Solar energy generating capacity prediction method, system and device and storage medium

Abstract

The invention relates to a solar generating capacity prediction method, system and device and a storage medium, and the method comprises the steps: obtaining real-time illumination data through a geographic information system platform, and carrying out the localization correction of the illumination data through combining with the geographic environment of a target device, and obtaining the corrected illumination data; acquiring real-time meteorological data, and performing comparative calculation based on the real-time meteorological data and the historical illumination data to generate an illumination data correction factor; based on the correction factor, adjusting deviation in the corrected illumination data to obtain optimized illumination data; predicting the optimized illumination data by using a convolutional neural network, and generating an illumination intensity prediction value in a preset time period; and inputting the illumination intensity prediction value into the solar power generation prediction model, and calculating to obtain predicted solar power generation. The method has the effect of improving the generating capacity prediction precision of the portable solar equipment.

Description

Technical Field

[0001] The present application relates to the technical field of solar photovoltaic energy management, and particularly to a method, system, device and storage medium for predicting solar power generation. Background Art

[0002] Currently, with the rise of green energy, solar power generation technology has been widely applied in various scenarios, especially in small and portable solar devices. Such devices are usually used in households, outdoor camping, mobile communication, remote monitoring and other occasions, and have gradually become the first choice of users due to their convenience and mobility.

[0003] Existing small solar devices, especially portable and lightweight devices, often rely on simplified light intensity prediction models. These devices usually adopt simple estimation methods based on meteorological data, ignoring the dynamic changes of device location and environmental factors, such as the influence of geographical location, altitude, meteorological conditions, etc. on light intensity. In addition, traditional light intensity prediction methods usually rely on historical data and fixed models, and are easily interfered by weather fluctuations and environmental changes in actual use, resulting in large errors in prediction results.

[0004] The above-mentioned prior art solutions have the following defects: The light prediction method of existing small solar devices fails to fully consider the changes in device location and the influence of real-time meteorological conditions on light intensity. Especially when the device moves, there is a lack of a dynamic adjustment mechanism, resulting in poor accuracy of light intensity prediction. Therefore, there is room for improvement. Summary of the Invention

[0005] In order to improve the prediction accuracy of power generation of lightweight solar devices, the present application provides a method, system, device and storage medium for predicting solar power generation.

[0006] The first invention object of the present application is achieved through the following technical solutions: A method for predicting solar power generation, the method for predicting solar power generation includes: Obtain real-time light data by using a geographic information system platform, and perform local correction on the light data in combination with the geographical environment where the target device is located to obtain corrected light data; Collect real-time meteorological data, and perform comparison calculation based on the real-time meteorological data and historical light data to generate a light data correction factor; Based on the correction factor, adjust the deviation in the corrected light data to obtain optimized light data; Use a convolutional neural network to predict the optimized light data and generate predicted light intensity values within a preset time period; input the predicted light intensity values into a solar power generation prediction model to calculate the predicted solar power generation.

[0007] By adopting the above technical solution, by using a geographic information system platform to obtain real-time light data and locally correcting the light data in combination with the geographical environment where the target device is located, the light data can more accurately reflect the light conditions in the actual environment where the device is located, thereby improving the reliability of the light data; by collecting real-time meteorological data and calculating and comparing based on the real-time meteorological data and historical light data to generate a light data correction factor, the light data can be dynamically supplemented and adjusted according to meteorological changes, thereby overcoming the problem that a static model cannot adapt to environmental changes; by adjusting the deviation in the corrected light data based on the correction factor, the errors caused by different environmental and meteorological conditions can be eliminated, and more accurate light data can be obtained, further improving the accuracy of subsequent predictions; by using a convolutional neural network to predict the optimized light data, the non-linear features in the data can be fully exploited to generate more accurate predicted light intensity values, thereby providing more accurate input data for power generation prediction; by inputting the predicted light intensity values into a solar power generation prediction model, seamless connection from light prediction to power generation can be achieved, and finally the predicted solar power generation can be accurately calculated, thereby effectively improving the usage efficiency and reliability of small solar devices.

[0008] In one example, this application can be further configured as: the use of a geographic information system platform to obtain real-time light data and locally correct the light data in combination with the geographical location where the target device is located, and the obtained corrected light data includes: Obtain the geographical information of the location where the target device is located, and the geographical information includes longitude, latitude, and altitude; Based on the geographical information, use the geographic information system platform to extract real-time light data corresponding to the location where the target device is located from a meteorological data source; According to the geographical environment where the target device is located, use a local correction algorithm to adjust the real-time light data to obtain the corrected light data.

[0009] By adopting the above technical solution, by obtaining the geographical information of the location where the target device is located, including longitude, latitude and altitude, the geographical location of the device can be accurately determined, providing basic information for subsequent acquisition and correction of light data; by using the geographical information system platform to extract the real-time light data corresponding to the location where the target device is located from the meteorological data source based on the geographical information, it can be ensured that the obtained light data has location pertinence, thus reflecting the actual light conditions in the area where the device is located; by adjusting the real-time light data according to the geographical environment where the target device is located by using the localization correction algorithm, the error of the light data caused by geographical environment differences can be eliminated, making the obtained corrected light data more accurate and reliable, so as to provide more accurate light input for subsequent power generation prediction.

[0010] In one example, the present application can be further configured as: collecting the real-time meteorological data, and comparing and calculating based on the real-time meteorological data and historical light data to generate a light data correction factor, including: Obtaining the real-time meteorological data corresponding to the location where the target device is located, where the meteorological data includes temperature, humidity, air pressure, wind speed and cloud cover; Obtaining the historical light data and the corresponding historical meteorological data in the previous period as a training set; Analyzing the correlation between the historical meteorological data and the historical light data in the training set through a regression analysis model, establishing a mathematical model between meteorological conditions and light changes, and obtaining a meteorological deviation and a regression analysis model; Based on the real-time meteorological data, predicting the future light change trend through the meteorological deviation and the regression analysis model to generate the light data correction factor.

[0011] By adopting the above technical solution, by obtaining the real-time meteorological data corresponding to the location where the target device is located, the meteorological factors affecting the light intensity can be comprehensively grasped, thus providing necessary meteorological support for light data correction; by obtaining the historical light data and the corresponding historical meteorological data in the previous period as a training set, sufficient historical data support can be provided for the regression analysis model to ensure the reliability and accuracy of the model when it is established; by analyzing the correlation between the historical meteorological data and the historical light data in the training set through the regression analysis model, a mathematical model between meteorological conditions and light changes can be established, thus revealing the influence law of meteorological conditions on the light intensity, and obtaining a meteorological deviation and a regression analysis model; by predicting the future light change trend based on the real-time meteorological data by using the meteorological deviation and the regression analysis model, the change of the light intensity can be predicted in advance, and the corresponding light data correction factor can be generated, so as to effectively adjust the deviation of the light data and improve the accuracy of light prediction.

[0012] In one example, the present application can be further configured as follows: Based on the real-time meteorological data, predicting the future light change trend through the meteorological deviation and regression analysis model, and generating the light data correction factor includes: Calculating the deviation between the real-time meteorological data and the historical meteorological data in the meteorological deviation and regression analysis model to obtain the meteorological deviation; Analyzing the correlation between the meteorological deviation and the historical light data in the meteorological deviation and regression analysis model to obtain the prediction result; Generating the corresponding light data correction factor based on the prediction result.

[0013] By adopting the above technical solution, by calculating the deviation between the real-time meteorological data and the historical meteorological data in the meteorological deviation and regression analysis model, the difference between the current meteorological conditions and the historical meteorological conditions can be quantified, so as to determine the influence degree of the meteorological conditions on the light data; by analyzing the correlation between the meteorological deviation and the historical light data in the meteorological deviation and regression analysis model, the specific influence of the meteorological change on the light intensity can be revealed, and the prediction result based on the meteorological data and the light data can be obtained; by generating the corresponding light data correction factor based on the prediction result, the deviation in the light data can be adjusted, and the accuracy of the light intensity prediction can be improved, so as to provide more accurate light data support for the subsequent prediction of the solar power generation.

[0014] In one example, the present application can be further configured as follows: Using the convolutional neural network to predict the optimized light data to generate the light intensity prediction value within a preset time period includes: Inputting the optimized light data into the input layer of the convolutional neural network, where the convolutional neural network includes at least one convolutional layer, pooling layer and fully connected layer; In the convolutional layer of the convolutional neural network, using the convolutional kernel to perform convolutional operation on the optimized light data to extract the local features in the optimized light data to obtain the feature map; In the pooling layer of the convolutional neural network, performing dimensionality reduction processing on the feature map to obtain the pooled feature map; In the fully connected layer of the convolutional neural network, performing non-linear transformation on the pooled feature map through the activation function: f(x) = max(0, x) to generate the light intensity prediction value within the preset time period.

[0015] By adopting the above technical solutions, by inputting the optimized light data into the input layer of the convolutional neural network, the light data can be converted into a format suitable for processing by the deep learning model, providing a basis for subsequent feature extraction and prediction. By using a convolutional kernel to perform a convolutional operation on the optimized light data in the convolutional layer, local features in the data can be effectively extracted to help the model identify key patterns and changing trends in the light data; by performing dimensionality reduction on the feature map in the pooling layer, the complexity and dimensionality of the data can be reduced, improving the computational efficiency while keeping the key feature information unchanged, thereby enhancing the generalization ability of the model; by performing a non-linear transformation on the pooled feature map in the fully connected layer, complex relationships in the data can be captured and accurate light intensity prediction values can be generated, providing accurate light intensity prediction results for the subsequent estimation of solar power generation.

[0016] In one example, the present application can be further configured as follows: inputting the light intensity prediction value into a solar power generation prediction model, and calculating the predicted solar power generation includes: Obtaining parameter data of the target device, where the parameter data includes the total area, efficiency, and temperature coefficient of the solar panel; Based on the light intensity prediction value and the parameter data of the target device, using the calculation formula in the solar power generation prediction model: P(t) = η·A·I(t)·(1 - β·(T - T ref )), the predicted solar power generation is obtained, where P(t) is the predicted solar power generation within time t, η is the conversion efficiency of the target device, A is the total area of the solar panel, I(t) is the light intensity prediction value at time t, β is the temperature coefficient, T is the real-time temperature of the solar panel, and T ref is the reference temperature.

[0017] By adopting the above technical solutions, by obtaining parameter data of the target device, including the total area, efficiency, and temperature coefficient of the solar panel, necessary device information can be provided for the estimation of solar power generation to ensure the accuracy of the calculation; by using the calculation formula in the solar power generation prediction model based on the light intensity prediction value and the device parameter data, the expected solar power generation can be calculated according to the real-time light intensity, device characteristics, and temperature information, thereby achieving an accurate prediction of future power generation, being able to more precisely reflect the actual power generation situation, and supporting optimized energy management and resource allocation.

[0018] The second inventive object of the present application is achieved by the following technical solutions: A solar power generation prediction system, the solar power generation prediction system includes: A correction module, configured to obtain real-time illumination data by using a geographic information system platform, and perform local correction on the illumination data in combination with the geographical environment where the target device is located to obtain corrected illumination data; A calculation module, configured to collect real-time meteorological data, and perform comparison and calculation based on the real-time meteorological data and historical illumination data to generate an illumination data correction factor; An optimization module, configured to adjust the deviation in the corrected illumination data based on the correction factor to obtain optimized illumination data; A prediction module, configured to use a convolutional neural network to predict the optimized illumination data to generate a predicted illumination intensity value within a preset time period; A power generation prediction module, configured to input the predicted illumination intensity value into a solar power generation prediction model to calculate the predicted solar power generation.

[0019] By adopting the above technical solution, by using a geographic information system platform to obtain real-time illumination data and performing local correction on the illumination data in combination with the geographical environment where the target device is located, the illumination data can more accurately reflect the illumination situation in the actual environment where the device is located, thereby improving the reliability of the illumination data; by collecting real-time meteorological data and performing comparison and calculation based on the real-time meteorological data and historical illumination data to generate an illumination data correction factor, the illumination data can be dynamically supplemented and adjusted according to meteorological changes, thereby overcoming the problem that a static model cannot adapt to environmental changes; by adjusting the deviation in the corrected illumination data based on the correction factor, the errors caused by different environments and meteorological conditions can be eliminated to obtain more accurate illumination data, further improving the accuracy of subsequent predictions; by using a convolutional neural network to predict the optimized illumination data, the non-linear features in the data can be fully mined to generate a more accurate predicted illumination intensity value, thereby providing more accurate input data for power generation prediction; by inputting the predicted illumination intensity value into a solar power generation prediction model, seamless connection from illumination prediction to power generation can be achieved, and finally the predicted solar power generation can be accurately calculated, thereby effectively improving the usage efficiency and reliability of small solar devices.

[0020] The above object three of the present application is achieved by the following technical solution: A computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the steps of the above solar power generation prediction method are implemented.

[0021] The above object four of the present application is achieved by the following technical solution: A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the above-mentioned solar power generation prediction method are implemented.

[0022] In summary, the present application includes the following beneficial technical effects: 1. By adopting the above technical solution, by using a geographic information system platform to obtain real-time light data and localizing and correcting the light data in combination with the geographical environment where the target device is located, the light data can more accurately reflect the light conditions in the actual environment where the device is located, thereby improving the reliability of the light data; by collecting real-time meteorological data and calculating and comparing the real-time meteorological data with historical light data to generate a light data correction factor, the light data can be dynamically supplemented and adjusted according to meteorological changes, thereby overcoming the problem that a static model cannot adapt to environmental changes. 2. By adjusting the deviation in the corrected light data based on the correction factor, the errors caused by different environmental and meteorological conditions can be eliminated, and more accurate light data can be obtained, further improving the accuracy of subsequent predictions; by using a convolutional neural network to predict the optimized light data, the non-linear features in the data can be fully exploited to generate a more accurate predicted light intensity value, thereby providing more accurate input data for power generation prediction; by inputting the predicted light intensity value into a solar power generation prediction model, seamless connection from light prediction to power generation can be achieved, and finally the predicted solar power generation can be accurately calculated, thereby effectively improving the usage efficiency and reliability of small solar devices. Description of the Drawings

[0023] Figure 1 is a flowchart of a solar power generation prediction method in an embodiment of the present application; Figure 2 is an implementation flowchart of step S10 in a solar power generation prediction method in an embodiment of the present application; Figure 3 is an implementation flowchart of step S20 in a solar power generation prediction method in an embodiment of the present application; Figure 4 is an implementation flowchart of step S24 in a solar power generation prediction method in an embodiment of the present application; Figure 5 is an implementation flowchart of step S40 in a solar power generation prediction method in an embodiment of the present application; Figure 6 is an implementation flowchart of step S50 in a solar power generation prediction method in an embodiment of the present application; Figure 7It is a principle block diagram of a solar power generation prediction system in an embodiment of the present application; Figure 8 It is a schematic diagram of the device in an embodiment of the present application. Specific embodiments

[0024] The present application will be further described in detail below with reference to the accompanying drawings.

[0025] In one embodiment, as Figure 1 shown, the present application discloses a method for predicting solar power generation, which specifically includes the following steps: S10: Obtain real-time illumination data using a geographic information system platform, and localize and correct the illumination data in combination with the geographical environment where the target device is located to obtain the corrected illumination data.

[0026] Specifically, first obtain the geographical information of the target device, such as the longitude, latitude, altitude, etc. of the device. Through this information, real-time illumination data corresponding to this location can be obtained from the geographic information system platform. These data are usually provided by satellites or ground meteorological stations and contain illumination intensity information for each time period. Next, according to the geographical environment characteristics of the target device, use a localization correction algorithm to adjust the illumination data, considering factors that affect the illumination intensity, such as the terrain (such as mountains, hills, etc.) in the area where the device is located, the occlusion of surrounding buildings, and weather conditions. In particular, altitude, terrain undulation, reflectivity, etc. will affect the transmission path of the illumination intensity, so as to adjust the data to make it more in line with the actual illumination intensity received by the target device, thereby obtaining the corrected illumination data.

[0027] S20: Collect real-time meteorological data, and generate an illumination data correction factor based on the comparison and calculation of the real-time meteorological data and historical illumination data.

[0028] Specifically, collect meteorological data of the area where the target device is located in real time through meteorological sensors or meteorological data services, including meteorological factors such as temperature, humidity, air pressure, wind speed, cloud cover, etc. Meteorological conditions directly affect the intensity and change of illumination. Real-time meteorological data provides context information for the change of illumination under the current weather conditions. Then, based on the comparison of historical illumination data and real-time meteorological data, use statistical methods or regression models to analyze the correlation between the two. Usually, historical meteorological data and illumination data are used as the training set, and a mathematical model is established to describe the relationship between meteorological changes and illumination changes. The regression analysis model helps to understand how meteorological factors affect the illumination intensity, thereby generating an illumination data correction factor to correct the deviation in the illumination data and improve its prediction accuracy.

[0029] S30: Based on the correction factor, adjust the deviation in the corrected illumination data to obtain optimized illumination data.

[0030] Specifically, the correction factor represents the magnitude and direction of the light intensity deviation under the current meteorological conditions. Based on this factor, the corrected light data is weighted and adjusted. The adjustment process is usually carried out through mathematical calculations. For example, the weighted average or linear regression model is used to combine the correction factor with the light data to adjust the corrected data so that it can more accurately reflect the actual light situation under the current meteorological conditions. The adjusted data is the optimized light data. This process can effectively reduce the errors caused by environmental factors, thereby improving the subsequent prediction accuracy. If a certain meteorological factor (such as a high cloud cover rate) causes a decrease in light intensity, the correction factor will take this into account and lower the light data. Conversely, it will be raised. This process can be carried out by setting thresholds or automatically adjusting the factor size, and according to the actual situation in a specific environment, the corrected data is precisely adjusted to make it closer to the true light intensity, and finally the optimized light data is obtained.

[0031] S40: Use a convolutional neural network to predict the optimized light data and generate the predicted light intensity values within a preset time period.

[0032] Specifically, the optimized light data is used as the input of the convolutional neural network. The convolutional neural network consists of multiple convolutional layers, pooling layers, and fully connected layers. First, in the convolutional layer, the convolutional kernel performs a convolutional operation on the optimized light data to extract the local features in the data. These local features include information such as the periodic changes of light and the influence of weather changes. The convolutional operation will generate a feature map. Then, in the pooling layer, the pooling operation reduces the dimension of the feature map. Common pooling methods include max pooling and average pooling. By this method, the dimension of the data is reduced, but the most important features are retained. Finally, in the fully connected layer, the pooled features are non-linearly transformed through activation functions such as ReLU or Sigmoid to generate the predicted light intensity values within a preset time period. The predicted values reflect the change trend of the light intensity in this area in the future time period, providing data support for the subsequent power generation prediction.

[0033] S50: Input the predicted light intensity values into the solar power generation prediction model to calculate the predicted solar power generation.

[0034] Specifically, the predicted light intensity value will be input into the solar power generation prediction model together with other parameters of the target device (such as the total area of the solar panels, conversion efficiency, temperature coefficient, etc.). This model is usually implemented through a calculation formula, where P(t) = η * A * I(t) * (1 - β * (T - Tref)). Here, P(t) represents the predicted solar power generation within time t, η is the conversion efficiency of the solar panels, A is the total area of the solar panels, I(t) is the predicted light intensity value at time t, β is the temperature coefficient, T is the real-time temperature, and Tref is the reference temperature. Based on these parameters, through calculation using the formula, the predicted solar power generation can be obtained, and finally, the power generation capacity of the solar system within a certain predetermined time period can be obtained, providing a basis for the use and optimization of the solar panels.

[0035] By adopting the above technical solution, by using the geographic information system platform to obtain real-time light data and localizing and correcting the light data in combination with the geographical environment where the target device is located, the light data can more accurately reflect the light conditions in the actual environment where the device is located, thereby improving the reliability of the light data; by collecting real-time meteorological data and calculating and comparing the real-time meteorological data with historical light data to generate a light data correction factor, the light data can be dynamically supplemented and adjusted according to meteorological changes, thereby overcoming the problem that the static model cannot adapt to environmental changes; by adjusting the deviation in the corrected light data based on the correction factor, the errors caused by different environmental and meteorological conditions can be eliminated, and more accurate light data can be obtained, further improving the accuracy of subsequent predictions; by using a convolutional neural network to predict the optimized light data, the non-linear features in the data can be fully exploited to generate a more accurate predicted light intensity value, thereby providing more accurate input data for power generation prediction; by inputting the predicted light intensity value into the solar power generation prediction model, seamless connection from light prediction to power generation can be achieved, and finally, the predicted solar power generation can be accurately calculated, thereby effectively improving the use efficiency and reliability of small solar devices.

[0036] In one embodiment, as Figure 2 shown, in step S10, that is, using the geographic information system platform to obtain real-time light data and localizing and correcting the light data in combination with the geographical location of the target device to obtain the corrected light data, specifically including: S11: Obtain the geographical information of the location where the target device is located, and the geographical information includes longitude, latitude, and altitude.

[0037] Specifically, first determine the specific location where the device is located through GPS positioning or other positioning means (such as Wi-Fi positioning, base station positioning, etc.), obtain the longitude and latitude information of this location, and further use the reference information provided by the measuring device or map data to determine the altitude of the device. The altitude is usually calculated by combining with the digital elevation model (DEM) in the geographic information system. These geographic information constitute the three-dimensional position data where the device is located, providing an accurate spatial reference for subsequent acquisition of light data. This information is crucial for subsequent adjustment of light data because different geographical locations and heights will have different impacts on light intensity. For example, the light intensity in areas with higher altitudes may be relatively stronger, and surrounding mountains or buildings may also block the light. S12: Based on the geographic information, use the geographic information system platform to extract the real-time light data corresponding to the location of the target device from the meteorological data source.

[0038] Specifically, according to the longitude, latitude and altitude information of the device, by calling the geographic information system platform, access the meteorological data source (such as satellite remote sensing data, meteorological station data or commercial meteorological service platform) to obtain the real-time light data of the location where the target device is located. The light data is usually collected and uploaded to the cloud in real time by meteorological stations or remote sensing satellites, covering information such as solar radiation intensity, sunshine duration, and climate factors (such as cloud cover). The light intensity data is calculated based on the intensity of solar radiation and weather conditions, usually expressed in watts per square meter (W / m 2 ) and the real-time light data will be updated according to different time periods and meteorological conditions. For example, when the device is located in an open area on a sunny day, the light intensity will be stronger, while in a cloudy or overcast environment, the light intensity will be weaker.

[0039] S13: According to the geographical environment where the target device is located, use the local correction algorithm to adjust the real-time light data to obtain the corrected light data.

[0040] Specifically, using the localization correction algorithm, the acquired real-time illumination data is corrected by analyzing the surrounding environmental factors at the location of the device. These environmental factors include the occlusion or reflection effects of terrain (such as mountains, hills), buildings, vegetation, etc. on illumination. For example, in mountainous areas, due to the occlusion of mountains, the direct illumination intensity may be reduced during certain time periods, or in places with higher altitudes, more sunlight may be received due to the thinner atmosphere. In addition, the shadows of surrounding buildings or trees also affect the illumination intensity. Therefore, it is necessary to use the localization correction algorithm to adjust these effects. The correction algorithm can be based on parameters such as the terrain data, building distribution, and vegetation coverage of the area, combined with the illumination data in the meteorological data source, and through techniques such as weighted average method, linear regression model, or interpolation method based on geographical information data, these factors are taken into account to make the corrected illumination data more in line with the actual illumination intensity received by the target device. For example, by calculating the influence of building shadows, the illumination values in the shadow areas are adjusted to more accurately reflect the actual illumination conditions of the target device.

[0041] In one embodiment, as Figure 3 shown, in step S20, that is, real-time meteorological data is collected, and based on the comparison and calculation of the real-time meteorological data and historical illumination data, an illumination data correction factor is generated, which specifically includes: S21: Obtain the real-time meteorological data corresponding to the location of the target device. The meteorological data includes temperature, humidity, air pressure, wind speed, and cloud cover.

[0042] Specifically, by obtaining the geographical location of the target device (such as longitude and latitude information), the real-time meteorological data corresponding to the location of the target device can be obtained from meteorological data service providers (such as meteorological bureaus, meteorological satellite data platforms, or meteorological sensor networks). These meteorological data are usually updated in real time to provide the most accurate weather conditions. The real-time meteorological data includes temperature (in degrees Celsius), humidity (relative humidity), air pressure (commonly sea-level air pressure), wind speed (in meters per second), and cloud cover (expressed as a percentage). For example, through the weather forecast API, these data can be obtained within hourly or minute intervals, thus providing the necessary meteorological input conditions for subsequent illumination intensity prediction. During the acquisition process of meteorological data, the geographical characteristics of the target device location are usually taken into account, such as the influence of altitude on temperature and air pressure, etc.

[0043] S22: Obtain the historical illumination data and the corresponding historical meteorological data for the previous period as the training set.

[0044] Specifically, by obtaining historical light data, using a meteorological service platform or historical meteorological data sources, and combining timestamp information, historical meteorological data and light data corresponding to the location and time period of the target device are extracted. These data usually cover light intensity data (unit: lx) within the past several days or hours and meteorological data (temperature, humidity, air pressure, wind speed, cloud cover, etc.) during the corresponding periods. For example, if the target device is located in a certain city, hourly meteorological data for the past week can be obtained through the meteorological service platform, and the corresponding light intensity data for these time periods can be extracted accordingly. These data can be used as the input training set for the regression model. The historical light data and historical meteorological data form a time series data set containing multi-dimensional features. The construction of the training set needs to ensure the integrity of data quality and the alignment of time for subsequent correlation analysis between meteorology and light.

[0045] S23: Analyze the correlation between the historical meteorological data and historical light data in the training set through a regression analysis model, establish a mathematical model between meteorological conditions and light changes, and obtain the meteorological deviation and the regression analysis model.

[0046] Specifically, analyze the correlation between the historical meteorological data and historical light data in the training set through a regression analysis model, establish a mathematical model between meteorological conditions and light changes, and obtain the meteorological deviation and the regression analysis model. Adopt regression analysis methods (such as linear regression, polynomial regression, or support vector machine regression, etc.), use the historical meteorological data in the training set as the independent variable and the historical light data as the dependent variable. The goal of regression analysis is to establish a mathematical model between meteorological factors and light changes. For example, the model can predict the change trend of light intensity under given meteorological conditions such as temperature and humidity. Specifically, through regression analysis of the training set, a regression equation can be obtained. For example, light intensity (Y) = a × temperature (T) + b × humidity (H) + c × air pressure (P) + d × wind speed (W), where a, b, c, and d are regression coefficients. These regression coefficients are automatically adjusted by the algorithm and optimized by the least squares method to make the prediction results of the model closest to the historical data, and finally the meteorological deviation and the regression analysis model are obtained.

[0047] S24: Based on real-time meteorological data, predict the future light change trend through the meteorological deviation and the regression analysis model, and generate a light data correction factor.

[0048] Specifically, the obtained real-time meteorological data (e.g., current temperature, humidity, air pressure, etc.) are input into the established regression analysis model. The model makes predictions based on these real-time input data to generate a predicted value of the light intensity at the current time point. Then, this predicted value is compared with the actually measured light data to calculate the prediction error. Subsequently, the light intensity is corrected according to the meteorological deviation (i.e., the residual of the model) to generate a correction factor. The light data correction factor is the corrected value of the light intensity obtained through the model. This correction factor specifically reflects the impact of changes in meteorological conditions on the light intensity. For example, when the temperature is relatively high, the light intensity may be stronger than expected, while when the cloud cover increases, the light intensity may be lower. The correction factor can be expressed as a deviation adjustment amount, such as the difference between the predicted light intensity value and the corrected light intensity value. Based on these factors, the light intensity data can be adjusted more accurately, making the prediction of light data more in line with the actual situation.

[0049] In one embodiment, as Figure 4 shown, in step S24, that is, based on the real-time meteorological data, the future light change trend is predicted through the meteorological deviation and the regression analysis model to generate a light data correction factor, which specifically includes: S241: Calculate the deviation between the real-time meteorological data and the historical meteorological data in the meteorological deviation and the regression analysis model to obtain the meteorological deviation.

[0050] Specifically, by calculating the deviation between the real-time meteorological data and the historical meteorological data in the meteorological deviation and the regression analysis model, the real-time meteorological data includes T current (t), H current (t), P current (t), W current (t), C current (t), and the historical meteorological data comes from the records of a previous period, such as the data in the past week or month. These historical data are T historical (t), H historical (t), P historical (t), W historical (t), C historical (t). By calculating the deviation of each item of meteorological data, that is, the difference between the real-time meteorological data and the historical meteorological data, for example: ΔT(t) = T current (t) - T historical (t); ΔH(t) = H current (t) - H historical (t); ΔP(t) = P current (t) - P historical (t); ΔW(t) = W current (t) - W historical (t); ΔC(t) = Ccurrent (t)-C historical (t), these deviations reflect the differences between real-time meteorological conditions and historical conditions, and these differences will be used for subsequent calibration of the prediction model to help determine the relationship between actual light changes and historical light data.

[0051] S242: Analyze the correlation between meteorological deviations and historical light data in the meteorological deviation and regression analysis model to obtain the prediction result.

[0052] Specifically, based on the correlation between the calculated meteorological deviations and historical light data in the regression analysis model, correlation analysis methods such as Pearson correlation coefficient and Spearman rank correlation can be used to calculate the correlation between meteorological deviations and historical light data. For example, calculate the relationship between current meteorological conditions (temperature, humidity, etc.) and historical light data, so as to deduce the correlation between them: Correlation(ΔT(t),I historical (t)); Correlation(ΔH(t),I historical (t)); Correlation(ΔP(t),I historical (t)), by analyzing these correlations, a set of regression analysis results can be obtained, reflecting the predicted relationship between changes in meteorological conditions (temperature, humidity, etc.) and light intensity. For example, through a multiple regression model, the following regression equation can be obtained: I pred (t) = β 0 + β 1 * ΔT(t) + β 2 * ΔH(t) + β 3 * ΔP(t) + β 4 * ΔW(t) + β 5 * ΔC(t), where I pred (t) represents the light intensity predicted according to real-time meteorological data, and β 0 , β 1 , …, β 5 are the model coefficients obtained by regression analysis, representing the relationship between various meteorological deviations and light data. Through the regression analysis results, the future light intensity can be accurately predicted and the impact of meteorological changes on light intensity can be understood.

[0053] S243: Generate corresponding light data correction factors based on the prediction results.

[0054] Specifically, according to the prediction results of the above regression model, the light data correction factors can be generated in the following way. First, use the predicted light value I pred (t) obtained from the regression analysis model, and the actually measured light intensity value I actual(t), calculate the deviation between the predicted light intensity value and the actual value, thereby generating a correction factor: This correction factor represents the deviation ratio between the actual light intensity and the predicted light intensity. When the correction factor is positive, it indicates that the actual light intensity is greater than the predicted value, and the predicted light data needs to be increased; when the correction factor is negative, it indicates that the actual light intensity is lower than the predicted value, and the predicted data needs to be reduced. Through the correction factor, the prediction result of the regression analysis model can be dynamically adjusted to make the predicted value of the light intensity more accurate.

[0055] In one embodiment, as Figure 5 shown, in step S40, that is, a convolutional neural network is used to predict the optimized light data to generate a predicted value of the light intensity within a preset time period, specifically including: S41: Input the optimized light data into the input layer of the convolutional neural network. The convolutional neural network includes at least one convolutional layer, a pooling layer, and a fully connected layer.

[0056] Specifically, first, the optimized light data is used as the input and input into the input layer of the convolutional neural network (CNN). The optimized light data usually exists in the form of a two-dimensional matrix, where each data item represents the light intensity at a certain moment or a certain location. For example, if the optimized light data contains the light intensity data for a 24-hour period, the data structure may be a 24×1 array or matrix. This data is passed through the input layer to the subsequent layers of the CNN. Here, the input data enters the convolutional layer in the form of a matrix and starts feature extraction and processing. The size and dimension of the data matrix are defined in the input layer and serve as the basis for subsequent processing. The convolutional neural network usually includes at least three layers of structures: the convolutional layer, the pooling layer, and the fully connected layer, which are responsible for feature extraction, dimensionality reduction, and the generation of the final prediction result respectively.

[0057] S42: In the convolutional layer of the convolutional neural network, use the convolution kernel to perform a convolution operation on the optimized light data to extract the local features in the optimized light data and obtain a feature map.

[0058] Specifically, in the convolutional layer, a set of convolutional kernels are used to perform convolutional operations on the input optimized illumination data. The convolutional kernels are a set of small matrices with fixed weights, used to scan the input data. Each convolutional kernel performs operations on a local region of the data to extract local features. For example, assume the input optimized illumination data is a 24×1 matrix, and the size of the convolutional kernel may be set to 3×1. Then, in a sliding window manner, the convolutional kernel is applied to each part of the input data, and after dot product operations, the output result is obtained. During each convolutional operation, the convolutional kernel slides over the data and extracts local illumination change features, such as the fluctuations of illumination intensity within a short period. Each convolutional operation generates a new matrix (i.e., the feature map), which reflects the local feature information in the optimized illumination data, such as the variation pattern of illumination intensity during a certain time period. The feature map obtained after convolutional operations is used for subsequent pooling operations and classification.

[0059] S43: In the pooling layer of the convolutional neural network, dimensionality reduction is performed on the feature map to obtain the pooled feature map.

[0060] Specifically, in the pooling layer, dimensionality reduction is performed on the feature map output by the convolutional layer to reduce the dimension of the feature map and enhance the computational efficiency and robustness of the model. The pooling layer usually uses methods such as max pooling or average pooling for dimensionality reduction. In max pooling, assume the size of the feature map is 24×1 and a 3×1 pooling window is used. Then, each time pooling is performed, the maximum value in the pooling window is selected as the output. For example, if the data in the pooling window is [1, 2, 3], then 3 is output as the result after pooling. The pooling layer covers the entire feature map by sliding the pooling window, and each time it slides, the maximum value in the pooling window is taken, thereby generating the pooled feature map. The pooled feature map has a smaller size compared to the original feature map while retaining important feature information, such as the trend of local illumination change.

[0061] S44: In the fully connected layer of the convolutional neural network, the pooled feature map is non-linearly transformed through the activation function: f(x) = max(0, x) to generate the predicted illumination intensity value within the preset time period.

[0062] Specifically, in the fully connected layer, the pooled feature map output by the pooling layer is flattened and transformed into a one-dimensional vector, which is then processed by fully connected neurons. In this layer, each neuron is connected to all neurons in the previous layer. By calculating the weighted sum and performing a non-linear transformation through an activation function, such as the ReLU (Rectified Linear Unit) activation function. The role of the activation function is to transform the linear model into a non-linear model, enabling the convolutional neural network to learn more complex light intensity prediction patterns. Through the calculation of the fully connected layer, a light intensity prediction value is finally generated, representing the light intensity within a preset time period. For example, when predicting the light intensity within the next 24 hours, 24-hour light prediction values can be generated through the fully connected layer and used as the final output. This process is trained through methods such as weight update and gradient descent to optimize the network's prediction ability during the training process.

[0063] In one embodiment, as Figure 6 shown, in step S50, the light intensity prediction value is input into the solar power generation prediction model, and the predicted solar power generation is calculated, specifically including: S51: Obtain the parameter data of the target device. The parameter data includes the total area, efficiency, and temperature coefficient of the solar panel.

[0064] Specifically, first, the basic parameter data of the solar panel needs to be obtained. These parameter data usually come from the technical manual of the solar panel or the settings during installation. For example, assume that the target device uses a solar panel with an area of 20m 2 . Its conversion efficiency is 18%, and the temperature coefficient is -0.4% / °C, which means that for every 1°C increase in the temperature of the solar panel, the output power will decrease by 0.4%. These parameters will be used as input data in the calculation model to accurately predict the power generation of the solar panel. Specifically, the area A and efficiency η of the solar panel are fixed parameters, and the temperature coefficient β is closely related to the environmental conditions and the working state of the panel. Usually, the real-time temperature T is obtained through a sensor.

[0065] S52: Based on the light intensity prediction value and the parameter data of the target device, use the calculation formula in the solar power generation prediction model: P(t) = η·A·I(t)·(1 - β·(T - T ref )), to obtain the predicted solar power generation, where P(t) is the predicted solar power generation within time t, η is the conversion efficiency of the target device, A is the total area of the solar panel, I(t) is the light intensity prediction value at time t, β is the temperature coefficient, T is the real-time temperature of the solar panel, and T ref is the reference temperature.

[0066] Specifically, based on the predicted light intensity value I(t) and the parameter data of the target device, the established solar power generation prediction model is used to calculate the predicted solar power generation P(t). The formula in this calculation model is as follows: P(t) = η·A·I(t)·(1 - β·(T - T ref ref)), where P(t) is the predicted solar power generation within time t, η is the conversion efficiency of the target device, A is the total area of the solar panels, I(t) is the predicted light intensity value at time t, β is the temperature coefficient, T is the real-time temperature of the solar panels, Tref is the reference temperature. For example, if the reference temperature Tref is 25 °C and the actual temperature T is 35 °C, the temperature difference ΔT is 10 °C, then according to the temperature coefficient β, the power generation efficiency of the solar panels will be negatively affected to a certain extent. Assuming β is -0.4% / °C, the correction factor for power generation is 1 + (-0.4%)×10 = 1 - 4% = 0.96. That is, under this temperature condition, the actual power generation capacity of the solar panels will decrease by 4%. Finally, substituting all the parameters into the formula, the predicted solar power generation P(t) at a specific moment t is obtained. For example, assuming the predicted light intensity value I(t) is 800 W / m 2 ², the area A of the solar panels is 20 m 2 ², the conversion efficiency η is 18%, the real-time temperature T is 35 °C, the reference temperature Tref is 25 °C, and the temperature coefficient β is -0.4% / °C, then the calculation process of the predicted solar power generation P(t) is as follows: P(t) = 18% * 20 * 800 * (1 - 4%) = 0.18 * 20 * 800 * 0.96 = 276.48 W. Therefore, under these conditions, the predicted solar power generation is 276.48 watts. Through this calculation process, the solar power generation at a future moment can be accurately predicted based on the predicted light intensity value, the parameters of the solar panels, and the real-time temperature data.

[0067] It should be understood that the magnitudes of the sequence numbers of the steps in the above embodiments do not mean the order of execution. The order of execution of each process should be determined according to its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.

[0068] In one embodiment, a solar power generation prediction system is provided, and this solar power generation prediction system corresponds one-to-one with the solar power generation prediction method in the above embodiment. As Figure 7 shown, this solar power generation prediction system includes a correction module, a calculation module, an optimization module, a prediction module, and a power generation prediction module. The detailed description of each functional module is as follows: The correction module is used to obtain real-time light data by using a geographic information system platform, and localize and correct the light data in combination with the geographical environment where the target device is located to obtain the corrected light data; A calculation module, configured to collect real-time meteorological data, and perform comparison calculations based on the real-time meteorological data and historical light data to generate a light data correction factor; An optimization module, configured to adjust the deviation in the corrected light data based on the correction factor to obtain optimized light data; A prediction module, configured to use a convolutional neural network to predict the optimized light data to generate a predicted light intensity value within a preset time period; A power generation prediction module, configured to input the predicted light intensity value into a solar power generation prediction model to calculate the predicted solar power generation.

[0069] Optionally, the correction module includes: A geographic information acquisition sub-module, configured to acquire the geographic information of the location where the target device is located, and the geographic information includes longitude, latitude, and altitude; A real-time light determination sub-module, configured to extract the real-time light data corresponding to the location where the target device is located from a meteorological data source by using a geographic information system platform based on the geographic information; A light correction sub-module, configured to adjust the real-time light data by using a localization correction algorithm according to the geographic environment where the target device is located to obtain the corrected light data.

[0070] Optionally, the calculation module includes: A real-time meteorological acquisition sub-module, configured to acquire the real-time meteorological data corresponding to the location where the target device is located, and the meteorological data includes temperature, humidity, air pressure, wind speed, and cloud cover; A training set sub-module, configured to acquire the historical light data and the corresponding historical meteorological data within a previous period of time as a training set; a modeling sub-module, configured to analyze the correlation between the historical meteorological data and the historical light data in the training set through a regression analysis model, establish a mathematical model between meteorological conditions and light changes, and obtain a meteorological deviation and a regression analysis model; A correction factor generation sub-module, configured to predict the future light change trend based on the real-time meteorological data through the meteorological deviation and the regression analysis model to generate a light data correction factor.

[0071] Optionally, the correction factor generation sub-module includes: A deviation calculation unit, configured to calculate the deviation between the real-time meteorological data and the historical meteorological data in the meteorological deviation and the regression analysis model to obtain a meteorological deviation; An analysis unit, configured to analyze the correlation between the meteorological deviation and the historical light data in the meteorological deviation and the regression analysis model to obtain a prediction result; A factor generation unit, configured to generate a corresponding light data correction factor based on the prediction result.

[0072] Optionally, the prediction module includes: An input sub-module for inputting the optimized illumination data into the input layer of the convolutional neural network, where the convolutional neural network includes at least one convolutional layer, pooling layer, and fully connected layer; A convolutional sub-module for performing a convolution operation on the optimized illumination data using a convolution kernel in the convolutional layer of the convolutional neural network to extract local features in the optimized illumination data and obtain a feature map; A pooling sub-module for performing a dimensionality reduction process on the feature map in the pooling layer of the convolutional neural network to obtain a pooled feature map; A fully connected sub-module for performing a non-linear transformation on the pooled feature map through an activation function: f(x) = max(0, x) in the fully connected layer of the convolutional neural network to generate a predicted illumination intensity value within a preset time period.

[0073] Optionally, the power generation prediction module includes: A parameter acquisition sub-module for acquiring parameter data of the target device, where the parameter data includes the total area, efficiency, and temperature coefficient of the solar panels; A power generation calculation sub-module for obtaining the predicted solar power generation based on the predicted illumination intensity value and the parameter data of the target device, using the calculation formula in the solar power generation prediction model: P(t) = η·A·I(t)·(1 - β·(T - T ref )), where P(t) is the predicted solar power generation within time t, η is the conversion efficiency of the target device, A is the total area of the solar panels, I(t) is the predicted illumination intensity value at time t, β is the temperature coefficient, T is the real-time temperature of the solar panels, and T ref is the reference temperature.

[0074] For the specific limitations of a solar power generation prediction system, reference can be made to the limitations of a solar power generation prediction method in the above text, which will not be elaborated here. Each module in the above solar power generation prediction system can be implemented in whole or in part through software, hardware, and their combination. The above modules can be embedded in the processor of the computer device in hardware form or be independent of it, or be stored in the memory of the computer device in software form, so as to facilitate the processor to call and execute the operations corresponding to the above respective modules.

[0075] In one embodiment, a computer device is provided. The computer device can be a server, and its internal structure diagram can be as Figure 8As shown in the figure. The computer device includes a processor, a memory, a network interface, and a database connected via a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, it implements a method for predicting solar power generation.

[0076] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the computer program, the following steps are implemented: Use a geographic information system platform to obtain real-time sunlight data, and localize and correct the sunlight data in combination with the geographical environment where the target device is located to obtain corrected sunlight data; Collect real-time meteorological data, and generate a sunlight data correction factor based on a comparison and calculation of the real-time meteorological data and historical sunlight data; Based on the correction factor, adjust the deviation in the corrected sunlight data to obtain optimized sunlight data; Use a convolutional neural network to predict the optimized sunlight data and generate a predicted value of sunlight intensity within a preset time period; Input the predicted value of sunlight intensity into a solar power generation prediction model to calculate the predicted solar power generation.

[0077] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by the processor, the following steps are implemented: Use a geographic information system platform to obtain real-time sunlight data, and localize and correct the sunlight data in combination with the geographical environment where the target device is located to obtain corrected sunlight data; Collect real-time meteorological data, and generate a sunlight data correction factor based on a comparison and calculation of the real-time meteorological data and historical sunlight data; Based on the correction factor, adjust the deviation in the corrected sunlight data to obtain optimized sunlight data; Use a convolutional neural network to predict the optimized sunlight data and generate a predicted value of sunlight intensity within a preset time period; Input the predicted value of sunlight intensity into a solar power generation prediction model to calculate the predicted solar power generation.

[0078] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, storage, database, or other medium used in the embodiments provided in the present application can include non-volatile and / or volatile memories. Non-volatile memories can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memories can include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM), etc.

[0079] Those skilled in the art can clearly understand that for the convenience and brevity of description, only the above division of each functional unit and module is used as an example. In actual applications, the above functions can be allocated to different functional units and modules according to needs, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above.

[0080] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should all be included in the protection scope of the present application.

Claims

1. A method for predicting solar power generation, characterized in that: The solar power generation prediction method comprises: Using a geographic information system platform to obtain real-time illumination data, and performing local correction on the illumination data in combination with the geographical environment where the target device is located, to obtain corrected illumination data; Collecting real-time meteorological data, and performing comparison and calculation based on the real-time meteorological data and historical light data to generate a light data correction factor; Based on the correction factor, adjusting the deviation in the corrected illumination data to obtain optimized illumination data; Using a convolutional neural network to predict the optimized lighting data, generating a predicted value of the lighting intensity within a preset time period; The light intensity prediction value is input into a solar power generation prediction model to calculate the predicted solar power generation.

2. A solar power generation prediction method according to claim 1, characterized in that: The method of obtaining real-time illumination data by using a geographic information system platform and performing local correction on the illumination data in combination with the geographical location of the target device to obtain the corrected illumination data includes: Acquire geographic information of the location of the target device, the geographic information including longitude, latitude and altitude; Based on the geographic information, extracting real-time illumination data corresponding to the location of the target device from a meteorological data source using the geographic information system platform; According to the geographical environment where the target device is located, the real-time illumination data is adjusted using a localized correction algorithm to obtain the corrected illumination data.

3. A solar power generation prediction method according to claim 1, characterized in that: The collecting of real-time meteorological data and comparing and calculating the real-time meteorological data with historical illumination data to generate illumination data correction factors include: Acquire real-time meteorological data corresponding to the location of the target device, the meteorological data including temperature, humidity, air pressure, wind speed, and cloud coverage; Obtain historical light data and corresponding historical meteorological data in the previous period as a training set; Analyze the correlation between the historical meteorological data and the historical light data in the training set by using a regression analysis model, establish a mathematical model between meteorological conditions and light changes, and obtain a meteorological deviation and a regression analysis model; Based on the real-time meteorological data, the future illumination change trend is predicted through the meteorological deviation and regression analysis model to generate the illumination data correction factor.

4. A solar power generation prediction method according to claim 3, characterized in that: The method of predicting the future illumination change trend based on the real-time meteorological data through the meteorological deviation and regression analysis model to generate the illumination data correction factor includes: Calculating the deviation between the real-time meteorological data and the historical meteorological data in the meteorological deviation and regression analysis model to obtain the meteorological deviation; Analyzing the correlation between the meteorological deviation and the historical illumination data in the meteorological deviation and regression analysis model to obtain a prediction result; Based on the prediction result, a corresponding illumination data correction factor is generated.

5. A solar power generation prediction method according to claim 1, characterized in that: The method of using a convolutional neural network to predict the optimized illumination data and generate a predicted value of illumination intensity within a preset time period includes: Inputting the optimized illumination data into an input layer of the convolutional neural network, wherein the convolutional neural network includes at least one convolutional layer, a pooling layer, and a fully connected layer; In the convolution layer of the convolutional neural network, a convolution operation is performed on the optimized illumination data using a convolution kernel to extract local features in the optimized illumination data to obtain a feature map; In the pooling layer of the convolutional neural network, the feature map is subjected to dimensionality reduction processing to obtain a pooled feature map; In the fully connected layer of the convolutional neural network, the pooled feature map is nonlinearly transformed through the activation function: f(x)=max(0,x) to generate a predicted value of light intensity within the preset time period.

6. A solar power generation prediction method according to claim 1, characterized in that: The step of inputting the light intensity prediction value into the solar power generation prediction model to calculate the predicted solar power generation includes: Acquiring parameter data of the target device, the parameter data including a total area, efficiency, and temperature coefficient of a solar cell panel; Based on the predicted light intensity and the parameter data of the target device, the calculation formula in the solar power generation prediction model is used: P(t) = η·A·I(t)·(1-β·(TT ref )), the predicted solar power generation is obtained, wherein P(t) is the predicted solar power generation within time t, η is the conversion efficiency of the target device, A is the total area of ​​the solar panel, I(t) is the predicted value of the light intensity at time t, β is the temperature coefficient, T is the real-time temperature of the solar panel, T ref is the reference temperature.

7. A solar power generation prediction system, characterized in that: The solar power generation prediction system comprises: A correction module, used to obtain real-time illumination data using a geographic information system platform, and to perform local correction on the illumination data in combination with the geographical environment where the target device is located, to obtain corrected illumination data; A calculation module, used to collect real-time meteorological data, and compare and calculate the real-time meteorological data with historical light data to generate a light data correction factor; An optimization module, configured to adjust the deviation in the corrected illumination data based on the correction factor to obtain optimized illumination data; A prediction module, used to predict the optimized illumination data using a convolutional neural network to generate a predicted value of illumination intensity within a preset time period; The power generation prediction module is used to input the light intensity prediction value into the solar power generation prediction model to calculate the expected solar power generation.

8. A solar power generation prediction system according to claim 7, characterized in that: The correction module comprises: A geographic information acquisition submodule is used to acquire geographic information of the location of the target device, wherein the geographic information includes longitude, latitude and altitude; Determine a real-time illumination submodule, which is used to extract real-time illumination data corresponding to the location of the target device from a meteorological data source based on the geographic information and using the geographic information system platform; The illumination correction submodule is used to adjust the real-time illumination data using a localized correction algorithm according to the geographical environment where the target device is located to obtain the corrected illumination data.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the computer program, the steps of the solar power generation prediction method according to any one of claims 1 to 6 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of a solar power generation prediction method as claimed in any one of claims 1 to 6 are implemented.

Images