A smart dual-drive photovoltaic module cleaning system

Abstract

This application belongs to the field of photovoltaic power plant technology, specifically an intelligent dual-drive photovoltaic module cleaning system, including: a photovoltaic module array, cleaning equipment, drive equipment, a wireless gateway, and a host computer; wherein, the cleaning equipment is placed vertically on the photovoltaic module array and can move on the photovoltaic modules; the drive equipment includes two identical sets of upper drive equipment and lower drive equipment, wherein the upper drive equipment is connected to the upper end of the cleaning equipment, and the lower drive equipment is connected to the lower end of the cleaning equipment; the wireless gateway remotely transmits the data of the drive equipment to the host computer, and the host computer uses a software system to display the operating status and operating errors of the drive equipment, and to control the operating mode of the drive equipment, so as to synchronize the upper and lower drive equipment, making module cleaning faster, more convenient, and less costly, and improving the power generation of the photovoltaic system.

Description

Technical Field

[0001] This application belongs to the field of photovoltaic power station technology, specifically a smart dual-drive photovoltaic module cleaning system. Background Technology

[0002] Currently, the main cleaning methods used in photovoltaic power stations include manual cleaning, cleaning truck cleaning, automatic cleaning systems, and dry cleaning. Manual cleaning typically uses soft brooms, mops, and cleaning cloths with clean water. High-pressure water guns may also be used, but care must be taken to control the water pressure to avoid damaging the photovoltaic modules. For stubborn stains, neutral detergents may be used. Manual cleaning is simple to operate, requires no special equipment, and is effective for cleaning localized stains and small power stations. However, it is labor-intensive, has low cleaning efficiency, and the cleaning effect heavily depends on the conscientiousness of the construction workers. Cleaning truck cleaning is divided into waterless cleaning and water-based cleaning. It is suitable for ground-mounted power stations that are relatively regular, not concentrated, and have large gaps between strings. Its advantage is high single-cycle cleaning efficiency, but it is not suitable if the distance between rows of strings is not wide enough for the cleaning truck to move. Automated cleaning systems mainly include roller brush cleaning, spray cleaning, and robotic cleaning. These systems can clean automatically according to preset programs without manual intervention, offering high cleaning efficiency, timed cleaning, and maintaining the cleanliness of the module surfaces. Standardized operation reduces the risk of damage to the modules. However, these systems are expensive, require significant investment, and necessitate regular maintenance. For modules with unusual shapes or installation locations, the cleaning effect may be less than ideal. Dry cleaning primarily uses specialized dry cleaning cloths or brushes, typically made of soft materials that won't scratch the photovoltaic module surface. Dry cleaning requires no water, making it suitable for water-scarce areas or water-sensitive environments. It doesn't generate wastewater, making it environmentally friendly, and is simple and convenient to operate. However, it may not be as effective as water washing for removing sticky stains, and dry cleaning cloths or brushes need to be replaced regularly, increasing costs. Therefore, photovoltaic power station cleaning faces numerous challenges, including low cleaning efficiency, high economic costs, significant threats to personnel safety, and poor adaptability to automated equipment. Summary of the Invention

[0003] The technical problem to be solved by this application is to provide an intelligent dual-drive photovoltaic module cleaning system, which solves the problems of low efficiency, high cost, high safety risks, and limitations of traditional cleaning modes, as well as the limitations of cleaning technology and equipment.

[0004] This application is implemented as follows: an intelligent dual-drive photovoltaic module cleaning system includes: a photovoltaic module array, a cleaning device, a drive device, a wireless gateway, and a host computer; wherein, The cleaning equipment is placed vertically on the photovoltaic module array and can move on the photovoltaic modules; The drive device includes two identical upper drive device and lower drive device, wherein the upper drive device is connected to the upper end of the cleaning device and the lower drive device is connected to the lower end of the cleaning device. The wireless gateway remotely transmits the data from the drive device to the host computer. The host computer uses software to display the operating status and errors of the drive device, and to control the operation mode of the drive device, so that the upper and lower drive devices can be synchronized.

[0005] Furthermore, the host computer controls the drive equipment to perform cleaning according to a predetermined time and a predetermined walking speed, and to shut down and return according to a predetermined time; The system acquires the running speed, running time, and position of the upper and lower ends of the cleaning device from the drive equipment, determines the actual running error of the upper and lower ends of the cleaning device, predicts the running status of the cleaning device at the next moment, and adjusts the running speed in real time to achieve real-time adjustment of the error.

[0006] Furthermore, the host computer determines the actual operating error at the upper and lower ends of the cleaning equipment, including: Let there be a first threshold s, a second threshold s1, and a third threshold s2, where s2>s1. Let the cleaning operation time be t. Obtain the actual distance z1(t) between the upper end of the cleaning device and the starting point, and the actual distance z2(t) between the lower end of the cleaning device and the starting point. The running speed v1 of the upper drive device and the running speed v2 of the lower drive device are also obtained. The expected running position of the cleaning equipment is calculated as zz1(t) = t * v1; The expected running position of the lower end of the cleaning equipment is calculated as zz2(t) = t * v2; Determine if If the actual operating position of the cleaning equipment is normal, then it is normal; otherwise, it is abnormal. Determine if If the actual operating position of the lower end of the cleaning equipment is normal, then it is normal; otherwise, it is abnormal. Determine if If so, the actual operating error at the top and bottom of the cleaning equipment is within the normal range.

[0007] Furthermore, when the host computer determines that the actual operating error at the upper and lower ends of the cleaning equipment exceeds the normal range, it performs fault handling on the drive equipment, including: when the actual operating error at the upper and lower ends of the cleaning equipment exceeds the specified threshold s1, the operating speed needs to be fine-tuned as follows: If s2>z1(t)-z2(t)>s1, according to z1(t)-z2(t)=t*(v1-v2), the running speed v1 of the upper drive device is adjusted so that v1=v2+(z1(t)-z2(t)) / t; If s2>z2(t)-z1(t)>s1, according to z2(t)-z1(t)=t*(v2-v1), the running speed v2 of the drive device is adjusted so that v2=v1+(z2(t)-z1(t)) / t; When a serious malfunction occurs, that is, when the actual operating error at the upper and lower ends of the cleaning equipment is very large and exceeds the specified threshold s2, i.e.: if If necessary, the drive device should be shut down.

[0008] Furthermore, predicting the operating status of the cleaning equipment at the next moment and adjusting the operating speed in real time enables real-time error correction. This includes using an LSTM network to predict faults in the operating status of the cleaning equipment, including two types of fault prediction: abnormal cleaning speed and excessive tilt angle. Specifically, this includes: Obtaining input and output samples: For time t, the input samples include the actual distance moved above the cleaning device. Actual running speed The actual distance the lower end of the cleaning equipment moves Actual running speed ;set up Given the input at time t, construct the input vector: ; The output vector corresponding to the input vector is the tilt angle of the cleaning device. tilt angle The calculation formula is: , ,in, The vertical length of the cleaning equipment. It is an intermediate variable. It refers to the tilt angle of the cleaning equipment; Determine the number of hidden nodes in the LSTM network and the dropout rate of the dropout layer, assuming: Output for the current time t. The output from the previous moment. : The cell state at the previous moment; Computing the forget gate; ,in It's the sigmoid function, which outputs a value between 0 and 1, where 1 indicates that everything is retained and 0 indicates that everything is forgotten. The weight matrix representing the forget gate, This is the bias vector for the forget gate; Calculate the input gate: First calculate the updated weights. ,in Let be the weight matrix of the input gate. The bias vector is the input gate; then the candidate new memory is calculated: , Output a value between -1 and 1, where, The weight matrix for the new memory. This is the bias vector for the new memory. For new memories; Update cell state based on weights, forget gate, and new memories: ; Calculate the output gate: First, calculate the output weights: ,in Let the weight matrix be the output gate. This is the bias vector for the output gate. The previous output is used as the basis for calculation; then the current output is calculated. , for matrix operations, This indicates element-wise multiplication. The predicted output vector is then the calculated current output. ; Adjust the operating speed of the upper and lower drive devices based on the predicted output vector.

[0009] Furthermore, the driving device is a rope pulling system, a low-altitude unmanned aerial vehicle system, or a walking robot system.

[0010] Furthermore, the rope-type pulling system includes a bracket, rotating wheels, traction equipment, a motor, and a first control system. The bracket is located on both sides of the photovoltaic module array. The two ends of the traction equipment are respectively wound into the rotating wheels on both sides and connected to the cleaning equipment in the middle. The motor is connected to the rotating wheels to drive the rotating wheels to rotate. The traction equipment is a rope-type traction, chain-type traction, or track-type traction. The first control system is used to control the start and stop of the motor to realize the rotation and stop of the rotating wheel, record the rotation running time, the moving position of the cleaning equipment, and the speed of the upper and lower ends of the cleaning equipment, and transmit them to the host computer, and adjust the motor speed according to the instructions of the host computer.

[0011] Furthermore, the walking robot system includes a walking robot vehicle, a sliding support, pulleys, a robotic arm column, a translation motor, a rotation motor, a docking mechanism, a second control system, a first image recognition system, and a first high-definition camera. The sliding support is a support with a sliding track and is installed inside the walking robot vehicle. The lower end of the robotic arm column has pulleys, which move on the track driven by the translation motor. The upper end of the robotic arm column is connected to the cleaning equipment to propel the cleaning equipment to move on the photovoltaic module array for cleaning. The robotic arm column is used to propel the cleaning equipment by means of the walking vehicle. The translation motor controls the robotic arm column to move up and down and slide left and right. The rotation motor and the robotic arm... The upper end of the column is connected to a docking mechanism that connects to a rotating motor. During cleaning, the rotating motor docks with the rotating shaft on the cleaning equipment, thereby driving the rotating shaft on the cleaning equipment to rotate. The force of the walking robot vehicle propels the cleaning brush to move on the photovoltaic module array for cleaning. A high-definition camera is used to capture images of the docking status between the robotic arm column and the cleaning equipment. The first high-definition camera captures images of the docking status between the robotic arm column and the cleaning equipment and transmits them to the first image recognition system. The first image recognition system transmits the recognition results to the control system. The first image recognition system is used to identify the center position of the rotating shaft on the cleaning brush, so that the rotating motor docks with the rotating shaft on the cleaning brush.

[0012] Furthermore, the second control system is used to control the operation mode of the translation motor and the rotary motor, to control the start and stop of the translation motor and the rotary motor, to record the movement position of the cleaning equipment, and to encode and calculate the results and send them to the host computer.

[0013] Furthermore, the low-altitude unmanned aerial vehicle (UAV) system includes the UAV body, UAV control system, support column, rotating motor, docking mechanism, second high-definition camera, and second image recognition system. The support column is composed of a support and a column fixed together, and can be folded and placed inside the UAV body. When opened, the column stands below the UAV and contacts the cleaning equipment. It is used to propel the cleaning brush on the photovoltaic module array for cleaning by using the forward flight force of the UAV. The rotating motor is placed inside the lower end of the column, and the docking mechanism is connected to the rotating motor, connecting the rotating motor to the rotating shaft on the cleaning equipment to drive the rotating shaft on the cleaning equipment to rotate. The second high-definition camera is used to capture images of the docking status between the robotic arm column and the cleaning equipment, and transmits them to the second image recognition system. The second image recognition system transmits the recognition results to the UAV control system. The second image recognition system is used to identify the center position of the rotating shaft on the cleaning brush, so that the rotating motor docks with the rotating shaft on the cleaning brush. The UAV control system is used to control the flight direction of the UAV body, control the start and stop of the rotating motor, record the movement position of the cleaning equipment and the speed of the upper and lower ends of the cleaning equipment and send them to the host computer, and adjust the operating speed of the low-altitude UAV system according to the instructions of the host computer.

[0014] Furthermore, the cleaning device is a cleaning brush structure with a planetary gear system, including a planetary gear system installed on one or both sides of the rotating shaft. The planetary gear system is used to rotate the rotating shaft when the traction device moves. The rotating shaft passes through the roller brush, and the two ends of the rotating shaft are sleeved and fixed to the roller brush wheel. The planetary gear system meshes with the roller brush wheel. The planetary gear system includes three small gears. The roller brush wheel is located in the middle position of the three small gears. The outer teeth of the three small gears mesh with the track wheel. The outer teeth of the track wheel mesh and travel on the track. The center of the small gear is connected by a tripod, and the tripod is fixed on the rotating shaft. One end of the traction device is connected to the rotating shaft.

[0015] Compared with the prior art, the advantages of this application are as follows: The advantages of this cleaning system are faster cleaning speed, lighter weight, and lower cost. Existing photovoltaic module cleaning systems generally fall into four categories: fully automatic horizontal cleaning robots on the module, crawling fully automatic cleaning robots on the module, walking cleaning robots with roller brushes, and drone-mounted cleaning robots with roller brushes. The disadvantages of fully automatic horizontal cleaning robots on the module are their excessive weight, which can easily damage the photovoltaic modules, and especially their high cost. Crawling fully automatic cleaning robots on the module have disadvantages such as slow cleaning speed, frequent charging requirements, and limited module tilt angles. Walking cleaning robots with roller brushes are too bulky and costly. Drone-mounted cleaning robots with roller brushes have disadvantages such as slow cleaning speed, frequent charging requirements, and high power consumption. Therefore, this intelligent dual-drive photovoltaic module cleaning system overcomes the above disadvantages, making module cleaning faster, more convenient, and lower in cost, thereby increasing the power generation of the photovoltaic system. Attached Figure Description

[0016] Figure 1 A structural block diagram of an intelligent dual-drive photovoltaic module cleaning system provided in this application embodiment; Figure 2 This is a schematic diagram of the rope pulling system provided in the embodiments of this application; Figure 3 This is a schematic diagram of the structure of the walking robot system provided in the embodiments of this application; Figure 4 This is a schematic diagram of the structure of a low-altitude unmanned aerial vehicle system provided in an embodiment of this application; Figure 5 This is a schematic diagram of the tilt angle of the cleaning equipment provided in the embodiments of this application; Figure 6This is a schematic diagram of the cleaning device provided in the embodiments of this application; Figure 7 This is a schematic diagram of a cleaning device with a planetary gear system provided in an embodiment of this application. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0018] See Figure 1 The illustrated intelligent dual-drive photovoltaic module cleaning system includes a photovoltaic module array, cleaning equipment, drive equipment, power supply system, wireless gateway, and host computer. The photovoltaic module array consists of multiple photovoltaic modules. The cleaning equipment is placed vertically on the photovoltaic module array and can move on the photovoltaic modules for cleaning. The dual-drive equipment consists of two identical drive units, one connected to the upper end of the cleaning equipment (called the upper drive unit) and the other connected to the lower end (called the lower drive unit). The wireless gateway remotely transmits data from the dual-drive equipment to the host computer. The host computer uses software to display the operating status and errors of the dual-drive equipment and to control its operation, ensuring synchronization between the upper and lower drive units. One implementation of the power supply system consists of photovoltaic modules and a battery; the photovoltaic modules charge the battery, and the battery powers the motor.

[0019] The driving device can be a rope pulling system, a low-altitude drone system, or a walking robot system.

[0020] In one embodiment, see Figure 6 The cleaning device shown includes two side supports 61, pulleys 64 located below the supports 61, and a rotating shaft 62 passing through the supports 61. A brush 63 is sleeved on the rotating shaft 62, and the rotating shaft 62 rotates under the drive of the up and down driving device.

[0021] like Figure 2As shown, the rope-driven pulling system comprises a bracket 11, rotating wheels 12, a traction device 13, a motor 15, and a first control system 16. The bracket 11 is located on both sides of the photovoltaic module array 14. The two ends of the traction device 13 are respectively wound into the rotating wheels 12 on both sides, and the middle is connected to the cleaning device. The motor 15 is connected to the rotating wheels 12 to drive their rotation. The power supply system is connected to the motor to supply power. The first control system includes control software and a control device. The control device is connected to the motor and includes a first data acquisition module, a first control module, and a first communication module. The first data acquisition module sends the acquired data to the first control module. The first control module encodes and calculates the data and sends the results to the communication module, which then transmits them to the wireless gateway. The control software controls the motor's start and stop to achieve the rotation and stop of the rotating wheels, records the rotation time, the moving position of the cleaning device, and the speed of the upper and lower ends of the cleaning device. The bracket is located on both sides of the photovoltaic module array, and the rotating wheels 12, motor 15, power supply system, and first control device are all placed on the bracket. The traction device 13 can be a rope traction, chain traction, or track traction.

[0022] See Figure 7 As shown, another type of cleaning device with a planetary gear system is provided, including: a planetary gear system installed on one or both sides of a rotating shaft 78. The planetary gear system is used to rotate the rotating shaft when the traction device 13 moves. The rotating shaft passes through a roller brush 75 and is fixed to the roller brush. The two ends of the rotating shaft are sleeved and fixed to roller brush wheels 76. The planetary gear system meshes with the roller brush wheels 76. The planetary gear system includes three pinions. The roller brush wheel 76 is located in the middle position of the three pinions. The outer sides of the three pinions mesh with a track wheel 74. The outer sides of the track wheel mesh and travel on a track 71. The center of the three pinions is connected to a tripod, which is connected to the rotating shaft. One end of the traction device is connected to the rotating shaft 78 after being fixed in direction by an auxiliary wheel 73. When the traction device pulls the rotating shaft forward, the rotating shaft, being fixed to the roller brush, drives the roller brush forward. At this time, the track wheel meshes with the track gear, and the track wheel rotates. The rotation of the track wheel drives the three small gears of the planetary gear system to rotate, and the three small gears drive the roller brush wheel to rotate. Finally, the roller brush wheel drives the roller brush connected to the rotating wheel to rotate, thus realizing the forward and rotating motion of the roller brush.

[0023] One implementation structure of the first data acquisition module includes a first distance acquisition module, a first time acquisition module, and a first running speed acquisition module. The first distance acquisition module is used to measure the actual distance between the cleaning device and the starting point. The first time acquisition module is used to measure the start-up time and running time of the driving device. The first running speed acquisition module is used to measure the running speed of the driving device. Another way to implement a rope-type traction system is to replace the traction device with a chain structure made up of short steel segments connected together.

[0024] Another way to implement a rope-type pulling system is to loop a rope or a chain through the rotating wheels at both ends to form a closed loop. The motor drives the rotating wheels to rotate, so that the cleaning equipment cleans the photovoltaic modules.

[0025] Another way to implement a rope-driven system is to share a control system between the upper and lower drive devices. This system can be used to directly control the operating error of the dual drive motors, enabling the upper and lower drive devices to synchronize.

[0026] In one embodiment, such as Figure 3 As shown, the walking robot system comprises a walking robot vehicle 22, a sliding bracket 23, pulleys, a robotic arm column 24, a translation motor 21, a first rotary motor 25, a first docking mechanism 26, a second control system, a first high-definition camera, a first image recognition system, and a rotary motor. The sliding bracket is a bracket with a sliding track, placed inside the walking robot vehicle. The lower end of the robotic arm column has pulleys, which move on the sliding track driven by the translation motor. The upper end contacts the cleaning equipment, which is used to push the cleaning equipment to move on the photovoltaic module array for cleaning. The robotic arm column is used to push the cleaning equipment to move by means of the walking vehicle. The translation motor is used to control the robotic arm column to move up and down and slide left and right (in the direction perpendicular to the walking vehicle).

[0027] A rotating motor is connected to the upper end of the robotic arm column 24, and a first docking mechanism 26 is connected to the first rotating motor 25. During cleaning, the rotating motor docks with the rotating shaft on the cleaning device, thereby driving the rotating shaft to rotate. The force of the walking robot propels the cleaning brush across the photovoltaic module array for cleaning. The docking mechanism connects the upper end of the robotic arm column to the cleaning device. A first high-definition camera captures images of the docking status between the robotic arm column and the cleaning device and transmits them to a first image recognition system. The first image recognition system then transmits the recognition results to a second control system. The first image recognition system identifies the center position of the rotating shaft on the cleaning brush, enabling the first rotating motor to dock with the rotating shaft on the cleaning brush.

[0028] The first docking mechanism 26 includes magnets. Magnets are placed on the front end of the first rotating motor of the robotic arm column and the front end of the rotating shaft of the cleaning device, respectively, so that the rotating motor of the robotic arm column and the rotating shaft of the cleaning device are docked by the magnetic attraction. The rotating motor of the robotic arm column and the rotating shaft of the cleaning device are locked together by a locking structure. This makes it easier for the walking robot system and the cleaning device to dock.

[0029] The second control system controls the operation of the translation motor and the first rotary motor, including starting and stopping them, and recording the movement position of the cleaning equipment. It includes a second data acquisition module and a third control module. The second data acquisition module sends the acquired data to the third control module, which encodes and calculates the data before sending the results to the host computer. One implementation structure of the second data acquisition module includes a second distance acquisition module, a second time acquisition module, and a second running speed acquisition module. The second distance acquisition module is used to measure the actual distance between the cleaning device and the starting point. The second time acquisition module is used to measure the start-up time and running time of the driving device. The second running speed acquisition module is used to measure the running speed of the driving device. One implementation of the first docking mechanism is a snap-fit ​​method, including retaining rings and legs installed at both ends of the cleaning equipment, retaining rings and legs installed at the upper end of the robotic arm column, and a rotating bolt. When the cleaning equipment docks with the robotic arm column, the first rotary motor automatically rotates the rotating bolt, connecting the two parts together. When it is necessary to separate the cleaning equipment from the robotic arm column, the first rotary motor automatically rotates the rotating bolt in the opposite direction, thus automatically separating the two parts. Therefore, automatic docking, automatic rotation, and automatic separation are achieved between the first rotary motor and the rotating shaft.

[0030] like Figure 4 As shown, the low-altitude unmanned aerial vehicle (UAV) system consists of a UAV body 31, a UAV control system, a support column 32, a second rotating motor 33, a second docking mechanism 34, a second high-definition camera, and a second image recognition system. The support column is composed of a support frame and a column fixed together. It can be folded and placed inside the UAV body. When opened, the column stands below the UAV and can contact the cleaning brush. The UAV's forward flight propels the cleaning brush to move and clean the photovoltaic module array. The rotating motor is located inside the lower end of the column. The docking mechanism connects to the rotating motor, connecting it to the rotating shaft on the cleaning device, thereby driving the rotating shaft on the cleaning device to rotate.

[0031] The second docking mechanism 34 includes magnets. Magnets are placed on the front end of the rotating motor of the support column and the front end of the rotating shaft of the cleaning equipment, respectively, so that the rotating motor of the support column and the rotating shaft of the cleaning equipment are docked by the magnetic attraction. A locking structure is used to lock the rotating motor of the support column and the rotating shaft of the cleaning equipment together. This makes it easier for low-altitude UAV systems to dock with cleaning equipment. The drone control system controls the drone's flight direction, the start and stop of the rotating motors, and records the movement of the cleaning equipment. The second control device includes a third data acquisition module, a third control module, and a third communication module. The third data acquisition module sends the acquired data to the third control module, which encodes and calculates the data, sends the results to the communication module, then to the wireless gateway, and finally uploads it to the host computer.

[0032] One implementation structure of the third data acquisition module includes a second distance acquisition module, a second time acquisition module, and a second running speed acquisition module. The second distance acquisition module is used to measure the actual distance between the cleaning device and the starting point. The second time acquisition module is used to measure the start-up time and running time of the driving device. The second running speed acquisition module is used to measure the running speed of the driving device. The second high-definition camera captures images of the docking status between the support column and the cleaning equipment, transmitting these images to the image recognition system. The second image recognition system then performs image recognition and analysis, sending the results to the drone control system to direct the drone's movement, thereby achieving docking between the rotating motor and the cleaning equipment. The image recognition system identifies the center position of the rotating shaft on the cleaning brush, ensuring the rotating motor aligns with the rotating shaft on the cleaning brush.

[0033] The main function of the host computer is to enable the cleaning system to clean according to a predetermined time and a predetermined walking speed, as well as to shut down and return according to a predetermined time; to detect cleaning system faults in real time, to make timely fine adjustments to the operating errors of the upper and lower drive devices, and to shut down the cleaning system in time for serious faults. Specific implementation details and steps: S1 acquires actual operating parameter data through the data acquisition module. Let there be a first threshold s, a second threshold s1, and a third threshold s2, where s2 > s1. The specific values ​​of s, s1, and s2 are related to the actual scenario and are directly proportional to the length of the cleaning device. Let the cleaning operation time be t. The actual distance z1(t) between the upper end of the cleaning device and the starting point and the actual distance z2(t) between the lower end of the cleaning device and the starting point are obtained through the distance acquisition module. The running speed v1 of the upper drive device and the running speed v2 of the lower drive device are also obtained. S2 Fault Detection Fault detection requires determining whether the actual operating position of the cleaning equipment is correct and whether the actual operating errors of the upper and lower drive devices are within the threshold range. S21 determines whether the actual operating position of the cleaning equipment is normal: The expected running position of the cleaning equipment is zz1(t) = t * v1; The expected operating position of the lower end of the cleaning equipment is zz2(t)=t*v2; if If so, the actual operating position of the upper part of the cleaning equipment is normal; if If so, the actual operating position of the lower end of the cleaning equipment is normal; Calculation of actual operating error at the upper and lower ends of the S22 sweeping equipment: if If so, the actual operating error at the upper and lower ends of the cleaning equipment is within the normal range; S3 troubleshooting S31 If the actual operating error at the upper and lower ends of the cleaning equipment exceeds the specified threshold s1, the operating speed needs to be fine-tuned as follows: If s2>z1(t)-z2(t)>s1, according to z1(t)-z2(t)=t*(v1-v2), the running speed v1 of the upper drive device is adjusted so that v1=v2+(z1(t)-z2(t)) / t; If s2>z2(t)-z1(t)>s1, according to z2(t)-z1(t)=t*(v2-v1), the running speed v2 of the drive device is adjusted so that v2=v1+(z2(t)-z1(t)) / t; S32 When a serious fault occurs, that is, when the actual operating error at the upper and lower ends of the cleaning equipment is very large and exceeds the specified threshold s2, i.e.: if If necessary, the drive equipment should be shut down immediately. S4 Fault Prediction The actual operating error of the upper and lower ends of the cleaning equipment needs to be within the normal range. In order to reduce the operating error of the dual-drive equipment, the prediction based on LSTM neural network deep learning can predict the operating state of the cleaning equipment at the next moment based on the network input, the current state and the memory information. Thus, by adjusting the operating speed in real time, the error can be adjusted in real time to avoid the cleaning equipment tilting too much.

[0034] Using LSTM networks to predict faults in the operation of cleaning equipment, including two types of faults: abnormal cleaning speed and excessive tilt angle.

[0035] The calculation steps are as follows: S41 acquires input and output samples First, collect a sufficient number of input and output samples. For time t, the input samples include the actual moving distance of the upper part of the cleaning device. running speed The actual distance the lower end of the cleaning equipment moves running speed ; set up Given the input at time t, construct the input vector: , The output vector corresponding to the above input vector is the tilt angle of the cleaning device. tilt angle The calculation formula is: , , in, The vertical length of the cleaning equipment. It is an intermediate variable. This refers to the tilt angle of the cleaning equipment. Tilting angle See Figure 5 The image shows the angle between a normal cleaning device and a tilted cleaning device.

[0036] Determine the number of hidden nodes in the LSTM network and the dropout rate of the dropout layer.

[0037] set up: Output for the current time t. The output from the previous moment (in a hidden state). The cell state at the previous moment (memory state). S42 calculates the forget gate: ,in It is the sigmoid function, which outputs a value between 0 and 1, where 1 means to retain everything and 0 means to forget everything. The weight matrix representing the forget gate, This is the bias vector for the forget gate.

[0038] S43 Calculates the input gate: First calculates and updates the weights. ,in Let be the weight matrix of the input gate. is the bias vector of the input gate.

[0039] Recalculate candidate new memories: , Output a value between -1 and 1, where The weight matrix for the new memory. This is the bias vector for the new memory. For new memories.

[0040] S44 updates the cell state based on weights, the forgetting gate, and new memories: ; S45 calculates the output gate: First, calculate the output weights: ,in Let the weight matrix be the output gate. This is the bias vector for the output gate. This is the previous output.

[0041] Recalculate the current output: , for matrix operations, This represents element-wise multiplication. The predicted output vector is then the current output calculated above. .

[0042] The LSTM network is trained using the Adams optimization algorithm, with the optimization objective being to minimize the following loss function, which is described below: ,in, For the predicted current moment, and These represent the cleaning equipment in the first stage. The actual observed and predicted values ​​of the tilt angle at any given time.

[0043] For the sigmoid logical function: , For hyperbolic tangent function: .

[0044] The above process is actually calculated using the MATLAB software LSTM network toolbox. The input and output samples are divided into two parts: weight training samples and validation samples. The number of hidden nodes in the LSTM unit structure is determined. First, the weight training samples are input into the LSTM network to obtain the weights that minimize the loss function. Then, the accuracy of the network calculation is verified by using validation samples. The results are verified, and this application can provide precise control for the cleaning system.

[0045] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. An intelligent dual-drive photovoltaic module cleaning system, characterized in that, include: Photovoltaic module arrays, cleaning equipment, drive equipment, wireless gateways, and host computers; among them, The cleaning equipment is placed vertically on the photovoltaic module array and can move on the photovoltaic modules; The drive device includes two identical upper drive device and lower drive device, wherein the upper drive device is connected to the upper end of the cleaning device and the lower drive device is connected to the lower end of the cleaning device. The wireless gateway remotely transmits the data from the drive device to the host computer. The host computer uses software to display the operating status and errors of the drive device, and to control the operation mode of the drive device, so that the upper and lower drive devices can be synchronized.

2. The intelligent dual-drive photovoltaic module cleaning system according to claim 1, characterized in that, The host computer controls the drive equipment to perform cleaning according to a predetermined time and a predetermined walking speed, and to shut down and return according to a predetermined time. The system acquires the running speed, running time, and position of the upper and lower ends of the cleaning device from the drive equipment, determines the actual running error of the upper and lower ends of the cleaning device, predicts the running status of the cleaning device at the next moment, and adjusts the running speed in real time to achieve real-time adjustment of the error.

3. The intelligent dual-drive photovoltaic module cleaning system according to claim 2, characterized in that, The host computer determines the actual operating error at the upper and lower ends of the cleaning equipment, including: Let there be a first threshold s, a second threshold s1, and a third threshold s2, where s2>s1. Let the cleaning operation time be t. Obtain the actual distance z1(t) between the upper end of the cleaning device and the starting point, and the actual distance z2(t) between the lower end of the cleaning device and the starting point. The running speed v1 of the upper drive device and the running speed v2 of the lower drive device are also obtained. The expected running position of the cleaning equipment is calculated as zz1(t) = t * v1; The expected running position of the lower end of the cleaning equipment is calculated as zz2(t) = t * v2; Determine if If the actual operating position of the cleaning equipment is normal, then it is normal; otherwise, it is abnormal. Determine if If the actual operating position of the lower end of the cleaning equipment is normal, then it is normal; otherwise, it is abnormal. Determine if If so, the actual operating error at the top and bottom of the cleaning equipment is within the normal range.

4. The intelligent dual-drive photovoltaic module cleaning system according to claim 3, characterized in that, When the host computer determines that the actual operating error at the upper and lower ends of the cleaning device exceeds the normal range, it performs fault handling on the drive device, including: when the actual operating error at the upper and lower ends of the cleaning device exceeds the specified threshold s1, the operating speed needs to be fine-tuned as follows: If s2>z1(t)-z2(t)>s1, according to z1(t)-z2(t)=t*(v1-v2), the running speed v1 of the upper drive device is adjusted so that v1=v2+(z1(t)-z2(t)) / t; If s2>z2(t)-z1(t)>s1, according to z2(t)-z1(t)=t*(v2-v1), the running speed v2 of the drive device is adjusted so that v2=v1+(z2(t)-z1(t)) / t; When a serious malfunction occurs, that is, when the actual operating error at the upper and lower ends of the cleaning equipment is very large and exceeds the specified threshold s2, i.e.: if If necessary, the drive device should be shut down.

5. The intelligent dual-drive photovoltaic module cleaning system according to claim 3, characterized in that, Predicting the operating status of the cleaning equipment in the next moment and adjusting the operating speed in real time to adjust errors in real time includes: using an LSTM network to predict faults in the operating status of the cleaning equipment, including two types of fault prediction: abnormal cleaning speed and excessive tilt angle. Specifically, this includes: Obtaining input and output samples: For time t, the input samples include the actual distance moved above the cleaning device. Actual running speed The actual distance the lower end of the cleaning equipment moves Actual running speed ;set up Given the input at time t, construct the input vector: ; The output vector corresponding to the input vector is the tilt angle of the cleaning device. tilt angle The calculation formula is: , ,in, The vertical length of the cleaning equipment. It is an intermediate variable. It refers to the tilt angle of the cleaning equipment; Determine the number of hidden nodes in the LSTM network and the dropout rate of the dropout layer, assuming: Output for the current time t. The output from the previous time step. : The cell state at the previous moment; Computing the forget gate; ,in It's the sigmoid function, which outputs a value between 0 and 1, where 1 indicates that everything is retained and 0 indicates that everything is forgotten. The weight matrix representing the forget gate, This is the bias vector for the forget gate; Calculate the input gate: first calculate the updated weights. ,in Let be the weight matrix of the input gate. The bias vector is the input gate; then the candidate new memory is calculated: , Output a value between -1 and 1, where, The weight matrix for the new memory. This is the bias vector for the new memory. For new memories; Update cell state based on weights, forget gate, and new memories: ; Calculate the output gate: First, calculate the output weights: ,in Let the weight matrix be the output gate. This is the bias vector for the output gate. The previous output is used as the basis for calculation; then the current output is calculated. , for matrix operations, This indicates element-wise multiplication. The predicted output vector is then the calculated current output. ; Adjust the operating speed of the upper and lower drive devices based on the predicted output vector.

6. The intelligent dual-drive photovoltaic module cleaning system according to claim 1, characterized in that, The driving device can be a rope-type pulling system, a walking robot system, or a low-altitude unmanned aerial vehicle system, either individually or in combination.

7. The intelligent dual-drive photovoltaic module cleaning system according to claim 6, characterized in that, The rope-type pulling system includes a bracket, rotating wheels, traction equipment, a motor, and a first control system. The bracket is located on both sides of the photovoltaic module array. The two ends of the traction equipment are respectively wound into the rotating wheels on both sides and connected to the cleaning equipment in the middle. The motor is connected to the rotating wheels to drive the rotating wheels to rotate. The traction equipment is a rope traction, chain traction, or track traction. The first control system is used to control the start and stop of the motor to realize the rotation and stop of the rotating wheel, record the rotation running time, the moving position of the cleaning equipment, and the speed of the upper and lower ends of the cleaning equipment, and transmit them to the host computer, and adjust the motor speed according to the instructions of the host computer.

8. The intelligent dual-drive photovoltaic module cleaning system according to claim 6, characterized in that, The walking robot system includes a walking robot vehicle, a sliding bracket, pulleys, a robotic arm column, a translation motor, a rotation motor, a docking mechanism, a second control system, a first image recognition system, and a first high-definition camera. The sliding bracket is a bracket with a sliding track and is installed inside the walking robot vehicle. The lower end of the robotic arm column has pulleys, which are driven by the translation motor to move on the sliding track. The translation motor is used to control the robotic arm column to move up and down and slide left and right. The rotation motor is connected to the upper end of the robotic arm column. The docking mechanism is connected to the rotation motor and is used to dock the rotation motor with the rotating shaft on the cleaning device during cleaning, thereby driving the rotating shaft on the cleaning device to rotate. The forward force of the walking robot vehicle pushes the cleaning brush to move on the photovoltaic module array for cleaning. The first high-definition camera is used to capture images of the docking status between the robotic arm column and the cleaning device and transmit them to the first image recognition system. The first image recognition system transmits the recognition results to the second control system. The first image recognition system is used to identify the center position of the rotating shaft on the cleaning brush and to dock the rotation motor with the rotating shaft on the cleaning brush.

9. The intelligent dual-drive photovoltaic module cleaning system according to claim 8, characterized in that, The second control system is used to control the operation mode of the translation motor and the rotary motor, to control the start and stop of the translation motor and the rotary motor, to record the movement position of the cleaning equipment, and to encode and calculate the results and send them to the host computer.

10. The intelligent dual-drive photovoltaic module cleaning system according to claim 6, characterized in that, The low-altitude unmanned aerial vehicle (UAV) system includes the UAV body, UAV control system, support column, rotating motor, docking mechanism, second high-definition camera, and second image recognition system. The support column is composed of a support frame and a column fixed together, and can be folded and placed inside the UAV body. When opened, the column stands below the UAV and contacts the cleaning equipment. It is used to propel the cleaning brush on the photovoltaic module array for cleaning by leveraging the forward flight force of the UAV. The rotating motor is located inside the lower end of the column, and the docking mechanism is connected to the rotating motor. During cleaning, the rotating motor docks with the rotating shaft on the cleaning equipment, thereby driving the rotating shaft to rotate. The second high-definition camera captures images of the docking status between the robotic arm-type column and the cleaning equipment, and transmits them to the second image recognition system. The second image recognition system transmits the recognition results to the UAV control system. The second image recognition system identifies the center position of the rotating shaft on the cleaning equipment, enabling the rotating motor to dock with the rotating shaft on the cleaning equipment.

11. The intelligent dual-drive photovoltaic module cleaning system according to claim 10, characterized in that, The UAV control system is used to control the flight direction of the UAV body, control the start and stop of the rotating motor, record the movement position of the cleaning equipment and the speed of the upper and lower ends of the cleaning equipment and send them to the host computer, and adjust the operating speed of the low-altitude UAV system according to the instructions of the host computer.

12. The intelligent dual-drive photovoltaic module cleaning system according to claim 1, characterized in that, The cleaning device is a cleaning brush structure with a planetary gear system, including a planetary gear system installed on one or both sides of the rotating shaft. The planetary gear system is used to rotate the rotating shaft when the traction device moves. The rotating shaft passes through the roller brush, and the two ends of the rotating shaft are sleeved and fixed to the roller brush wheel. The planetary gear system meshes with the roller brush wheel. The planetary gear system includes three small gears. The roller brush wheel is located in the middle position of the three small gears. The outer teeth of the three small gears mesh with the track wheel. The outer teeth of the track wheel mesh and move on the track. The center of the small gear is connected by a tripod, and the tripod is fixed on the rotating shaft. One end of the traction device is connected to the rotating shaft.

Images