Photovoltaic power generation power prediction method and device considering power limitation, equipment and medium

Abstract

The invention relates to the technical field of photovoltaic power generation power prediction, and discloses a photovoltaic power generation power prediction method, device, equipment and medium considering power limitation, and the photovoltaic power generation power prediction method considering power limitation comprises the steps: carrying out the clustering analysis of first power data and first irradiance data of a sample day, and obtaining a power limitation moment set; the power limiting moment set is a moment set in which the first irradiance data is greater than a preset irradiance threshold value and the first power data is smaller than a preset power threshold value; selecting a preset number of target similar days from the plurality of historical days; using the second power data to update the power data corresponding to the power limiting moment set in the first power data of the sample day to obtain target power data; and predicting the power data of the day to be predicted according to the target power data and the first irradiance data. According to the invention, by correcting the data at the power limiting moment, the effectiveness of the target power data is enhanced, and the accuracy of photovoltaic power generation power prediction is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of photovoltaic power generation prediction, and in particular to a method, device, equipment and medium for predicting photovoltaic power generation considering power curtailment. Background Art

[0002] Photovoltaic power generation utilizes the photovoltaic effect at semiconductor interfaces to convert sunlight directly into electricity. Photovoltaic power generation offers significant advantages, including being a clean, renewable energy source that is environmentally friendly and reduces emissions of harmful gases like carbon dioxide. However, photovoltaic power generation suffers from intermittent, random, and volatile performance, posing a series of challenges to the safe operation of power grids.

[0003] One approach to addressing this issue is to predict photovoltaic power generation. Specifically, this approach uses a neural network model based on historical power data. However, this method uses inaccurate historical power data. For example, in the event of artificial power rationing, the historical power data may not reflect the power trend under normal conditions. This results in inaccurate predictions from the neural network model, which do not reflect the actual photovoltaic power generation. Summary of the Invention

[0004] In view of this, the present invention provides a method, device, equipment and medium for predicting photovoltaic power generation considering power curtailment, so as to solve the problem of inaccurate prediction results output by the neural network model when predicting photovoltaic power generation in the related art.

[0005] In a first aspect, the present invention provides a method for predicting photovoltaic power generation power considering power restriction, comprising: obtaining first power data of a sample day and first irradiance data of the sample day, performing cluster analysis on the first power data and the first irradiance data to obtain a power restriction moment set of the sample day; the power restriction moment set is a moment set when the first irradiance data is greater than a preset irradiance threshold and the first power data is less than a preset power threshold; selecting a target preset number of target similar days from multiple historical days based on the second irradiance data of multiple historical days and the first irradiance data of the sample day; using the second power data of the target preset number of target similar days, updating the power data corresponding to the power restriction moment set in the first power data of the sample day to obtain target power data; predicting the power data of the day to be predicted based on the target power data and the meteorological data of the day to be predicted to obtain a target photovoltaic power generation power prediction value.

[0006] The present invention can accurately find the power-restriction moment set by performing cluster analysis on the first power data and the first irradiance data of the sample day, that is, the moment when the first irradiance data is greater than the preset irradiance threshold but the first power data is less than the preset power threshold. The present invention selects a target preset number of target similar days from multiple historical days based on the second irradiance data of multiple historical days and the first irradiance data of the sample day, and filters similar days based on the irradiance data, so as to find historical conditions similar to the lighting conditions of the sample day, making subsequent reference data more targeted and effective. The present invention uses the second power data of the target preset number of target similar days to update the power data corresponding to the power-restriction moment set in the first power data of the sample day to obtain the target power data, eliminate the power-restriction data, avoid the abnormal data caused by artificial power restriction, and replace the power-restriction data by similar days, thereby enhancing the effectiveness of the target power data and improving the accuracy of subsequent power prediction. Based on the processed target power data and first irradiance data, the present invention predicts power data for the forecasted day to obtain a predicted target photovoltaic power generation value. By comprehensively considering multiple factors, the present invention improves the accuracy of photovoltaic power generation predictions, facilitating the rational planning of power generation plans for photovoltaic power plants and optimizing grid scheduling. Compared with related technologies, the present invention considers the possibility of artificial power curtailment, thereby improving the accuracy of power data predictions for the forecasted day.

[0007] In an optional embodiment, cluster analysis is performed on the first power data and the first irradiance data to obtain a set of power-limiting moments on the sample day, including: taking the first power data at each moment of the sample day as the first sample point, taking the first irradiance data at each moment of the sample day as the second sample point, determining a first truncation distance of multiple first sample points, and determining a second truncation distance of multiple second sample points; determining a first local density of each first sample point based on the first truncation distance, determining a first relative distance of each first sample point based on the first local density, determining a second local density of each second sample point based on the second truncation distance, and determining a second relative distance of each second sample point based on the second local density; determining a first cluster center based on the first local density and the first relative distance, clustering the multiple first sample points based on the first cluster center to obtain multiple first cluster groups, and determining a second cluster center based on the second local density and the second relative distance. Center, cluster multiple second sample points according to the second cluster center to obtain multiple second cluster groups; select a first representative point in each first cluster group, use each first representative point as the first initial cluster center, and obtain multiple first target cluster groups by updating and iterating the first initial cluster center and the first membership matrix; select a second representative point in each second cluster group, use each second representative point as the second initial cluster center, and obtain multiple second target cluster groups by updating and iterating the second initial cluster center and the second membership matrix; among the multiple first target cluster groups, select multiple third target cluster groups whose first irradiance data are all greater than the preset irradiance threshold, and among the multiple second target cluster groups, select multiple fourth target cluster groups whose first power data are all less than the preset power threshold; and use the set of multiple moments in the third target cluster group and in the fourth target cluster group as the power-limiting moment set.

[0008] In an optional embodiment, a target preset number of target similar days are selected from multiple historical days based on the second irradiance data of multiple historical days and the first irradiance data of the sample day, including: obtaining the second irradiance data of multiple historical days, and screening a preset number of similar days from multiple historical days based on the Euclidean distance between the second irradiance data of each historical day and the first irradiance data of the sample day; screening a target preset number of target similar days from the preset number of similar days based on the correlation coefficient between the third irradiance data of each similar day and the first irradiance data of the sample day; the target preset number is less than the preset number.

[0009] The present invention screens similar days by calculating the Euclidean distance between the irradiance data of historical days and sample days. The Euclidean distance can measure the "distance" of data in space. The closer the distance, the smaller the numerical difference in the irradiance data and the more similar the distribution characteristics. It can quickly screen a preset number of similar days from a large number of historical days, narrow the scope of subsequent analysis, and improve screening efficiency. Based on the similar days initially screened out, the present invention further calculates the correlation coefficient between the irradiance data of similar days and sample days. The correlation coefficient can measure the degree of linear correlation between the two sets of data. The higher the correlation coefficient, the more consistent the change trend between the two. The target similar day with the most similar irradiance change trend to the sample day can be accurately selected from the similar days initially screened out, making the data subsequently used for power data update and prediction more targeted and reliable.

[0010] In an optional embodiment, multiple target similar days are selected from multiple historical days based on the second irradiance data of multiple historical days and the first irradiance data of the sample day, and the method also includes: aligning the first time series data corresponding to the second irradiance data of each historical day with the second time series data corresponding to the first irradiance data of the sample day; determining the dynamic programming distance between the first time series data of each historical day after the alignment and the aligned second time series data; and selecting multiple target similar days from multiple historical days based on the multiple dynamic programming distances.

[0011] The present invention aligns the time series corresponding to the irradiance data of historical days and sample days, which can eliminate the differences caused by inconsistent time scales. After the time series data are aligned, the distance between the time series data of each historical day and the sample day is calculated. The distance can quantify the degree of difference between the data. The smaller the distance, the more similar the pattern of irradiance change over time between the historical day and the sample day is. Target similar days with highly similar irradiance change patterns to the sample day can be accurately screened out from multiple historical days.

[0012] In an optional embodiment, the second power data of a target preset number of target similar days is used to update the power data corresponding to the power-restriction moment set in the first power data of the sample day to obtain the target power data, including: extracting the second power data of each power-restriction moment in the power-restriction moment set on each target similar day, performing weighted averaging on the corresponding second power data of each power-restriction moment on each target similar day to obtain the third power data of each power-restriction moment; replacing the power data corresponding to each power-restriction moment in the first power data of the sample day with the third power data of the corresponding moment to obtain the target power data.

[0013] In an optional embodiment, the power data of the day to be predicted is predicted based on the target power data and the meteorological data of the day to be predicted to obtain a target photovoltaic power generation power prediction value, including: inputting the target power data and the meteorological data of the day to be predicted into a trained photovoltaic power generation power prediction model to obtain the target photovoltaic power generation power prediction value; the photovoltaic power generation power prediction model is a neural network model constructed based on a converter and a multi-layer perceptron; the input of the photovoltaic power generation power prediction model is the target power data and meteorological data, and the output of the photovoltaic power generation power prediction model is the target photovoltaic power generation power prediction value.

[0014] In a second aspect, the present invention provides a device for predicting photovoltaic power generation power taking into account power restriction, including: a cluster analysis module for obtaining the first power data of a sample day and the first irradiance data of the sample day, performing cluster analysis on the first power data and the first irradiance data to obtain a power restriction moment set of the sample day; the power restriction moment set is a moment set when the first irradiance data is greater than a preset irradiance threshold and the first power data is less than a preset power threshold; a similar day selection module for selecting a target preset number of target similar days from multiple historical days based on the second irradiance data of multiple historical days and the first irradiance data of the sample day; a data update module for updating the power data corresponding to the power restriction moment set in the first power data of the sample day using the second power data of the target preset number of target similar days to obtain target power data; a power prediction module for predicting the power data of the day to be predicted based on the target power data and the meteorological data of the day to be predicted to obtain a target photovoltaic power prediction value.

[0015] In a third aspect, the present invention provides a computer device comprising: a memory and a processor, the memory and the processor being communicatively connected to each other, the memory storing computer instructions, and the processor executing the computer instructions to execute the method for predicting photovoltaic power generation power considering power rationing according to the above-mentioned first aspect or any corresponding embodiment thereof.

[0016] In a fourth aspect, the present invention provides a computer-readable storage medium having computer instructions stored thereon, the computer instructions being used to enable a computer to execute the method for predicting photovoltaic power generation power taking into account power curtailment according to the first aspect or any corresponding embodiment thereof.

[0017] In a fifth aspect, the present invention provides a computer program product comprising computer instructions for causing a computer to execute the method for predicting photovoltaic power generation with consideration of power curtailment according to the first aspect or any corresponding embodiment thereof. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the specific embodiments of the present invention or the technical solutions in related technologies, the following briefly introduces the drawings required for use in the specific embodiments or related technical descriptions. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0019] Figure 1 4 is a flow chart of a method for predicting photovoltaic power generation considering power curtailment according to an embodiment of the present invention.

[0020] Figure 2 4 is a flow chart of another method for predicting photovoltaic power generation considering power curtailment according to an embodiment of the present invention.

[0021] Figure 3 4 is a flow chart of another method for predicting photovoltaic power generation considering power curtailment according to an embodiment of the present invention.

[0022] Figure 4 4 is a structural block diagram of a device for predicting photovoltaic power generation considering power curtailment according to an embodiment of the present invention.

[0023] Figure 5 Schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0024] To make the purpose, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without making creative efforts shall fall within the scope of protection of the present invention.

[0025] Photovoltaic power generation is characterized by intermittent, random, and volatile characteristics, which pose a series of challenges to the safe operation of the power grid. As photovoltaic power stations increase in the proportion of power grid power, photovoltaic power prediction systems have become increasingly important. Currently, historical data is often used to train power prediction models, and the accuracy of training data is crucial for model training.

[0026] Current training data often includes instances of artificial power curtailment. During these periods, actual power may not conform to normal power trends. This can lead to inaccurate predictions from trained models. Related technologies fail to consider the impact of curtailment on PV power, resulting in low PV power prediction accuracy.

[0027] The embodiment of the present invention provides a method for predicting photovoltaic power generation considering power curtailment, which enhances the validity of target power data by correcting data at the time of power curtailment, thereby achieving the effect of improving the accuracy of photovoltaic power generation prediction.

[0028] According to an embodiment of the present invention, an embodiment of a method for predicting photovoltaic power generation considering power curtailment is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0029] In this embodiment, a method for predicting photovoltaic power generation considering power curtailment is provided, which can be used for computer equipment. Figure 1 FIG. 1 is a flow chart of a method for predicting photovoltaic power generation considering power curtailment according to an embodiment of the present invention. Figure 1 As shown, the process includes the following steps:

[0030] Step S101, obtain the first power data of the sample day and the first irradiance data of the sample day, perform cluster analysis on the first power data and the first irradiance data, and obtain the power-limiting time set of the sample day; the power-limiting time set is the time set when the first irradiance data is greater than the preset irradiance threshold and the first power data is less than the preset power threshold.

[0031] Among them, the sample day is a plurality of specific dates pre-selected for photovoltaic power generation analysis and prediction; the first power data is data related to the power actually generated by the photovoltaic power station; and the first irradiance data is the solar radiation intensity data received at the location of the photovoltaic power station.

[0032] In some optional embodiments, the first power data of the sample day and the first irradiance data of the sample day are obtained from the photovoltaic power station monitoring system. The power monitoring equipment equipped in the photovoltaic power station can collect the inverter output power in real time. These devices record the power data at a certain time interval (such as every minute, every 15 minutes) and store it in the local server or cloud database of the photovoltaic power station monitoring system. The embodiment of the present invention can obtain the first power data by filtering the corresponding database according to the date and time of the sample day; irradiance sensors (such as silicon photodiodes, thermocouple sensors) installed at appropriate locations around the photovoltaic power station can convert the received solar radiation energy into electrical signals. After signal processing, the data acquisition system collects and converts it into digital signals for storage at a specific sampling frequency (such as once per minute), and extracts it from the storage device according to the sample day to obtain the first irradiance data.

[0033] In some optional embodiments, a DPC (Density Peak Clustering) algorithm is used to perform preliminary clustering on the first power data of the sample day and the first irradiance data of the sample day, and representative points of each cluster (such as core points or a combination of core points and boundary points) are extracted from the results of the DPC algorithm as the initial clustering centers of the FCM (Density Peak Clustering) algorithm. After the initial clustering centers are determined, the FCM algorithm is used for subsequent clustering processes. By iteratively updating the membership matrix and clustering centers, the FCM algorithm can find a fuzzy division of the data so that the data points in the same cluster are as similar as possible, and the difference between the data points of different clusters is as large as possible. The moments in the high irradiance cluster and the low actual power cluster are regarded as power-limited moments, otherwise they are non-power-limited moments.

[0034] Specifically, a cluster analysis is performed on the first power data and the first irradiance data to obtain a set of power-limiting moments on the sample day, including: taking the first power data at each moment of the sample day as the first sample point, taking the first irradiance data at each moment of the sample day as the second sample point, determining a first truncation distance of multiple first sample points, and determining a second truncation distance of multiple second sample points; determining a first local density of each first sample point according to the first truncation distance, determining a first relative distance of each first sample point according to the first local density, determining a second local density of each second sample point according to the second truncation distance, and determining a second relative distance of each second sample point according to the second local density; determining a first cluster center according to the first local density and the first relative distance, clustering multiple first sample points according to the first cluster center to obtain multiple first cluster groups, determining a second cluster center according to the second local density and the second relative distance, and Clustering multiple second sample points according to the second cluster center to obtain multiple second cluster groups; selecting a first representative point in each first cluster group, taking each first representative point as the first initial cluster center, and obtaining multiple first target cluster groups by updating and iterating the first initial cluster center and the first membership matrix; selecting a second representative point in each second cluster group, taking each second representative point as the second initial cluster center, and obtaining multiple second target cluster groups by updating and iterating the second initial cluster center and the second membership matrix; selecting multiple third target cluster groups whose first irradiance data are all greater than a preset irradiance threshold from the multiple first target cluster groups, and selecting multiple fourth target cluster groups whose first power data are all less than a preset power threshold from the multiple second target cluster groups; and taking a set of multiple moments in the third target cluster group and in the fourth target cluster group as a power restriction moment set.

[0035] The preset irradiance threshold can be set according to the actual situation. The preset irradiance threshold should be set larger so that the selected first irradiance data is as large as possible. For example, the preset irradiance threshold can be 500W / m 2 ; The preset power threshold can be set according to actual conditions. The preset power threshold should be set smaller so that the selected first power data is as small as possible. For example, the preset power threshold can be 300KW.

[0036] The DPC algorithm of the embodiment of the present invention can quickly determine the cluster center and the approximate cluster structure, and the FCM algorithm can further optimize the clustering results based on the DPC algorithm and refine the attribution of data points. The combination of the two not only takes advantage of the DPC algorithm's strong adaptability to data distribution and the automatic determination of the number of clusters, but also gives play to the strengths of the FCM algorithm in considering data fuzziness and iterative optimization, and more accurately extracts the power-limiting time set from the power and irradiance data. A single algorithm may behave unstable when faced with data noise, outliers, etc. The combination of the DPC algorithm and the FCM algorithm can enhance the overall robustness of the algorithm. Noise may be mixed in during the photovoltaic data collection process. The combined algorithm can suppress noise interference to a certain extent and more reliably identify the true power-limiting time.

[0037] Step S102 : selecting a preset number of target similar days from the plurality of historical days according to the second irradiance data of the plurality of historical days and the first irradiance data of the sample day.

[0038] The historical day is a pre-selected date before the day to be predicted; the target preset number can be set according to actual needs. For example, the target preset number can be 3.

[0039] In some optional embodiments, a target preset number of target similar days is selected from multiple historical days based on the second irradiance data of multiple historical days and the first irradiance data of the sample day, including: using the Euclidean distance algorithm to determine the distance matrix between the second irradiance data of each historical day and the first irradiance data of the sample day, and screening a preset number of similar days from multiple historical days based on the distance matrix between the second irradiance data of each historical day and the first irradiance data of the sample day; using the Pearson correlation coefficient algorithm to determine the correlation coefficient between the third irradiance data of each similar day and the first irradiance data of the sample day, and screening a target preset number of target similar days from the preset number of similar days based on the correlation coefficient between the third irradiance data of each similar day and the first irradiance data of the sample day; the target preset number is less than the preset number.

[0040] The preset number can be set according to actual conditions, and for example, the preset number can be 10. First, based on the Euclidean distance matrix, 10 similar days with the irradiance most similar to the sample day are found, and then based on the correlation matrix, the three most similar days are selected from these 10 similar days as the final target similar days.

[0041] In some optional embodiments, based on the second irradiance data of multiple historical days and the first irradiance data of the sample day, a target preset number of target similar days are selected from multiple historical days, including: based on the DTW (Dynamic TimeWarping) algorithm, determining the dynamic programming distance between the second irradiance data of each historical day and the first irradiance data of the sample day, and selecting a target preset number of target similar days based on the dynamic programming distance.

[0042] Step S103 : using the second power data of a target preset number of target similar days, update the power data corresponding to the power-cutoff time set in the first power data of the sample day to obtain target power data.

[0043] Among them, the power data corresponding to the power-limiting moments in a target preset number of similar days are weighted averaged to obtain the power data of each power-limiting moment, and the power data of each power-limiting moment is used to replace the power data of the corresponding power-limiting moment in the first power data of the sample day, so as to obtain the target power data based on the replaced first power data.

[0044] Specifically, the second power data of a target preset number of target similar days is used to update the power data corresponding to the power-restriction moment set in the first power data of the sample day to obtain the target power data, including: extracting the second power data of each power-restriction moment in the power-restriction moment set on each target similar day, and performing weighted averaging on the corresponding second power data of each power-restriction moment on each target similar day to obtain the third power data of each power-restriction moment; replacing the power data corresponding to each power-restriction moment in the first power data of the sample day with the third power data of the corresponding moment to obtain the target power data.

[0045] Step S104 , predicting the power data of the day to be predicted based on the target power data and the meteorological data of the day to be predicted, and obtaining a target photovoltaic power prediction value.

[0046] In some optional embodiments, the power data of the day to be predicted is predicted based on the target power data and the meteorological data of the day to be predicted to obtain a target photovoltaic power generation power prediction value, including: inputting the target power data and the meteorological data of the day to be predicted into a trained photovoltaic power generation power prediction model to obtain the target photovoltaic power generation power prediction value; the photovoltaic power generation power prediction model is a neural network model constructed based on a converter and a multi-layer perceptron; the input of the photovoltaic power generation power prediction model is the target power data and meteorological data, and the output of the photovoltaic power generation power prediction model is the target photovoltaic power generation power prediction value.

[0047] The meteorological data may be irradiance data and humidity data. In an embodiment of the present invention, the meteorological data (irradiance and humidity) at 96 time points on the day to be predicted and the target power data at 96 time points on the sample day are input into a trained photovoltaic power prediction model to obtain a predicted target photovoltaic power value for the day to be predicted. The 96 time points in a day may be obtained every 15 minutes, starting from a certain moment, for a total of 96 time points per day.

[0048] In some optional implementations, the photovoltaic power generation prediction model is a neural network model constructed based on a transformer and a multi-layer perceptron (MLP).

[0049] In an embodiment of the present invention, the photovoltaic power generation prediction model uses the output of the Transformer encoder as the input of the MLP. The MLP focuses on the local features of the input data and is more suitable for the problem of using similar daily power data to predict future power in an embodiment of the present invention. It can better capture short-term trends and influence relationships in historical time series.

[0050] Among them, the structure of Transformer is as follows: input layer: maps the input feature vector (dimension is 96) to a new embedding space through linear transformation and nonlinear activation function; position encoding layer: adds unique information to each position in the input sequence; two Transformer encoder layers: one of the core components of the Transformer model, responsible for processing the input sequence, each layer contains a self-attention mechanism and a feedforward neural network, as well as layer normalization and dropout; feature embedding layer: maps the original features to the embedding space. The structure of MLP is as follows: normalization layer (Layer Normalization, LayerNorm): normalizes the input features; linear layer (Linear): performs linear transformation on the normalized input; regularization layer (Dropout): randomly sets part of the input unit to 0 to prevent the model from overfitting; activation function (Activation_Function): introduces nonlinearity so that the model can learn complex patterns; the second linear layer (Linear): performs another linear transformation on the output of the activation function to prepare for output.

[0051] In some optional embodiments, the method for predicting photovoltaic power generation considering power restrictions also includes a training process of a photovoltaic power generation prediction model. The training process of the photovoltaic power generation prediction model includes: inputting the training meteorological data (irradiance and humidity) at 96 time points on the training day to be predicted and the target training power data at 96 time points on the training sample day into the photovoltaic power generation prediction model to train the photovoltaic power generation prediction model.

[0052] The photovoltaic power generation power prediction method provided by this embodiment, taking power curtailment into account, can accurately identify a set of power curtailment moments by performing cluster analysis on the first power data and first irradiance data of a sample day, namely, moments when the first irradiance data exceeds a preset irradiance threshold but the first power data is less than a preset power threshold. This embodiment of the present invention selects a preset number of target similar days from multiple historical days based on the second irradiance data of multiple historical days and the first irradiance data of the sample day. By screening similar days based on the irradiance data, it can identify historical conditions with similar lighting conditions to the sample day, making subsequent reference data more targeted and effective. This embodiment of the present invention uses the second power data of the preset number of target similar days to update the power data corresponding to the power curtailment moment set in the first power data of the sample day to obtain target power data, eliminate power curtailment data, and avoid abnormal data caused by artificial power curtailment. By replacing the power curtailment data with similar days, the effectiveness of the target power data is enhanced, improving the accuracy of subsequent power predictions. This embodiment of the present invention predicts power data for the forecasted day based on the processed target power data and first irradiance data, obtaining a predicted target photovoltaic power generation value. This comprehensive consideration of multiple factors improves the accuracy of photovoltaic power generation predictions, facilitating the rational planning of power generation plans for photovoltaic power plants and optimized grid scheduling. Compared to related technologies, this embodiment of the present invention takes into account the possibility of artificial power curtailment, improving the accuracy of power data predictions for the forecasted day.

[0053] In this embodiment, a method for predicting photovoltaic power generation considering power curtailment is provided, which can be used for computer equipment. Figure 2 FIG. 1 is a flow chart of another method for predicting photovoltaic power generation considering power curtailment according to an embodiment of the present invention. Figure 2 As shown, the process includes the following steps:

[0054] Step S201: Obtain the first power data and the first irradiance data of the sample day, perform cluster analysis on the first power data and the first irradiance data, and obtain the power-limiting time set of the sample day; the power-limiting time set is the time set when the first irradiance data is greater than the preset irradiance threshold and the first power data is less than the preset power threshold. For details, please refer to Figure 1 Step S101 of the illustrated embodiment will not be described in detail here.

[0055] Step S202 : selecting a preset number of target similar days from the plurality of historical days based on the second irradiance data of the plurality of historical days and the first irradiance data of the sample day.

[0056] Specifically, the above step S202 includes:

[0057] Step S2021 , obtaining second irradiance data of multiple historical days, and screening a preset number of similar days from the multiple historical days based on the Euclidean distance between the second irradiance data of each historical day and the first irradiance data of the sample day.

[0058] Wherein, a plurality of Euclidean distances are obtained according to the distance between the first vector corresponding to the second irradiance data of each historical day and the second vector corresponding to the first irradiance data of the sample day.

[0059] For example, the formula for determining the Euclidean distance is:

[0060]

[0061] Among them, D k is the Euclidean distance between the second irradiance data of the kth historical day and the first irradiance data of the sample day, x kl is the lth irradiance characteristic value of the kth historical day, y k is the lth irradiance characteristic value of the sample day, and L is the total number of irradiance characteristic values.

[0062] In some optional embodiments, based on the Euclidean distance between the second irradiance data of each historical day and the first irradiance data of the sample day, a preset number of similar days are screened from multiple historical days, including: selecting the first preset number of similar days sorted in ascending order from the multiple Euclidean distances.

[0063] For example, among a plurality of Euclidean distances, similar days corresponding to the top 10 Euclidean distances sorted in ascending order are selected.

[0064] Step S2022 , based on the correlation coefficient between the third irradiance data of each similar day and the first irradiance data of the sample day, screening a target preset number of target similar days from a preset number of similar days; the target preset number is less than the preset number.

[0065] Among them, the target preset number can be 3.

[0066] In some optional embodiments, based on the correlation coefficient between the third irradiance data of each similar day and the first irradiance data of the sample day, a second preset number of target similar days are screened from a first preset number of similar days, including: determining the first mean of the third irradiance data of each similar day, and determining the second mean of multiple meteorological data of the day to be predicted; determining the covariance based on the third irradiance data of each similar day, the first mean, the meteorological data of the day to be predicted, and the second mean; determining the first standard deviation based on the third irradiance data of each similar day and the first mean; determining the second standard deviation based on the meteorological data of the day to be predicted and the second mean; obtaining the target standard deviation based on the product of the first standard deviation and the second standard deviation; obtaining multiple correlation coefficients based on the quotient of the covariance of each similar day and the target standard deviation of each similar day; and screening out a target preset number of target similar days from the preset number of similar days.

[0067] For example, the calculation process of the correlation coefficient is:

[0068]

[0069] Among them, r i is the correlation coefficient between the third irradiance data of the i-th similar day and the first irradiance data of the sample day, J is the total number of irradiance data, R ij is the third irradiance data of the jth similar day, is the first mean of the third irradiance data of the i-th similar day, M j is the first irradiance data of the jth sample day, is the second mean of the first irradiance data of the sample day

[0070] In some optional implementations, screening out a target preset number of target similar days from a preset number of similar days includes: selecting a target preset number of target similar days that are ranked first in descending order from a plurality of correlation coefficients.

[0071] For example, among multiple correlation coefficients, target similar days corresponding to the top three correlation coefficients ranked in descending order are selected.

[0072] Step S203 : using the second power data of a target preset number of target similar days, update the power data corresponding to the power-cutoff time set in the first power data of the sample day to obtain target power data.

[0073] Specifically, the above step S203 includes:

[0074] Step S2031: extract the second power data of each power-cutting moment in the power-cutting moment set on each target similar day, and perform weighted averaging on the second power data corresponding to each power-cutting moment on each target similar day to obtain the third power data of each power-cutting moment.

[0075] Among them, the corresponding second power data of each power-restriction moment in each target similar day is weighted averaged, including: obtaining a target preset number of weight coefficients; the weight coefficient is used to reflect the degree of influence of the second power data of each power-restriction moment in the power-restriction moment set of each target similar day; the second power data of each power-restriction moment in each target similar day is multiplied by the corresponding weight coefficient to obtain the product of the target preset number, and the products of the target preset numbers are summed to obtain the third power data of each power-restriction moment; wherein, the sum of the weight coefficients of the target preset number is 1.

[0076] In some optional embodiments, obtaining a weight coefficient includes: selecting a target preset number of target training similar days from multiple training historical days based on the first training irradiance data of multiple training historical days and the second training irradiance data of the training sample day, taking the first training power data of each power-limiting moment in the power-limiting moment set on each target training similar day as input, taking the actual power data as output, constructing a linear regression equation, solving the linear regression equation, and determining the weight coefficient corresponding to each target similar day.

[0077] Step S2032: Replace the power data corresponding to each power-limiting moment in the first power data of the sample day with the third power data at the corresponding moment to obtain target power data.

[0078] In the embodiment of the present invention, the first power data after replacement is used as the target power data.

[0079] Step S204: predict the power data of the day to be predicted based on the target power data and the meteorological data of the day to be predicted, and obtain the target photovoltaic power prediction value. Figure 1 Step S104 of the illustrated embodiment will not be described in detail here.

[0080] The present embodiment provides a method for predicting photovoltaic power generation power taking into account power restrictions. Similar days are screened by calculating the Euclidean distance between the irradiance data of historical days and sample days. The Euclidean distance can measure the "distance" of data in space. The closer the distance, the smaller the numerical difference of the irradiance data and the more similar the distribution characteristics. A preset number of similar days can be quickly screened out from a large number of historical days, thereby narrowing the scope of subsequent analysis and improving screening efficiency. Based on the similar days preliminarily screened out, the embodiment of the present invention further calculates the correlation coefficient between the irradiance data of similar days and sample days. The correlation coefficient can measure the degree of linear correlation between the two groups of data. The higher the correlation coefficient, the more consistent the change trend between the two. The target similar day that is most similar to the irradiance change trend of the sample day can be accurately selected from the similar days preliminarily screened out, making the data subsequently used for power data update and prediction more targeted and reliable.

[0081] In this embodiment, a method for predicting photovoltaic power generation considering power curtailment is provided, which can be used for computer equipment. Figure 3 FIG. 1 is a flow chart of another method for predicting photovoltaic power generation considering power curtailment according to an embodiment of the present invention. Figure 3 As shown, the process includes the following steps:

[0082] Step S301: Obtain the first power data and the first irradiance data of the sample day, perform cluster analysis on the first power data and the first irradiance data, and obtain the power-limiting time set of the sample day; the power-limiting time set is the time set when the first irradiance data is greater than the preset irradiance threshold and the first power data is less than the preset power threshold. For details, please see Figure 1 Step S101 of the illustrated embodiment will not be described in detail here.

[0083] Step S302 : selecting a preset number of target similar days from the plurality of historical days according to the second irradiance data of the plurality of historical days and the first irradiance data of the sample day.

[0084] Specifically, the above step S302 includes:

[0085] Step S3021 : aligning the first time series data corresponding to the second irradiance data of each historical day with the second time series data corresponding to the first irradiance data of the sample day.

[0086] Step S3022 : determining the dynamic programming distance between the aligned first time series data of each historical day and the aligned second time series data.

[0087] Step S3023: Select multiple target similar days from multiple historical days based on multiple dynamic programming distances.

[0088] Among the multiple distances, a preset number of target similarity days that are ranked first in ascending order are selected. For example, among the multiple distances, target similarity days corresponding to the top three distances that are ranked first in ascending order are selected.

[0089] In an embodiment of the present invention, similarity calculation is performed based on the DTW algorithm. The core idea of ​​the DTW algorithm is to find an optimal matching path between time series, so that the two sequences are dynamically aligned in time, thereby calculating the distance between them. The smaller the distance, the higher the similarity. For the second irradiance data of each historical day and the first irradiance data of the sample day, due to factors such as weather changes, the time points of irradiance changes may not be completely consistent. For example, the irradiance peak of the historical day occurs several time points later than that of the sample day, but the overall change trend is similar. The DTW algorithm can find an optimal matching method between the two time series through dynamic programming methods to adapt to this time misalignment.

[0090] Step S303: Using the second power data of the target preset number of target similar days, the power data corresponding to the power-cutoff time set in the first power data of the sample day is updated to obtain the target power data. Figure 1 Step S103 of the illustrated embodiment will not be described in detail here.

[0091] Step S304: predict the power data of the day to be predicted based on the target power data and the meteorological data of the day to be predicted, and obtain the target photovoltaic power prediction value. Figure 1 Step S104 of the illustrated embodiment will not be described in detail here.

[0092] The photovoltaic power generation power prediction method considering power rationing in an embodiment of the present invention aligns the time series corresponding to the irradiance data of historical days and sample days, which can eliminate the differences caused by inconsistent time scales. After the time series data are aligned, the distance between the time series data of each historical day and the sample day is calculated. The distance can quantify the degree of difference between the data. The smaller the distance, the more similar the pattern of irradiance change over time between the historical day and the sample day is. Target similar days that are highly similar to the irradiance change pattern of the sample day can be accurately screened out from multiple historical days.

[0093] This embodiment also provides a device for predicting photovoltaic power generation considering power curtailment. This device is used to implement the above-mentioned embodiments and preferred embodiments, and the details already described will not be repeated. As used below, the term "module" may refer to a combination of software and / or hardware that implements a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, implementation using hardware, or a combination of software and hardware, is also possible and contemplated.

[0094] This embodiment provides a photovoltaic power prediction device taking power restrictions into consideration. Figure 4 Shown, including:

[0095] The cluster analysis module 401 is used to obtain the first power data of the sample day and the first irradiance data of the sample day, perform cluster analysis on the first power data and the first irradiance data, and obtain the power-limiting time set of the sample day; the power-limiting time set is the time set when the first irradiance data is greater than the preset irradiance threshold and the first power data is less than the preset power threshold.

[0096] The similar day selection module 402 is configured to select a preset number of target similar days from the plurality of historical days based on the second irradiance data of the plurality of historical days and the first irradiance data of the sample day.

[0097] The data updating module 403 is configured to update the power data corresponding to the power-cutoff time set in the first power data of the sample day by using the second power data of a target preset number of target similar days to obtain target power data.

[0098] The power prediction module 404 is used to predict the power data of the to-be-predicted day based on the target power data and the meteorological data of the to-be-predicted day, and obtain a target photovoltaic power prediction value.

[0099] In some optional implementations, the cluster analysis module 401 includes:

[0100] The truncation distance determination unit is used to use the first power data at each moment of the sample day as the first sample point and the first irradiance data at each moment of the sample day as the second sample point to determine the first truncation distance of multiple first sample points and the second truncation distance of multiple second sample points.

[0101] The relative distance determining unit is configured to determine a first local density of each first sample point based on the first cutoff distance, determine a first relative distance of each first sample point based on the first local density, determine a second local density of each second sample point based on the second cutoff distance, and determine a second relative distance of each second sample point based on the second local density.

[0102] The first cluster analysis unit is used to determine a first cluster center based on the first local density and the first relative distance, cluster the multiple first sample points based on the first cluster center to obtain multiple first cluster groups, determine a second cluster center based on the second local density and the second relative distance, cluster the multiple second sample points based on the second cluster center to obtain multiple second cluster groups.

[0103] The second cluster analysis unit is used to select a first representative point in each first cluster group, use each first representative point as a first initial cluster center, and obtain multiple first target cluster groups by updating and iterating the first initial cluster center and the first membership matrix; select a second representative point in each second cluster group, use each second representative point as a second initial cluster center, and obtain multiple second target cluster groups by updating and iterating the second initial cluster center and the second membership matrix.

[0104] The target cluster selection unit is used to select multiple third target cluster groups from multiple first target cluster groups whose first irradiance data are all greater than a preset irradiance threshold, and to select multiple fourth target cluster groups from multiple second target cluster groups whose first power data are all less than a preset power threshold.

[0105] The power-restriction time determination unit is configured to use a set of multiple time periods in the third target cluster grouping and in the fourth target cluster grouping as a power-restriction time set.

[0106] In some optional implementations, the similar day selection module 402 includes:

[0107] The similar day selection unit is used to obtain the second irradiance data of multiple historical days, and screen a preset number of similar days from the multiple historical days based on the Euclidean distance between the second irradiance data of each historical day and the first irradiance data of the sample day.

[0108] The target similar day selection unit is used to screen a preset number of target similar days from a preset number of similar days according to the correlation coefficient between the third irradiance data of each similar day and the first irradiance data of the sample day; the target preset number is less than the preset number.

[0109] In some optional implementations, the similar day selection module 402 includes:

[0110] The alignment processing unit is used to align the first time series data corresponding to the second irradiance data of each historical day with the second time series data corresponding to the first irradiance data of the sample day.

[0111] The distance determination unit is used to determine the dynamic programming distance between the first time series data of each historical day after the alignment process and the aligned second time series data.

[0112] The target similar day selection unit is used to select multiple target similar days from multiple historical days according to multiple dynamic programming distances.

[0113] In some optional implementations, the data updating module 403 includes:

[0114] The third power data determination unit is used to extract the second power data of each power-limiting moment in the power-limiting moment set on each target similar day, and perform weighted averaging on the corresponding second power data of each power-limiting moment on each target similar day to obtain the third power data of each power-limiting moment.

[0115] The data updating unit is configured to replace the power data corresponding to each power-limiting moment in the first power data of the sample day with the third power data at the corresponding moment to obtain the target power data.

[0116] In some optional implementations, the power prediction module 404 includes:

[0117] The power prediction unit is used to input the target power data and the meteorological data of the day to be predicted into the trained photovoltaic power generation power prediction model to obtain the target photovoltaic power generation power prediction value; the photovoltaic power generation power prediction model is a neural network model constructed based on a converter and a multi-layer perceptron; the input of the photovoltaic power generation power prediction model is the target power data and meteorological data, and the output of the photovoltaic power generation power prediction model is the target photovoltaic power generation power prediction value.

[0118] The further functional description of each of the above modules and units is the same as that of the above corresponding embodiments and will not be repeated here.

[0119] The photovoltaic power generation prediction device considering power curtailment in this embodiment is presented in the form of a functional unit, where the unit refers to an ASIC (Application Specific Integrated Circuit) circuit, a processor and memory that executes one or more software or fixed programs, and / or other devices that can provide the above functions.

[0120] The embodiment of the present invention also provides a computer device having the above Figure 4 The device for predicting photovoltaic power generation considering power curtailment is shown.

[0121] See also Figure 5 , Figure 5 is a structural diagram of a computer device provided by an optional embodiment of the present invention, such as Figure 5As shown, the computer device includes: one or more processors 10, memory 20, and interfaces for connecting various components, including high-speed interfaces and low-speed interfaces. Various components utilize different buses to communicate with each other and can be installed on a common mainboard or installed in other ways as needed. The processor can process the instructions executed in the computer device, including instructions stored in the memory or on the memory to display the graphical information of the GUI on an external input / output device (such as, a display device coupled to the interface). In some optional embodiments, if necessary, multiple processors and / or multiple buses can be used together with multiple memories and multiple memories. Equally, multiple computer devices can be connected, and each device provides part of the necessary operations (for example, as a server array, a group of blade servers, or a multi-processor system). Figure 5 A processor 10 is taken as an example.

[0122] The processor 10 may be a central processing unit, a network processor, or a combination thereof. The processor 10 may further include a hardware chip. The hardware chip may be an application-specific integrated circuit, a programmable logic device, or a combination thereof. The programmable logic device may be a complex programmable logic device, a field programmable gate array, a general purpose array logic, or any combination thereof.

[0123] The memory 20 stores instructions that can be executed by at least one processor 10, so that the at least one processor 10 executes the method shown in the above embodiment.

[0124] The memory 20 may include a program storage area and a data storage area, wherein the program storage area may store an operating system and application programs required for at least one function; the data storage area may store data created based on the use of the computer device, etc. In addition, the memory 20 may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some optional embodiments, the memory 20 may optionally include a memory remotely located relative to the processor 10, and these remote memories may be connected to the computer device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0125] The memory 20 may include a volatile memory, such as a random access memory; the memory may also include a non-volatile memory, such as a flash memory, a hard disk or a solid-state drive; the memory 20 may also include a combination of the above types of memory.

[0126] The computer device further includes a communication interface 30 for the computer device to communicate with other devices or a communication network.

[0127] The embodiment of the present invention also provides a computer-readable storage medium. The above-mentioned method according to the embodiment of the present invention can be implemented in hardware, firmware, or implemented as a computer code that can be recorded in a storage medium, or implemented as a computer code that is originally stored in a remote storage medium or a non-temporary machine-readable storage medium and downloaded through a network and will be stored in a local storage medium, so that the method described herein can be stored in such software processing on a storage medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware. Among them, the storage medium can be a magnetic disk, an optical disk, a read-only storage memory, a random access memory, a flash memory, a hard disk or a solid-state drive, etc.; further, the storage medium can also include a combination of the above-mentioned types of memory. It can be understood that a computer, a processor, a microprocessor controller or programmable hardware includes a storage component that can store or receive software or computer code. When the software or computer code is accessed and executed by a computer, a processor or hardware, the method shown in the above embodiment is implemented.

[0128] A portion of the present invention may be applied as a computer program product, such as a computer program instruction, which, when executed by a computer, can call or provide the method and / or technical solution according to the present invention through the operation of the computer. Those skilled in the art should understand that the form in which the computer program instruction exists in a computer-readable medium includes, but is not limited to, a source file, an executable file, an installation package file, etc. Accordingly, the way in which the computer program instruction is executed by the computer includes, but is not limited to: the computer directly executes the instruction, or the computer compiles the instruction and then executes the corresponding compiled program, or the computer reads and executes the instruction, or the computer reads and installs the instruction and then executes the corresponding installed program. Here, the computer-readable medium may be any available computer-readable storage medium or communication medium that can be accessed by the computer.

[0129] Although the embodiments of the present invention have been described with reference to the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the present invention. Such modifications and variations are all within the scope defined by the appended claims.

Claims

1. A method for predicting photovoltaic power generation considering power curtailment, characterized in that: The method comprises: Obtaining first power data of a sample day and first irradiance data of the sample day, performing cluster analysis on the first power data and the first irradiance data, and obtaining a power-limiting time set for the sample day; the power-limiting time set is a time set when the first irradiance data is greater than a preset irradiance threshold and the first power data is less than a preset power threshold; Selecting a preset number of target similar days from the plurality of historical days according to the second irradiance data of the plurality of historical days and the first irradiance data of the sample day; Using the second power data of the target preset number of target similar days, the power data corresponding to the power-limiting time set in the first power data of the sample day is updated to obtain target power data; The power data of the day to be predicted is predicted based on the target power data and the meteorological data of the day to be predicted to obtain a target photovoltaic power prediction value.

2. The method according to claim 1, characterized in that Performing cluster analysis on the first power data and the first irradiance data to obtain a power restriction time set for the sample day includes: Taking the first power data at each moment of the sample day as a first sample point, taking the first irradiance data at each moment of the sample day as a second sample point, determining a first truncation distance for a plurality of the first sample points, and determining a second truncation distance for a plurality of the second sample points; Determine a first local density of each of the first sample points according to the first cutoff distance, determine a first relative distance of each of the first sample points according to the first local density, determine a second local density of each of the second sample points according to the second cutoff distance, and determine a second relative distance of each of the second sample points according to the second local density; Determining a first cluster center based on the first local density and the first relative distance, clustering the plurality of first sample points based on the first cluster center to obtain a plurality of first cluster groups; determining a second cluster center based on the second local density and the second relative distance, clustering the plurality of second sample points based on the second cluster center to obtain a plurality of second cluster groups; Selecting a first representative point in each of the first cluster groups, using each of the first representative points as a first initial cluster center, and obtaining a plurality of first target cluster groups by iteratively updating the first initial cluster centers and the first membership matrix; selecting a second representative point in each of the second cluster groups, using each of the second representative points as a second initial cluster center, and obtaining a plurality of second target cluster groups by iteratively updating the second initial cluster centers and the second membership matrix; Selecting, from among the plurality of first target cluster groups, a plurality of third target cluster groups whose first irradiance data are all greater than a preset irradiance threshold, and selecting, from among the plurality of second target cluster groups, a plurality of fourth target cluster groups whose first power data are all less than a preset power threshold; A set of multiple time instants that are in the third target cluster grouping and in the fourth target cluster grouping is used as the power-limiting time instant set.

3. The method according to claim 1 or 2, characterized in that The step of selecting a preset number of target similar days from the plurality of historical days based on the second irradiance data of the plurality of historical days and the first irradiance data of the sample day comprises: Acquire second irradiance data of a plurality of historical days, and select a preset number of similar days from the plurality of historical days based on the Euclidean distance between the second irradiance data of each historical day and the first irradiance data of the sample day; According to the correlation coefficient between the third irradiance data of each of the similar days and the first irradiance data of the sample day, a target preset number of target similar days is screened from the preset number of similar days; the target preset number is less than the preset number.

4. The method according to claim 1 or 2, characterized in that The step of selecting a plurality of target similar days from the plurality of historical days based on the second irradiance data of the plurality of historical days and the first irradiance data of the sample day further includes: Aligning the first time series data corresponding to the second irradiance data of each historical day with the second time series data corresponding to the first irradiance data of the sample day; Determine a dynamic programming distance between the first time series data of each of the historical days after alignment and the aligned second time series data; A plurality of target similar days are selected from the plurality of historical days according to the plurality of dynamic programming distances.

5. The method according to claim 1 or 2, characterized in that The step of updating the power data corresponding to the power-limiting time set in the first power data of the sample day by using the second power data of the target preset number of target similar days to obtain target power data includes: Extracting the second power data of each power-cutting moment in the power-cutting moment set on each target similar day, and performing weighted averaging on the second power data corresponding to each power-cutting moment on each target similar day to obtain the third power data of each power-cutting moment; The power data corresponding to each power-limiting moment in the first power data of the sample day is replaced with the third power data at the corresponding moment to obtain the target power data.

6. The method according to claim 1 or 2, characterized in that The step of predicting the power data of the day to be predicted based on the target power data and the meteorological data of the day to be predicted to obtain a target photovoltaic power prediction value includes: The target power data and the meteorological data of the day to be predicted are input into the trained photovoltaic power generation power prediction model to obtain the target photovoltaic power generation power prediction value; the photovoltaic power generation power prediction model is a neural network model constructed based on a converter and a multi-layer perceptron; the input of the photovoltaic power generation power prediction model is the target power data and the meteorological data, and the output of the photovoltaic power generation power prediction model is the target photovoltaic power generation power prediction value.

7. A device for predicting photovoltaic power generation considering power curtailment, characterized in that: The device comprises: a cluster analysis module, configured to obtain first power data of a sample day and first irradiance data of the sample day, perform cluster analysis on the first power data and the first irradiance data, and obtain a set of power-limiting moments for the sample day; the set of power-limiting moments being a set of moments when the first irradiance data is greater than a preset irradiance threshold and the first power data is less than a preset power threshold; a similar day selection module, configured to select a preset number of target similar days from a plurality of historical days based on the second irradiance data of the plurality of historical days and the first irradiance data of the sample day; A data updating module, configured to update the power data corresponding to the power-limiting time set in the first power data of the sample day by using the second power data of the target preset number of target similar days to obtain target power data; The power prediction module is used to predict the power data of the day to be predicted based on the target power data and the meteorological data of the day to be predicted, so as to obtain a target photovoltaic power prediction value.

8. A computer device, characterized in that: include: A memory and a processor, wherein the memory and the processor are communicatively connected to each other, the memory stores computer instructions, and the processor executes the method for predicting photovoltaic power generation power considering power curtailment according to any one of claims 1 to 6 by executing the computer instructions.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a computer to execute the method for predicting photovoltaic power generation considering power curtailment according to any one of claims 1 to 6.

10. A computer program product, characterized in that The method comprises computer instructions for causing a computer to execute the method for predicting photovoltaic power generation considering power curtailment according to any one of claims 1 to 6.

Images