Method, system, readable storage medium and computer program product for cleaning a photovoltaic array
By coordinating sensor data acquisition and model diagnostics, the cleaning needs of photovoltaic arrays are dynamically assessed, and optimized cleaning tasks are generated. This solves the problems of low efficiency, high cost, and insufficient accuracy of existing photovoltaic array cleaning methods, and achieves efficient and safe cleaning results.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- JIANGXI MECHANICAL & ELECTRICAL VOCATIONAL & TECH COLLEGE
- Filing Date
- 2026-05-13
- Publication Date
- 2026-06-26
AI Technical Summary
Existing photovoltaic array cleaning methods suffer from low cleaning efficiency, high cost, lack of accurate pollution detection capabilities, and lack of verification and optimization of the cleaning process, resulting in untimely or excessive cleaning, and an inability to effectively distinguish between pollution and electrical faults.
By collaboratively deploying electrical, visual, infrared, and lighting sensors to collect data at the same timestamp, a multi-source dataset is constructed. An improved model is used for data diagnosis, electrical and status feature data are extracted, demand indices are calculated, cleaning tasks are generated, and cleaning robots are used for secondary verification and optimization of cleaning paths.
It enables automated and high-precision identification of photovoltaic array faults and contamination characteristics, dynamically assesses cleaning needs and priorities, avoids ineffective operations, and ensures the accuracy and efficiency of cleaning operations.
Smart Images
Figure CN122274995A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of data processing technology, and in particular to a method, system, readable storage medium, and computer for cleaning photovoltaic arrays. Background Technology
[0002] As a core component of a solar power generation system, the cleanliness of the photovoltaic array directly affects its photoelectric conversion efficiency and power generation revenue. In natural environments, dust, sand, bird droppings, fallen leaves, and other foreign matter gradually adhere to the surface of the photovoltaic panels, creating a shading effect that leads to localized temperature increases (hot spot effect), reduced output power, and even permanent damage to the modules. Therefore, regularly cleaning the photovoltaic array is a necessary measure to ensure its efficient and stable operation.
[0003] Currently, the main cleaning methods for photovoltaic arrays include manual cleaning, fixed spray cleaning, and automatic cleaning devices based on timed or simple sensor triggering. However, these traditional methods have the following prominent problems: 1. Low cleaning efficiency and high cost: Manual cleaning relies on experience and judgment, and the cleaning cycle is fixed. It cannot be dynamically adjusted according to the actual degree of pollution, resulting in untimely or excessive cleaning, which wastes human and water resources.
[0004] 2. Lack of accurate pollution detection capability: Existing automated systems are mostly based on a single sensor (such as a light attenuation sensor) or simple visual judgment, which makes it difficult to accurately quantify dust distribution density, identify the type of foreign matter (such as the difference in the impact of adhesive pollutants and loose dust), and also cannot effectively distinguish between power reduction caused by pollution and other factors such as electrical faults and shadow occlusion.
[0005] 3. Lack of verification and optimization in the cleaning process: The cleaning robot lacks a secondary confirmation mechanism for the target area before performing the task, which can easily lead to poor cleaning results due to positioning deviation or environmental changes; the cleaning path planning also does not fully consider multi-robot collaboration, energy constraints and task priorities, resulting in limited overall system intelligence. Summary of the Invention
[0006] Therefore, the purpose of this invention is to provide a method, system, readable storage medium, and computer for cleaning photovoltaic arrays, so as to at least solve the shortcomings of the above-mentioned technologies.
[0007] This invention proposes a cleaning method for photovoltaic arrays, comprising: Data is collected using sensor devices deployed on the photovoltaic array at the same timestamp, and the collected data is standardized to obtain the corresponding multi-source dataset. Data diagnostics are performed on the multi-source dataset to extract electrical feature data and state feature data from the multi-source dataset, respectively. Collect demand data for the photovoltaic array, and calculate the demand index for the photovoltaic array based on the demand data, the electrical characteristic data, and the state characteristic data; Based on the demand index, a corresponding cleaning task is generated, and based on the cleaning task, the corresponding cleaning robot is controlled to travel to the corresponding work area, and the work area is verified a second time by the cleaning robot. If the secondary verification passes, a corresponding cleaning path is generated based on the cleaning task, and the cleaning robot is controlled to clean according to the cleaning path.
[0008] Furthermore, the sensor devices include electrical sensor devices, visual sensor devices, infrared sensor devices, and light sensor devices. The steps of using the sensor devices deployed on the photovoltaic array to collect data at the same timestamp and standardizing the collected data to obtain the corresponding multi-source dataset include: The electrical sensor device is used to acquire time-series data sequences of electrical operation within several sampling periods at a fixed sampling frequency. A moving average filter is then applied to the time-series data sequences to suppress noise, resulting in an enhanced time-series feature vector. The visual sensor device and the infrared sensor device are used to collect data from the photovoltaic array along a preset path to obtain registered visible light images and infrared images, and distortion correction, illumination normalization and temperature calibration are performed on the visible light images and the infrared images. The light sensor device emits infrared modulated light and projects it onto the photovoltaic array, and receives the reflected light intensity distribution of the photovoltaic array based on the infrared modulated light. The dust density distribution data of the photovoltaic array is then mapped based on the reflected light intensity distribution.
[0009] Furthermore, the step of performing data diagnostics on the multi-source dataset to extract electrical feature data and state feature data from the multi-source dataset includes: Key feature vectors are extracted from the enhanced time-series feature vectors, and the key feature vectors are input into a pre-trained diagnostic model for data diagnosis to obtain electrical feature data; The registered visible light image, the infrared image, and the dust density distribution data are stitched together along the channel dimension to obtain the corresponding fusion tensor. The fusion tensor is input into the improved target detection model to obtain the corresponding state feature data.
[0010] Furthermore, the step of collecting demand data for the photovoltaic array and calculating the demand index for the photovoltaic array based on the demand data, the electrical characteristic data, and the state characteristic data includes: The photovoltaic array is divided into several management units based on the work zones of the cleaning robots in the photovoltaic array; Based on the power generation plan, electricity price period, and environmental conditions, assign corresponding time-varying weighting coefficients to the demand data, electrical characteristic data, and state characteristic data; The demand index of each management unit is calculated based on the time-varying weighting coefficient, the demand data of each management unit, the electrical characteristic data, and the state characteristic data.
[0011] Furthermore, the formula for calculating the demand index is as follows:
[0012] In the formula, , , , These are time-varying weighting coefficients. Indicates the first in the photovoltaic array The current operating power of each management unit. Indicates the first Each management unit is under the current irradiance. and temperature The theoretical maximum output power is as follows. Indicates the first Average dust density distribution data for each management unit Indicates the first The dust density distribution threshold for each management unit The first in the state feature data The area of the foreign object Indicates the first The influence weight of each foreign object Indicates the first The total area of each management unit Indicating the first in electrical characteristic data Electrical fault coefficient of each management unit Indicates the first Electrical fault confidence level of each management unit.
[0013] The present invention also proposes a cleaning system for photovoltaic arrays, comprising: The data processing module is used to collect data using the sensor devices deployed on the photovoltaic array at the same timestamp, and to standardize the collected data to obtain the corresponding multi-source dataset. The data diagnostic module is used to perform data diagnostics on the multi-source dataset to extract electrical feature data and state feature data from the multi-source dataset respectively. The data calculation module is used to collect the demand data of the photovoltaic array and calculate the demand index of the photovoltaic array based on the demand data, the electrical characteristic data and the state characteristic data. The data verification module is used to generate corresponding cleaning tasks based on the demand index, and control the corresponding cleaning robot to drive to the corresponding work area based on the cleaning task, and perform secondary verification of the work area through the cleaning robot. An array cleaning module is used to generate a corresponding cleaning path based on the cleaning task if the secondary verification passes, and to control the cleaning robot to clean according to the cleaning path.
[0014] Furthermore, the sensor device includes an electrical sensor device, a vision sensor device, an infrared sensor device, and a light sensor device, and the data processing module is specifically used for: The electrical sensor device is used to acquire time-series data sequences of electrical operation within several sampling periods at a fixed sampling frequency. A moving average filter is then applied to the time-series data sequences to suppress noise, resulting in an enhanced time-series feature vector. The visual sensor device and the infrared sensor device are used to collect data from the photovoltaic array along a preset path to obtain registered visible light images and infrared images, and distortion correction, illumination normalization and temperature calibration are performed on the visible light images and the infrared images. The light sensor device emits infrared modulated light and projects it onto the photovoltaic array, and receives the reflected light intensity distribution of the photovoltaic array based on the infrared modulated light. The dust density distribution data of the photovoltaic array is then mapped based on the reflected light intensity distribution.
[0015] Furthermore, the data diagnostic module is specifically used for: Key feature vectors are extracted from the enhanced time-series feature vectors, and the key feature vectors are input into a pre-trained diagnostic model for data diagnosis to obtain electrical feature data; The registered visible light image, the infrared image, and the dust density distribution data are stitched together along the channel dimension to obtain the corresponding fusion tensor. The fusion tensor is input into the improved target detection model to obtain the corresponding state feature data.
[0016] Furthermore, the data calculation module is specifically used for: The photovoltaic array is divided into several management units based on the work zones of the cleaning robots in the photovoltaic array; Based on the power generation plan, electricity price period, and environmental conditions, assign corresponding time-varying weighting coefficients to the demand data, electrical characteristic data, and state characteristic data; The demand index of each management unit is calculated based on the time-varying weighting coefficient, the demand data of each management unit, the electrical characteristic data, and the state characteristic data.
[0017] The present invention also proposes a storage medium storing a computer program that, when executed by a processor, implements the above-described method for cleaning photovoltaic arrays.
[0018] The present invention also proposes a computer, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the above-described method for cleaning photovoltaic arrays.
[0019] The photovoltaic array cleaning method, system, readable storage medium, and computer of this invention collect data at the same timestamp through the coordinated deployment of different sensors, thereby constructing a multi-dimensional, highly synchronous multi-source dataset. An improved model is used to diagnose the multi-source dataset, proposing electrical characteristic data reflecting electrical health status and state characteristic data reflecting surface contamination status, achieving automated and high-precision identification of fault and contamination characteristics. Time-varying weight coefficients are constructed based on external factors, and corresponding demand indices are built using these coefficients to dynamically assess the immediate cleaning needs and priorities of each management unit in the photovoltaic array, optimizing cleaning timing in conjunction with power generation economic benefits. Cleaning tasks are automatically generated based on the demand index. After the cleaning robot arrives at the work area of the cleaning task, a secondary verification is performed to confirm task matching, avoiding invalid or erroneous operations. An optimization algorithm is used to generate an efficient and collaborative cleaning path, ensuring the accuracy, safety, and overall efficiency of the cleaning operation. Attached Figure Description
[0020] Figure 1 This is a flowchart of the cleaning method for a photovoltaic array according to the first embodiment of the present invention; Figure 2 This is a structural block diagram of the cleaning system for the photovoltaic array in the second embodiment of the present invention; Figure 3 This is a structural block diagram of the computer in the third embodiment of the present invention.
[0021] The following detailed description, in conjunction with the accompanying drawings, will further illustrate the present invention. Detailed Implementation
[0022] To facilitate understanding of the present invention, a more complete description will be given below with reference to the accompanying drawings. Several embodiments of the invention are illustrated in the drawings. However, the invention can be implemented in many different forms and is not limited to the embodiments described herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete.
[0023] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the specification of this invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.
[0024] Example 1 Please see Figure 1 The image shows a method for cleaning a photovoltaic array according to a first embodiment of the present invention, the method specifically including steps S101 to S104: S101, data is collected using sensor devices deployed on the photovoltaic array at the same timestamp, and the collected data is standardized to obtain the corresponding multi-source dataset; Furthermore, the sensor device includes an electrical sensor device, a vision sensor device, an infrared sensor device, and a light sensor device, and step S101 specifically includes steps S1011 to S1013: S1011, the electrical sensor device is used to collect time-series data sequences of electrical operation within several sampling periods at a fixed sampling frequency, and the time-series data sequences are subjected to moving average filtering to suppress noise, so as to obtain an enhanced time-series feature vector: S1012, the visual sensor device and the infrared sensor device are used to collect data from the photovoltaic array along a preset path to obtain a registered visible light image and an infrared image, and the visible light image and the infrared image are subjected to distortion correction, illumination normalization and temperature calibration. S1013, the light sensor device emits infrared modulated light and projects it onto the photovoltaic array, and receives the reflected light intensity distribution of the photovoltaic array based on the infrared modulated light, and maps the dust density distribution data of the photovoltaic array according to the reflected light intensity distribution.
[0025] In practice, electrical sensor devices are used at a fixed sampling frequency. collection A sequence of timing data of electrical operation within a sampling period, wherein the electrical sensor device includes, but is not limited to, current sensors, voltage sensors, radiometers, and temperature sensors:
[0026] In the formula, , , They represent The output voltage, output current, and output power of the photovoltaic array at that time. , They represent At any given moment, ambient irradiance and backplate temperature; Furthermore, the obtained time-series data sequence is subjected to moving average filtering to suppress noise, and its first-order difference features are calculated to form an enhanced time-series feature vector.
[0027] Specifically, visual and infrared sensors are used to collect data from the photovoltaic array along a preset path to obtain registered visible light and infrared images. The visual and infrared sensors can be mounted on a drone or placed around the photovoltaic array to identify the photovoltaic array by scanning, and to perform distortion correction, illumination normalization, and temperature calibration on the visible light and infrared images. Furthermore, the laser emitter of the light sensor device emits infrared modulated light of a preset wavelength (808nm in this embodiment) towards the photovoltaic array. The receiver of the light sensor device receives the light reflected by the photovoltaic array to obtain the corresponding reflected light intensity distribution data. The reflected light intensity distribution data is then inverted using a calibrated mapping model to obtain two-dimensional dust density distribution data.
[0028] In the formula, Indicates the intensity of emitted light. Indicates from the launch point to the receiving point geometrical attenuation factor The reflectivity function of the dust layer The transmittance function of the dust layer is calculated using the specific extinction coefficient of the dust. This indicates the reflectivity of the photovoltaic array. This indicates ambient background noise.
[0029] S102, perform data diagnosis on the multi-source dataset to extract electrical feature data and state feature data from the multi-source dataset respectively; Furthermore, step S102 specifically includes steps S1021 to S1023: S1021, extract key feature vectors from the enhanced time-series feature vectors, and input the key feature vectors into the pre-trained diagnostic model for data diagnosis to obtain electrical feature data; S1022, The registered visible light image, the infrared image, and the dust density distribution data are stitched together in the channel dimension to obtain the corresponding fusion tensor; S1023, The fusion tensor is input into the improved target detection model to obtain the corresponding state feature data.
[0030] In specific implementation, a preset feature extraction algorithm (in this embodiment, the feature extraction algorithm includes, but is not limited to, Tsfresh algorithm, random tree algorithm, Tsfeatures algorithm, TODS algorithm, etc.) is used to extract key feature vectors from the enhanced time series feature vectors. The extracted key feature vectors include the maximum power point voltage, current, open circuit voltage, short circuit voltage, and curve feature values, etc. In this embodiment, the pre-trained diagnostic model uses the global search mechanism in the firefly algorithm to update the position of individual sparrows in the sparrow search algorithm, reducing the algorithm from getting stuck in local minima in the later stage. The position of all sparrows and the optimal sparrow is updated by perturbation using the firefly algorithm. Furthermore, the population is initialized based on sinusoidal chaotic mapping theory to ensure uniform distribution within the feasible region. The number of iterations and the ratio of predators to intruders are set. Fitness values are calculated and sorted. The positions of predators, intruders, and guards are updated using the sparrow search algorithm. Fitness values are calculated and sparrow positions are updated. The perturbation method of the firefly algorithm is used to update sparrow positions. Fitness values are calculated and sparrow positions are updated. This process is repeated until the maximum number of iterations is reached to construct the diagnostic model. Key feature vectors are input into the diagnostic model for data diagnosis to obtain electrical feature data. Specifically, the registered visible light image, infrared image, and dust density distribution data are concatenated along the channel dimension to obtain the corresponding fusion tensor. A depthwise separable convolution module replaces the standard 3x3 convolution of the YOLOv5s network. In the intermediate layer of the backbone of this YOLOv5s network, a 1x1 convolution adaptively weights and fuses the feature map of the dust accumulation channel with the RGB and Thermal feature maps. A convolutional attention module is introduced, where global average pooling and global max pooling are performed on the input feature map. The resulting output is then fed into a multilayer perceptron mechanism for shared weight learning to generate channel attention weights. The input feature map is multiplied channel by channel to obtain the corresponding channel attention map through feature weighting. Simultaneously, average pooling and max pooling are performed on the channel dimension of the input feature map to generate two spatial feature maps. The two feature maps are concatenated and then passed through a convolutional layer to generate a spatial attention map. The convolutional layer uses the Sigmoid activation function, has a 7x7 kernel, and outputs 1 channel. The spatial attention map and the channel attention map are fused to output the final feature map. A decoupled head structure is adopted, with the classification branch and the regression branch having independent shallow networks, thereby constructing an improved object detection model. The obtained fused tensor is input into the improved object detection model to obtain the corresponding state feature data.
[0031] S103, Collect the demand data of the photovoltaic array, and calculate the demand index of the photovoltaic array based on the demand data, the electrical characteristic data and the state characteristic data; Furthermore, step S103 specifically includes steps S1031 to S1033: S1031, the photovoltaic array is divided into several management units based on the work zoning of the cleaning robot in the photovoltaic array; S1032, assign corresponding time-varying weighting coefficients to the demand data, the electrical characteristic data and the state characteristic data according to the power generation plan, electricity price period and environmental conditions; S1033, calculate the demand index of each management unit based on the time-varying weighting coefficient, the demand data of each management unit, the electrical characteristic data, and the state characteristic data.
[0032] In practical implementation, the photovoltaic array is divided into several management units based on the work zones of the cleaning robots within the photovoltaic array. Demand data for the photovoltaic array is collected, and time-varying weight coefficients are assigned to the demand data, electrical characteristic data, and status characteristic data according to the power generation plan, electricity price period, and environmental conditions. The demand index for each management unit is calculated based on the time-varying weight coefficients, the demand data, electrical characteristic data, and status characteristic data of each management unit. The formula for calculating the demand index is:
[0033] In the formula, , , , These are time-varying weighting coefficients. Indicates the first in the photovoltaic array The current operating power of each management unit. Indicates the first Each management unit is under the current irradiance. and temperature The theoretical maximum output power is as follows. Indicates the first Average dust density distribution data for each management unit Indicates the first The dust density distribution threshold for each management unit The first in the state feature data The area of the foreign object Indicates the first The influence weight of each foreign object Indicates the first The total area of each management unit Indicating the first in electrical characteristic data Electrical fault coefficient of each management unit Indicates the first Electrical fault confidence level of each management unit.
[0034] S104, Generate a corresponding cleaning task based on the demand index, and control the corresponding cleaning robot to drive to the corresponding work area based on the cleaning task, and perform secondary verification of the work area through the cleaning robot; In practice, the demand index of all management units of the photovoltaic array is calculated and compared with a pre-set demand index threshold. If more than half of the demand indices are greater than the threshold, a cleaning task is generated. The cleaning task includes the geographical coordinates of the management unit, the specific target bounding box to be cleaned, and the recommended cleaning mode based on the type and density of dust accumulation (such as "dry brush", "low-pressure water wash", "high-pressure water wash + brushing").
[0035] Furthermore, based on the cleaning task, the corresponding cleaning robot is controlled to travel to the corresponding work area. The work area is scanned by the vision positioning module and LiDAR installed on the cleaning robot. Its coordinates, area anomaly identification results, and surface images are used to determine whether the work area meets the priority in the cleaning task. If it does, the cleaning robot is marked as the object responsible for this cleaning task.
[0036] S105, if the secondary verification passes, a corresponding cleaning path is generated based on the cleaning task, and the cleaning robot is controlled to clean according to the cleaning path.
[0037] In practical implementation, if the secondary verification is successful, the improved ant colony algorithm or genetic algorithm is used to allocate tasks to multiple robots based on the task priority, the location information of the management unit, the current location of the cleaning robot, and the battery level. A cleaning path is generated, and the cleaning robot is controlled to clean according to the cleaning path.
[0038] In summary, the photovoltaic array cleaning method in the above embodiments of the present invention constructs a multi-dimensional, highly synchronous multi-source dataset by coordinating the deployment of different sensors to collect data at the same timestamp. An improved model is used to diagnose the multi-source dataset, proposing electrical characteristic data reflecting electrical health status and state characteristic data reflecting surface contamination status, thereby achieving automated and high-precision identification of fault and contamination characteristics. Time-varying weight coefficients are constructed based on external factors, and corresponding demand indices are built using these coefficients to dynamically assess the immediate cleaning needs and priorities of each management unit in the photovoltaic array, optimizing cleaning timing in conjunction with power generation economic benefits. Cleaning tasks are automatically generated based on the demand index. After the cleaning robot arrives at the work area of the cleaning task, a secondary verification is performed to confirm task matching, avoiding invalid or erroneous operations. An optimization algorithm is used to generate an efficient and collaborative cleaning path, ensuring the accuracy, safety, and overall efficiency of the cleaning operation.
[0039] Example 2 In another aspect, this invention also proposes a cleaning system for photovoltaic arrays; please refer to [link / reference needed]. Figure 2 The image shows a cleaning system for a photovoltaic array according to a second embodiment of the present invention. The system includes: The data processing module 11 is used to collect data using the sensor devices deployed on the photovoltaic array at the same timestamp, and to standardize the collected data to obtain the corresponding multi-source dataset. Data diagnostic module 12 is used to perform data diagnostics on the multi-source dataset to extract electrical feature data and state feature data from the multi-source dataset respectively; The data calculation module 13 is used to collect the demand data of the photovoltaic array and calculate the demand index of the photovoltaic array based on the demand data, the electrical characteristic data and the state characteristic data. The data verification module 14 is used to generate corresponding cleaning tasks based on the demand index, and control the corresponding cleaning robot to drive to the corresponding work area based on the cleaning task, and perform secondary verification of the work area through the cleaning robot. The array cleaning module 15 is used to generate a corresponding cleaning path based on the cleaning task if the secondary verification passes, and to control the cleaning robot to clean according to the cleaning path.
[0040] Furthermore, the sensor device includes an electrical sensor device, a vision sensor device, an infrared sensor device, and a light sensor device, and the data processing module 11 is specifically used for: The electrical sensor device is used to acquire time-series data sequences of electrical operation within several sampling periods at a fixed sampling frequency. A moving average filter is then applied to the time-series data sequences to suppress noise, resulting in an enhanced time-series feature vector. The visual sensor device and the infrared sensor device are used to collect data from the photovoltaic array along a preset path to obtain registered visible light images and infrared images, and distortion correction, illumination normalization and temperature calibration are performed on the visible light images and the infrared images. The light sensor device emits infrared modulated light and projects it onto the photovoltaic array, and receives the reflected light intensity distribution of the photovoltaic array based on the infrared modulated light. The dust density distribution data of the photovoltaic array is then mapped based on the reflected light intensity distribution.
[0041] Furthermore, the data diagnostic module 12 is specifically used for: Key feature vectors are extracted from the enhanced time-series feature vectors, and the key feature vectors are input into a pre-trained diagnostic model for data diagnosis to obtain electrical feature data; The registered visible light image, the infrared image, and the dust density distribution data are stitched together along the channel dimension to obtain the corresponding fusion tensor. The fusion tensor is input into the improved target detection model to obtain the corresponding state feature data.
[0042] Furthermore, the data calculation module 13 is specifically used for: The photovoltaic array is divided into several management units based on the work zones of the cleaning robots in the photovoltaic array; Based on the power generation plan, electricity price period, and environmental conditions, assign corresponding time-varying weighting coefficients to the demand data, electrical characteristic data, and state characteristic data; The demand index of each management unit is calculated based on the time-varying weighting coefficient, the demand data of each management unit, the electrical characteristic data, and the state characteristic data.
[0043] The functions or operation steps implemented by the above modules and units are largely the same as those in the above method embodiments, and will not be repeated here.
[0044] The photovoltaic array cleaning system provided in this embodiment of the invention has the same implementation principle and technical effects as the aforementioned method embodiment. For the sake of brevity, any parts not mentioned in the system embodiment can be referred to the corresponding content in the aforementioned method embodiment.
[0045] Example 3 This invention also proposes a computer, please refer to [link / reference]. Figure 3 The computer shown in the third embodiment of the present invention includes a memory 10, a processor 20, and a computer program 30 stored in the memory 10 and executable on the processor 20. When the processor 20 executes the computer program 30, it implements the above-described method for cleaning photovoltaic arrays.
[0046] The memory 10 includes at least one type of storage medium, such as flash memory, hard disk, multimedia card, card-type memory (e.g., SD or DX memory), magnetic memory, magnetic disk, optical disk, etc. In some embodiments, the memory 10 can be an internal storage unit of a computer, such as the computer's hard disk. In other embodiments, the memory 10 can be an external storage device, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, etc. Furthermore, the memory 10 can include both internal and external storage units of the computer. The memory 10 can be used not only to store application software and various types of data installed on the computer, but also to temporarily store data that has been output or will be output.
[0047] In some embodiments, the processor 20 may be an electronic control unit (ECU), a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chip, used to run program code stored in the memory 10 or process data, such as executing access restriction programs.
[0048] It should be pointed out that, Figure 3 The structure shown does not constitute a limitation on the computer. In other embodiments, the computer may include fewer or more components than shown, or combine certain components, or have different component arrangements.
[0049] This invention also proposes a storage medium storing a computer program that, when executed by a processor, implements the photovoltaic array cleaning method described above.
[0050] Those skilled in the art will understand that the logic and / or steps represented in the flowcharts or otherwise described herein, for example, can be considered as a ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can mean any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.
[0051] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.
[0052] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.
[0053] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0054] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this patent application should be determined by the appended claims.
Claims
1. A method for cleaning a photovoltaic array, characterized in that, include: Data is collected using sensor devices deployed on the photovoltaic array at the same timestamp, and the collected data is standardized to obtain the corresponding multi-source dataset. Data diagnostics are performed on the multi-source dataset to extract electrical feature data and state feature data from the multi-source dataset, respectively. Collect demand data for the photovoltaic array, and calculate the demand index for the photovoltaic array based on the demand data, the electrical characteristic data, and the state characteristic data; Based on the demand index, a corresponding cleaning task is generated, and based on the cleaning task, the corresponding cleaning robot is controlled to travel to the corresponding work area, and the work area is verified a second time by the cleaning robot. If the secondary verification passes, a corresponding cleaning path is generated based on the cleaning task, and the cleaning robot is controlled to clean according to the cleaning path.
2. The cleaning method for a photovoltaic array according to claim 1, characterized in that, The sensor devices include electrical sensor devices, visual sensor devices, infrared sensor devices, and light sensor devices. The steps of using the sensor devices deployed on the photovoltaic array to collect data at the same timestamp and then standardizing the collected data to obtain the corresponding multi-source dataset include: The electrical sensor device is used to acquire time-series data sequences of electrical operation within several sampling periods at a fixed sampling frequency. A moving average filter is then applied to the time-series data sequences to suppress noise, resulting in an enhanced time-series feature vector. The visual sensor device and the infrared sensor device are used to collect data from the photovoltaic array along a preset path to obtain registered visible light images and infrared images, and distortion correction, illumination normalization and temperature calibration are performed on the visible light images and the infrared images. The light sensor device emits infrared modulated light and projects it onto the photovoltaic array, and receives the reflected light intensity distribution of the photovoltaic array based on the infrared modulated light. The dust density distribution data of the photovoltaic array is then mapped based on the reflected light intensity distribution.
3. The cleaning method for photovoltaic arrays according to claim 2, characterized in that, The steps of performing data diagnostics on the multi-source dataset to extract electrical feature data and state feature data from the multi-source dataset include: Key feature vectors are extracted from the enhanced time-series feature vectors, and the key feature vectors are input into a pre-trained diagnostic model for data diagnosis to obtain electrical feature data; The registered visible light image, the infrared image, and the dust density distribution data are stitched together along the channel dimension to obtain the corresponding fusion tensor. The fusion tensor is input into the improved target detection model to obtain the corresponding state feature data.
4. The cleaning method for a photovoltaic array according to claim 2, characterized in that, The steps of collecting demand data for the photovoltaic array and calculating the demand index of the photovoltaic array based on the demand data, the electrical characteristic data, and the state characteristic data include: The photovoltaic array is divided into several management units based on the work zones of the cleaning robots in the photovoltaic array; Based on the power generation plan, electricity price period, and environmental conditions, assign corresponding time-varying weighting coefficients to the demand data, electrical characteristic data, and state characteristic data; The demand index of each management unit is calculated based on the time-varying weighting coefficient, the demand data of each management unit, the electrical characteristic data, and the state characteristic data.
5. The cleaning method for a photovoltaic array according to claim 4, characterized in that, The formula for calculating the demand index is as follows: In the formula, , , , These are time-varying weighting coefficients. Indicates the first in the photovoltaic array The current operating power of each management unit. Indicates the first Each management unit is under the current irradiance. and temperature The theoretical maximum output power is as follows. Indicates the first Average dust density distribution data for each management unit Indicates the first The dust density distribution threshold for each management unit The first in the state feature data The area of the foreign object Indicates the first The influence weight of each foreign object Indicates the first The total area of each management unit Indicating the first in electrical characteristic data Electrical fault coefficient of each management unit Indicates the first Electrical fault confidence level of each management unit.
6. A cleaning system for a photovoltaic array, characterized in that, include: The data processing module is used to collect data using the sensor devices deployed on the photovoltaic array at the same timestamp, and to standardize the collected data to obtain the corresponding multi-source dataset. The data diagnostic module is used to perform data diagnostics on the multi-source dataset to extract electrical feature data and state feature data from the multi-source dataset respectively. The data calculation module is used to collect the demand data of the photovoltaic array and calculate the demand index of the photovoltaic array based on the demand data, the electrical characteristic data and the state characteristic data. The data verification module is used to generate corresponding cleaning tasks based on the demand index, and control the corresponding cleaning robot to drive to the corresponding work area based on the cleaning task, and perform secondary verification of the work area through the cleaning robot. An array cleaning module is used to generate a corresponding cleaning path based on the cleaning task if the secondary verification passes, and to control the cleaning robot to clean according to the cleaning path.
7. The cleaning system for photovoltaic arrays according to claim 6, characterized in that, The sensor devices include electrical sensor devices, visual sensor devices, infrared sensor devices, and light sensor devices. The data processing module is specifically used for: The electrical sensor device is used to acquire time-series data sequences of electrical operation within several sampling periods at a fixed sampling frequency. A moving average filter is then applied to the time-series data sequences to suppress noise, resulting in an enhanced time-series feature vector. The visual sensor device and the infrared sensor device are used to collect data from the photovoltaic array along a preset path to obtain registered visible light images and infrared images, and distortion correction, illumination normalization and temperature calibration are performed on the visible light images and the infrared images. The light sensor device emits infrared modulated light and projects it onto the photovoltaic array, and receives the reflected light intensity distribution of the photovoltaic array based on the infrared modulated light. The dust density distribution data of the photovoltaic array is then mapped based on the reflected light intensity distribution.
8. The cleaning system for a photovoltaic array according to claim 7, characterized in that, The data diagnostic module is specifically used for: Key feature vectors are extracted from the enhanced time-series feature vectors, and the key feature vectors are input into a pre-trained diagnostic model for data diagnosis to obtain electrical feature data; The registered visible light image, the infrared image, and the dust density distribution data are stitched together along the channel dimension to obtain the corresponding fusion tensor. The fusion tensor is input into the improved target detection model to obtain the corresponding state feature data.
9. A readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the program implements the cleaning method for the photovoltaic array as described in any one of claims 1 to 5.
10. A computer comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the cleaning method for the photovoltaic array as described in any one of claims 1 to 5.