustproof and rainproof; the brush head stain cleaning tank and track grooves set in the energy storage charging compartment can realize the positioning of the photovoltaic cleaning robot, Through two sets of pressure sensors, it can be judged whether the charging plug is fully inserted into the charging slot, so that the photovoltaic cleaning robot can be accurately docked.

4) Using the stain recognition method based on machine vision, and using the Dijkstra algorithm for path planning, the fixed-point cleaning of stains is realized, and the impact of photovoltaic cleaning robots on photovoltaic panel power generation is reduced.

5) Realized the intelligent operation and maintenance of the photovoltaic system, integrating monitoring and cleaning, the whole system has a clear division of labor and clear process, realized the centralized and remote management of photovoltaic operation and maintenance, greatly reduced human labor, and improved the efficiency of photovoltaic operation and maintenance work efficiency.

6) The energy storage charging compartment can regularly clean the cleaning brush of the photovoltaic cleaning robot, so as to realize the self-cleaning of the device, prolong the service life of the device, and reduce manual operation.

7) Fixed-point cleaning and comprehensive cleaning can be realized. These two methods minimize the impact of robot cleaning on the power generation efficiency of photovoltaic panels.

Description of drawings

Below in conjunction with accompanying drawing and embodiment the present invention will be further described:

Fig. 1 is a schematic diagram of a partial structure of an unmanned aerial vehicle in the present invention.

Fig. 2 is the front view of the drone in the present invention.

Fig. 3 is a front view of the hanging ring in the present invention.

Fig. 4 is a front view of the controllable hook in the present invention.

Fig. 5 is a front view of the telescoping mechanism in the present invention.

Fig. 6 is a top view of the bottom plate of the present invention.

Fig. 7 is a schematic diagram of the state when the controllable hook is suspended in the present invention.

Fig. 8 is a schematic diagram of the state when the controllable hook is released in the present invention.

Fig. 9 is a top view of the photovoltaic cleaning robot in the present invention.

Fig. 10 is a bottom view of the photovoltaic cleaning robot in the present invention.

Fig. 11 is a schematic structural diagram of a photovoltaic cleaning robot in the present invention.

Fig. 12 is a top view of the energy storage charging bin in the present invention.

Fig. 13 is a schematic structural view of the end cap in the present invention.

Fig. 14 is a schematic structural diagram of the actuator in the present invention.

Fig. 15 is a front view of the charging plug in the present invention.

Fig. 16 is the stain identification process in the present invention.

Fig. 17 is the stain recognition result (1) in the present invention.

Fig. 18 is the stain recognition result (2) in the present invention.

Fig. 19 is a system block diagram of the present invention.

In the figure: drone 1, drone control part 1.1, signal receiver 1.2, machine vision recognition camera 1.3, photovoltaic cleaning robot 2, robot body 2.1, walking mechanism 2.2, cleaning mechanism 2.3, disc-shaped cleaning brush 2.3 .1, long strip auxiliary cleaning brush 2.3.2, vacuum device 2.3.3, charging tank 2.4, energy storage charging bin 3, bin body 3.1, end cover 3.2, actuator 3.3, rack 3.3.1, limit position Groove 3.3.2, gear 3.3.3, motor 3.3.4, brush head stain cleaning tank 3.4, charging plug 3.5, pressure sensor 3.6, track groove 3.7, cloud platform 4, controllable hook 5, hook body 5.1, tension sensor 5.2, integrated control module 5.4, locking clip 5.5, spring 5.6, telescopic mechanism 5.7, push-pull electromagnet 5.7.1, lever 5.7.2, bracket 5.7.3, push rod 5.7.4, hanging ring 6, screw hole 6.1 , bottom plate 7.

Detailed ways

As shown in Figure 1, a photovoltaic cleaning system combined with a drone includes a drone 1, a photovoltaic cleaning robot 2, an energy storage charging bin 3, and a cloud platform 4. Cloud platform 4 includes computer, ground station control software (MixxionPlaner), real-time image transmission signal and signal transceiver device (Pixhawk) and supporting receiver, etc., which play the role of control center. The UAV 1, the photovoltaic cleaning robot 2, the energy storage charging bin 3 and the cloud platform 4 are connected through a signal transceiver device to realize data transmission and command transmission. The UAV 1 and the photovoltaic cleaning robot 2 are connected through a combined hook to realize the positioning and transportation of the photovoltaic cleaning robot.

As shown in Figure 1-2, the UAV 1 is not much different from the normal unmanned aerial vehicle (DJI Matrice 600). Due to the need for stable flight and safety of the unmanned aerial vehicle, this solution adopts a six-rotor structure, which is different from the market The commonly used four-rotor structure is different. The six-rotor structure is relatively more powerful and can carry higher loads; with a wider wheelbase and more rotors, it can fly more stably and has a certain risk buffer capability. (with single-axis propeller stop protection function).

In addition to modules such as the drone control part 1.1, the first signal transceiver module 1.2, and the voltage regulation module, the electronic equipment carried by the drone 1 is also equipped with an ultrasonic distance sensor (XL-MaxSonar EZ0), a machine vision recognition camera 1.3 (AT-OV9281), each pair of horizontal axes of UAV 1 is equipped with the above-mentioned ultrasonic module, visual recognition module, and infrared module, and all the above-mentioned expansion modules can be connected to the flight controller through the PIN line.

As shown in Figure 3-4, the combined hook is composed of a controllable hook 5 and a hanging ring 6. The controllable hook 5 is connected to the bottom plate 7 of the UAV 1 through the screw hole on the top, and the hanging ring 6 is connected to the photovoltaic system through the screw hole at the bottom. The top of the body of the cleaning robot 2 is connected; when working, the UAV 1 drives the controllable hook 5 to penetrate the hanging ring 6 from the upper part; the integrated control module of the controllable hook 5 is connected to the UAV control part of the UAV 1, The integrated control module makes the telescopic mechanism 5.7 expand and contract by identifying the signal of the UAV control part and the tension sensor, thereby realizing the locking and releasing of the controllable hook 5 .

As shown in Figure 3, there are 4 screw holes at the bottom of the hanging ring 6, and there are 3 sets of screw holes 6.1 on the top of the corresponding photovoltaic cleaning robot 2, each group has 4 screw holes, the hanging ring 6 and the photovoltaic cleaning robot 2 The car body is connected by M2.5 cylinder head hexagon socket screws.

As shown in Fig. 4, there are four screw holes on the top of the controllable hook 5, and the bottom plate 7 of the corresponding UAV 1 has 3 sets of screw holes, each group has 4 screw holes, and the controllable hook 5 uses M2. 5. Connect the socket head cap screws with the bottom plate of the UAV 1.

As shown in Figure 4, the controllable hook includes a hook body 5.1, a hook 5.3 is installed at the lower end of the hook body 5.1 through a tension sensor 5.2, and an integrated control module 5.4 is installed on the upper part of the hook body 5.1. The core of the integrated control module 5.4 is an intelligent The control module, its pins are connected with tension sensor PBCL-03, push-pull electromagnet, and the flight controller of the drone; the upper part of the hook body 5.1 is opposite to the hook 5.3 and is provided with a locking clip 5.5, and the locking clip 5.5 is under normal conditions The spring 5.6 is pressed upwards and forms an opening with the hook 5.3. After the photovoltaic cleaning robot 2 is hung, the locking clip 5.5 moves down through the telescopic mechanism 5.7 and forms a closed-loop structure with the hook.

As shown in Figure 5, the telescopic mechanism 5.7 includes a push-pull electromagnet 5.7.1, the driving part of the push-pull electromagnet 5.7.1 is fixed with a pin shaft, and the pin shaft is freely inserted into the waist-shaped hole of the lever 5.7.2, The pin shaft can move left and right along the waist hole. The part of the pin shaft protruding out of the waist-shaped hole is limited by a limit ring to prevent the pin shaft from breaking away from the waist-shaped hole. One end of the lever 5.7.2 is hinged with the bracket 5.7.3, and the other end is hinged with the push rod 5.7.4. The push rod 5.7.4 is slidably arranged on one side of the support 5.7.3 through the chute and the lower end is in contact with the locking clip 5.5. The push-pull electromagnet 5.7.1 is electrically connected with the integrated control module 5.4, and the power-on or power-off is completed through the integrated control module 5.4.

As shown in Figure 7-8, when the UAV 1 returns to the energy storage and charging compartment 3, it hovers at the starting position and sends a signal to the cloud platform 4. When the cloud platform 4 receives the signal, it orders no one to Machine 1 flies towards the slanted upward uniform speed of link 6, so just can make link 5.3 pass link 6. When the hook 5.3 is not hanging objects, the locking clip 5.5 keeps the hook 5.3 always in an open state due to the action of the spring 5.6. When the photovoltaic cleaning robot 2 is suspended, the tension sensor 5.2 (PBCL-03) sends a force to the integrated control module 5.4 of the controllable hook 5 due to the gravity of the photovoltaic cleaning robot 2 and the force generated by the acceleration of the photovoltaic cleaning robot 2 when it is moving. After sending an electrical signal and receiving the electrical signal, the integrated control module 5.4 controls the push-pull electromagnet 5.7.1 to energize, the push-pull electromagnet 5.7.1 drives the lever 5.7.2 to rotate clockwise, and the push rod 5.7.4 moves down to drive the lock The clip 5.5 moves down, allowing the teeth at the bottom of the locking clip 5.5 to engage with the teeth of the hook 5.3, so as to prevent the suspended photovoltaic cleaning robot 2 from falling due to high winds or other conditions when flying.

The above-mentioned integrated control module 5.4 is connected with the drone control part (Pixhawk) of the drone 1 . When the drone 1 hangs the photovoltaic cleaning robot 2 to a fixed position and prepares to release the photovoltaic cleaning robot 2, the cloud platform 4 sends a signal to the drone control part of the drone 1, and the drone control part controls the integrated control module 5.4 to push and pull Type electromagnet 5.7.1 is powered off, and push rod 5.7.4 moves up, and spring 5.6 releases elastic potential energy and lock clip 5.5 is withdrawn again and presses against tightly, thereby makes controllable hook 5 bottom openings. Simultaneously, the UAV 1 moves to the lower left side opposite to the head of the controllable hook 5 to make the controllable hook 5 break away from the hanging ring 6, so that the delivery is successful. The schematic diagram of the hanging and delivery process is shown in Figure 7-8.

As shown in FIG. 9 , the photovoltaic cleaning robot 2 includes a robot body 2.1, a traveling mechanism 2.2, and a cleaning mechanism 2.3.

The body of the robot body 2.1 is provided with a control circuit board inside. The control circuit board is a PCB circuit board integrated with electronic components such as an integrated controller, a second signal transceiver module (Quectel BC28), and an ESC. The module (LM-2596S) is connected to the battery (GP12170). The second signal transceiving module is communicatively connected with the signal transceiving device of the cloud platform 4 .

The walking mechanism 2.2 is a crawler chassis, and the turning of the robot is realized by the speed difference of the crawlers on both sides in the process of advancing. The crawler design allows the robot to run stably on a photovoltaic panel with a certain inclination. A series of protrusions are also designed on the crawler, and rubber is installed on the protrusion to prevent slipping. The use of the crawler design can also prevent the robot from encountering The aluminum sides and gaps of the photovoltaic panels are sometimes toppled over.

As shown in Figure 11, the cleaning mechanism 2.3 includes a disc-shaped cleaning brush 2.3.1, a strip-shaped auxiliary cleaning brush 2.3.2, and a vacuum between the disc-shaped cleaning brush and the strip-shaped auxiliary cleaning brush. Apparatus 2.3.3. Disc type cleaning hairbrush 2.3.1 links to each other with the built-in brushless motor in the outrigger of robot body 2.1 front end, and brushless motor is connected to the control circuit board inside the nacelle by wiring.

Preferably, in order to adapt to a variety of working environments, a self-tightening quick-release buckle design is adopted on the connection between the disc-shaped cleaning brush 2.3.1 and the driving power.

Preferably, the disc-shaped cleaning brush 2.3.1 and the brush head of the elongated auxiliary cleaning brush 2.3.2 are selected from materials such as hair brush, sponge brush, cloth brush, etc., and adopt spiral, Shapes such as rings can be quickly converted according to actual needs, improving the work efficiency of the robot.

Preferably, as shown in FIG. 10 , a charging slot 2.4 is provided on the bottom surface of the robot body 2.1. The charging slot 2.4 is used for later insertion into the charging plug 3.5 for charging.

Preferably, the photovoltaic cleaning robot 2 mainly uses 3D printing and carbon fiber workpieces, which enhances the reliability of the robot while reducing the cost and weight; The possibility of damage due to impact between the photovoltaic cleaning robot 2 and the photovoltaic panel, and between the photovoltaic cleaning robot 2 and the energy storage charging bin 3.

As shown in Figure 9, when the photovoltaic cleaning robot is moving, the disc-shaped cleaning brush 2.3.1 on the left rotates clockwise under the drive of the brushless motor in the outrigger, and the disc-shaped cleaning brush 2.3. 1 Rotate counterclockwise, so that most of the dust is concentrated in the middle of the two brushes. At this time, the vacuum device 2.3.3 of the robot makes the air pressure difference between the storage bin and the outside world, so that the dust is sucked into the dust storage of the robot In the cabin, the dust collection device is connected to the front end of the cabin of the robot, and the dust storage cabin is installed inside the cabin close to the vacuum cleaner; only the two disc-shaped cleaning brushes 2.3. During the cleaning process, there may be situations where it is not cleaned. The long strip-shaped auxiliary cleaning brush 2.3.2 at the tail plays the role of "finishing" and cleans the dust that cannot be collected by the dust collection device 2.3.3.

When the photovoltaic cleaning robot 2 finishes cleaning along the path, it stops at the end point and sends a completion signal to the cloud platform 4. After receiving the signal, the cloud platform 4 orders the UAV 1 to start and fly away from the charging bin 3 , and carry out the inspection according to the inspection method and path at the beginning, collect the pictures of the cleaned photovoltaic panels and upload them to the cloud platform 4, the cloud platform 4 uses the open source computer vision library OpenCV for image recognition, and then uploads the recognition results to the background , where there are still stains are marked and displayed on the visualization panel in the background. The panel shows which area of the photovoltaic power station the stain belongs to and which photovoltaic panel it is located in. It is clear and clear. At the same time, the cloud Platform 4 will give an alarm prompt and wait for manual further processing; the staff can choose whether to transfer the robot back to cleaning according to the image results, or adopt manual cleaning methods (such as oil stains and other difficult-to-clean stains).

After completing all the operations, the photovoltaic cleaning robot 2 will stop at the end position and send a retraction signal to the cloud platform 4. After receiving the signal, the cloud platform 4 will order the drone 1 to transport the photovoltaic cleaning robot 2 to the energy storage for charging. In bin 3. (The transportation process is the same as before, so I won’t repeat it here).

As shown in Figure 12, the energy storage and charging bin 3 includes a bin body 3.1, which is closed in the form of two open end covers 3.2 to prevent dust, rainwater, etc. The robot 2 and the energy storage and charging bin 3 themselves caused damage.

As shown in Figures 13-14, the opening and closing of the end cover 3.2 is realized by the actuator 3.3 connected to the end cover 3.2 and the warehouse body 3.1. The actuator 3.3 includes a rack 3.3.1 fixed on the back of the end cover 3.2 , the rack 3.3.1 is placed up and down in the limit groove 3.3.2 of the frame of the warehouse body 3.1, the outer side of the rack 3.3.1 meshes with the gear 3.3.3, and the gear 3.3.3 is driven by the motor 3.3.4.

The executive mechanism 3.3 is installed at the four corners of the energy storage and charging bin 3; when it is necessary to manually change the battery for the drone 1, the end cover 3.2 can also be opened with the manual switch on the front of the energy storage charging bin; the surface of the end cover 3.2 is still Photovoltaic battery panels 3.8 are arranged, and the wiring of the photovoltaic battery panels is connected to the energy storage and charging bin 3 through the hollow rack 3.3.1, so that the self-power supply of the energy storage and charging bin 3 can be realized when the light is sufficient, and the power consumption of the system can be reduced cost.

After the drone puts the robot back into the charging energy storage bin, it will land in the preset landing area next to the nearest energy storage bin, and perform energy replenishment through manual battery replacement/wireless charging/wired charging.

The energy storage and charging compartment 3 includes a development board based on STM32 as the controller, and the opening and closing part covered with the solar photovoltaic panel uses the GW4058 DC motor as the power to drive this part, and the drive of the built-in cleaning module uses the GW4632 motor as the power. In terms of energy storage, LC-P12100ST lead-acid batteries are used for energy storage. The GW4058 and GW4632 motors are connected to the control board module of the charging compartment through the control line, and are interactively controlled through the display screen with touch function (based on the GT911 driver chip).

As shown in Figure 12, the energy storage and charging bin 3 is equipped with a brush head stain cleaning tank 3.4, a charging plug 3.5, and a pressure sensor 3.6. The brush head stain cleaning tank 3.4, the charging plug 3.5, and the pressure sensor 3.6 are arranged at the bottom of the charging compartment according to the body size of the photovoltaic cleaning robot 2.

When the drone 1 carries the photovoltaic cleaning robot 2 and returns to the top, it will automatically correct the orientation of the drone 1, so that the position of the photovoltaic cleaning robot 2 is aligned with the positions of the grooves in the energy storage and charging compartment 3. After alignment, The drone 1 will descend at a constant speed, and put the photovoltaic cleaning robot 2 into the energy storage charging bin 3. The charging plug 3.5 corresponds to the charging slot at the bottom of the photovoltaic cleaning robot 2, and adopts traditional plug-in charging.

As shown in Fig. 15, a shrinkable protective cover 3.5.1 is provided outside the charging plug 3.5 in the charging compartment to prevent problems such as poor contact caused by foreign objects. The electrode 3.5.2 of the charging column is provided with a shrinkable protective cover 3.5.1. The protective cover 3.5.1 is supported by a spring 3.5.3 and the baffle 3.5.4 is installed in the shell 3.5.5. The shrinkable protective cover 3.5.1 can shrink 3.5.2 to the bottom of the electrode. In this way, when the electrode 3.5.2 is inserted into the charging slot 2.4, the charging slot 2.4 presses down the shrinkable protective cover 3.5.1 to make it shrink to the bottom of the electrode 3.5.2, and the electrode 3.5.2 is inserted into the charging slot 2.4 alone. When the photovoltaic cleaning robot 2 is transferred away from the energy storage charging bin 3, the charging plug 3.5 will be separated from the charging slot 2.4, and the shrinkable protective cover 3.5.1 will pop up again to protect the electrode 3.5.2.

In the energy storage and charging compartment, there is also a stain cleaning tank 3.4, a track groove 3.7 with the same size as the track wheel of the photovoltaic cleaning robot 2, and a pressure sensor 3.6 is also installed in the groove. The stain cleaning tank 3.4 and the groove 3.7 are arranged according to the size of the photovoltaic cleaning robot. The charging plug 3.5 is inserted into the charging slot 2.4 (because the charging plug can be fully inserted into the charging slot only when all parts of the robot enter the groove), and the track groove 3.7 is also equipped with a pressure sensor 3.6. When the robot aligns with the stain cleaning tank 3.4 and the track groove 3.7, and parks smoothly in the energy storage and charging compartment 3, the pressure sensors 3.6 on both sides will feel the same pressure and will move to the robot. The cloud platform 4 sends a stop signal, and the cloud platform 4 then orders the UAV 1 to reset and stop. When the photovoltaic cleaning robot 2 is not accurately docked in the stain cleaning tank 3.4 and the groove 3.7, and the charging plug 3.5 is not fully inserted into the charging tank 2.4, the pressure sensors 3.6 on both sides are unbalanced, and the signal will not be sent out. 1 will readjust the position of the photovoltaic cleaning robot 2 so that the photovoltaic cleaning robot 2 can dock precisely. When the pressure of the pressure sensors 3.6 on both sides is the same, after 5 minutes, the energy storage and charging bin 3 will start the cleaning device in the stain cleaning tank 3.4, and the motors at both ends of the cleaning device will drive the cleaning brush to clean the brush head of the robot; Open the energy storage and charging bin 3 with the manual switch, and inspect the conditions inside the photovoltaic cleaning robot 2 and the energy storage charging bin 3 to see if the brush head needs to be replaced, so as to ensure the cleanliness of the brush head and the cleaning efficiency.

A photovoltaic cleaning method combined with a drone, comprising the following steps:

Step 1. According to the size of the photovoltaic power station and the input cost of photovoltaic cleaning, the photovoltaic power station can be divided into several areas, and each area is assigned a set of cleaning devices.

Step 2. During operation, the cloud platform 4 sends a start signal to the energy storage and charging bin 3 and the UAV 1 in each area according to the manually set time period. After the receiver of the energy storage and charging bin 3 receives the signal, it will store the energy The end cover of the charging compartment 3 is opened; at the same time, the drone 1 is powered on. The cloud platform 4 performs code matching with the receiver on the drone 1 by sending radio electromagnetic waves. After the code matching is successful, the drone control part (flight control) connected to the receiver will process the information sent by the cloud platform 4, and the flight control First, a preliminary self-inspection will be performed on UAV 1 to ensure the safety of take-off. After the self-inspection is completed, the flight controller outputs the specified power to the motor through the voltage regulation module (ESC), and the UAV 1 will lift vertically to the specified height and go to the predetermined starting point planned by the cloud platform 4.

Step 3. After arriving at the starting point, UAV 1 will align the head of the drone with the photovoltaic panels in the area according to the onboard sensor, and fly along the direction of the photovoltaic panels in the area at a speed of 0.5m/s. When it reaches the predetermined image planned by the cloud platform At the collection point, UAV 1 will collect images of photovoltaic panels through the machine vision camera attached below, and the collection results will be uploaded to cloud platform 4 through the receiver in real time; during this period, cloud platform 4 will monitor the UAV airborne The feedback data of the ultrasonic distance sensor ensures that the vertical distance between the flight inspection height of the UAV 1 and the photovoltaic panel is kept at about 3m, so as to take into account the flight safety of the UAV 1 and the integrity of the collected pictures.

Step 4: After the cloud platform 4 receives the image from the inspection of the UAV 1, it will use the open source vision library OpenCV to perform stain recognition. The recognition process is shown in Figure 16, and the recognition effect is shown in Figure 17-18.

Step 5, the cloud platform 4 will establish a coordinate system in a series of pixel points of the image, mark the identified pollutants according to the results of stain identification and record the coordinates of the pollutants in the coordinate system; complete the identification process Finally, the cloud platform 4 presents all the pollutant marker points in the entire photovoltaic panel area according to their coordinates, and classifies the adjacent pollutant marker points as a node, and plans out the The initial cleaning path of the photovoltaic cleaning robot.

Considering that there are gaps or protrusions between series and parallel photovoltaic panels that will affect the operation stability and cleaning efficiency of the photovoltaic cleaning robot during the cleaning process, the cloud platform will detect where the metal frame of the photovoltaic panel or the gap passes through the initial path. Re-plan and try to make the robot run according to the horizontal and vertical arrangement of photovoltaic panels.

Due to the barrel effect of photovoltaic panels when generating electricity, the surface of any solar cell or component in a series branch in the solar panel matrix is blocked by dust or foreign matter, and the blocked solar cell or component is not only If it cannot work normally, it will also be used as a load to consume the energy produced by other illuminated solar cell components. The energy generated will also cause the temperature of the solar cells or modules to continue to rise, which will eventually melt the solder joints, destroy the packaging material (if there is no bypass diode protection), and even cause the entire matrix to fail. Although the robot will not block a certain part of the photovoltaic panel for a long time during its travel, for the cleaning effect, the slower moving speed will more or less affect the power generation of the photovoltaic panel. Therefore, the cloud platform is planning the path In the process of cleaning, the long cleaning path along the series branch will be avoided, thereby reducing the impact on the power generation efficiency of the battery panel during the cleaning process; through the above steps, the cloud platform will complete the stain identification of the photovoltaic panel and the path of the cleaning robot planning.

Step 6: After the cloud platform 4 plans the cleaning path of the photovoltaic cleaning robot 2, the cloud platform 4 sends a return command to the drone 1, and when the flight controller of the drone 1 receives the command, it returns at a speed of 25m/s Energy storage and charging bin 3, and use the combined hook to move the photovoltaic cleaning robot 2 away from the charging bin, the UAV 1 hangs the robot vertically to 5m, and transports the photovoltaic cleaning robot 2 to the cloud platform 4 planning at a speed of 15m/s At the starting point of the path; after arriving at the starting point, the UAV 1 will point the head towards the starting direction of the cleaning path according to the onboard sensor, and descend slowly and uniformly at a speed of 0.8m/s, and put the photovoltaic cleaning robot 2 on the photovoltaic panel superior. After the delivery is completed, the drone returns to the energy storage and charging compartment 3 .

Step 7. When the photovoltaic cleaning robot 2 reaches the photovoltaic panel, it cleans along the path planned by the cloud platform 4 until the cleaning is completed. When the photovoltaic cleaning robot 2 finishes cleaning along the path, it stops at the end point and sends a completion signal to the cloud platform 4. After receiving the signal, the cloud platform 4 orders the drone 1 to start and fly away from the energy storage and charging bin 3. And carry out the inspection according to the inspection method and path at the beginning;

Step 8: The UAV 1 collects the picture of the cleaned photovoltaic panel and uploads it to the cloud platform 4. The cloud platform 4 uses the open source computer vision library OpenCV to perform image recognition, and then uploads the recognition result to the background, and places where there are still stains are marked It comes out and is displayed on the visualization panel in the background. The panel shows which area of the photovoltaic power station the stain belongs to and which photovoltaic panel it is located in, which is clear and clear.

At the same time, the cloud platform will give an alarm prompt and wait for further manual processing; the staff can choose whether to return the robot to cleaning according to the image results; or adopt manual cleaning methods (such as oil stains and other difficult-to-clean stains).

Step 9. After completing all the operations, the photovoltaic cleaning robot 2 will stop at the end position and send a retraction signal to the cloud platform 4. After receiving the signal, the cloud platform 4 will order the drone 1 to transport the photovoltaic cleaning robot 2 to the In the energy storage charging compartment. (The transportation process is the same as before, so I won’t repeat it here).

The photovoltaic cleaning system described in this application has high on-site adaptability. The management personnel of photovoltaic power plants can choose a photovoltaic cleaning robot different from this application according to the characteristics of stains and the estimated investment cost of photovoltaic cleaning. Unmanned aerial vehicles can still be used. , the function of this system can be realized by simply modifying the combined hook; at the same time, the brush head of the photovoltaic cleaning robot described in this application is detachable, so that the staff can choose the brush with the most reasonable shape and material according to the characteristics of the stain. head to improve cleaning efficiency;

With the cloud platform described in this application, users can set the number of outbound inspections, inspection time, and inspection range of the drone by themselves; when the photovoltaic panel generates power under sufficient light, the robot will Marching cleaning will definitely affect the efficiency of photovoltaic panel power generation. Therefore, this system can adopt fixed-point cleaning mode during the day. Carry out large-scale cleaning to improve the overall photovoltaic power generation efficiency. The staff can set the cleaning mode and cleaning cycle on the cloud platform according to the local light intensity distribution (such as fixed-point cleaning twice a day, comprehensive cleaning once a week). In summary, the system described in this application has high field adaptability.

The photovoltaic cleaning system described in this application has strong scalability and can not only meet the requirements of photovoltaic cleaning but also complete other tasks. For example, the cleaning robot can be directly installed on the six-rotor unmanned aerial vehicle. Nowadays, there are more and more factories integrating photovoltaic buildings, and the height of the buildings is generally high. There are safety and efficiency problems in the manual cleaning of the factory glass. Therefore, using this The device can easily clean the glass of the factory building. Combining the camera carried by the drone and the intelligent algorithm of the cloud platform, the drone will clean the glass according to the path planned by the cloud platform to improve cleaning efficiency.

Claims (9)

1. A photovoltaic cleaning method combined with an unmanned aerial vehicle, characterized in that: comprising the following steps: Step 1. Divide the photovoltaic power station into several areas according to the area of the photovoltaic power station and the input cost of photovoltaic cleaning, and assign a set of cleaning devices to each area; Step 2. During operation, the cloud platform (4) sends a start signal to the energy storage and charging bin (3) and the drone (1) in each area according to the manually set time period; the energy storage and charging bin (3) receives the signal Finally, the end cover (3.2) is opened; at the same time, the UAV (1) is powered on and goes to the scheduled departure point planned by the cloud platform (4); Step 3. After arriving at the starting point, the UAV (1) flies along the direction of photovoltaic panels in the area at a certain speed; when it reaches the scheduled image collection point planned by the cloud platform (4), the UAV (1) passes through The machine vision camera (1.3) collects images of photovoltaic panels, and the collected results will be uploaded to the cloud platform through the receiver in real time; The vertical distance between the flight inspection height of the man-machine and the photovoltaic panel is kept at a certain height, so as to take into account the flight safety of the drone and the integrity of the collected pictures; Step 4. After the cloud platform (4) receives the image from the inspection of the drone (1), it will use the open source vision library to identify stains; Step 5. The cloud platform (4) will establish a coordinate system in a series of pixel points of the image, mark the identified pollutants and record the coordinates of the pollutants in this coordinate system according to the results of stain identification; After the identification process, the cloud platform (4) presents all the pollutant marker points in the entire photovoltaic panel area according to their coordinates, and classifies the adjacent pollutant marker points as a node, through Dijkstra in graph theory Algorithm to plan the first-knowledge cleaning path of the photovoltaic cleaning robot (2); Step 6. After the cloud platform (4) plans the cleaning path of the photovoltaic cleaning robot (2), the cloud platform (4) sends a return command to the drone (1). When the flight controller of the drone (1) receives After the command, return to the energy storage charging bin (3) at a certain speed, and use the combination hook to transfer the photovoltaic cleaning robot (2) away from the energy storage charging bin (3); after arriving at the starting point, the drone (1) will clean the photovoltaic The robot is placed on the photovoltaic panel; after the delivery is completed, the drone (1) returns to the charging compartment; Step 7. When the photovoltaic cleaning robot (2) arrives at the photovoltaic panel, it cleans along the path planned by the cloud platform (4) until the cleaning is completed; Step 8. When the photovoltaic cleaning robot (2) finishes cleaning along the path, it stops at the end point and sends a completion signal to the cloud platform (4). After receiving the signal, the cloud platform (4) orders the drone (1 ) to start, fly away from the energy storage charging compartment (3), and conduct inspections according to the inspection method and path at the beginning; Step 9. The UAV (1) collects the picture of the cleaned photovoltaic panel and uploads it to the cloud platform (4). The cloud platform (4) uses the open source computer vision library for image recognition, and then uploads the recognition result to the background. The stained place is marked and displayed on the visualization panel in the background, which shows which area of the photovoltaic power station the stained spot belongs to and which photovoltaic panel it is located in; at the same time, the cloud platform (4) will carry out Alarm prompt, waiting for manual further processing; staff can choose whether to transfer the robot back to cleaning according to the image results; or take manual cleaning; Step 10. After completing all the operations, the photovoltaic cleaning robot (2) will stop at the end position and send a retraction signal to the cloud platform (4). After receiving the signal, the cloud platform (4) will command the drone (1 ) Transport the photovoltaic cleaning robot (2) to the energy storage charging bin (3). 2. A photovoltaic cleaning method combined with a drone according to claim 1, characterized in that: the cleaning device in the first step includes a drone (1), a photovoltaic cleaning robot (2), a storage Energy charging compartment (3) and cloud platform (4), wherein: The UAV (1) adopts the existing conventional unmanned aerial vehicle, and an ultrasonic distance sensor, a machine vision recognition camera (1.3), an infrared monitoring probe and a power monitoring module are additionally installed on the unmanned aerial vehicle; The photovoltaic cleaning robot (2) adopts a crawler-type robot body (2.1), the robot body (2.1) is equipped with a control circuit board, and the robot body (2.1) is equipped with a cleaning mechanism (2.3); The energy storage and charging bin (3) includes a bin body (3.1) and an end cover (3.2). The bin body (3.1) is provided with a charging plug (3.5), and the charging plug (3.5) is connected to the photovoltaic cleaning robot (2). The slot (2.4) is plugged in; the end cover (3.2) is set left and right and driven to open or close by the actuator (3.3); The cloud platform (4) communicates with the drone (1), the photovoltaic cleaning robot (2), and the energy storage charging bin (3) for unified management. 3. A photovoltaic cleaning method combined with a drone according to claim 2, characterized in that: the combined hook in the step 6 includes a controllable hook (5) installed at the bottom of the drone (1) ) and the hanging ring (6) installed on the top of the photovoltaic cleaning robot (2), the controllable hook (5) and the hanging ring (6) are multiple groups and correspond to each other. 4. A photovoltaic cleaning method combined with a drone according to claim 3, characterized in that: the controllable hook (5) includes a hook body (5.1), and the lower end of the hook body (5.1) passes through a tension sensor (5.2) The hook (5.3) is installed, and the integrated control module (5.4) is installed on the top of the hook body (5.1). 5.5) Under normal conditions, press upward through the spring (5.6) and form an opening with the hook (5.3). After the photovoltaic cleaning robot (2) is hung, the locking clip (5.5) moves down through the telescopic mechanism (5.7) and connects with the hook (5.3). The hooks form a closed loop structure. 5. A photovoltaic cleaning method combined with a drone according to claim 4, characterized in that: the telescopic mechanism (5.7) includes a push-pull electromagnet (5.7.1), a push-pull electromagnet (5.7 .1) The driving part is installed in the waist hole of the lever (5.7.2) through the pin shaft; one end of the lever (5.7.2) is hinged with the bracket (5.7.3), and the other end is hinged with the push rod (5.7.4) , the push rod (5.7.4) is slidably set on one side of the bracket (5.7.3) through the chute and the lower end is in contact with the locking clip (5.5). 6. A photovoltaic cleaning method combined with drones according to claim 2, characterized in that: the cleaning mechanism (2.3) includes a disc-shaped cleaning brush (2.3.1), a strip-shaped auxiliary Cleaning brushes (2.3.2) and dust collection devices (2.3.3); the disc-shaped cleaning brushes (2.3.1) are arranged at left and right intervals at the front end of the robot body (2.1) and are driven by motors; long strip-shaped auxiliary cleaning brushes (2.3.2) Arranged at the tail of the robot body (2.1); the dust suction device (2.3.3) is located between the disc-shaped cleaning brush (2.3.1) and the strip-shaped auxiliary cleaning brush (2.3.2) and the opening Towards the disc-shaped cleaning brush (2.3.1). 7. A photovoltaic cleaning method combined with a drone according to claim 2, characterized in that: the actuator (3.3) includes a rack (3.3.1) fixed on the back of the end cover (3.2) , the rack (3.3.1) is placed up and down in the limit groove (3.3.2) of the frame of the bin body (3.1), the outer side of the rack (3.3.1) meshes with the gear (3.3.3), and the gear (3.3.3 ) driven by a motor (3.3.4). 8. A photovoltaic cleaning method combined with an unmanned aerial vehicle according to claim 2, characterized in that: brush head stain cleaning tanks (3.4), track grooves (3.7) are installed in the warehouse body (3.1) ), the brush head stain cleaning tank (3.4), track groove (3.7) correspond to the walking mechanism (2.2) and cleaning mechanism (2.3) of the photovoltaic cleaning robot (2); the brush head stain cleaning tank (3.4) A cleaning device is arranged inside, and a pressure sensor (3.6) is arranged inside the track groove (3.7). 9. A photovoltaic cleaning method combined with a drone according to claim 2, characterized in that: a shrinkable protective cover (3.5.1) is provided on the outside of the charging plug (3.5), and the shrinkable protective cover (3.5.1 ) is fixed on the baffle (3.5.4), and the lower end of the baffle (3.5.4) is installed in the shell (3.5.5) through the spring (3.5.3).

Images