Mountain photovoltaic power station self-cleaning system based on intelligent regulation and control

Abstract

The invention belongs to the technical field of photovoltaic power station cleaning, and particularly relates to a mountain photovoltaic power station self-cleaning system based on intelligent regulation and control. The system comprises a sensor module, an intelligent cleaning strategy regulation and control module, an intelligent decision module and a cleaning equipment control module. The sensor module is used for monitoring weather conditions and pollution conditions in real time and collecting related data; the intelligent cleaning strategy regulation and control module comprises an intelligent cleaning strategy regulation and control algorithm and dynamically adjusts a cleaning strategy according to real-time weather and pollution conditions; the intelligent decision-making module predicts future cleaning requirements according to historical data and current environment conditions through deep learning and data analysis, and automatically adjusts a cleaning strategy; and the cleaning equipment control module regulates and controls operation of the cleaning equipment according to the cleaning strategy. According to the system, manual intervention can be reduced, the operation cost is reduced, the cleaning efficiency is improved, the problem of excessive cleaning or insufficient cleaning is avoided, and therefore the power generation efficiency of the photovoltaic panel is improved.

Description

Technical Field

[0001] This invention belongs to the field of photovoltaic power station cleaning technology, specifically relating to a self-cleaning system for a mountain photovoltaic power station. Background Technology

[0002] In existing technologies, the cleaning of photovoltaic power plants typically relies on manual or simple mechanical cleaning methods. Manual cleaning is not only labor-intensive and inefficient, but also causes significant damage to the panels, increasing cleaning costs and labor expenses. Furthermore, traditional mechanical cleaning devices often lack intelligent control, failing to adjust cleaning strategies in a timely manner based on weather changes, pollution levels, and panel conditions. This can easily lead to cleaning frequencies that are too high or too low, impacting the long-term operational efficiency and energy savings of the photovoltaic power plant. Moreover, with climate change and the increase in extreme weather events, existing cleaning technologies cannot flexibly cope with cleaning needs under special weather conditions such as sandstorms and snow accumulation. Therefore, a more intelligent and adaptive cleaning solution is urgently needed to improve the self-cleaning efficiency of photovoltaic power plants, reduce energy waste, and lower operating costs. Summary of the Invention

[0003] The purpose of this invention is to propose a self-cleaning system for mountain photovoltaic power stations based on intelligent regulation, so as to reduce human intervention and improve the overall operating efficiency of photovoltaic power stations.

[0004] The self-cleaning system for mountain photovoltaic power stations based on intelligent regulation provided by this invention includes: a sensor module, an intelligent cleaning strategy regulation module, an intelligent decision-making module, and a cleaning equipment control module; wherein:

[0005] The sensor module is used to monitor weather conditions and pollution levels in real time and collect relevant data.

[0006] This sensor module includes various sensors, such as weather sensors, pollution sensors, and photovoltaic panel pollution monitoring sensors. The weather conditions include temperature, humidity, wind speed, precipitation, and light intensity, while the pollution conditions include dust, sand, dirt, and snow. All collected data is transmitted in real time to the intelligent cleaning strategy control module via the MQTT protocol for multi-dimensional analysis, ensuring data transmission latency of ≤100ms, providing real-time data support for subsequent cleaning strategy adjustments.

[0007] The intelligent cleaning strategy control module designs an intelligent cleaning strategy control algorithm based on data collected by the sensor module, that is, dynamically adjusts the cleaning strategy according to real-time weather and pollution conditions; the specific process is as follows:

[0008] (1) Weather data analysis: The system analyzes meteorological data (wind speed, temperature, humidity, precipitation, etc.) in real time and predicts future weather changes. For example, in hot and dry weather, the system will increase the cleaning frequency to avoid the accumulation of dust and sand; in rainy weather, the system will automatically delay cleaning and use the rain to clean the panel naturally.

[0009] (2) Pollutant analysis: The system monitors the pollution status of the photovoltaic panel through pollution sensors. When pollutants (such as dust, sand, sewage, snow, etc.) accumulate to a certain level, the system automatically starts the cleaning program and selects the most suitable cleaning method, such as spray cleaning or wind cleaning.

[0010] (3) Dynamically adjust cleaning frequency: Based on weather forecasts and pollution monitoring data, the system will dynamically adjust the cleaning frequency and method. In environments with no precipitation for a long period of time, the system may increase the cleaning frequency, while in high humidity or precipitation weather, the system will delay cleaning to reduce the energy consumption of cleaning.

[0011] (4) Cleaning priority: The algorithm determines the cleaning priority based on the degree of pollution and weather changes. For example, in the case of a sandstorm, the system will prioritize cleaning the panels to avoid the impact of dust on the solar panels; in the case of snow, the system will decide the cleaning operation based on the degree of snow accumulation.

[0012] The intelligent decision-making module uses deep learning and data analysis to predict future cleaning needs based on historical data and current environmental conditions, and automatically adjusts the cleaning strategy. For example, if the power plant's power generation efficiency is low for a certain period of time and the weather is dry and rainless, the system will automatically increase the cleaning frequency to ensure that the panels do not reduce efficiency due to excessive accumulation of pollutants.

[0013] The cleaning equipment control module regulates the operation of the cleaning equipment according to the cleaning strategy; including:

[0014] (1) Spray system: In dry, highly polluted weather, the system removes surface dirt by spraying cleaning equipment.

[0015] (2) Ultrasonic cleaning device: In dusty weather, the system will activate the ultrasonic cleaning device to remove dust by blowing it away.

[0016] (3) Automatic scraper system: When there is snow or hard dirt, the system can start the scraper equipment for cleaning.

[0017] The self-cleaning system for mountain photovoltaic power stations based on intelligent regulation of this invention is specifically implemented using a combination of various models and algorithms, mainly including decision tree models, support vector machine (SVM) models, and neural network models. These models combine environmental data, photovoltaic panel status, and cleaning equipment control information to achieve intelligent regulation of the cleaning strategy. Specifically, it is implemented using the following programming languages ​​and technology stack:

[0018] 1. Machine Learning Frameworks:

[0019] Scikit-learn: Used to build and train traditional machine learning models such as support vector machines (SVM), decision trees, and random forests.

[0020] TensorFlow / Keras: Used for developing deep learning models, especially neural network models used in intelligent decision engines.

[0021] XGBoost: Used to implement efficient gradient boosting decision trees (GBDT), optimize cleaning strategies, and dynamically adjust decision processes.

[0022] 2. Data Processing and Analysis:

[0023] Pandas: Used for data processing, supporting the cleaning, merging, and analysis of raw data collected from sensor modules.

[0024] NumPy: Used for numerical computation, optimizing data processing, and feature engineering steps.

[0025] 3. Real-time data communication and control:

[0026] MQTT protocol: used for real-time data transmission and control command issuance in the clean system within a photovoltaic power plant, ensuring the system responds to environmental changes in real time.

[0027] ROS (Robot Operating System): Used for real-time control of cleaning equipment, especially robotic devices used to perform cleaning tasks.

[0028] 4. Front-end and visualization platform:

[0029] JavaScript / HTML / CSS: Used to develop visual interfaces to implement remote monitoring platforms, allowing power station managers to view cleaning status and adjust strategies.

[0030] D3.js: Used for data visualization, enabling graphical display of real-time monitoring data.

[0031] The main technical features and functional advantages of this invention are as follows:

[0032] The system of this invention can intelligently determine cleaning needs based on real-time environmental data, automatically adjust cleaning strategies, reduce manual intervention, and lower operating costs.

[0033] The system of this invention uses intelligent analysis and decision-making to adjust the cleaning strategy according to the actual situation, making the cleaning operation more efficient and avoiding the problems of over-cleaning or under-cleaning.

[0034] The system of this invention can improve the power generation efficiency of photovoltaic panels by reducing the accumulation of pollutants, thus ensuring the efficient operation of the power plant.

[0035] By dynamically adjusting cleaning strategies and optimizing the use of cleaning equipment, the system can effectively save energy and reduce unnecessary energy consumption during the cleaning process. Attached Figure Description

[0036] Figure 1 This is a flowchart of the present invention. Detailed Implementation

[0037] The present invention will be further described below through specific embodiments.

[0038] Example 1:

[0039] In this embodiment, when the mountain photovoltaic power station faces sandstorm weather, the system effectively solves the problem of sandstorm pollution on the photovoltaic panels through an intelligent cleaning strategy. The system collects meteorological data, pollution data, and photovoltaic panel status data every 10 minutes from 00:00 to 24:00. The specific data collected is as follows:

[0040] .

[0041] Based on the above data, the system determines the photovoltaic panel's pollution level to be "severely polluted" (pollution concentration ≥ 0.8 g / m² is considered severely polluted), and initiates a first-level cleaning priority, prioritizing the use of the ultrasonic cleaning device (wind-powered cleaning mode). The initial cleaning strategy is set as follows: cleaning frequency once per hour, each cleaning session lasting 3 minutes, and wind output intensity of 5 m / s (corresponding to a device power of 300W).

[0042] During the cleaning process, the system dynamically optimizes its strategy every 30 minutes based on real-time feedback data.

[0043] (1) When the dust concentration in the 6-hour average drops to 0.5 g / m³ and the dust concentration on the panel drops to 0.8 g / m², the cleaning frequency is adjusted to once every 2 hours, the wind power output intensity is reduced to 4 m / s (equipment power 220W), and the single working time is shortened to 2 minutes;

[0044] (2) When the dust concentration in the 12-hour average drops to 0.35 g / m³ and the dust concentration on the panel drops to 0.5 g / m², the cleaning frequency is adjusted to once every 3 hours, the wind power output intensity is maintained at 4 m / s, and the single working time is 2 minutes.

[0045] (3) When the average dust concentration in the 24-hour period drops to 0.2 g / m³ and the dust concentration on the panel drops to 0.3 g / m², the regular cleaning frequency is restored (twice a day, 2 minutes each time, with a wind output intensity of 3 m / s).

[0046] Through intelligent cleaning and control, the dust concentration was stabilized at 0.2 g / m³ from an initial 0.8 g / m³, and the power generation efficiency of the photovoltaic panels increased from 82% to 97%, an improvement of approximately 15%. Furthermore, the system's adaptive cleaning strategy avoids unnecessary frequent cleaning operations, effectively reducing energy consumption. Compared to traditional manual cleaning methods (cleaning twice a day, consuming 1.2 kWh each time), this system's daily cleaning energy consumption is only 0.84 kWh, saving approximately 30% of cleaning energy. Simultaneously, no on-site manual operation is required, reducing labor costs by approximately 30%, thereby improving the overall operating efficiency and economic benefits of the photovoltaic power station.

[0047] Example 2:

[0048] In this embodiment, addressing the snow removal issue at a photovoltaic power station in a cold region (latitude 45°N, average winter temperature -10°C) during snowfall, the system collects meteorological data, snow accumulation data, and photovoltaic panel contamination data every 15 minutes from 08:00 to 20:00. The specific data collected is as follows:

[0049] .

[0050] Based on the above data, the system determined the snow accumulation level of the photovoltaic panel to be "moderate snow accumulation" (snow thickness ≥ 5cm is considered moderate snow accumulation) and the pollution level to be "light pollution," thus initiating a level two cleaning priority and activating the automatic scraper system. The initial cleaning strategy was set as follows: cleaning frequency once every 2 hours, scraper operating speed 0.3 m / s, operating pressure 0.2 MPa (corresponding to equipment power 500W), and single cleaning duration 5 minutes.

[0051] During the cleaning process, the system dynamically optimizes its strategy every hour based on real-time feedback data.

[0052] (1) When the snow thickness drops to 5cm and the pollution concentration drops to 0.4 g / m² in the 4-hour average, the cleaning frequency is adjusted to once every 3 hours, the scraper running speed is reduced to 0.25 m / s, the operating pressure is reduced to 0.15 MPa (equipment power 350W), and the cleaning time is shortened to 3 minutes.

[0053] (2) When the snow thickness drops to 2 cm and the pollution concentration drops to 0.2 g / m² in the 8-hour average, the cleaning frequency is adjusted to once every 6 hours, the scraper running speed is maintained at 0.25 m / s, the operating pressure is 0.15 MPa, and the cleaning time is 3 minutes.

[0054] (3) When the snow thickness drops to 0cm and the pollution concentration drops to 0.1 g / m² in the 12-hour average, stop the scraper cleaning operation and switch to the regular monitoring mode (collect data once every 6 hours, and restart cleaning if the snow thickness is ≥2cm).

[0055] After 12 hours of intelligent cleaning, the snow on the photovoltaic panel surface was completely removed (snow thickness 0cm), the pollution concentration decreased from the initial 0.6 g / m² to 0.1 g / m², and the power generation efficiency increased from 65% to 95%, an improvement of approximately 30%. Compared with traditional manual cleaning methods (cleaning once a day, consuming 2.5 kWh of electricity per cleaning, and costing 200 yuan per cleaning), this intelligent cleaning system consumes only 1.625 kWh of energy per day, saving approximately 35% of energy consumption. Simultaneously, it eliminates the need for manual outdoor work, reducing labor costs by approximately 35%, and avoids excessive cleaning that could damage the panel coating (traditional manual scraping involves significant pressure fluctuations, easily causing scratches on the coating), thus improving the system's operational stability and economic efficiency.

Claims

1. A self-cleaning system for a mountain photovoltaic power station based on intelligent control, characterized in that, include: The system comprises a sensor module, an intelligent cleaning strategy control module, an intelligent decision-making module, and a cleaning equipment control module; among which: The sensor module is used to monitor weather conditions and pollution levels in real time and collect relevant data. This sensor module includes multiple sensors: a weather sensor, a pollution sensor, and a photovoltaic panel pollution monitoring sensor; the weather conditions include temperature, humidity, wind speed, precipitation, and light intensity; the pollution conditions include dust, sand, dirt, and snow; all collected data is transmitted in real time to the intelligent cleaning strategy control module for analysis; The intelligent cleaning strategy control module designs an intelligent cleaning strategy control algorithm based on data collected by the sensor module, that is, dynamically adjusts the cleaning strategy according to real-time weather and pollution conditions; the specific process is as follows: (1) Weather data analysis: Real-time analysis of meteorological data and prediction of future weather changes; including increasing cleaning frequency in hot and dry weather to avoid dust and sand accumulation; and automatically delaying cleaning in rainy weather to use rainfall to clean the panel naturally. (2) Pollutant analysis: The pollution status of the photovoltaic panel is monitored by pollution sensors; when pollutants, including dust, sand, sewage and snow, accumulate to a certain level, the cleaning program is automatically started and the most suitable cleaning method is selected for cleaning. (3) Dynamically adjust cleaning frequency: Based on weather forecasts and pollution monitoring data, dynamically adjust the frequency and method of cleaning; including increasing the cleaning frequency in environments with no precipitation for a long time; and delaying cleaning in high humidity or precipitation weather to reduce energy consumption for cleaning. (4) Cleaning priority: Based on the degree of pollution and weather changes, a cleaning priority shall be determined; including, in the case of sandstorms, priority shall be given to cleaning the panels to avoid the impact of dust on the solar panels; in the case of snow, the cleaning operation shall be determined according to the degree of snow accumulation. The intelligent decision-making module uses deep learning and data analysis to predict future cleaning needs based on historical data and current environmental conditions, and automatically adjusts the cleaning strategy. This includes automatically increasing the cleaning frequency if the power plant's power generation efficiency is low and the weather is dry and rainless for a certain period of time, so as to ensure that the panel does not reduce efficiency due to excessive accumulation of pollutants. The cleaning equipment control module regulates the operation of the cleaning equipment according to the cleaning strategy; including: (1) Spray system: In dry, highly polluted weather, surface dirt is removed by spray cleaning equipment; (2) Ultrasonic cleaning device: In dusty weather, the ultrasonic cleaning device is activated to remove dust by blowing it away with wind. (3) Automatic scraper system: When there is snow or hard dirt, the scraper equipment is activated for cleaning.

2. The self-cleaning system for mountain photovoltaic power stations based on intelligent regulation according to claim 1, characterized in that, The implementation employs a combination of models and algorithms, including decision tree models, support vector machine models, and neural network models. These models combine environmental data, photovoltaic panel status, and cleaning equipment control information to achieve intelligent regulation of the cleaning strategy. The following programming languages ​​and technology stacks are used for implementation: (1) Machine learning framework: Scikit-learn: Used to build and train traditional machine learning models, including support vector machines, decision trees, and random forest models; TensorFlow / Keras: Used for developing deep learning models, especially neural network models used in intelligent decision engines; XGBoost: Used to implement efficient gradient boosting decision trees (GBDT), optimize cleaning strategies and dynamically adjust decision processes; (2) Data processing and analysis: Pandas: Used for data processing, supporting the cleaning, merging, and analysis of raw data collected from sensor modules; NumPy: Used for numerical computation, optimizing data processing and feature engineering steps; (3) Real-time data communication and control: MQTT protocol: used for real-time data transmission and control command issuance in the clean system within the photovoltaic power plant, ensuring the system responds to environmental changes in real time; ROS: Real-time control of cleaning equipment, especially robotic devices that perform cleaning tasks; (4) Front-end and visualization platform: JavaScript / HTML / CSS: Used to develop visual interfaces and implement remote monitoring platforms, allowing power station managers to view cleaning status and adjust strategies; D3.js: Used for data visualization, enabling graphical display of real-time monitoring data.

Images