Distributed photovoltaic cluster ultra-short-term accurate power prediction method

Abstract

The invention relates to the technical field of power prediction, in particular to a distributed photovoltaic cluster ultra-short-term accurate power prediction method, which comprises the following steps: acquiring historical power generation data and real-time meteorological information; establishing a cloud layer shielding model based on the historical power generation data and the real-time meteorological information; actual meteorological parameters of an area where the distributed photovoltaic cluster is located are collected; analyzing the actual meteorological parameters based on the cloud layer shielding model to obtain power fluctuation influence factors; performing ultra-short-term prediction on the power output of the distributed photovoltaic cluster according to the power fluctuation influence factor; the method has the effect of improving the ultra-short-term precise power prediction precision of the distributed photovoltaic cluster.

Description

Technical Field

[0001] The present application relates to the technical field of power prediction, and in particular to a method for ultra-short-term accurate power prediction of distributed photovoltaic clusters. Background Art

[0002] With the continuous increase in the installed capacity of distributed photovoltaics, its proportion in the power grid has been increasing year by year. To ensure the safe and stable operation of the power grid, it is necessary to accurately predict the power generation of distributed photovoltaic clusters to optimize the power grid dispatching arrangement and power balance.

[0003] However, compared with centralized photovoltaic power stations, distributed photovoltaics are installed in a more scattered manner, with different installed capacities, lacking local meteorological measurement devices, unable to obtain meteorological data such as local irradiance, and there is no effective numerical weather forecast data for ultra-short-term power prediction. How to conveniently and accurately obtain the ultra-short-term distributed photovoltaic power generation prediction is of great significance for the optimal operation of distributed photovoltaic power stations, the dispatching of new power systems, and the safe, stable and economic operation of the power grid. Summary of the Invention

[0004] In order to improve the accuracy of ultra-short-term accurate power prediction of distributed photovoltaic clusters, the present application provides a method for ultra-short-term accurate power prediction of distributed photovoltaic clusters.

[0005] In a first aspect, the present application provides a method for ultra-short-term accurate power prediction of distributed photovoltaic clusters, adopting the following technical solutions: A method for ultra-short-term accurate power prediction of distributed photovoltaic clusters, Obtain historical power generation data and real-time meteorological information; Establish a cloud occlusion model based on the historical power generation data and the real-time meteorological information; Collect the actual meteorological parameters of the area where the distributed photovoltaic cluster is located; Analyze the actual meteorological parameters based on the cloud occlusion model to obtain a power fluctuation influence factor; Perform an ultra-short-term prediction on the power output of the distributed photovoltaic cluster according to the power fluctuation influence factor.

[0006] By adopting the above technical solution, combining historical power generation data and real-time meteorological information to establish a cloud occlusion model. Since cloud occlusion is one of the main factors causing fluctuations in photovoltaic power generation, the prediction based on the cloud occlusion model can significantly improve the accuracy of power prediction and make the prediction results closer to the actual power generation situation. The cloud occlusion model is established based on the statistical analysis of historical power generation data and meteorological information, considering the complex relationship between the key factor of cloud occlusion and photovoltaic power generation. Therefore, by using this model to analyze the actual meteorological parameters, and by predicting the power fluctuation impact factor and accordingly making a very short-term prediction of the power output of the photovoltaic cluster, the fluctuations in grid voltage and frequency caused by photovoltaic power generation fluctuations can be reduced. It can more accurately capture the key factors affecting photovoltaic power generation, thereby improving the prediction accuracy.

[0007] In another possible implementation manner, establishing the cloud occlusion model based on historical power generation data and real-time meteorological information includes: Clean the historical power generation data to determine and remove abnormal power generation records; Extract features from the cloud distribution information in the real-time meteorological information using an image processing algorithm; Conduct a correlation analysis by combining historical meteorology and historical power generation data to obtain reference indicators; Train a machine learning model based on the reference indicators to obtain an occlusion coefficient model.

[0008] In another possible implementation manner, analyzing the actual meteorological parameters based on the cloud occlusion model to obtain a power fluctuation impact factor includes: Collect real-time meteorological data in the area where the photovoltaic power station is located; Establish a cloud occlusion model according to the meteorological data; Use the cloud occlusion model to calculate the attenuation degree of cloud occlusion on ground solar radiation; Determine the attenuation degree as the impact factor of power fluctuation.

[0009] In another possible implementation manner, establishing the cloud occlusion model according to the meteorological data includes: Obtain the cloud image and solar position in the area where the photovoltaic power station is located; Combine the cloud image with the meteorological data to determine cloud information, where the cloud information includes the height, thickness and type of the cloud; Construct a position model of the cloud in three-dimensional space according to the solar position, the height, thickness and type of the cloud.

[0010] In another possible implementation manner, establishing the cloud occlusion model based on historical power generation data and real-time meteorological information includes: Obtain the power generation per hour within the historical period; Collect meteorological factor data corresponding to the time period of the power generation; Draw a relationship curve graph of photovoltaic power generation and the degree of occlusion at different time periods; and Use the statistical regression analysis method to obtain the function expression R1 between the occlusion rate and the actual power generation from the curve graph.

[0011] In another possible implementation manner, the drawing of the relationship curve graph of photovoltaic power generation and the degree of occlusion at different time periods specifically refers to drawing the variation law of the cloud cover coefficient and the photovoltaic power generation intensity, including: Measure the maximum power generation output P of the distributed photovoltaic system in the natural state without any cloud cover under sunny weather conditions max ; Calculate the average photovoltaic output power P within each time period based on the historical period data average And record the cloud layer distribution status; Calculate the cloud cover ratio C cr =C c / C total ; where C c is the effective area actually covering the surface of the photovoltaic module, and C total is the total area of this region; and Based on the formula R1: Calculate the predicted output power P predicted =P max (1 - C cr ) for predicted power calculation.

[0012] In another possible implementation manner, the analysis of the actual meteorological parameters based on the cloud occlusion model includes: Detect the light intensity every 5 minutes within the target time period, and calculate the ratio I of its instantaneous illuminance to the peak value in the same historical period R ; Perform smoothing processing on I R to reduce errors, and match it with the shielding effect correction coefficient k in the current cloud amount state to obtain the corrected shielding rate C cr′ , where C cr′ =(k ± λ)I R , and λ represents a preset proportional adjustment factor; Check the range of the shielding effect correction value: 0 ≤ C cr′ ≤ C total ; Use the corrected shielding rate to replace the initial C cr value and substitute it into the predicted power formula for final prediction, so as to obtain an accurate fluctuation influence factor.

[0013] Second aspect, the present application provides a device for ultra-short-term accurate power prediction of a distributed photovoltaic cluster, adopting the following technical solution: The device for ultra-short-term accurate power prediction of a distributed photovoltaic cluster includes: An information acquisition module, configured to acquire historical power generation data and real-time meteorological information; A model establishment module, configured to establish a cloud occlusion model based on the historical power generation data and the real-time meteorological information; A parameter acquisition module, configured to acquire actual meteorological parameters of the area where the distributed photovoltaic cluster is located; A factor acquisition module, configured to analyze the actual meteorological parameters based on the cloud occlusion model to obtain power fluctuation influence factors; A prediction module, configured to perform ultra-short-term prediction on the power output of the distributed photovoltaic cluster according to the power fluctuation influence factors.

[0014] In another possible implementation manner, when the model establishment module establishes a cloud occlusion model based on the historical power generation data and the real-time meteorological information, it is specifically configured to: Clean the historical power generation data, determine abnormal power generation records and remove them; Extract features from the cloud distribution information in the real-time meteorological information by using an image processing algorithm; Perform correlation analysis by combining historical meteorology and historical power generation data to obtain reference indicators; Train a machine learning model according to the reference indicators to obtain an occlusion coefficient model.

[0015] In another possible implementation manner, when the factor acquisition module analyzes the actual meteorological parameters based on the cloud occlusion model to obtain power fluctuation influence factors, it is specifically configured to: Collect meteorological data of the area where the photovoltaic power station is located in real time; Establish a cloud occlusion model according to the meteorological data; Use the cloud occlusion model to calculate the attenuation degree of cloud occlusion on ground solar radiation; Determine the attenuation degree as the influence factor of power fluctuation.

[0016] In another possible implementation manner, when the factor acquisition module establishes a cloud occlusion model according to the meteorological data, it is specifically configured to: Acquire the cloud image and the sun position of the area where the photovoltaic power station is located; Combine the cloud image with the meteorological data to determine cloud information, where the cloud information includes the height, thickness and type of the cloud; According to the sun position, the height, thickness and type of the cloud, construct a position model of the cloud in three-dimensional space.

[0017] In another possible implementation, when the model establishment module establishes the cloud occlusion model based on historical power generation data and real-time meteorological information, it specifically is used for: Obtain the power generation amount per hour within the historical period; Collect meteorological factor data corresponding to the time period of the power generation amount; Draw a relationship curve graph of photovoltaic power generation and occlusion degree at different time periods; and Use the statistical regression analysis method to obtain the function expression R1 between the occlusion rate and the actual power generation amount from the curve graph.

[0018] In another possible implementation, when the model establishment module specifically draws the change rule of the cloud occlusion coefficient and the photovoltaic power generation intensity in the relationship curve graph of photovoltaic power generation and occlusion degree at different time periods, it specifically is used for: Measure the maximum power generation output P of the distributed photovoltaic system in the natural state without any cloud occlusion under sunny weather conditions max ; Calculate the average photovoltaic output power P within each time period based on historical period data average And record the cloud distribution status; Calculate the cloud occlusion ratio C cr =C c / C total ; where C c is the effective area actually covering the surface of the photovoltaic module, and C total is the total area of this region; and Based on formula R1: Calculate the predicted output power P predicted =P max (1 - C cr ) for predicted power calculation.

[0019] In another possible implementation, when the factor acquisition module analyzes the actual meteorological parameters based on the cloud occlusion model, it specifically is used for: Detect the light intensity every 5 minutes within the target time period, and calculate the ratio I of its instantaneous illuminance to the peak value in the same historical period R ; Perform smoothing processing on I R to reduce errors, and match it with the shielding effect correction coefficient k in the current cloud amount state to obtain the corrected shielding rate C cr′ , where C cr′ =(k ± λ)I R , and λ represents a preset proportional adjustment factor; Check the range of the shielding effect correction value: 0 ≤ C cr′ ≤ C total ; Use the calibrated occlusion rate to replace the initial C cr Substitute the value into the predicted power formula for final prediction, so as to obtain an accurate fluctuation impact factor.

[0020] In a third aspect, the present application provides an electronic device, adopting the following technical solution: An electronic device, which includes: At least one processor; A memory; At least one application program, where at least one application program is stored in the memory and is configured to be executed by at least one processor. At least one is configured to: execute the distributed photovoltaic cluster ultra-short-term precise power prediction method shown in any possible implementation manner of the first aspect.

[0021] In a fourth aspect, the present application provides a computer-readable storage medium, adopting the following technical solution: A computer-readable storage medium, when the computer program is executed in a computer, causes the computer to execute the distributed photovoltaic cluster ultra-short-term precise power prediction method described in any item of the first aspect.

[0022] In summary, the present application includes at least one of the following beneficial technical effects: Combining historical power generation data and real-time meteorological information, a cloud occlusion model is established. Since cloud occlusion is one of the main factors causing fluctuations in photovoltaic power generation, the prediction based on the cloud occlusion model can significantly improve the accuracy of power prediction and make the prediction result closer to the actual power generation situation. The cloud occlusion model is established based on the statistical analysis of historical power generation data and meteorological information, considering the complex relationship between the key factor of cloud occlusion and photovoltaic power generation. Therefore, by using this model to analyze actual meteorological parameters, predicting the power fluctuation impact factor and accordingly making an ultra-short-term prediction of the power output of the photovoltaic cluster, it is possible to reduce the fluctuations in grid voltage and frequency caused by photovoltaic power generation fluctuations, and can more accurately capture the key factors affecting photovoltaic power generation, thereby improving the prediction accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 is a schematic flowchart of the distributed photovoltaic cluster ultra-short-term precise power prediction method in an embodiment of the present application.

[0024] Figure 2 is a schematic flowchart of the distributed photovoltaic cluster ultra-short-term precise power prediction device in an embodiment of the present application.

[0025] Figure 3 is a schematic structural diagram of an electronic device in an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0026] The following will further elaborate on this application in conjunction with the appended Figures 1-3 drawings.

[0027] After reading this specification, those skilled in the art can make modifications to this embodiment as needed without making creative contributions, but as long as they are within the scope of the claims of this application, they are protected by the patent law.

[0028] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions in the embodiments of this application will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of this application. Obviously, the described embodiments are some, but not all, of the embodiments of this application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in this application without making creative efforts fall within the scope of protection of this application.

[0029] In addition, the term "and / or" herein is merely a description of the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B may represent: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " herein generally represents an "or" relationship between the associated objects before and after, unless otherwise specified.

[0030] The following will further describe the embodiments of this application in detail in conjunction with the drawings of the specification.

[0031] The embodiments of this application provide a method for ultra-short-term accurate power prediction of a distributed photovoltaic cluster, which is executed by an electronic device. The electronic device can be a server or a terminal device. Among them, the server can be an independent physical server, a server cluster or a distributed system composed of multiple physical servers, or a cloud server providing cloud computing services. The terminal device can be a smart phone, a tablet computer, a laptop computer, a desktop computer, etc., but is not limited thereto. The terminal device and the server can be directly or indirectly connected through wired or wireless communication methods, which are not limited in the embodiments of this application. As Figure 1 shown, the method includes: step S101, step S102, step S103, step S104, and step S105, where step S101, obtaining historical power generation data and real-time meteorological information; step S102, establishing a cloud occlusion model based on the historical power generation data and the real-time meteorological information; step S103, collecting actual meteorological parameters of the area where the distributed photovoltaic cluster is located; step S104, analyzing the actual meteorological parameters based on the cloud occlusion model to obtain a power fluctuation impact factor; Step S105, perform ultra-short-term prediction on the power output of the distributed photovoltaic cluster according to the power fluctuation impact factor.

[0032] In the embodiment of the present application, historical power generation data is exported from the monitoring system of the distributed photovoltaic cluster. Among them, the historical power generation data includes the power generation records of each day (or each hour). Real-time meteorological information is collected by various methods such as meteorological observation stations, satellite remote sensing, and radar detection, including cloud height, cloud thickness, cloud amount, solar radiation intensity, temperature, humidity, wind speed, etc. The historical power generation data is matched with the real-time meteorological information of the corresponding time period to ensure that each power generation data point has corresponding meteorological data support. Statistical analysis, machine learning, or deep learning methods are used to establish a cloud occlusion model according to the relationship between the historical power generation data and the meteorological data. This model should be able to reflect the impact of the cloud occlusion degree on the photovoltaic power generation.

[0033] Algorithms such as support vector machine (SVM), neural network (such as BPNN), and random forest can be considered for modeling, and model parameters can be optimized by methods such as cross-validation and grid search. Meteorological observation equipment is deployed in the area where the distributed photovoltaic cluster is located to collect real-time meteorological parameters, including cloud height, cloud thickness, cloud amount, solar radiation intensity, temperature, humidity, wind speed, etc. Based on the cloud occlusion model, the actual meteorological parameters are analyzed to calculate power fluctuation impact factors such as the cloud occlusion degree. These factors will be used as the input for subsequent power prediction. Select a suitable prediction model according to the actual situation, such as a time series analysis model, a machine learning model, etc. It should be noted that the power fluctuation impact factor is used as the input of the prediction model, combined with the historical power generation data and the real-time meteorological information, so as to accurately perform ultra-short-term prediction on the photovoltaic power generation in the future for a period of time.

[0034] Preferably, establishing a cloud occlusion model based on historical power generation data and real-time meteorological information includes: Clean the historical power generation data to determine and remove abnormal power generation records; Extract features from the cloud distribution information in the real-time meteorological information by using an image processing algorithm; Conduct a correlation analysis by combining historical meteorology and historical power generation data to obtain a reference index; Train a machine learning model according to the reference index to obtain an occlusion coefficient model.

[0035] In the embodiments of the present application, historical power generation data is imported from the monitoring system of the distributed photovoltaic cluster. This data usually includes dates, timestamps, and corresponding power generation records. The data is initially inspected to identify possible outliers, such as records with sudden surges or drops in power generation. Statistical methods (such as mean, median, standard deviation, etc.) or rules based on business logic are used to determine abnormal power generation records. These outliers are removed or corrected to ensure the accuracy of the data. Image data of the real-time cloud distribution is obtained from meteorological satellites, radars, or ground observation stations.

[0036] Image processing algorithms (such as edge detection, texture analysis, shape recognition, etc.) are applied to extract features from the cloud distribution images. These features may include cloud thickness, cloud amount, cloud type, etc. The extracted features are quantified into numerical data for subsequent analysis. The cleaned historical power generation data is integrated with the historical meteorological data for the corresponding time period (including the cloud features extracted from the images). Statistical methods (such as correlation coefficient, covariance, etc.) are used to calculate the correlation between historical power generation and each meteorological factor (including cloud features). Determine which meteorological factors have a significant impact on power generation, and use these factors as reference indicators. According to the results of the correlation analysis, select the meteorological factors (including cloud features) that have a significant impact on power generation as the input features of the model. Use the historical power generation data as the output target of the model. Select a suitable machine learning algorithm to train the occlusion coefficient model. Commonly used algorithms include linear regression, random forest, gradient boosting decision tree (GBDT), neural network, etc. Select a suitable algorithm according to the characteristics of the problem and the properties of the data. It should be noted that when training the model, the prepared data needs to be divided into a training set and a test set. Use the training set data to train the machine learning model, and optimize the performance of the model by adjusting the model parameters. During the training process, methods such as cross-validation can be used to evaluate the generalization ability of the model. Then apply the trained occlusion coefficient model to the real-time meteorological information to predict the occlusion coefficient based on the real-time cloud features and other meteorological factors. Combine the occlusion coefficient with other factors (such as the efficiency and temperature of photovoltaic modules, etc.) to perform a very short-term prediction of the power output of the distributed photovoltaic cluster.

[0037] Preferably, analyzing the actual meteorological parameters based on the cloud occlusion model to obtain a power fluctuation impact factor includes: Real-time collect the meteorological data of the area where the photovoltaic power station is located; Establish a cloud occlusion model according to the meteorological data; Use the cloud occlusion model to calculate the attenuation degree of cloud occlusion on the ground solar radiation; Determine the attenuation degree as the power fluctuation impact factor.

[0038] In the embodiments of the present application, meteorological data of the area where the photovoltaic power station is located is collected in real time by various methods such as meteorological observation stations, satellite remote sensing, and radar detection. The data should include cloud height, cloud thickness, cloud amount, solar radiation intensity, temperature, humidity, wind speed, etc. A data receiving system is established to receive meteorological data from different data sources in real time. The received data is preprocessed, including data cleaning, format conversion, and standardization, to ensure the accuracy and consistency of the data. A machine learning model (such as a neural network, random forest, etc.) is selected or trained, and this model can predict the cloud occlusion situation based on meteorological data. Historical meteorological data and the corresponding cloud occlusion situations (which may be obtained by manual annotation or back-calculation from historical power generation) are used to train the model. The meteorological data collected in real time is input into the trained cloud occlusion model for real-time prediction to obtain the current cloud occlusion situation.

[0039] According to the prediction results of cloud occlusion (such as cloud thickness, cloud amount, etc.), the attenuation degree of cloud on ground solar radiation is calculated using a physical model or empirical formula. The attenuation degree is usually expressed as the reduction ratio of solar radiation intensity after passing through the cloud. The predicted value of cloud occlusion is substituted into the attenuation model to calculate the attenuation degree of the current cloud on ground solar radiation. The calculated attenuation degree is directly used as the influencing factor of power fluctuation. This factor reflects the degree of reduction in photovoltaic power generation caused by cloud occlusion. When predicting the photovoltaic power generation amount, the influencing factor is multiplied by the theoretical power generation amount under the condition of no occlusion to obtain the predicted power generation amount considering cloud occlusion.

[0040] Suppose the real-time meteorological data of a certain photovoltaic power station at a certain moment shows that the cloud thickness is 500 meters and the cloud amount is 60% (partial occlusion). The cloud thickness and cloud amount are input into the cloud occlusion model, and the model predicts that the current cloud occlusion situation is "moderate occlusion". According to the empirical formula or physical model, "moderate occlusion" is converted into the attenuation degree of ground solar radiation, which is assumed to be 30%. The attenuation degree of 30% is determined as the influencing factor of power fluctuation. Suppose the theoretical power generation amount under the condition of no occlusion is 1000 kW·h, then the predicted power generation amount considering cloud occlusion is 1000×(1 - 0.3)=700 kW·h.

[0041] Preferably, establishing the cloud occlusion model according to the meteorological data includes: Obtaining the cloud image and the sun position of the area where the photovoltaic power station is located; Combining the cloud image with the meteorological data to determine cloud information, where the cloud information includes the height, thickness, and type of the cloud; According to the sun position, the height, thickness, and type of the cloud, constructing a position model of the cloud in three-dimensional space.

[0042] In the embodiments of the present application, the current position information of the sun is obtained by astronomical calculations or directly from meteorological data sources, including the altitude angle and azimuth angle of the sun. Image processing techniques (such as edge detection, texture analysis, pattern recognition, etc.) are applied to analyze cloud images to identify different features and layers of the clouds. Through image processing algorithms, we can extract information such as the boundaries, shapes, and textures of the clouds from the images. Using 3D modeling techniques, the height, thickness, and type information of the clouds are converted into geometric models in 3D space. The position information of the sun is integrated into the 3D cloud model to analyze the occlusion of solar radiation by the clouds. By calculating the intersection points or intersection regions between the sun's rays and the cloud model, the attenuation degree of solar radiation by the clouds can be evaluated.

[0043] For example, the altitude angle of the sun is 45 degrees, the azimuth angle is 135 degrees, the height of stratus clouds is about 2000 meters, and the thickness is about 500 meters; the height of cumulus clouds is about 1500 meters. According to the height and thickness information of the stratus clouds, a cuboid or custom-shaped cloud model is placed in 3D space, with its bottom located 2000 meters above the ground and the height (or thickness) being 500 meters. Adjust the transparency of the cloud model to simulate the actual visual effect of the clouds. Then output the result of the cloud occlusion model.

[0044] Preferably, establishing the cloud occlusion model based on historical power generation data and real-time meteorological information includes: Obtaining the power generation per hour during a historical period; Collecting meteorological factor data corresponding to the time period of the power generation; Plotting a relationship curve graph of photovoltaic power generation and occlusion degree at different time periods; and Using the statistical regression analysis method to obtain a function expression R1 between the occlusion rate and the actual power generation from the curve graph.

[0045] In the embodiments of the present application, historical power generation data is exported from the monitoring system of the photovoltaic power station. Select the power generation data per hour within a specific time period (such as one year, half a year, or one quarter). Obtain the meteorological data corresponding to the time period from meteorological stations, satellite remote sensing, or meteorological data service providers. Collect meteorological factors that may affect photovoltaic power generation, such as cloud thickness, solar radiation intensity, temperature, humidity, wind speed, etc. Estimate the cloud occlusion degree of each time period according to indicators such as cloud thickness or solar radiation intensity in the meteorological data. Use data visualization tools (such as Excel, Matplotlib in Python, etc.) to plot a scatter plot or line graph with the occlusion degree as the abscissa and the actual power generation as the ordinate. Among them, the occlusion rate = 1 - actual solar radiation intensity / maximum possible solar radiation intensity, and the form of the linear regression model is: power generation = β 0 +β 1 × occlusion rate + ε, where β0 is the intercept, and β 1 is the slope, and ε is the error term.

[0046] Preferably, the curve graph showing the relationship between photovoltaic power generation and the degree of occlusion at different time periods is specifically a graph showing the variation law of the cloud cover coefficient and the photovoltaic power generation intensity, including: Measure the maximum power generation output P of the distributed photovoltaic system under natural conditions without any cloud cover on a sunny day max ; Calculate the average photovoltaic output power P within each time period based on historical time period data average and record the cloud distribution status; Calculate the cloud cover ratio C cr = C c / C total ; where C c is the effective area actually covering the surface of the photovoltaic module, and C total is the total area of this region; and Based on formula R1: Calculate the predicted output power P predicted = P max (1 - C cr ).

[0047] In the embodiment of the present application, assume that the maximum power generation output P of a distributed photovoltaic system located at a certain place max is known to be 1000 kW.

[0048] The specific data is as follows: Time <![CDATA[Actual output power P average (kW)]]> <![CDATA[Cloud cover area C c (m²)]]> <![CDATA[Total area C total (m²)]]> 08:00 300 1000 10000 09:00 450 800 10000 10:00 600 500 10000 11:00 800 200 10000 12:00 950 50 10000 13:00 880 200 10000 14:00 650 600 10000 ... ... ... ... Calculate the cloud cover ratio C cr For each time period, calculate the cloud cover ratio: C cr = C c / C total .

[0049] For example, for the data at 12:00: C cr = 50 / 10000 = 0.005 Predicted power calculation: Use the formula P predicted = P max × (1 - C cr ) to calculate the predicted output power.

[0050] For example, for 12:00: P predicted = 1000 × (1 - 0.005) = 995 kW Draw the curve graph; Using data visualization tools (such as Excel, Matplotlib in Python, etc.), one of the following two curve graphs can be plotted: Relationship between actual output power and time: The horizontal axis is time, and the vertical axis is P average .

[0051] Relationship between predicted output power (or actual output power) and cloud cover ratio: The horizontal axis is cloud cover ratio C cr , and the vertical axis is P predicted (or P average ). To more clearly show the relationship, time can be used as color coding or data point labels.

[0052] Assume that the relationship graph between predicted output power and cloud cover ratio is plotted: Horizontal axis: Cloud cover ratio C cr (for example, from 0 to 0.1, with an interval of 0.01) Vertical axis: Predicted output power P predicted（ or use P according to actual data average ) Then, for each time period, calculate C cr and P predicted (or P average ), plot the corresponding points in the graph and connect them into a line.

[0053] Preferably, the analysis of actual meteorological parameters based on the cloud occlusion model includes: Detect the light intensity within every 5 minutes during the target time period, and calculate the ratio I of its instantaneous illuminance to the peak value in the same historical period R ; Smooth I R to reduce errors, and match it with the shielding effect correction coefficient k under the current cloud amount state to obtain the corrected shielding rate C cr′ , where C cr′ = (k ± λ)I R , and λ represents a preset proportional adjustment factor; Check the range of the shielding effect correction value: 0 ≤ C cr′ ≤ C total ; Use the corrected shielding rate to replace the initial C cr value and substitute it into the predicted power formula for final prediction, so as to obtain an accurate fluctuation influence factor.

[0054] In the embodiment of the present application, assume that the detected light intensity data is as follows (unit: kW / m²): Time Light intensity 11:30 600 11:35 620 ... ... 12:25 850 12:30 830 Calculate the ratio I of the instantaneous illuminance to the peak value in the same historical period R : Assume that the peak light intensity at the same time of last year (12:00 noon on the same day last year) was 1000 kW / m². Then, for each time point, calculate I R : For example, for 11:30, I R = 600 / 1000 = 0.6 Smooth I R Perform smoothing Assume that we use a simple moving average method (averaging every three points) for smoothing. The smoothed I R value will be used as the new I Rsmoothed .

[0055] Match with the shading effect correction coefficient k under the current cloud cover status; Use the formula C cr′ =(k ± λ)I Rsmoothed Calculate the corrected shading rate.

[0056] For the smoothed I Rsmoothed = 0.65, then C cr′ =(0.6 + 0.05)×0.65 = 0.4225 Check the range of the shading effect correction value: Ensure that C cr′ is within 0 ≤ C cr′ ≤ 1 (because C total can be regarded as 1 here, that is, the fully covered area of the photovoltaic module). If C cr′ exceeds this range, adjustment is required.

[0057] Substitute the corrected shading rate into the predicted power formula: Use the formula P predicted = P max ×(1 - C cr′ ) for the final prediction.

[0058] For example, C cr′ = 0.4225, then P predicted = 1000×(1 - 0.4225) = 577.5 kW.

[0059] The above embodiments introduce the ultra-short-term accurate power prediction method for distributed photovoltaic clusters from the perspective of the method process. The following embodiments introduce the ultra-short-term accurate power prediction device 20 for distributed photovoltaic clusters from the perspective of virtual modules or virtual units. For details, see the following embodiments.

[0060] The embodiments of the present application provide a device 20 for ultra-short-term accurate power prediction of distributed photovoltaic clusters, as Figure 2As shown in the figure, the device 20 for ultra-short-term accurate power prediction of the distributed photovoltaic cluster specifically includes: An information acquisition module 201, configured to acquire historical power generation data and real-time meteorological information; A model establishment module 202, configured to establish a cloud occlusion model based on the historical power generation data and the real-time meteorological information; A parameter acquisition module 203, configured to acquire actual meteorological parameters of the area where the distributed photovoltaic cluster is located; A factor acquisition module 204, configured to analyze the actual meteorological parameters based on the cloud occlusion model to obtain power fluctuation influence factors; A prediction module 205, configured to perform ultra-short-term prediction on the power output of the distributed photovoltaic cluster according to the power fluctuation influence factors.

[0061] Preferably, when the model establishment module 202 establishes a cloud occlusion model based on the historical power generation data and the real-time meteorological information, it is specifically configured to: Clean the historical power generation data, determine abnormal power generation records and remove them; Extract features from the cloud distribution information in the real-time meteorological information by using an image processing algorithm; Perform correlation analysis on the historical meteorological and historical power generation data to obtain reference indicators; Train a machine learning model according to the reference indicators to obtain an occlusion coefficient model; Preferably, when the factor acquisition module 204 analyzes the actual meteorological parameters based on the cloud occlusion model to obtain power fluctuation influence factors, it is specifically configured to: Collect real-time meteorological data of the area where the photovoltaic power station is located; Establish a cloud occlusion model according to the meteorological data; Use the cloud occlusion model to calculate the attenuation degree of cloud occlusion on ground solar radiation; Determine the attenuation degree as the influence factor of power fluctuation.

[0062] Preferably, when the factor acquisition module 204 establishes a cloud occlusion model according to the meteorological data, it is specifically configured to: Obtain the cloud image and the sun position of the area where the photovoltaic power station is located; Combine the cloud image with the meteorological data to determine cloud information, where the cloud information includes the height, thickness and type of the cloud; Construct a position model of the cloud in three-dimensional space according to the sun position, the height, thickness and type of the cloud.

[0063] Preferably, when establishing the cloud occlusion model based on historical power generation data and real-time meteorological information, the model establishment module 202 is specifically configured to: Obtain the power generation amount per hour within a historical period; Collect meteorological factor data corresponding to the time period of the power generation amount; Draw a relationship curve graph of photovoltaic power generation and occlusion degree at different time periods; and Use the statistical regression analysis method to obtain the function expression R1 between the occlusion rate and the actual power generation amount from the curve graph.

[0064] In a possible implementation manner of the embodiment of the present application, when the model establishment module 202 specifically draws the change rule of the cloud occlusion coefficient and the photovoltaic power generation intensity when drawing the relationship curve graph of photovoltaic power generation and occlusion degree at different time periods, it is specifically configured to: Measure the maximum power generation output P of the distributed photovoltaic system in the natural state without any cloud occlusion under sunny weather conditions max ; Calculate the average photovoltaic output power P within each time period based on historical period data average And record the cloud distribution status; Calculate the cloud occlusion ratio C cr = C c / C total ; where C c is the effective area actually covering the surface of the photovoltaic module, and C total is the total area of this region; and Based on formula R1: Calculate the predicted output power P predicted = P max (1 - C cr ) for predicted power calculation.

[0065] Preferably, when analyzing the actual meteorological parameters based on the cloud occlusion model, the factor acquisition module 204 is specifically configured to: Detect the light intensity every 5 minutes within the target time period, and calculate the ratio I of its instantaneous illuminance to the peak value in the same historical period R ; Perform smoothing processing on I R to reduce errors, and match it with the shielding effect correction coefficient k in the current cloud amount state to obtain the corrected shielding rate C cr′ , where C cr′ =(k ± λ)I R , and λ represents a preset proportional adjustment factor; Check the range of the shielding effect correction value: 0 ≤ C cr′ ≤ C total ; Use the corrected shielding rate to replace the initial C crSubstitute the value into the predicted power formula for final prediction, so as to obtain an accurate fluctuation influence factor.

[0066] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the foregoing method embodiments, and will not be elaborated herein.

[0067] An electronic device is provided in an embodiment of the present application, such as Figure 3 shown Figure 3 The electronic device 30 shown in the figure includes: a processor 301 and a memory 303. Among them, the processor 301 and the memory 303 are connected, such as through a bus 302. Optionally, the electronic device 30 may further include a transceiver 304. It should be noted that in actual applications, the transceiver 304 is not limited to one, and the structure of the electronic device 30 does not constitute a limitation on the embodiments of the present application.

[0068] The processor 301 may be a CPU (Central Processing Unit, central processing unit), a general-purpose processor, a DSP (Digital Signal Processor, data signal processor), an ASIC (Application Specific Integrated Circuit, application-specific integrated circuit), an FPGA (Field Programmable Gate Array, field programmable gate array), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It can implement or execute various exemplary logical blocks, modules, and circuits described in combination with the disclosure of the present application. The processor 301 may also be a combination that implements a computing function, such as a combination including one or more microprocessors, a combination of a DSP and a microprocessor, etc.

[0069] The bus 302 may include a path for transmitting information between the above components. The bus 302 may be a PCI (Peripheral Component Interconnect, peripheral component interconnect standard) bus or an EISA (Extended Industry Standard Architecture, extended industry standard architecture) bus, etc. The bus 302 may be divided into an address bus, a data bus, a control bus, etc. For the sake of convenience of representation, Figure 3 only a thick line is used to represent it in the figure, but it does not mean that there is only one bus or one type of bus.

[0070] The memory 303 can be a ROM (Read Only Memory), or other types of static storage devices that can store static information and instructions, a RAM (Random Access Memory), or other types of dynamic storage devices that can store information and instructions. It can also be an EEPROM (Electrically Erasable Programmable Read Only Memory), a CD-ROM (Compact Disc Read Only Memory), or other optical disc storage, optical disc storage (including compact discs, laser discs, optical discs, digital versatile discs, Blu-ray discs, etc.), magnetic disk storage media, or other magnetic storage devices, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer, but is not limited thereto.

[0071] The memory 303 is used to store the application program code for executing the solution of this application, and is controlled by the processor 301 for execution. The processor 301 is used to execute the application program code stored in the memory 303 to implement the content shown in the foregoing method embodiments.

[0072] Among them, the electronic device includes but is not limited to: mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Tablet Computers), PMPs (Portable Multimedia Players), in-vehicle terminals (such as in-vehicle navigation terminals), etc., and fixed terminals such as digital TVs, desktop computers, etc. It can also be a server, etc. Figure 3 The shown electronic device is only an example and should not impose any limitations on the functions and usage scope of the embodiments of this application.

[0073] An embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored. When it runs on a computer, the computer can execute the corresponding content in the foregoing method embodiment. Compared with the related art, in the embodiment of the present application, a cloud occlusion model is established by combining historical power generation data and real-time meteorological information. Since cloud occlusion is one of the main factors causing fluctuations in photovoltaic power generation, the prediction based on the cloud occlusion model can significantly improve the accuracy of power prediction and make the prediction result closer to the actual power generation situation. The cloud occlusion model is established based on the statistical analysis of historical power generation data and meteorological information, considering the complex relationship between the key factor of cloud occlusion and photovoltaic power generation. Therefore, by using this model to analyze actual meteorological parameters and predicting the power fluctuation influence factor and accordingly performing ultra-short-term prediction on the power output of the photovoltaic cluster, the grid voltage and frequency fluctuations caused by photovoltaic power generation fluctuations can be reduced. The key factors affecting photovoltaic power generation can be captured more accurately, thereby improving the prediction accuracy.

[0074] It should be understood that although the steps in the flowchart of the accompanying drawings are shown in sequence according to the indication of the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless there is a clear indication in this article, the execution of these steps has no strict order restriction, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed alternately or alternately with at least a part of other steps or sub-steps or stages of other steps.

[0075] The above are only some embodiments of the present application. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present application, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present application.

Claims

1. Ultra-short-term accurate power prediction method for distributed photovoltaic clusters, characterized in that: include: Obtain historical power generation data and real-time weather information; Establishing a cloud cover model based on the historical power generation data and the real-time meteorological information; Collect actual meteorological parameters in the area where the distributed photovoltaic cluster is located; Analyzing the actual meteorological parameters based on the cloud cover model to obtain a power fluctuation influencing factor; An ultra-short-term prediction is made on the power output of the distributed photovoltaic cluster according to the power fluctuation influencing factor.

2. The distributed photovoltaic cluster ultra-short-term accurate power prediction method according to claim 1 is characterized in that: The cloud cover model is established based on historical power generation data and real-time meteorological information, including: Cleaning the historical power generation data, determining and removing abnormal power generation records; Extracting features from the cloud distribution information in the real-time meteorological information using an image processing algorithm; Combine historical meteorological and historical power generation data to conduct correlation analysis to obtain reference indicators; The machine learning model is trained according to the reference index to obtain an occlusion coefficient model.

3. The distributed photovoltaic cluster ultra-short-term accurate power prediction method according to claim 1 is characterized in that: The analyzing the actual meteorological parameters based on the cloud cover model to obtain the power fluctuation influencing factor includes: Collect meteorological data in real time in the area where the photovoltaic power station is located; Establishing a cloud cover model based on the meteorological data; Using the cloud shielding model, calculate the attenuation degree of solar radiation on the ground caused by cloud shielding; The attenuation degree is determined as a power fluctuation influence factor.

4. The distributed photovoltaic cluster ultra-short-term accurate power prediction method according to claim 3 is characterized in that: The step of establishing a cloud cover model according to the meteorological data comprises: Obtaining a cloud image and a sun position in an area where the photovoltaic power station is located; Combining the cloud layer image with the meteorological data to determine cloud layer information, the cloud layer information including the height, thickness and type of the cloud layer; A position model of the cloud layer in three-dimensional space is constructed according to the sun position, the height, thickness and type of the cloud layer.

5. The distributed photovoltaic cluster ultra-short-term accurate power prediction method according to claim 1 is characterized in that: The cloud cover model is established based on historical power generation data and real-time meteorological information, including: Get the hourly power generation in the historical period; Collecting meteorological factor data for a period corresponding to the power generation; Draw a graph of the relationship between photovoltaic power generation and shading degree at different time periods; and The function expression R1 between the shading rate and the actual power generation is obtained from the curve graph using the statistical regression analysis method.

6. The distributed photovoltaic cluster ultra-short-term accurate power prediction method according to claim 5 is characterized in that: The drawing of the relationship curve between photovoltaic power generation and shading degree in different time periods specifically draws the variation law of cloud cover coefficient and photovoltaic power generation intensity, including: Measure the maximum power output P of the distributed photovoltaic system under clear weather conditions without any cloud cover. max ; Calculate the average photovoltaic output power P in each time period based on historical period data average And record the cloud distribution; Calculate the cloud cover ratio C cr =C c / C total ; where C c is the effective area actually covering the surface of the photovoltaic module, C total is the total area of ​​the region; and Based on formula R1: Estimated output power P predicted =P max (1-C cr ) to calculate the estimated power.

7. The distributed photovoltaic cluster ultra-short-term accurate power prediction method according to claim 1 is characterized in that: The analysis of actual meteorological parameters based on the cloud cover model includes: Detect the light intensity every 5 minutes during the target period and calculate the ratio of its instantaneous illumination to the historical peak value during the same period I R ; to I R Smoothing is performed to reduce the error, and the corrected shielding rate C is obtained by matching it with the shielding effect correction coefficient k under the current cloud cover state. cr′ , where C cr′ =(k±λ)I R , λ represents the preset proportional adjustment factor; Inspection shielding effect correction value range: 0≤C cr′ ≤ C total ; Use the corrected shielding ratio instead of the initial C cr The value is substituted into the estimated power formula for final prediction, thereby obtaining the accurate fluctuation impact factor.

8. A device for ultra-short-term accurate power prediction of distributed photovoltaic clusters, characterized in that: include: Information acquisition module, used to obtain historical power generation data and real-time meteorological information; A model building module, used to build a cloud cover model based on the historical power generation data and the real-time meteorological information; Parameter collection module, used to collect actual meteorological parameters in the area where the distributed photovoltaic cluster is located; A factor acquisition module, used for analyzing the actual meteorological parameters based on the cloud cover model to obtain a power fluctuation influencing factor; The prediction module is used to perform ultra-short-term prediction of the power output of the distributed photovoltaic cluster according to the power fluctuation influencing factor.

9. An electronic device, characterized in that: It includes: at least one processor; Memory; At least one application, wherein the at least one application is stored in the memory and configured to be executed by the at least one processor, and the at least one application is used to execute the ultra-short-term precise power prediction method for a distributed photovoltaic cluster according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed in a computer, the computer is caused to execute the distributed photovoltaic cluster ultra-short-term accurate power prediction method according to any one of claims 1 to 7.

Images