Photovoltaic visualization control method based on big data

Abstract

The invention discloses a photovoltaic visualization control method based on big data, and relates to the technical field of information, and the method comprises the following steps: obtaining module temperature parameters, and achieving temperature visualization through filtering smoothing; acquiring a photovoltaic panel image, and analyzing light intensity distribution to realize radiation resource distribution visualization; establishing a mapping relation between the temperature and radiation group member distribution and the historical generating capacity by using a linear regression prediction model, and predicting the generating capacity in a future period; performing fault early warning by using a user-based collaborative filtering algorithm; calculating the generating capacity of the photovoltaic station at different altitudes and inclination angles, and evaluating the theoretical generating upper limit of the station; based on a GIS map, an optimal site position is selected, a site layout is designed, and photovoltaic visual management is carried out; according to the method, global monitoring and scheduling are realized, and the efficiency and the management level of the photovoltaic power generation system are improved.

Description

Technical Field

[0001] The present invention relates to the field of information technology, and particularly to a photovoltaic visualization control method based on big data. Background Art

[0002] With the continuous expansion of the scale of photovoltaic power generation systems and the improvement of their intelligence level, the existing photovoltaic management methods face the following problems. First, the existing management methods cannot accurately and timely warn of fault modes, resulting in untimely handling of faults, which affects the stability and reliability of photovoltaic power generation. In addition, due to the failure to detect faults in a timely manner, the magnitude of fault risks cannot be evaluated, further affecting the operational safety of the system. Human factors may also lead to the failure to detect faults in a timely manner. For example, operation and maintenance personnel may not pay sufficient attention to the system operation status, or may overlook potential fault signs due to being busy with work. In addition, human errors, incorrect operations or negligence may also cause faults to occur. If a large amount of data collected in the system is not fully analyzed and processed, the signs of faults may be submerged in the vast amount of data. In this case, even if there are abnormal or warning signals, it is difficult to extract effective information from the data, thus unable to detect faults in a timely manner. Second, the existing methods lack comprehensive consideration of the influence of altitude and inclination angle on the power generation of photovoltaic sites, and cannot accurately evaluate the theoretical power generation upper limit of the sites. This leads to blindness in the site selection and design of photovoltaic fields, which may cause waste of resources and reduction of benefits.

[0003] For example, Chinese Patent CN117578721A, with a publication date of February 20, 2024, discloses a three-dimensional visualization system for a photovoltaic power station, belonging to the field of intelligent monitoring and management technology of photovoltaic power stations based on digital twins, and realizing the system field of integrated three-dimensional visualization monitoring and warning. It includes a solution for a data acquisition module, a design solution for an intelligent video analysis module, and a design solution for a photovoltaic panel fire prevention monitoring module. In the present invention, through technical means such as automated data acquisition, intelligent video analysis, and big data analysis, comprehensive monitoring and management of the photovoltaic power station are realized, enabling operation managers to obtain accurate data information in a timely manner and being able to actively handle abnormal situations, improving the operation efficiency and resource utilization rate of the photovoltaic power station. Through the invention of the intelligent monitoring and management technology of the photovoltaic power station, it helps to improve the utilization efficiency of clean energy and reduce the dependence on traditional energy. However, this method is not comprehensive enough and cannot perfectly achieve photovoltaic management. Summary of the Invention

[0004] The technical problem to be solved by the present invention is: the technical problem that the prior art cannot well realize photovoltaic management. A photovoltaic visualization control method based on big data is proposed, which can effectively realize photovoltaic management.

[0005] To solve the above technical problems, the technical solution adopted by the present invention is as follows: A photovoltaic visualization control method based on big data, comprising the following steps: S1: Obtain the module temperature parameter and realize temperature visualization through filtering and smoothing; S2: Obtain the photovoltaic panel image and analyze the light intensity distribution to realize the visualization of the radiation resource distribution; S3: Use a linear regression prediction model to establish a mapping relationship between the temperature, radiation component distribution and historical power generation, and predict the power generation in the future period; S4: Use a user-based collaborative filtering algorithm for fault warning; S5: Calculate the power generation of photovoltaic sites at different altitudes and inclinations, and evaluate the theoretical power generation upper limit of the sites; S6: Based on the GIS map, select the best site location, design the site layout, and perform photovoltaic visualization management.

[0006] Preferably, the step S1 includes the following content: Obtain the temperature signal through the change of the resistance value of the photosensitive resistor and convert it into a temperature value; if the converted data is abnormal and exceeds the set range, perform median filtering to remove the abnormal data; smooth the temperature curve through a moving average filter to eliminate noise; judge whether the smoothed data is within the normal range, and if not, set it to the default value; draw a temperature change curve according to the processed temperature data; judge whether the temperature data exceeds the threshold three times continuously to confirm whether the temperature sensor is faulty; if it is confirmed that there is a fault, switch to the standby photosensitive resistor and continue to obtain the temperature parameter.

[0007] Preferably, the step S2 includes the following content: Use a camera to obtain the real-time image of the photovoltaic panel and perform denoising and calibration preprocessing; on the preprocessed image, use the Sobel edge detection algorithm to obtain a comprehensive edge intensity image; use the SIFT feature extraction algorithm to extract key points and descriptors in the comprehensive edge image, perform feature matching or clustering on the descriptors of the key points to obtain the light intensity distribution information; based on the obtained light intensity features, calculate the irradiance intensity per unit area of each region; perform color mapping on the irradiance intensity per unit area to realize the visual display of the light intensity distribution.

[0008] Preferably, the specific steps of using the Sobel edge detection algorithm to obtain a comprehensive edge intensity image are as follows: Apply the Sobel operator in the horizontal and vertical directions to perform convolution operations on the preprocessed image respectively to obtain the edge intensity images in the horizontal and vertical directions, merge the edge intensity images in the horizontal and vertical directions into a single edge intensity image, and obtain a comprehensive edge intensity image by taking the square root of the sum of squares.

[0009] Preferably, the step S3 includes the following: using the acquired sample data as input, constructing a linear regression prediction model between temperature and power generation to obtain the mapping relationship between power generation and temperature and radiation resource distribution; based on the existing temperature data and radiation resource distribution data, combining with the linear regression prediction model, predicting the predicted power generation in a future period; comparing the predicted power generation with the actual power generation to calculate the prediction error; and evaluating the credibility of the prediction result according to the accuracy of the linear regression model and the stability of historical data.

[0010] Preferably, the step S4 includes the following: obtaining historical temperature and power generation data, constructing a collaborative filtering model in combination with the data characteristics of temperature and power generation to establish a user-item matrix; using the historical temperature and power generation data to train the collaborative filtering model and evaluating the model using cross-validation; performing association rule mining on the trained collaborative filtering model to determine the association rules between temperature and power generation; obtaining the correlation pattern between temperature and power generation and the influence law of temperature anomalies on power generation by mining frequent item sets and association rules; identifying and classifying fault modes according to the mined association rules; according to the determined fault modes, sending out warning signals in a timely manner and visually displaying the degree of fault risk to the operators.

[0011] Preferably, the step S5 includes the following: obtaining the altitude and tilt data of the photovoltaic site and the actual power generation of the corresponding photovoltaic site; visually displaying the actual power generation of the photovoltaic sites at different altitudes and tilts using a line chart or a scatter plot; plotting a heat map or a bubble chart to show the power generation distribution at different altitudes and tilts; counting the number of sites at different altitudes and tilts and showing the proportion of the number of sites at each altitude or tilt in the form of a pie chart or a bar chart to further analyze the site distribution; according to the results of the heat map or the bubble chart, selecting the combination of altitude and tilt with the best power generation; setting simulation parameters according to the selected best combination to calculate the theoretical power generation; comparing the calculated power generation with the original power generation to evaluate the improvement of power generation; and making further optimizations and adjustments according to the evaluation results to improve the power generation efficiency of the photovoltaic site.

[0012] Preferably, the step S6 includes the following: evaluating the illumination and terrain indicators based on the GIS map to determine the ideal site selection area for the photovoltaic site; selecting the optimal area to layout the photovoltaic power station site according to the evaluation results; collecting and summarizing the real-time monitoring data of the photovoltaic power station for global monitoring; dynamically predicting and optimizing the system operation plan based on the monitoring data, and automatically issuing control instructions through self-analysis and judgment.

[0013] Preferably, the linear regression prediction model is that the predicted power generation is equal to the sum of 0.6 times the temperature and 0.05 times the radiation resource distribution minus 10; the prediction error calculation formula is that the prediction error is equal to the difference between the predicted power generation and the actual power generation.

[0014] Preferably, the photovoltaic visualization management adopts a modular and microservices architecture, utilizes open RESTful interfaces, and supports the access and use of third-party algorithms and modes.

[0015] The substantial effect of the present invention is that the present invention designs a photovoltaic visualization control method based on big data. This method analyzes the association rules between temperature and power generation data, timely warns of fault modes, and gives visual prompts for fault risks; this method calculates the power generation of photovoltaic sites at different altitudes and inclinations, evaluates the theoretical power generation upper limit of the sites, and conducts site resource analysis through graphical expression; based on the GIS map, selects the best site location, and designs the site layout to achieve photovoltaic visualization management; by integrating all technologies, this method realizes global monitoring and scheduling, and improves the efficiency and management level of the photovoltaic power generation system. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 is a flowchart of a photovoltaic control of the present invention.

[0017] Figure 2 is a flowchart of a temperature signal processing of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0018] The following will further specifically describe the specific embodiments of the present invention through specific embodiments and in conjunction with the drawings.

[0019] Embodiment 1: A photovoltaic visualization control method based on big data, as Figure 1As shown, first, the temperature parameters of the module are obtained in real time, filtered and smoothed to achieve visual monitoring of the temperature. The temperature signal is obtained through the change in the resistance value of the photoresistor and converted into a temperature value. If the converted data is abnormal and exceeds the set range, median filtering is performed to remove the abnormal data. The temperature curve is smoothed by a moving average filter to eliminate noise. It is judged whether the smoothed data is within the normal range, and if not, it is set to the default value. According to the processed temperature data, a temperature change curve graph is drawn to achieve visual monitoring of the temperature change. By judging whether the temperature data exceeds the threshold value three times continuously, it is confirmed whether the temperature sensor is faulty. If a fault is confirmed, the standby photoresistor is switched to continue obtaining the temperature parameters. For example, the original temperature data obtained is [15.3, 12.1, 33.2, 23.5], and the converted temperature range is set to 0 - 30 °C. By judging, it is found that the third data point 33.2 °C exceeds the normal range, so the 5-point median filter is used to replace this point with the median value of 15.3 °C. Then, the simple moving average filter is used to calculate the 5-point moving average [14.2, 13.9, 15.3, 23.0, 23.5], and the abnormal jumps in the smoothed temperature curve are eliminated. It is judged that the filtered data is within the normal range of 0 - 30 °C. Then, based on the filtered data, the matplotlib library is used to draw a temperature change line graph, with the horizontal axis being time and the vertical axis being the temperature value, to achieve the visualization of the temperature change process of the photovoltaic system. By judging that the temperature data such as [32.5, 33.6, 34.2] exceeds the set key temperature of 30 °C three times continuously, it can be judged that the temperature probe is faulty, and then the standby PT100 chip is switched in real time for temperature backup detection, and the new temperature data is written into the visual monitoring system to make the system work continuously and avoid monitoring interruption.

[0020] The camera is used to obtain the image of the photovoltaic panel, analyze the light intensity distribution, obtain the irradiance intensity per unit area, and visualize the radiation resource distribution. The camera is used to obtain the real-time image of the photovoltaic panel and perform preprocessing such as denoising and calibration. On the preprocessed image, the Sobel edge detection algorithm is used, and the Sobel operators in the horizontal and vertical directions are respectively applied for convolution operations to obtain the edge intensity images in the horizontal and vertical directions. The edge intensity images in the horizontal and vertical directions are combined into a single edge intensity image, usually by calculating the sum of squares and then taking the square root to obtain the comprehensive edge intensity image. The SIFT feature extraction algorithm is used to extract key points and descriptors in the comprehensive edge image, and the descriptors of the key points are feature-matched or clustered to obtain the light intensity distribution information, including the light intensity change around the key points, the intensity and direction of the edges. Based on the obtained light intensity features, the irradiance intensity per unit area of each region is calculated. The irradiance intensity per unit area is color-mapped to achieve the visual display of the light intensity distribution. A standardized data interface is provided to support integration with other system development.

[0021] Adjust the lens parameters to ensure the stability of image acquisition quality. For example, the pixel size of the real-time image of the photovoltaic panel obtained by the camera is 1280x720. After denoising and calibration preprocessing, the preprocessed image is obtained. Perform a convolution operation using the Sobel operator in the horizontal direction to obtain the edge intensity image in the horizontal direction. At a certain pixel point P in the image, the convolution result of the Sobel operator in the horizontal direction is -10. Similarly, perform a convolution operation using the Sobel operator in the vertical direction to obtain the edge intensity image in the vertical direction. At the pixel point P in the image, the convolution result of the Sobel operator in the vertical direction is 5. Combine the edge intensity images in the horizontal and vertical directions into a single edge intensity image, and obtain the comprehensive edge intensity image by taking the square root of the sum of squares. At the pixel point P in the image, the comprehensive edge intensity is 15. Use the SIFT feature extraction algorithm to extract key points and descriptors in the comprehensive edge image. 10 key points are extracted in the image, and their descriptors are calculated. Through feature matching or clustering, obtain the light intensity distribution information. During the feature matching process, 4 key points are matched, and their light intensity information is obtained. Based on the obtained light intensity features, calculate the irradiance intensity per unit area of each region. In a certain region, the light intensities of the 4 key points are 10, 15, 20, and 25 respectively. Map the irradiance intensity per unit area to colors to achieve a visual display of the light intensity distribution. According to a certain mapping rule, map 10 to red, 15 to green, 20 to blue, and 25 to yellow. Provide a standardized data interface to support integration with other system development. Adjust the lens parameters to ensure the stability of image acquisition quality. Adjust parameters such as exposure time and focus distance to obtain clear and accurate images.

[0022] Use a linear regression prediction model to establish a mapping relationship between temperature, radiation resource distribution, and historical power generation data, realize the prediction of power generation in a future period of time, and display the prediction results. Obtain datasets of temperature, radiation resource distribution, and historical power generation as sample data. Use the obtained sample data as input to construct a linear regression prediction model between temperature and power generation, and obtain the mapping relationship between power generation, temperature, and radiation resource distribution. According to the existing temperature data and radiation resource distribution data, combined with the linear regression prediction model, predict the predicted power generation in a future period of time. Compare the predicted power generation with the actual power generation, and calculate the prediction error. Evaluate the credibility of the prediction results based on the accuracy of the linear regression model and the stability of historical data. Display information such as time, temperature, radiation resource distribution, predicted power generation, and prediction error in a chart or report.

[0023] For example, a dataset of temperature, radiation resource distribution, and historical power generation in City A is obtained as sample data. This data includes the temperature, radiation resource distribution, and power generation per hour, and records 100 consecutive hours of data. First, a linear regression algorithm is used to construct a prediction model between temperature and power generation. The obtained model is: power generation = 0.6 * temperature + 0.05 * radiation resource distribution - 10. Now, it is desired to predict the power generation within the next 10 hours. According to the weather forecast data of City A, the predicted temperatures for the next 10 hours are 25°C, 26°C, 24°C, 27°C, 23°C, 24°C, 22°C, 23°C, 21°C, 20°C. The predicted radiation resource distributions are 300 watts per square meter, 350 watts per square meter, 320 watts per square meter, 380 watts per square meter, 310 watts per square meter, 330 watts per square meter, 300 watts per square meter, 290 watts per square meter, 280 watts per square meter, 270 watts per square meter. According to the linear regression model, the predicted power generation within the next 10 hours can be calculated. For the first hour, with a temperature of 25°C and a radiation resource distribution of 300 watts per square meter, the predicted power generation = 0.6 * 25 + 0.05 * 300 - 10 = 15 megawatts. Similarly, the predicted power generation for the next 9 hours can be calculated. Next, the predicted power generation is compared with the actual power generation to calculate the prediction error. The prediction error for each hour can be calculated, which is the difference between the predicted power generation and the actual power generation. For the first hour, the prediction error = predicted power generation - actual power generation = 15 megawatts - 12 megawatts = 3 megawatts. Similarly, the prediction errors for the next 9 hours can be calculated.

[0024] Finally, the credibility of the prediction results can be evaluated based on the accuracy of the linear regression model and the stability of the historical data. If the prediction error is small and close to zero, it indicates that the model has good accuracy. Additionally, if the gap between the historical power generation and the predicted power generation is small, it indicates that the model has a good fitting effect on the past data, thereby increasing the stability of the prediction results. The above data and results can be presented in the form of charts or reports. A line chart showing the change of power generation over time is drawn, and the predicted power generation and the actual power generation are marked on the chart. At the same time, a bar chart showing the change of prediction error over time is drawn. Intuitively display the prediction results and evaluate the accuracy and credibility of the model.

[0025] Use the user-based collaborative filtering algorithm to analyze the association rules between temperature and power generation data, complete the timely warning of fault modes, and provide visual cues for fault risks. Obtain historical temperature and power generation data, and construct a collaborative filtering model in combination with the data characteristics of temperature and power generation. Establish a user-item matrix, where users represent different temperature data, items represent different power generation data, and the elements in the matrix are the association degrees between the corresponding temperature and power generation. Use the historical temperature and power generation data to train the collaborative filtering model, and evaluate the model using cross-validation. Mine the association rules for the trained collaborative filtering model to determine the association rules between temperature and power generation. By mining frequent item sets and association rules, obtain the correlation patterns between temperature and power generation, as well as the influence law of temperature anomalies on power generation. Identify and classify the fault modes according to the mined association rules. According to the determined fault modes, issue warning signals in a timely manner, and visually display the degree of fault risk to the operators. Continuously monitor the temperature and power generation data to ensure the accuracy and timeliness of warnings and cues.

[0026] For example, there is a dataset of historical temperature and power generation. The dataset contains the temperature on 100 different dates and the corresponding power generation. These data can be used to build a collaborative filtering model. First, a user-item matrix needs to be established. The users are different temperature data, and the items are different power generation data. The elements in the matrix represent the degree of association between the corresponding temperature and power generation. If 10 temperature data and 10 power generation data are selected, the size of the user-item matrix will be 10x10, with a total of 100 elements. The historical temperature and power generation data can be used to fill this matrix and calculate the corresponding degree of association. Next, this matrix can be used to train the collaborative filtering model. During the training process, the model will learn the association rules between temperature and power generation. The model will find the correlation patterns and influence laws between temperature and power generation according to the data characteristics in the matrix. Once the model training is completed, cross-validation can be used to evaluate the model. The dataset can be divided into a training set and a test set. The training set is used to train the model, and then the test set is used to evaluate the accuracy and performance of the model. Through association rule mining, the association rules between temperature and power generation can be determined. It may be found that when the temperature rises, the power generation increases; or when the temperature drops to a certain extent, the power generation decreases. According to the mined association rules, fault patterns can be identified and classified. If the temperature rises or drops abnormally to a certain extent, and the power generation does not change accordingly, then there may be a fault. Once the fault pattern is identified, a warning signal can be sent, and the degree of fault risk can be visually displayed to the operator. Color coding or charts can be used to represent the severity of the fault. Finally, the temperature and power generation data need to be continuously monitored to ensure the accuracy and timeliness of the warning and prompt. The dataset can be updated daily or hourly, and the model can be retrained to maintain the accuracy of the model.

[0027] Calculate the power generation of photovoltaic sites at different altitudes and inclinations, evaluate the theoretical upper limit of power generation at the sites, and conduct graphical representation of site resource analysis. Obtain the altitude and inclination data of the photovoltaic sites and the actual power generation of the corresponding photovoltaic sites. Visualize the actual power generation of photovoltaic sites at different altitudes and inclinations, using line charts or scatter plots, with altitude as the horizontal axis and power generation as the vertical axis, to show the variation of power generation at different inclinations. Use altitude and inclination as the coordinate axes and power generation as the color intensity or bubble size to draw a heat map or bubble chart to show the power generation distribution at different altitudes and inclinations. Count the number of sites at different altitudes and inclinations, and show the proportion of the number of sites at each altitude or inclination in the form of a pie chart or bar chart to further analyze the site distribution. According to the results of the heat map or bubble chart, select the combination of altitude and inclination with the best power generation. According to the selected best combination, set simulation parameters, including the conversion efficiency of the photovoltaic panel and the atmospheric attenuation coefficient, and conduct theoretical power generation calculation. Compare the calculated power generation with the original power generation to evaluate the improvement of power generation. According to the evaluation results, further optimization and adjustment can be carried out to improve the power generation efficiency of the photovoltaic site.

[0028] For example, obtain the altitude and inclination data of the photovoltaic sites and compare the actual power generation at different inclinations. The following data is collected: Photovoltaic site A has an altitude of 1000 meters, an inclination of 30 degrees, and an actual power generation of 2000 kWh; Photovoltaic site B has an altitude of 1200 meters, an inclination of 30 degrees, and an actual power generation of 1800 kWh; Photovoltaic site C has an altitude of 1000 meters, an inclination of 40 degrees, and an actual power generation of 2200 kWh; Photovoltaic site D has an altitude of 1200 meters, an inclination of 40 degrees, and an actual power generation of 2100 kWh. Use a line chart or scatter plot to visually display the variation of power generation at different inclinations. The horizontal axis represents altitude, and the vertical axis represents power generation. Each point represents a photovoltaic site. There are two points on the horizontal axis representing altitudes of 1000 meters and 1200 meters respectively, and the points on the vertical axis represent the actual power generation. Among them, the two points at an inclination of 30 degrees represent 2000 kWh and 1800 kWh respectively, and the two points at an inclination of 40 degrees represent 2200 kWh and 2100 kWh respectively. In addition, a heat map or bubble chart can also be used to show the power generation distribution at different altitudes and inclinations. The values on the coordinate axes represent altitude and inclination, and the color intensity or bubble size represents the power generation amount. The bubble size corresponding to the position of 1000 meters altitude and 30 degrees inclination represents an actual power generation of 2000 kWh, and the bubble size corresponding to the position of 1200 meters altitude and 40 degrees inclination represents an actual power generation of 2100 kWh. To further analyze the site distribution, count the number of sites at different altitudes and inclinations, and use a pie chart or bar chart to show the proportion of the number of sites at each altitude or inclination.

[0029] The number of stations at an altitude of 1000 meters accounts for 50% of the total number of stations, and the number of stations at an altitude of 1200 meters accounts for 50% of the total number of stations. According to the results of the heat map or bubble chart, select the combination of the best altitude and inclination angle for power generation. The power generation of the station at an altitude of 1200 meters and an inclination angle of 40 degrees is relatively high, and this can be considered a relatively good combination. According to the selected best combination, set simulation parameters, such as the conversion efficiency of the photovoltaic panel and the atmospheric attenuation coefficient, and calculate the theoretical power generation. With a conversion efficiency of 90% and an atmospheric attenuation coefficient of 0.8, the calculated theoretical power generation under this combination is 2400 kWh. Next, compare the calculated power generation with the original power generation to evaluate the improvement of power generation. The original power generation was 2100 kWh, while the calculated theoretical power generation is 2400 kWh, and the power generation has increased by 14%. According to the evaluation results, further optimization and adjustment are carried out, such as changing the inclination angle or altitude, to improve the power generation efficiency of the photovoltaic station.

[0030] Based on the GIS map, select the best site location, design the station layout, conduct photovoltaic visualization management, and achieve global monitoring and scheduling. Based on the GIS map, evaluate the illumination and terrain indicators to determine the ideal site selection area for the photovoltaic station. According to the evaluation results, select the optimal area to layout the photovoltaic power station sites. Collect and summarize the real-time monitoring data of the photovoltaic power station for global monitoring. Based on the monitoring data, dynamically predict and optimize the system operation plan, and automatically issue control instructions through independent analysis and judgment. Adopt a microservice system architecture to achieve flexible configuration and function expansion, support data docking with third-party systems, and ensure safety and reliability. Evaluate the system effect and continuously iterate and update the model algorithm. Continuously improve the power station construction plan to ensure efficient use of resources. For example, use the ArcGIS geographic information system software to conduct spatial analysis and evaluation modeling on indicators such as the topography and landform, and sunshine duration of a certain area, judge the best site selection area for the photovoltaic station, and mark it on the map. Layout 2 photovoltaic power stations with a scale of 100 billion watts in the planned area, named Station A and Station B respectively. Install sensors such as light intensity, temperature and humidity, and video inside and at the boundary of each power station to monitor the operation status of the power station in real time.

[0031] These monitoring data are transmitted to the intelligent management platform in real time through the Internet, and are marked, stored, analyzed and processed. The platform software can generate a real-time visual monitoring interface for the operation of the power station. Based on the analysis of historical operation data using statistical models, a prediction model for the power generation of the power station is established. Then, combined with external factors such as weather and electricity price, the neural network algorithm is used to dynamically determine the optimal operation plan for the next moment. The platform will automatically issue control instructions, such as adjusting the Rotate angle of some photovoltaic modules, to optimize the power generation efficiency of the system in real time. At the same time, a modular and microservices architecture is adopted, and the open RESTful interface is used to support the access and use of third-party algorithms and models. And based on the OPCUA protocol, data interaction is carried out with the power trading center and the meteorological information system. The management and control effects of the system are also regularly evaluated, and the model is continuously optimized using incremental learning to ensure the continuous and efficient operation of the power station.

[0032] The above embodiments only represent several implementation manners of the present invention, and the description thereof is relatively specific and detailed, but it should not be construed as a limitation on the scope of the invention patent. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present invention, several modifications and improvements can still be made, and these all belong to the protection scope of the present invention.

Claims

1. A photovoltaic visualization control method based on big data, characterized in that It includes the following steps: S1: Obtain the module temperature parameter and achieve temperature visualization through filtering and smoothing; S2: Obtain the photovoltaic panel image and analyze the light intensity distribution to achieve the visualization of radiation resource distribution; S3: Use the linear regression prediction model to establish the mapping relationship between the temperature, radiation component distribution and historical power generation, and predict the power generation in the future period; S4: Use the user-based collaborative filtering algorithm for fault warning; S5: Calculate the power generation of photovoltaic sites at different altitudes and inclinations, and evaluate the theoretical power generation upper limit of the sites; S6: Based on the GIS map, select the best site location, design the site layout, and conduct photovoltaic visualization management.

2. The photovoltaic visualization control method based on big data according to claim 1, wherein The step S1 includes the following contents: Obtain the temperature signal through the change of the resistance value of the photoresistor and convert it into a temperature value; if the converted data is abnormal and exceeds the set range, perform median filtering to remove the abnormal data; smooth the temperature curve through a moving average filter to eliminate noise; judge whether the smoothed data is within the normal range, and if not, set it to the default value; draw a temperature change curve according to the processed temperature data; judge whether the temperature data exceeds the threshold three times continuously to confirm whether the temperature sensor is faulty; if the fault is confirmed, switch to the backup photoresistor and continue to obtain the temperature parameter.

3. A photovoltaic visualization control method based on big data according to claim 1, characterized in that, The step S2 includes the following contents: Use the camera to obtain the real-time image of the photovoltaic panel and perform preprocessing of denoising and correction; on the preprocessed image, use the Sobel edge detection algorithm to obtain the comprehensive edge intensity image; use the SIFT feature extraction algorithm to extract key points and descriptors in the comprehensive edge image, perform feature matching or clustering on the descriptors of the key points to obtain the light intensity distribution information; based on the obtained light intensity features, calculate the irradiance intensity per unit area of each region; Perform color mapping on the irradiance intensity per unit area to achieve the visual display of the light intensity distribution.

4. A photovoltaic visualization control method based on big data according to claim 3, characterized in that, The specific steps of using the Sobel edge detection algorithm to obtain the comprehensive edge intensity image are as follows: Apply the Sobel operators in the horizontal and vertical directions to perform convolution operations on the preprocessed image respectively to obtain the horizontal and vertical edge intensity images, merge the horizontal and vertical edge intensity images into a single edge intensity image, and obtain the comprehensive edge intensity image by taking the square root of the sum of squares.

5. A photovoltaic visualization control method based on big data according to claim 1, characterized in that, The step S3 includes the following contents: Take the obtained sample data as input, construct a linear regression prediction model between the temperature and the power generation, and obtain the mapping relationship between the power generation and the temperature and radiation resource distribution; according to the existing temperature data and radiation resource distribution data, combine with the linear regression prediction model to predict the predicted power generation in the future period; compare the predicted power generation with the actual power generation and calculate the prediction error; Evaluate the credibility of the prediction result according to the accuracy of the linear regression model and the stability of the historical data.

6. A photovoltaic visualization control method based on big data according to claim 1 or 2 or 3, characterized in that, Step S4 includes the following: Obtain historical temperature and power generation data, construct a collaborative filtering model by combining the data characteristics of temperature and power generation, and establish a user-item matrix; Use the historical temperature and power generation data to train the collaborative filtering model, and evaluate the model using cross-validation; Conduct association rule mining on the trained collaborative filtering model to determine the association rules between temperature and power generation; By mining frequent itemsets and association rules, obtain the correlation pattern between temperature and power generation, as well as the influence law of temperature anomalies on power generation; Identify and classify fault modes according to the mined association rules; According to the determined fault modes, issue early warning signals in a timely manner, and visually display the degree of fault risk to the operators.

7. A photovoltaic visualization control method based on big data according to claim 1 or 2 or 3, characterized in that, Step S5 includes the following: Obtain the altitude and inclination data of the photovoltaic site and the actual power generation of the corresponding photovoltaic site; Visualize the actual power generation of photovoltaic sites at different altitudes and inclinations using line charts or scatter plots; Draw a heat map or bubble chart to show the power generation distribution at different altitudes and inclinations; Count the number of sites at different altitudes and inclinations, and display the proportion of the number of sites at each altitude or inclination in the form of a pie chart or bar chart to further analyze the site distribution; According to the results of the heat map or bubble chart, select the combination of altitude and inclination with the best power generation; According to the selected best combination, set simulation parameters to calculate the theoretical power generation; Compare the calculated power generation with the original power generation to evaluate the improvement of power generation; According to the evaluation results, conduct further optimization and adjustment to improve the power generation efficiency of the photovoltaic site.

8. A photovoltaic visualization control method based on big data according to claim 1 or 2 or 3, characterized in that, Step S6 includes the following: Based on the GIS map, evaluate the light and terrain indicators to determine the ideal site selection area for the photovoltaic site; According to the evaluation results, select the optimal area to layout the photovoltaic power station site; Collect and summarize the real-time monitoring data of the photovoltaic power station for global monitoring; Based on the monitoring data, dynamically predict and optimize the system operation plan, and automatically issue control instructions through independent analysis and judgment.

9. A photovoltaic visualization control method based on big data according to claim 5, characterized in that The linear regression prediction model is that the predicted power generation is equal to the difference obtained by subtracting 10 from the sum of 0.6 times the temperature and 0.05 times the radiation resource distribution; The prediction error calculation formula is that the prediction error is equal to the difference between the predicted power generation and the actual power generation.

10. A photovoltaic visualization control method based on big data according to claim 1 or 2 or 3, characterized in that, Photovoltaic visualization management adopts a modular and microservices architecture, and uses open RESTful interfaces to support the access and use of third-party algorithms and models.

Images