Digital twin-driven photovoltaic module monitoring and self-purification system

Abstract

The invention discloses a digital twin-driven photovoltaic module monitoring and self-purification system, belongs to the technical field of photovoltaic module cleaning, and realizes intelligent management and maintenance of a photovoltaic system through mutual cooperation of a sensing layer, an information capturing layer, an information transfer layer, an intelligent analysis layer, a decision support layer and an interactive experience layer. The sensing layer deploys various sensors to collect key data of the photovoltaic module; the information capturing layer transmits the data to a back-end processing system in real time; the information transfer layer ensures efficient transmission of data in the system; the intelligent analysis layer predicts a maintenance demand by using a machine learning algorithm, realizes predictive maintenance, and maintains the optimal working state of the component through an intelligent cleaning mechanism; the decision support layer provides decision support based on data analysis for an operator of the photovoltaic power station, and optimizes a maintenance plan and a productivity plan; and the interactive experience layer provides a visual management interface, so that the component state can be remotely monitored conveniently.

Description

Technical Field

[0001] The present invention belongs to the technical field of photovoltaic component cleaning, and specifically relates to a photovoltaic component monitoring and self-purification system driven by a digital twin. Background Art

[0002] Solar energy is an unlimited and continuous clean renewable energy. Photovoltaic systems can achieve effective conversion from solar energy to electrical energy. Photovoltaic panels, as one of the core components of solar power generation systems, are responsible for converting solar energy into electrical energy, because their role directly determines the power generation capacity and efficiency of the system. In actual use scenarios, photovoltaic panels are tilted and fixed for a long time. During long-term operation, they are easily affected by sand, dust, pollen and other substances in the environment, resulting in dust accumulation on photovoltaic panels. In addition, in cold areas, photovoltaic panels are also affected by snow accumulation. The above-mentioned dust and snow accumulation problems will cause the surface of photovoltaic panels to be covered, thereby reducing the photovoltaic panel's photoelectric conversion efficiency, affecting power generation and losing economic benefits.

[0003] Chinese Patent: CN117879473A, a patent application for a photovoltaic panel cleaning assembly and cleaning device, includes: a brush head, and also includes: a hanging board; a length adjustment structure is arranged at the bottom of the hanging board, the length adjustment structure has an output end, and the output end of the length adjustment structure is connected to the brush head; a tube body is arranged horizontally on the hanging board; a nozzle is arranged on the tube body, and a groove is opened at the bottom of the nozzle, and the groove makes the spraying range of the nozzle fan-shaped; a height adjustment structure is arranged on the hanging board, has an output end, and the output end of the height adjustment structure is connected to the tube body. By setting a reasonable nozzle spray angle and installation angle, the dust on the surface of the photovoltaic panel is washed away by the ejected fluid, and the sewage or dust is blocked from flowing back, avoiding secondary pollution. By combining the hanging board made of flexible material with the telescopic rod structure of the cleaning head, the bottom surface of the brush head can follow the shape of the photovoltaic panel surface, avoiding the low-position cleaning blind area and high-position jamming, and at the same time, the stubborn bottom dust on the surface of the photovoltaic panel can be effectively wiped off. However, this photovoltaic panel cleaning assembly and cleaning device cannot perform intelligent monitoring and self-cleaning on the photovoltaic panel surface. Summary of the invention

[0004] The purpose of the present invention is to address the problem that the existing photovoltaic panels are easily affected by the environment during operation, causing the surface of the photovoltaic panels to be covered, resulting in reduced photoelectric conversion efficiency. A digital twin-driven photovoltaic component monitoring and self-purification system is proposed. By setting up a photovoltaic component self-purification system driven by a digital twin, not only the power generation efficiency and reliability of the photovoltaic system are improved, the maintenance cost is reduced, but also the economic value of the entire system is significantly improved. From the perspective of engineering economics, the system significantly shortens the investment payback period by reducing manual intervention, optimizing energy output, and improving component performance, enhancing the financial attractiveness of the project, and providing solid technical support for the long-term operation of photovoltaic power stations.

[0005] To solve the above technical problems, the present invention provides a photovoltaic component monitoring and self-purification system driven by a digital twin, including: a perception layer, which monitors and cleans photovoltaic panels by installing sensors and a cleaning system on the photovoltaic panels; an information capture layer, which collects the environmental conditions and data of the photovoltaic panels in real time through sensors; an information transmission layer, which caches the collected data through wireless communication technology and data processing; an intelligent analysis layer, which processes data through cloud or edge computing technology, builds and updates a digital twin model in real time according to the design parameters of the photovoltaic panels to simulate the influence of environmental factors, and uses intelligent analysis algorithms to analyze and learn the data to predict the development trend of environmental factors; a decision support layer, which formulates a cleaning strategy based on the output results of the intelligent analysis layer and controls the cleaning system to clean the photovoltaic panels through a control system.

[0006] Preferably, the sensing layer monitors the surface conditions of the photovoltaic panels through image acquisition and processing, and uses image processing algorithms to analyze the acquired images to determine the degree of influence of environmental factors.

[0007] Preferably, in the information capture layer, data is collected at a frequency set by the sensor and transmitted through the sensor interface, and the analog signal collected by the sensor is converted into a digital signal and cached.

[0008] Preferably, the information transmission layer includes: Sa, transmitting the data of the information capture layer to the data collection node, and the data collection node performs preliminary processing and cache on the data; Sb, the data collection node packages the cached data and transmits it to the gateway device via the wireless communication network; Sc, the gateway device further processes and packages the received data and transmits it to the cloud or edge computing device via a wireless communication network; Sd, the cloud or edge computing device receives the data packet, unpacks and processes it, and stores the processed data in a database.

[0009] Preferably, the intelligent parsing layer includes: Sa, remove duplicates, process missing values, detect and clean up outliers, and convert formats of data transmitted by the information transmission layer; Sb. Store the preliminarily processed data, and use statistical analysis, time series analysis, cluster analysis, regression analysis, machine learning and deep learning methods to analyze the stored data; Sc. Obtain experimental data and combine it with physical laws to build dynamic models, physical models, data-driven models, and then build digital twin models; Sd, using machine learning and deep learning algorithms to train the data-driven model, using historical data to train the data-driven model, simulating and predicting the behavior of the photovoltaic module monitoring and self-purification system; Se. Use experimental data to verify the prediction accuracy of the digital twin model, and optimize the digital twin model parameters based on the verification results to obtain the optimal digital twin model; Sf. Based on the optimal digital twin model, the development trend of environmental factors is predicted.

[0010] Preferably, the decision support layer uses the prediction results provided by the intelligent analysis layer, rule-based algorithms, optimization algorithms, machine learning algorithms and deep learning algorithms to formulate and execute cleaning decisions, and controls the cleaning system to perform actual operations based on the decision results.

[0011] Preferably, the cleaning system includes an automatic cleaning device, wherein a spraying device is used to spray cleaning liquid and adjust the spraying amount according to environmental factors; a brushing device is used to physically clean the surface of the photovoltaic panel; a track movement system installs a track above the photovoltaic panel so that the cleaning device moves along the track; a sensor and control system are used to monitor the position of the cleaning device in real time; a pressure sensor monitors the pressure value during the cleaning process; and the control system controls the start, stop, movement and cleaning intensity of the cleaning device according to the instructions of the intelligent decision-making layer.

[0012] Compared with the prior art, the present invention has the following beneficial effects: 1. This solution not only improves the power generation efficiency and reliability of the photovoltaic system and reduces maintenance costs by setting up a photovoltaic module self-purification system based on digital twin drive, but also significantly improves the economic value of the entire system. From the perspective of engineering economics, the system significantly shortens the investment payback period by reducing manual intervention, optimizing energy output, and improving module performance, enhancing the financial attractiveness of the project and providing solid technical support for the long-term operation of photovoltaic power stations.

[0013] 2. This solution realizes the intelligent management and maintenance of photovoltaic systems through the cooperation and collaboration of the perception layer, information capture layer, information flow layer, intelligent analysis layer, decision support layer and interactive experience layer. The perception layer deploys various sensors to collect key data of photovoltaic components; the information capture layer transmits this data to the back-end processing system in real time; the information flow layer ensures the efficient transmission of data within the system; the intelligent analysis layer uses machine learning algorithms to predict maintenance needs, realize predictive maintenance, and maintain the best working state of components through intelligent cleaning mechanisms; the decision support layer provides decision support based on data analysis for operators of photovoltaic power stations, optimizes maintenance plans and capacity planning; the interactive experience layer provides a visual management interface to facilitate remote monitoring of component status. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] Other features, objects and advantages of the present invention will become more apparent by reading the detailed description of non-limiting embodiments made with reference to the following drawings. The drawings are only for the purpose of illustrating preferred embodiments and are not to be considered as limiting the present invention. Also, the same reference symbols are used throughout the drawings to represent the same parts.

[0015] Figure 1 It is a schematic diagram of the overall system of the present invention; Figure 2 is a physical layer block diagram of the present invention; Figure 3 It is a block diagram of the information capture layer of the present invention; Figure 4 It is a block diagram of the information flow layer of the present invention; Figure 5 It is a block diagram of the intelligent parsing layer of the present invention; Figure 6 It is a block diagram of the decision support layer of the present invention; Figure 7 This is a block diagram of the interactive experience layer of the present invention; Figure 8 This is a schematic diagram of the cleaning and overall installation of the bracket of the present invention; Fig. 9 This is a code diagram of using a genetic algorithm to optimize a process in an embodiment of the present invention. DETAILED DESCRIPTION

[0016] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific implementation method described herein is only an optimal embodiment of the present invention, which is only used to explain the present invention and does not limit the scope of protection of the present invention. All other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0017] Example 1: Figures 1 to 8 As shown, a photovoltaic module monitoring and self-purification system driven by a digital twin includes: Figure 2 As shown in the figure, in order to effectively monitor the dust and snow accumulation on the photovoltaic panels, the perception layer will install corresponding sensors on the photovoltaic panels and deploy high-resolution cameras at key locations to confirm the actual conditions of dust and snow accumulation. In addition, an automated cleaning system will be installed to ensure the cleanliness of the photovoltaic panels.

[0018] like Figure 3 As shown, the information capture layer: by integrating multiple sensors and high-resolution cameras, the environmental conditions around the photovoltaic panels and the status data of the panels themselves are collected in real time.

[0019] like Figure 4 As shown, the information transmission layer: select appropriate wireless communication technology, and perform data compression and encryption to ensure rapid, stable and secure data transmission.

[0020] like Figure 5 As shown, intelligent analysis layer: This layer uses advanced cloud or edge computing technology to process data from sensors and cameras and transmits it to the digital twin model. It builds and updates the digital twin model in real time according to the design parameters of the photovoltaic panels to simulate the impact of dust and snow accumulation. At the same time, it uses intelligent analysis algorithms based on machine learning and deep learning to analyze and learn historical data to predict the development trend of dust and snow accumulation.

[0021] like Figure 6 As shown, the decision support layer: formulates cleaning and snow removal strategies based on the output results of the intelligent analysis algorithm in the previous layer, including determining the frequency, optimal time and specific methods of cleaning and snow removal, and automatically activates the corresponding cleaning and snow removal devices through the control system to implement these strategies.

[0022] Specifically, the formulation of cleaning strategies includes frequency analysis. According to the climate, seasonal changes and pollution level of the region, the cleaning frequency under different environmental conditions is evaluated. For example, in dry areas, more frequent cleaning may be required to deal with dust, and in humid areas, natural rainfall may reduce the need for cleaning. Data-driven decision-making: Based on the data of historical power generation efficiency and cleaning frequency, the relationship between cleaning frequency and power generation efficiency is analyzed. Determine the cleaning time: First, determine the cleaning time, and dynamically adjust the cleaning frequency according to real-time monitoring data to ensure that cleaning is performed when it is most needed. Determine the cleaning time, such as cleaning after rain and cleaning after snow. Then, based on the early warning system, combined with weather warnings and meteorological data, establish an early warning system to notify in advance when cleaning may be required. The system automatically determines the cleaning time based on real-time data, such as automatically triggering the cleaning plan when the dust accumulation reaches the preset threshold; Finally, the cleaning method is executed. In the case of severe dust and snow accumulation, mechanical brushing and spraying are carried out, and the spraying system contains special cleaning agents to remove stubborn dirt without damaging the materials of photovoltaic modules. When the dust accumulation is general, mechanical brushing and spraying with clean water are selected. When the dust accumulation is light and cleaning is required, spraying with clean water is selected for flushing.

[0023] like Figure 7 As shown in the figure, the interactive experience layer: through an intuitive visual interface, it displays the monitoring data of dust and snow accumulation, the status of cleaning and snow removal, and the operation status of photovoltaic panels. Users can remotely access this interface through terminal devices such as personal computers and smart phones to view and manage the operation status of photovoltaic power stations in real time.

[0024] Specifically, in the perception layer, the appropriate type of photovoltaic panels is selected according to specific environmental conditions, and their performance in dust and snow environments is considered. By reasonably designing the installation angle and layout of the photovoltaic panels, the self-cleaning ability of the panels is optimized, thereby effectively reducing the accumulation of dust and snow; dust sensors with strong adaptability and high sensitivity are selected, and snow sensors with high precision and good weather resistance are selected; temperature, humidity, wind speed and light sensors are installed at appropriate locations around the photovoltaic panels to monitor the ambient temperature and the surface temperature of the photovoltaic panels, ambient humidity, wind speed and ambient light intensity to evaluate the wind's effect on the removal of dust and snow, as well as the impact of dust and snow on the light acceptance of the photovoltaic panels; high-resolution and weather-resistant cameras are selected to ensure that they can work normally under different weather conditions, and cameras are installed at appropriate locations of the photovoltaic panels to ensure that they can cover the entire surface of the panels. Through image acquisition and processing, the dust and snow accumulation on the surface of the photovoltaic panels are monitored, and the image processing algorithm is used to analyze the collected images to determine the degree of dust and snow accumulation; the automatic cleaning device consists of a mechanical brush and a water spray system. Choose a soft but good cleaning bristle material to avoid damage to the surface of the photovoltaic panel. Install the mechanical brush on the top or both sides of the photovoltaic panel to ensure that the entire panel surface can be fully covered. The water spray system uses a nozzle that can provide high-pressure water flow and is equipped with sufficient water source and pump system to ensure the cleaning effect. The water spray system is installed above the photovoltaic panel to ensure that the water flow can evenly cover the entire panel surface, and effectively clean the dust and dirt on the surface of the photovoltaic panel by spraying water.

[0025] Specifically, in the information capture layer, a suitable sensor interface is used to receive real-time monitoring data, each sensor collects data in real time at a set frequency, and the camera collects image data on the surface of the photovoltaic panel in real time at a set frequency. In the data conversion module, the analog signals collected by the sensor and camera are converted into digital signals to ensure the accuracy and consistency of the data. In order to avoid the occurrence of uncontrollable factors, the collected data is cached to ensure that it is not lost during data transmission.

[0026] Specifically, in the information flow layer, sensors and cameras collect data in real time and transmit the data to the data collection node. The data collection node performs preliminary processing and cache on the received data to avoid data loss and reduce transmission delay. Subsequently, the data collection node packages the cached data and transmits it to the gateway device through the wireless communication network. The gateway device further processes and packages the received data packets and transmits them to the cloud or edge computing device through the wireless communication network. The cloud or edge computing device receives these data packets, unpacks and processes them, and stores the processed data in the database for subsequent data analysis and intelligent decision-making. In addition, the entire data transmission process is monitored in real time to ensure the stability and reliability of data transmission, monitor the status of the transmission link, and promptly detect and handle any transmission failures to ensure the integrity and continuity of the data.

[0027] Specifically, in the intelligent analysis layer, data cleaning is the first step in data processing and analysis, with the goal of ensuring the accuracy and integrity of the data. Data is deduplicated, missing values ​​are processed, outliers are detected and cleaned, and format conversion is performed. The preliminarily processed data is stored for subsequent analysis and query. Data analysis is an in-depth analysis of the stored data to extract useful information and patterns. Various methods can be used, including statistical analysis, time series analysis, cluster analysis, regression analysis, machine learning, and deep learning.

[0028] Build digital twin models based on physical model building, data-driven model building, model verification and optimization, real-time simulation and prediction. Use physical laws and experimental data to build dynamic models of dust and snow accumulation, such as dust particle deposition model and snow melting model. Build physical models of dust and snow accumulation on photovoltaic panels to simulate their behavior under different environmental conditions. Use machine learning and deep learning algorithms to train data-driven models, such as time series prediction models and classification models, and use historical data to train data-driven models to simulate and predict system behavior.

[0029] Use experimental data and real data to verify the prediction accuracy of the model, and optimize the model parameters based on the verification results to improve the accuracy of the model. Verify the accuracy of the model and optimize the model. Use the digital twin model for real-time simulation and prediction to guide the cleaning and maintenance of photovoltaic panels. Optimize the cleaning strategy of photovoltaic panels based on dust and snow accumulation, reduce cleaning costs, and improve cleaning effects.

[0030] Use data analysis and digital twin models to predict the failure and aging of photovoltaic panels and formulate maintenance plans, such as regular inspections and component replacement. Establish an intelligent alarm and early warning system to issue alarm and early warning signals in a timely manner based on data analysis results and model predictions, such as excessive dust concentration and excessive snow thickness.

[0031] Specifically, in the decision support layer, the intelligent decision and execution layer is the final layer of the photovoltaic dust and snow monitoring and cleaning system, which is responsible for making and executing cleaning and maintenance decisions based on the results provided by the data processing and analysis layer to ensure the efficient operation of the photovoltaic system. The intelligent decision layer relies on rule algorithms, optimization algorithms, machine learning algorithms, deep learning algorithms, etc. to formulate the best cleaning and maintenance strategies, and control the cleaning system to perform actual operations based on the decision results. Figure 8 As shown, the cleaning system includes an automatic cleaning device, in which a spraying device is used to spray cleaning liquid, such as water or special cleaning agent, which can evenly cover the surface of the photovoltaic panel and adjust the spraying amount according to the dust and snow accumulation; a brushing device is used to physically brush the surface of the photovoltaic panel to remove stubborn dirt and snow, and needs to be soft and durable to avoid damaging the photovoltaic panel; the track moving system installs a track above the photovoltaic panel so that the cleaning device can move along the track to achieve large-area cleaning. In addition, it also includes sensors and control systems. The position sensor monitors the position of the cleaning device in real time to ensure that all photovoltaic panels are covered; the pressure sensor monitors the pressure during the cleaning process to prevent the cleaning device from causing damage to the photovoltaic panel; the control system controls the start, stop, movement and cleaning intensity of the cleaning device according to the instructions of the intelligent decision-making layer.

[0032] This layer also needs to conduct real-time monitoring and feedback to ensure the effectiveness of decision execution. Monitor the operating status of the photovoltaic system in real time to detect and handle abnormal situations in a timely manner; monitor the working status and cleaning effect of the cleaning device to ensure that the cleaning task is executed as planned; evaluate the maintenance effect based on the system performance after maintenance, and feed back the results to the intelligent decision-making layer to optimize the maintenance plan; evaluate the cleaning effect based on the cleanliness and power generation efficiency of the photovoltaic panel, and feed back the results to the intelligent decision-making layer to optimize the cleaning strategy.

[0033] Specifically, in the interactive experience layer, a real-time monitoring interface is provided, which can display real-time monitoring data such as the cleanliness of the photovoltaic panels, power generation efficiency, and environmental parameters, and present them to the user in a graphical manner, so that the user can intuitively understand the operating status of the photovoltaic system; the decision results generated by the intelligent decision-making layer and their execution status are displayed, including but not limited to cleaning plans, maintenance plans, and fault predictions, etc., and presented to the user through the decision display interface; a user interaction interface is set up to allow users to manually adjust cleaning and maintenance plans, and to view historical data and reports, thereby enhancing user participation and system transparency; regular reports are automatically generated, including the operating status of the photovoltaic system, cleaning and maintenance activities, fault handling, etc., to ensure that users can regularly obtain an overview of the system's operation; abnormal reports are automatically generated to record any abnormal conditions and their handling processes that occur during the operation of the system, so that users can be informed in a timely manner and take appropriate measures.

[0034] Embodiment 2: The present application also discloses a photovoltaic module monitoring and self-purification method driven by a digital twin, which is applicable to a photovoltaic module monitoring and self-purification system driven by a digital twin, and includes the following steps: Step 1: As attached Figure 2 , select photovoltaic panels and sensors suitable for specific environmental conditions, monitor the dust and snow accumulation on the surface of the photovoltaic panel through image acquisition and processing, use image processing algorithms to analyze the collected images and determine the degree of dust and snow accumulation; install mechanical brushes on the top or both sides of the photovoltaic panel to ensure that the entire panel surface can be fully covered. The water spray system uses a nozzle that can provide high-pressure water flow, and is equipped with sufficient water sources and pump systems to ensure the cleaning effect. The water spray system is installed above the photovoltaic panel to ensure that the water flow can evenly cover the entire panel surface. The bracket cleaning and overall installation are as shown in the attached Figure 8 .

[0035] Among them, photovoltaic panels can be selected from monocrystalline silicon, polycrystalline silicon, thin-film photovoltaic panels, etc.; dust sensors in the sensor network can be dust sensors that adapt to environmental changes, such as photoelectric type, laser type, etc.; snow sensors can be ultrasonic type, infrared type, etc.; temperature sensors can be temperature sensors with high precision and good stability, such as thermocouples, thermistors, etc.; humidity sensors can be humidity sensors suitable for outdoor environments, such as capacitive type, impedance type, etc.; wind speed sensors can be weather-resistant and accurate wind speed sensors, such as ultrasonic type, rotating cup type, etc.; light sensors can be light sensors with fast response speed and wide measurement range, such as photodiodes, photoresistors, etc.; cameras can be high-resolution, weather-resistant cameras to ensure that they can work normally under different weather conditions and that they can cover the entire surface of the panel.

[0036] Step 2: As attached Figure 3In the information capture layer, when the system starts, the processing sensor or camera will perform a self-test. Based on the real-time data measured by the sensor in step one, a dedicated sensor interface is used in step two to receive the real-time monitoring data. All sensors are guaranteed to work normally in various environments. Each sensor collects data in real time at a set frequency, such as dust concentration, snow thickness, temperature, humidity, wind speed, light intensity, etc. The camera collects image data on the surface of the photovoltaic panel in real time at a set frequency. In the data conversion module, the analog signals collected by the sensor and the camera are converted into digital signals to ensure the accuracy and consistency of the data. In order to avoid the generation of uncontrollable factors, the data acquisition module pre-processes the collected raw data, such as filtering, denoising, data correction, etc., to ensure the accuracy and consistency of the data, and temporarily stores the pre-processed data in the local buffer for subsequent transmission. The data acquisition module transmits the processed data to a remote server or cloud platform through a wireless communication network (such as LoRa, NB-IoT, 4G / 5G, etc.). During the transmission process, the real-time and reliability of data transmission are ensured to avoid data loss and transmission delay.

[0037] Step 3: As attached Figure 4In the information transmission layer, for the large number of distributed sensors used in this embodiment, a P2P model is adopted. In this model, data can be exchanged directly between sensors and cameras without going through a central server. This method can reduce latency and improve the response speed of the system. Each sensor and camera is connected to a data collection node by wired or wireless means. The data collection node is responsible for preliminary processing and caching the data collected in step 2. After the gateway device receives the raw data, it first needs to parse the data to identify the format, type and content of the data. This step usually involves parsing the data packet header to extract the payload of the data. To ensure the accuracy of the data, the gateway device will verify the received data. This includes checking the integrity, consistency and whether there are errors in the data. If there is a problem with the data, the gateway device may request retransmission or discard the erroneous data; after preprocessing, the gateway device will encapsulate the data into a data packet. The data packet usually contains a data header (used to identify the type and source of the data), a data body (containing the actual data content) and a checksum (used to ensure the integrity of the data). In order to improve transmission efficiency, the gateway device may compress the data packet. At the same time, in order to ensure the security of the data, the gateway device will also encrypt the data packet. Encryption can prevent data from being stolen or tampered with during transmission. During the packaging process, the gateway device may also add some metadata, such as timestamps, device IDs, etc. This information is helpful for subsequent data processing and analysis. After packaging, the gateway device will send the data packet to the collection network management node or remote server through the selected data transmission method and protocol (such as TCP / IP, UDP, etc.). During the transmission process, the gateway device will be responsible for the routing and forwarding of the data to ensure that the data packet can accurately reach the destination. To ensure reliable data transmission, the gateway device usually implements a confirmation and retransmission mechanism. When the data packet is successfully sent to the destination, the gateway device will receive confirmation information from the destination. If the data packet is lost or damaged during transmission, the gateway device will retransmit the data packet as needed and transfer the cached data to the gateway device; the gateway device supports multiple communication protocols, such as LoRa, NB-IoT, 4G / 5G, etc., to ensure that data can be stably transmitted in different network environments; adopt transmission protocols suitable for wireless communication, such as MQTT, CoAP, etc. These protocols are lightweight, low power consumption, and highly reliable, and are suitable for data transmission of IoT devices; in wired transmission environments, HTTP, HTTPS, FTP and other protocols can be used to ensure the security and reliability of data transmission; encrypt the transmitted data to ensure the security of the data during transmission, and encryption algorithms such as AES and RSA can be used; at the transmission protocol level, adopt secure transmission protocols, such as HTTPS, MQTTover TLS, to ensure the security of data during network transmission.

[0038] Step 4: As attached Figure 5 In the intelligent analysis layer, the data transmitted in step 3 is cleaned to ensure the accuracy and completeness of the data. The data is deduplicated according to the key fields such as the timestamp and sensor ID of the data. For a small number of missing values, interpolation and mean filling methods can be used to fill them. For data with a large number of missing values, deletion or marking can be considered. Statistical methods (such as standard deviation method, box plot method) or machine learning algorithms (such as isolation forest algorithm, LOF algorithm) are used to detect abnormal values, and they are processed according to specific circumstances (such as elimination, correction, marking, etc.). According to the predetermined standards, the unit, type, and format of the data are converted and standardized. The data after preliminary processing is stored for subsequent analysis and query. The database can be a relational database, a non-relational database, and a time series database. Then, the stored data is deeply analyzed to extract useful information and patterns. Various methods such as statistical analysis, time series analysis, cluster analysis, regression analysis, machine learning, and deep learning can be used.

[0039] Digital twin modeling is performed through design parameters and monitoring data, and a digital model of the physical system is established to simulate and predict the behavior of the system. Physical model construction: Based on physical laws and experimental data, a dynamic model of dust and snow accumulation is established, such as a dust particle deposition model and a snow melting model; a physical model of dust and snow accumulation on photovoltaic panels is established to simulate its behavior under different environmental conditions. Data-driven model construction: Machine learning and deep learning algorithms are used to train data-driven models, such as time series prediction models and classification models; model verification and optimization: Historical data is used to train data-driven models to simulate and predict the behavior of the system, and experimental data and real data are used to verify the prediction accuracy of the model, and model parameters are optimized based on the verification results to improve the accuracy of the model. Verify the accuracy of the model and optimize the model. Real-time simulation and prediction: Use digital twin models for real-time simulation and prediction to guide the cleaning and maintenance of photovoltaic panels. According to the dust and snow accumulation, the cleaning strategy of photovoltaic panels is optimized to reduce cleaning costs and improve cleaning effects.

[0040] Use data analysis and digital twin models to predict the failure and aging of photovoltaic panels and formulate maintenance plans, such as regular inspections and component replacement. Establish an intelligent alarm and early warning system to issue alarm and early warning signals in a timely manner based on data analysis results and model predictions, such as excessive dust concentration and excessive snow thickness.

[0041] Step 5: As attached Figure 6,The Decision Support Layer is the final layer of the PV dust and snow monitoring and cleaning ,system. It is responsible for making and executing cleaning and maintenance decisions ,based on the results provided by the data processing and ,analysis layer in step 4, to ensure the efficient operation of the PV ,system. ,The intelligent decision layer relies on rule algorithms, optimization algorithms, machine learning algorithms, ,deep learning algorithms, etc. to formulate the best cleaning and maintenance ,strategies, and control the cleaning system to perform actual operations ,according to the decision results.

[0042] The cleaning strategy was formulated based on a location in Inner Mongolia from October to December. The cleaning frequency is recommended to be fully cleaned once a month, especially in months with frequent sandstorms (such as October and November) to reduce the impact of dust on photovoltaic modules. The frequency is adjusted according to the weather forecast. If sandstorms or strong winds are predicted in the next few days, cleaning is required in advance. The specific schedule is as follows: In October, the first cleaning is carried out at the beginning of each month to ensure that the modules are in the best condition before the high-incidence season of sandstorms. If the weather conditions are not good (such as strong winds), the second cleaning can be arranged in the middle of the month. In November, the monthly cleaning frequency is maintained, especially at the beginning of the month. If it is found that the surface of the module is seriously dusty or sandstorms are predicted, additional cleaning is arranged. In December, a comprehensive cleaning is carried out at the beginning of the month to prepare for the low temperature and snow accumulation in winter. After cleaning, the power generation efficiency is monitored to evaluate the cleaning effect and provide a basis for the cleaning strategy for the next year. When the cleaning method is selected, mechanical brushing and detergent spraying are selected for heavy dust accumulation. If the dust accumulation is not heavy, it can be directly rinsed with water to reduce water consumption and improve cleaning efficiency. Aiming at the cleaning strategy of photovoltaic modules throughout the year, an intelligent optimization algorithm is used to optimize the cleaning strategy.

[0043] As attached Fig. 9 , taking the genetic algorithm optimization process as an example, taking the maximum power generation of photovoltaic modules as the optimization target, and the automatic cleaning frequency, automatic cleaning time, cleaning method, and cleaning power consumption of photovoltaic modules as the optimization independent variables, this pseudo code shows how to use genetic algorithms to optimize the cleaning strategy of photovoltaic modules to maximize power generation. The algorithm starts with a randomly generated initial population, and evaluates the effectiveness of the cleaning strategy by calculating the fitness of each individual. Subsequently, a new generation of population is generated by selecting excellent individuals for crossover and mutation, and iterates and optimizes continuously. In each generation, the algorithm records and updates the currently found optimal solution, and finally outputs the global optimal cleaning strategy and its corresponding maximum power generation. This method effectively combines the global search capability and randomness of the genetic algorithm to help find the optimal combination of cleaning strategies.

[0044] The cleaning system includes an automatic cleaning device, in which a spraying device is used to spray cleaning liquid, such as water or special cleaning agents, which can evenly cover the surface of the photovoltaic panel and adjust the spraying amount according to the dust and snow accumulation; a brushing device is used to physically brush the surface of the photovoltaic panel to remove stubborn dirt and snow, and needs to be soft and durable to avoid damaging the photovoltaic panel; a track moving system installs tracks above the photovoltaic panel so that the cleaning device can move along the track to achieve large-area cleaning. In addition, it also includes sensors and control systems. Position sensors monitor the position of the cleaning device in real time to ensure that all photovoltaic panels are covered; pressure sensors monitor the pressure during the cleaning process to prevent the cleaning device from causing damage to the photovoltaic panel; the control system controls the start, stop, movement and cleaning intensity of the cleaning device according to the instructions of the intelligent decision-making layer.

[0045] Step 6, as attached Figure 7 In the interactive experience layer, a real-time monitoring interface is provided to display key indicators of the photovoltaic system, such as total power generation, current power generation efficiency, dust and snow accumulation, etc., and real-time monitoring videos and data, such as the cleanliness of each photovoltaic panel, environmental parameters such as temperature, humidity, wind speed, etc., are displayed and presented to users in a graphical manner, so that users can intuitively understand the operating status of the photovoltaic system; the decision results generated by the intelligent decision-making layer and their execution status are displayed, including but not limited to cleaning plans, maintenance plans, and fault predictions, etc., and are presented to users through a decision display interface; a user interactive interface is set up to allow users to manually adjust cleaning and maintenance plans, and to view historical data and reports, thereby enhancing user participation and system transparency; regular reports are automatically generated, including the operating status of the photovoltaic system, cleaning and maintenance activities, fault handling, etc., to ensure that users can regularly obtain an overview of the system's operation; abnormal reports are automatically generated to record any abnormal conditions and their handling processes during the operation of the system, so that users can be informed in a timely manner and take corresponding measures.

[0046] Using the above steps, a 60MW photovoltaic array in a certain place in the Inner Mongolia Autonomous Region, a Class II solar energy resource area, was analyzed and compared, and Table 1 was obtained. The photovoltaic array installation area is located in the Gobi Desert area, with strong winds and sand. At the same time, it is accompanied by snowfall from November to April of the following year, which is greatly affected by dust and snow accumulation. The selected component specification is monocrystalline silicon bifacial 60WP. Based on PVsyst photovoltaic simulation software, the above photovoltaic system is simulated and analyzed to obtain the power generation data as follows.

[0047] Table 1 Analysis and comparison of a 60MW photovoltaic array in a certain place years Power generation of non-clean photovoltaic system (10,000 kWh) No clean power generation benefit (ten thousand yuan) Power generation of self-purification photovoltaic system (10,000 kWh) Self-purification power generation benefits (10,000 yuan) Increase ratio 2014 12114.21 5,330.2524 13,358.59 5,877.7896 +10.28% 2015 12152.51 5,335.2244 13,348.32 5,873.2608 +10.09% 2016 12,128.71 5,336.2324 13,324.77 5,862.8988 +9.86% 2017 12,135.91 5,339.8904 13,370.42 5,882.9848 +10.17% 2018 12,142.31 5,342.6164 13,379.52 5,886.9888 +10.20% 2019 12,148.71 5,345.3324 13,388.32 5,890.8608 +10.21% 2020 12,152.31 5,346.4164 13,397.09 5,895.7296 +10.24% 2021 12,155.91 5,347.6904 13,406.12 5,900.6928 +10.29% 2022 12,161.91 5,351.2364 13,415.12 5,902.6528 +10.31% 2023 12,168.11 5,353.9684 13,424.22 5,906.6568 +10.37% total 121,513.31 53,427.1596 133,712.52 58,787.6944 +10.04% From the data in the table, we can see that the power generation of the photovoltaic component monitoring and self-purification system driven by the digital twin in the past ten years is compared. The average AC grid-connected electricity price in Inner Mongolia is 0.44 yuan / kWh. It is calculated that the average increase in power generation in the past ten years is 10.04%, and the economic benefits have increased by about 53.6 million yuan.

[0048] It can be seen that the photovoltaic module self-purification system driven by digital twins not only improves the power generation efficiency and reliability of the photovoltaic system, reduces maintenance costs, but also significantly improves the economic value of the entire system. From the perspective of engineering economics, the system significantly shortens the investment payback period by reducing manual intervention, optimizing energy output, and improving module performance, enhancing the financial attractiveness of the project and providing solid technical support for the long-term operation of photovoltaic power stations.

[0049] The above specific embodiments are preferred embodiments of the present invention, and are not intended to limit the specific implementation scope of the present invention. The scope of the present invention includes but is not limited to the present specific embodiments. All equivalent changes made in accordance with the shape, structure, and method of the present invention are within the protection scope of the present invention.

Claims

1. A photovoltaic module monitoring and self-purification system driven by digital twins, characterized in that: include: The perception layer monitors and cleans the photovoltaic panels by installing sensors and cleaning systems on them. The information capture layer collects the environmental conditions and data of the photovoltaic panels in real time through sensors. The information transmission layer uses wireless communication technology and caches the collected data after data processing; The intelligent analysis layer processes data through cloud or edge computing technology, builds and updates the digital twin model in real time according to the design parameters of the photovoltaic panels to simulate the impact of environmental factors, and uses intelligent analysis algorithms to analyze and learn data to predict the development trend of environmental factors; The decision support layer, based on the output results of the intelligent analysis layer, formulates a cleaning strategy and controls the cleaning system to clean the photovoltaic panels through the control system.

2. According to claim 1, a photovoltaic module monitoring and self-purification system driven by digital twins is characterized in that: The sensing layer monitors the surface conditions of the photovoltaic panels through image acquisition and processing, and uses image processing algorithms to analyze the acquired images to determine the degree of influence of environmental factors.

3. According to claim 1, a photovoltaic module monitoring and self-purification system driven by digital twins is characterized in that: In the information capture layer, data is collected at a frequency set by the sensor and transmitted through the sensor interface, and the analog signal collected by the sensor is converted into a digital signal and cached.

4. A photovoltaic module monitoring and self-purification system driven by digital twins according to claim 3, characterized in that: The information transmission layer includes: Sa, transmitting the data of the information capture layer to the data collection node, and the data collection node performs preliminary processing and cache on the data; Sb, the data collection node packages the cached data and transmits it to the gateway device via the wireless communication network; Sc, the gateway device further processes and packages the received data and transmits it to the cloud or edge computing device via a wireless communication network; Sd, the cloud or edge computing device receives the data packet, unpacks and processes it, and stores the processed data in a database.

5. The photovoltaic module monitoring and self-purification system driven by digital twin according to claim 1, characterized in that: The intelligent parsing layer includes: Sa, remove duplicates, process missing values, detect and clean up outliers, and convert formats of data transmitted by the information transmission layer; Sb. Store the preliminarily processed data, and use statistical analysis, time series analysis, cluster analysis, regression analysis, machine learning and deep learning methods to analyze the stored data; Sc. Obtain experimental data and combine it with physical laws to build dynamic models, physical models, data-driven models, and then build digital twin models; Sd, using machine learning and deep learning algorithms to train the data-driven model, using historical data to train the data-driven model, simulating and predicting the behavior of the photovoltaic module monitoring and self-purification system; Se. Use experimental data to verify the prediction accuracy of the digital twin model, and optimize the digital twin model parameters based on the verification results to obtain the optimal digital twin model; Sf. Based on the optimal digital twin model, the development trend of environmental factors is predicted.

6. A photovoltaic module monitoring and self-purification system driven by digital twins according to claim 1, characterized in that: The decision support layer uses the prediction results provided by the intelligent analysis layer, rule-based algorithms, optimization algorithms, machine learning algorithms and deep learning algorithms to formulate and execute cleaning decisions, and controls the cleaning system to perform actual operations based on the decision results.

7. A photovoltaic module monitoring and self-purification system driven by digital twins according to claim 6, characterized in that: The cleaning system includes an automatic cleaning device, in which a spraying device is used to spray cleaning liquid and adjust the spraying amount according to environmental factors; a brushing device is used to physically clean the surface of the photovoltaic panel; a track movement system installs a track above the photovoltaic panel so that the cleaning device moves along the track; a sensor and control system are used to monitor the position of the cleaning device in real time; a pressure sensor monitors the pressure value during the cleaning process; and the control system controls the start, stop, movement and cleaning intensity of the cleaning device according to the instructions of the intelligent decision-making layer.

8. The photovoltaic module monitoring and self-purification system driven by digital twin according to claim 1, characterized in that: The cleaning strategy includes frequency analysis, data-driven decision making, determining cleaning timing, and cleaning method execution.

Images