Photovoltaic power generation power prediction method and device, equipment and storage medium
By performing correlation analysis and clustering of the historical power generation power and meteorological parameters of photovoltaic power stations, the photovoltaic power generation prediction model is optimized, and the problem of insufficient prediction accuracy in the existing technology is solved, and higher prediction accuracy is achieved.
Patent Information
- Application Number
- CN202411089599.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-09
- Publication Date
- 2024-12-06
AI Technical Summary
The existing photovoltaic power prediction methods have insufficient prediction accuracy in the short-term and ultra-short-term ranges, especially as the prediction parameters increase, the advantages of the backpropagation neural network (BP) model decrease, making it difficult to effectively improve the prediction capability.
By obtaining the historical power generation power and meteorological parameter information of the photovoltaic power station cluster, performing correlation analysis and clustering grouping, and after obtaining the target data, the model parameters of the photovoltaic power generation prediction model are tuned based on the training sample set to form the optimal power generation prediction model, and then make predictions.
It effectively reduces the prediction result error caused by the volatility of photovoltaic power generation and improves the prediction accuracy of photovoltaic power.
Smart Images

Figure CN119089227A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of power prediction of renewable energy power generation in power systems, and in particular to a photovoltaic power generation power prediction method, device, equipment and storage medium. Background Art
[0002] The output power of photovoltaic power generation is related to many factors such as sunshine, season and weather. It is intermittent, random and volatile. Large-scale photovoltaic power generation access will have a certain impact on the safe and stable operation of the power grid. If the power dispatching system can adjust the power of wind power generation, thermal power generation and hydropower generation according to the power of future photovoltaic power generation, it can better keep the overall power of the power grid stable. Therefore, it is necessary to predict the photovoltaic power generation power.
[0003] Photovoltaic power generation prediction is to predict the changes in photovoltaic power within a period of time in the future. For example, ultra-short-term prediction generally refers to the prediction of power changes within the next few hours, which is widely used in real-time monitoring of power grid operation; short-term prediction generally refers to the prediction of photovoltaic output changes within the next 7-10 days, which is widely used in power system electricity price calculation and power balance. At present, the power prediction of photovoltaic power generation is mainly aimed at short-term and ultra-short-term ranges, and the back propagation neural network (BP) is used to predict the power of photovoltaic power generation. The accuracy of the prediction model of this method needs to be improved. At the same time, as the number of prediction parameters increases, the advantage of the BP neural network will decrease, making it difficult to significantly improve the prediction ability of the model. Summary of the invention
[0004] In view of this, the present invention provides a photovoltaic power prediction method, device, equipment and storage medium, which can effectively reduce the large error in the prediction result caused by the volatility of photovoltaic power, and effectively improve the prediction accuracy of photovoltaic power.
[0005] According to one aspect of the present invention, an embodiment of the present invention provides a photovoltaic power generation power prediction method, the method comprising:
[0006] Acquire the power generation and meteorological parameter information of the photovoltaic power stations included in the photovoltaic power station cluster, and perform correlation analysis on the historical power generation and meteorological parameter information corresponding to each of the photovoltaic power stations to obtain a correlation analysis result;
[0007] Determine, according to the correlation analysis result, the meteorological parameter information at the moment when the correlation with the historical power generation reaches a preset correlation threshold, and cluster the photovoltaic power stations based on the correlation analysis result to obtain a cluster grouping result;
[0008] For each of the grouping results, data regularization is performed on the historical power generation of the photovoltaic power station contained in each of the groupings and the meteorological parameter information at the same time when the correlation with the historical power generation reaches a preset correlation threshold, so as to obtain target data;
[0009] The target data is used as a training sample set, and based on the training sample set, a model parameter tuning is performed on a photovoltaic power generation prediction model to obtain an optimal power generation prediction model;
[0010] Acquire future meteorological factor data, and input the future meteorological factor data into the optimal power generation prediction model to obtain photovoltaic power generation prediction data.
[0011] According to another aspect of the present invention, an embodiment of the present invention further provides a photovoltaic power generation prediction device, the device comprising:
[0012] A correlation determination module is used to obtain the power generation and meteorological parameter information of the photovoltaic power stations included in the photovoltaic power station cluster, and to perform correlation analysis on the historical power generation and meteorological parameter information corresponding to each of the photovoltaic power stations to obtain a correlation analysis result;
[0013] A grouping module, used to determine the meteorological parameter information at the moment when the correlation with the historical power generation reaches a preset correlation threshold according to the correlation analysis result, and cluster the photovoltaic power stations based on the correlation analysis result to obtain a cluster grouping result;
[0014] A data regularization module is used to regularize the historical power generation of the photovoltaic power station contained in each of the grouping results and the meteorological parameter information at the same time when the correlation with the historical power generation reaches a preset correlation threshold, so as to obtain target data;
[0015] An optimization module, used to use the target data as a training sample set, and to tune the model parameters of the photovoltaic power generation prediction model based on the training sample set to obtain an optimal power generation prediction model;
[0016] The prediction module is used to obtain future meteorological factor data, and input the future meteorological factor data into the optimal power generation prediction model to obtain photovoltaic power generation prediction data.
[0017] According to another aspect of the present invention, an embodiment of the present invention further provides an electronic device, the electronic device comprising:
[0018] at least one processor; and
[0019] a memory communicatively connected to the at least one processor; wherein,
[0020] The memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the photovoltaic power generation power prediction method described in any embodiment of the present invention.
[0021] According to another aspect of the present invention, an embodiment of the present invention further provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the photovoltaic power generation power prediction method described in any embodiment of the present invention when executed.
[0022] According to another aspect of the present invention, an embodiment of the present invention further provides a computer program product, characterized in that the computer program product includes a computer program, and when the computer program is executed by a processor, the photovoltaic power generation power prediction method described in any embodiment of the present invention is implemented.
[0023] The technical effect of the present invention lies in that, by performing correlation analysis on the historical power generation and meteorological parameter information corresponding to each photovoltaic power station, and clustering each photovoltaic power station based on the correlation analysis result to obtain a grouping result, for each grouping in the grouping result, the historical power generation and meteorological parameter information of the photovoltaic power station contained in each grouping are respectively regularized to obtain a training sample set, and based on the training sample set, the model parameters of the photovoltaic power generation prediction model are tuned to obtain an optimal power generation prediction model, so as to obtain photovoltaic power generation prediction data through the optimal power generation prediction model, thereby effectively reducing the large error in the prediction result caused by the volatility of photovoltaic power generation, and effectively improving the prediction accuracy of photovoltaic power.
[0024] It should be understood that the contents described in this section are not intended to identify the key or important features of the embodiments of the present invention, nor are they intended to limit the scope of the present invention. Other features of the present invention will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS
[0025] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0026] Figure 1 A flowchart of a photovoltaic power generation power prediction method provided by an embodiment of the present invention;
[0027] Figure 2A flowchart of another photovoltaic power generation prediction method provided by an embodiment of the present invention;
[0028] Figure 3 A structural block diagram of a photovoltaic power generation prediction device provided by an embodiment of the present invention;
[0029] Figure 4 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION
[0030] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only 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 ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.
[0031] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.
[0032] In one embodiment, Figure 1 A flowchart of a photovoltaic power prediction method provided for an embodiment of the present invention. This embodiment can be applied to situations when predicting the power generation of photovoltaic power stations in a photovoltaic power station cluster. The method can be executed by a photovoltaic power prediction device. The photovoltaic power prediction device can be implemented in the form of hardware and / or software. The photovoltaic power prediction device can be configured in an electronic device.
[0033] like Figure 1 As shown, the photovoltaic power generation power prediction method in this embodiment includes the following specific steps:
[0034] S110, obtaining historical power generation and meteorological parameter information of the photovoltaic power stations included in the photovoltaic power station cluster, and performing correlation analysis on the historical power generation and meteorological parameter information corresponding to each photovoltaic power station to obtain a correlation analysis result.
[0035] Among them, meteorological parameter information may include but is not limited to radiation illumination, temperature, humidity, air quality index, etc.
[0036] In this embodiment, the photovoltaic power station cluster includes multiple photovoltaic power stations, each photovoltaic power station corresponds to the corresponding power generation output power and meteorological parameters, so that the historical power generation power and meteorological parameter information of the photovoltaic power stations included in the photovoltaic power station cluster can be obtained. Therefore, the power generation power and meteorological parameter information corresponding to each photovoltaic power station are respectively analyzed for correlation to obtain the correlation analysis result between the power generation power and the meteorological parameter information. In this embodiment, the correlation analysis method can be specifically calculated by calculating the Pearson correlation coefficient to obtain the correlation analysis result, that is, the correlation strength between the power generation output power and meteorological parameters such as temperature, irradiance, and air pressure. It can be understood that the correlation between the power generation output power and meteorological parameters such as temperature, irradiance, and air pressure may be relatively strong or relatively weak.
[0037] S120, determining the meteorological parameter information at the moment when the correlation with the historical power generation reaches a preset correlation threshold according to the correlation analysis result, and clustering the photovoltaic power stations based on the correlation analysis result to obtain a cluster grouping result.
[0038] In this embodiment, the meteorological parameter information at the same moment can be referred to as meteorological factor data at the same moment, which can be understood as meteorological parameter information that has a greater correlation with historical power generation.
[0039] In this embodiment, since the output of the photovoltaic power station is greatly affected by the geographical terrain and climatic conditions, different weather conditions such as wind speed, irradiance, cloud cover, etc. will cause the output of the photovoltaic power station in the sub-area to vary greatly. For example, in areas such as rivers and lakes, beaches, wastelands, saline-alkali land, and sandy land, the temperature, humidity, and air quality index will be different, which will affect the power generation in these areas. Therefore, when performing power prediction, it is necessary to perform classification prediction based on the location of the photovoltaic power generation unit, which can effectively improve the accuracy of the prediction. Therefore, before using the prediction model for prediction calculation, the present invention designs a step of calculating the correlation between the output power and meteorological parameters of each photovoltaic power station in the photovoltaic power station cluster, clustering the photovoltaic power stations, and obtaining the grouping steps of the photovoltaic power stations.
[0040] In this embodiment, the meteorological parameter information at the same time when the correlation with the generated power reaches a preset correlation threshold is determined according to the correlation analysis result. For example, the correlation between the generated power and the meteorological parameters is calculated first, such as the correlation with the temperature, irradiance, and air pressure. After calculating the correlation, the meteorological parameters with higher correlation are selected for subsequent calculations. For example, if it is found that the correlation with the temperature and irradiance is higher, the temperature and irradiance data at the same time are obtained without adding the air pressure data for calculation.
[0041] In this embodiment, based on the above-mentioned correlation analysis results, each photovoltaic power station can be clustered and grouped to obtain a group grouping result. Specifically, the power generation power of each photovoltaic power station is standardized to form a data matrix, and the distance between each photovoltaic power station in the data matrix is calculated using a preset clustering algorithm to form a clustering distance matrix. The similarity between each photovoltaic power station is determined based on the clustering distance matrix and the correlation analysis results, and the photovoltaic power stations are classified based on the similarity and a preset similarity threshold to obtain a group grouping result.
[0042] S130, for each grouping in the grouping result, respectively regularize the historical power generation of the photovoltaic power station contained in each grouping and the meteorological parameter information at the same time when the correlation with the historical power generation reaches a preset correlation threshold, so as to obtain target data.
[0043] Among them, data regularization can be understood as the process of processing data on the historical power generation of the photovoltaic power station and the meteorological parameter information at the moment when the correlation with the historical power generation reaches a preset correlation threshold.
[0044] In this embodiment, for each grouping in the grouping result, the historical power generation of the photovoltaic power station contained in each grouping, and the meteorological parameter information at the same time when the correlation with the historical power generation reaches the preset correlation threshold, are obtained, and the power generation of the photovoltaic power station contained in each grouping and the meteorological parameter information at the same time when the correlation with the power generation reaches the preset correlation threshold are respectively regularized to obtain the target data. Specifically, all the data of a feature are statistically calculated to determine the first limit value, the second limit value and the limit interval, and the upper limit and the lower limit of the abnormal value are obtained according to the first limit value, the second limit value and the limit interval, and then it is determined whether all the data of a feature are within the range of the upper limit and the lower limit of the abnormal value. If they are within the range, they are determined as normal values. If the values are above the upper limit of the abnormal value and below the lower limit of the abnormal value, they are all abnormal values; the detected abnormal data are taken as missing values, and the filling values are calculated by interpolation method to fill them to obtain the processed target data.
[0045] S140: taking the target data as a training sample set, and optimizing the model parameters of the photovoltaic power generation prediction model based on the training sample set to obtain an optimal power generation prediction model.
[0046] In this embodiment, the historical power generation of the photovoltaic power station with regular data and the meteorological parameter information at the same time when the correlation with the historical power generation reaches the preset correlation threshold are used as a training data set, and the training data set is input into the photovoltaic power generation prediction model for calculation, and the photovoltaic power generation prediction model is evaluated for performance, and the photovoltaic power generation prediction model with the best parameters is determined. Specifically, the training sample set is input into the preset first power prediction model to obtain the initial power prediction value, the actual historical power generation is obtained, and the difference between the actual historical power generation and the initial power prediction value is determined, and the difference is used as correction data to form a correction data set, and the correction data set is input into the preset photovoltaic power error correction model to obtain the correction prediction value, and the initial prediction value is superimposed with the correction prediction value to obtain the photovoltaic power generation value after correction, and the photovoltaic power generation value after correction is used to train the photovoltaic power generation prediction model to determine the optimal photovoltaic power generation prediction model. Of course, in this embodiment, the photovoltaic power generation prediction model can be trained by other existing methods to determine the optimal photovoltaic power generation prediction model, and this embodiment is not limited here.
[0047] S150, obtaining future meteorological factor data, and inputting the future meteorological factor data into an optimal power generation prediction model to obtain photovoltaic power generation prediction data.
[0048] Among them, the future meteorological factor data can be understood as the meteorological factor data related to the photovoltaic power generation data to be predicted.
[0049] In this embodiment, by obtaining meteorological factor data related to the photovoltaic power generation data to be predicted, and then inputting the meteorological factor data related to the photovoltaic power generation data to be predicted into the optimal photovoltaic power generation prediction model, the photovoltaic power generation prediction data is obtained. It should be noted that if the meteorological factor data related to the photovoltaic power generation data to be predicted is complete and has no abnormalities, it can be directly input. If the meteorological factor data related to the photovoltaic power generation data to be predicted is true or abnormal, the meteorological factor data related to the photovoltaic power generation data to be predicted can be respectively data regularized, and then input into the optimal photovoltaic power generation prediction model after regularization to obtain photovoltaic power generation prediction data. In this embodiment, the method of data regularization is the same as the method of data regularization of the meteorological parameter information at the same time when the historical power generation of the photovoltaic power station contained in each group group and the correlation with the historical power generation reaches the preset correlation threshold, which is not described in this embodiment.
[0050] The technical solution of the embodiment of the present invention performs correlation analysis on the historical power generation and meteorological parameter information corresponding to each photovoltaic power station, and clusters each photovoltaic power station based on the correlation analysis result to obtain a grouping result. For each grouping in the grouping result, the historical power generation and meteorological parameter information of the photovoltaic power stations contained in each grouping are respectively regularized to obtain a training sample set. Based on the training sample set, the model parameters of the photovoltaic power generation prediction model are optimized to obtain an optimal power generation prediction model, so as to obtain photovoltaic power generation prediction data through the optimal power generation prediction model, thereby effectively reducing the large error in the prediction result caused by the volatility of photovoltaic power generation, and effectively improving the prediction accuracy of photovoltaic power.
[0051] In one embodiment, Figure 2 A flow chart of another photovoltaic power prediction method provided for an embodiment of the present invention. Based on the above embodiments, this embodiment clusters and groups each photovoltaic power station based on the correlation analysis result to obtain a group grouping result, and performs data regularization on the historical power generation of the photovoltaic power stations included in each grouping and the meteorological parameter information at the same time when the correlation with the historical power generation reaches a preset correlation threshold to obtain target data, and further refines the photovoltaic power prediction model by tuning the model parameters based on the training sample set to obtain the optimal power generation prediction model.
[0052] like Figure 2 As shown, the photovoltaic power generation power prediction method in this embodiment may specifically include the following steps:
[0053] S210, obtaining historical power generation and meteorological parameter information of the photovoltaic power stations included in the photovoltaic power station cluster, and performing correlation analysis on the historical power generation and meteorological parameter information corresponding to each photovoltaic power station to obtain a correlation analysis result.
[0054] S220. Determine, based on the correlation analysis result, the meteorological parameter information at the moment when the correlation with the historical power generation reaches a preset correlation threshold.
[0055] S230, standardize the historical power generation of each photovoltaic power station to form a data matrix.
[0056] The data matrix includes photovoltaic power stations and the corresponding historical power generation.
[0057] In this embodiment, the power generation of each photovoltaic power station is standardized to form a data matrix; wherein the data matrix includes the photovoltaic power station and the corresponding power generation. It can be understood that the historical power generation of each photovoltaic power station is standardized to form a data matrix, wherein the power data of the photovoltaic power station constitutes a row or a column of data, and the data of multiple power stations constitute a data matrix. Then, the Euclidean distance is used to calculate the inter-class distance between two power stations in the data matrix.
[0058] S240, using a preset clustering algorithm to calculate the distance between each photovoltaic power station in the data matrix to form a clustering distance matrix.
[0059] In this embodiment, a preset clustering algorithm is used to calculate the distance between each photovoltaic power station in the data matrix to form a cluster distance matrix. That is, each photovoltaic power station is regarded as a class, and the distance between each two power stations is calculated using a clustering algorithm to construct a cluster distance matrix. Among them, the preset clustering algorithm can be understood as dividing a data set into different classes or clusters according to a specific standard (such as distance), so that the similarity of data objects in the same cluster is as large as possible, and the difference of data objects not in the same cluster is as large as possible. That is, after clustering, data of the same class are clustered together as much as possible, and data of different classes are separated as much as possible. In this embodiment, the preset clustering algorithm can be improved K-Means, K-Medoids, Gaussian mixture model, spectral clustering and the like.
[0060] S250: Determine the similarity between the photovoltaic power stations based on the cluster distance matrix and the correlation analysis result.
[0061] In this embodiment, the similarity between the two power stations is calculated based on the distance between the two power stations in the cluster distance matrix and the correlation between the output power and meteorological parameters of the two power stations.
[0062] S260, classifying the photovoltaic power stations whose similarity reaches a preset similarity threshold to obtain a cluster grouping result, updating the number of categories, and determining the average distance between each photovoltaic power station in the cluster grouping result according to a preset weighted average method, and updating the cluster distance matrix according to the average distance.
[0063] In this embodiment, the photovoltaic power stations whose similarity reaches the preset similarity threshold are classified to obtain the group grouping result. It can be understood that the two categories with the highest similarity are merged as a new category, the number of categories is updated, the average distance between the new categories is analyzed by weighted average method, and the cluster distance matrix is updated, that is, the power stations with the same or highly similar power generation trends with respect to meteorological parameters are finally classified as one category. For example, if the power generation trends of two power stations with respect to meteorological parameters are the same or highly similar, then the two power stations are classified as one category, which makes it easier to use fewer power stations for power generation prediction and reduce the amount of data in the algorithm.
[0064] In this embodiment, the power data of each photovoltaic power station is standardized, and the overall data matrix is constructed. Each photovoltaic power station is regarded as a category, and the distance between each two power stations is calculated using a clustering algorithm to construct a clustering distance matrix. At the same time, the output power and meteorological parameters of each photovoltaic power station are correlated. The similarity of the two power stations is calculated based on the distance between the two power stations and the correlation between the output power and meteorological parameters of the two power stations. The two categories with the highest similarity are merged as a new category, the number of categories is updated, and the average distance between the new categories is analyzed using a weighted average method to update the clustering distance matrix; the previous step is continued until all categories are clustered into one category, and finally, the number of categories is determined to obtain the final clustering result of the photovoltaic power station. After the photovoltaic power stations are clustered and grouped using the above method, the data of the corresponding groups are processed separately when performing subsequent predictions, and finally the data of each group are superimposed to obtain the total data of the entire photovoltaic power station.
[0065] It should be noted that, during the actual operation of a photovoltaic power station, the recorded data usually contains some missing values and / or outliers. Therefore, it is necessary to pre-process the data of these missing values and / or outliers to improve the data quality during subsequent processing, so as to avoid large deviations in subsequent prediction results. Among them, outliers are usually data recorded for system abnormalities caused by hardware abnormalities in the system (such as poor sensor contact or sensor abnormalities, etc.). If the abnormal data in the data is not processed, the abnormal data will participate in the subsequent model training, which will cause the model to train the outliers, and then cause certain errors in the model output results.
[0066] Therefore, first check the abnormal data in the data, take the detected abnormal data as missing values, and use the interpolation method to calculate the filling value to fill in, so as to ensure the integrity of the data. As for the real missing values, they can be obtained by statistical counting, and the interpolation method can be used to repair these missing data. After the above processing, the data corresponding to the points of the missing values in the above are obtained by interpolation, and the data corresponding to the points of the abnormal values are first detected and deleted and then interpolated. Therefore, the data at these points are repaired data, which may produce certain deviations. Therefore, it is necessary to evaluate the interpolated data later to determine the deviation of the interpolated data. Therefore, the present invention also needs to mark these repaired data points so that these repaired data can be evaluated after the prediction model is established. Specifically, the specific steps for data regularization include: S270-S2100.
[0067] S270: Use the historical power generation and the meteorological parameter information at the same moment as feature data respectively, and count all recorded values of each feature data in the time series as target feature data of each feature data.
[0068] Among them, characteristic data can be understood as power generation, as well as temperature, irradiance and other characteristics in the meteorological parameter information at the same time, and target characteristic data refers to the values of all records of a certain characteristic data in a time series. The data can be obtained in the following ways: power generation can be obtained from power stations; meteorological factor data such as temperature and irradiance can be obtained from weather numerical forecasts. After obtaining, the values of all records of a certain characteristic data in a time series can be counted.
[0069] In this embodiment, for each grouping in the grouping results, the generated power of the photovoltaic power station contained in each grouping and the meteorological parameter information at the same time when the correlation with the generated power reaches a preset correlation threshold are obtained, and the generated power and the meteorological parameter information at the same time are respectively used as feature data, and all recorded values of each feature data in the time series are counted as target feature data.
[0070] S280, determine a first limit value, a second limit value and a limit spacing from the target feature data; wherein the first limit value is a value in the target feature data that reaches a first proportion, the second limit value is a value in all data that reaches a second proportion, and the limit spacing is the difference between the first limit value and the second limit value; the first proportion is less than the second proportion.
[0071] The first proportion and the second proportion are both preset proportions, and the first proportion is smaller than the second proportion. For example, the first proportion is the value at the 25% position of all data; the second proportion is the value at the 75% position of all data.
[0072] In this embodiment, all data of a feature are statistically calculated to obtain a first limit value A. 上 , the second limit value A 下 And the limit distance A 间距 ; Among them, a feature refers to a feature such as power generation, temperature, irradiance, etc., and all data refers to the values of all records of the feature in the time series. The first limit value A 上 The second limit value A is the value at the 25% position of all data. 下 It is the value at 75% of all data, and the limit interval is A 间距 is the difference between the first limit value and the second limit value. 上 , the second limit value A 下 And the limit distance A 间距 ; Among them, the first limit value A 上 The second limit value A is the value at the 25% position of all data. 下 It is the value at 75% of all data, and the limit interval is A 间距 is the difference between the first limit value and the second limit value.
[0073] S290, obtaining an upper limit of the abnormal value and a lower limit of the abnormal value according to the first limit value, the second limit value and the limit spacing; wherein the upper limit of the abnormal value is the sum of the first limit value and the limit spacing of a preset multiple; the lower limit of the abnormal value is the difference between the second limit value and the limit spacing of a preset multiple.
[0074] In this embodiment, the upper limit of the abnormal value and the lower limit of the abnormal value are obtained according to the first limit value, the second limit value and the limit interval; wherein the upper limit of the abnormal value is the sum of the first limit value and the limit interval of the preset multiple; the lower limit of the abnormal value is the difference between the second limit value and the limit interval of the preset multiple. Exemplarily, the upper limit of the abnormal value and the lower limit of the abnormal value are calculated; wherein the upper limit of the abnormal value is A 上 +1.5A 间距 , the lower limit of the outlier is A 下 -1.5A 间距 .
[0075] S2100. When the target feature data is within the range of the upper limit and the lower limit of the abnormal value, the target feature data is determined to be a normal value. When the target feature data is above the upper limit of the abnormal value and below the lower limit of the abnormal value, it is determined to be an abnormal value. The detected abnormal data is regarded as a missing value, and the filling value is calculated using the interpolation method to obtain the processed target data.
[0076] In this embodiment, when the target feature data is within the range of the upper and lower limits of the abnormal value, the target feature data is determined to be a normal value, and when the target feature data is above the upper limit of the abnormal value and below the lower limit of the abnormal value, the values are all abnormal values. The detected abnormal data is regarded as a missing value, and the interpolation method is used to calculate the filling value to fill it in order to obtain the processed target data. It can be understood that all the data of a feature are substituted into the calculation range formed by the upper and lower limits of the abnormal value, and the values above the upper limit of the abnormal value and below the lower limit of the abnormal value are all abnormal values; the detected abnormal data is regarded as a missing value, and the interpolation method is used to calculate the filling value to fill it in.
[0077] Through the above data processing, the various data output by the photovoltaic power generation system usually have units, such as irradiance in W / m2, temperature in °C, humidity in %, and air quality index as a dimensionless quantity. The data ranges of various data are also very different. If they are directly input into the subsequent prediction model without processing, the large-dimensional data will usually be greater than the small-dimensional data. At this time, the large-dimensional data will be more prominent in the prediction model, which will make it almost impossible for the prediction model to perform machine learning calculations on the small-dimensional data. Therefore, it is necessary to normalize the data to improve the training efficiency of the prediction model and avoid data distribution disorder caused by the dimension.
[0078] In this embodiment, when the prediction model is subsequently trained, the data in the training data set is regularized using the above method, that is, the data of some time periods are missing or interpolated due to anomalies, and are not the original real data. Therefore, when the photovoltaic power generation data and meteorological factor data are respectively regularized, the photovoltaic power generation data and meteorological factor data that have undergone data regularization are also marked.
[0079] When the photovoltaic power generation data of the photovoltaic power stations grouped in each category and the meteorological factor data at the same time are used as training data sets, and the training data sets are input into the photovoltaic power generation prediction model for calculation, the data at the marked time points are screened out, and the predicted photovoltaic power generation data are used to replace the data initially supplemented by the interpolation method. Finally, the photovoltaic power generation data of the photovoltaic power station after replacement and the meteorological factor data at the same time are used as training data sets, and the training data sets are input into the photovoltaic power generation prediction model for secondary calculation. The performance of the trained photovoltaic power generation prediction model is evaluated using the secondary calculation structure at this time.
[0080] S2110 , input the training sample set into a preset first power prediction model to obtain an initial power prediction value.
[0081] In this embodiment, since photovoltaic power generation is greatly affected by external factors, its time series will show a certain volatility. Directly taking the original power data for prediction based on the support vector machine (SVM) regression model will produce a large error. Therefore, before using the SVM regression model for prediction, it is necessary to process the photovoltaic power generation signal to solve the problem of large prediction error of the SVM single model. Based on the above considerations, the present invention uses a decomposition algorithm to first decompose the fluctuating photovoltaic power generation original power signal into multiple periodic components to solve the problem of local characteristic oscillation.
[0082] In this embodiment, the training sample set is input into a preset first power prediction model to obtain an initial power prediction value. Specifically, a preset decomposition algorithm is used to decompose the power generation power of the photovoltaic power station in the training sample set to obtain at least two intrinsic mode components (IMF) and corresponding at least two residual components (RES). Each IMF and the corresponding RES are respectively formed into a training set to establish an SVM regression model corresponding to each IMF and the corresponding RES. The preset sparrow search algorithm is used to optimize the penalty parameter c and the kernel function parameter g of each SVM regression model to obtain each optimized optimal SVM regression model. Each optimal SVM regression model is used as a preset first power prediction model to perform power prediction on a preset test set to obtain a corresponding photovoltaic power generation value, and each photovoltaic power generation value is superimposed as an initial prediction value.
[0083] Through the above scheme, the present invention decomposes the historical photovoltaic power data into intrinsic mode components and residual components of different frequencies, inputs each IMF component into the prediction model respectively, and finally reconstructs the prediction results of each frequency component and residual component to obtain the final prediction result, which can effectively reduce the large error in the prediction result caused by the volatility of photovoltaic power generation.
[0084] In one embodiment, S2110 may include: S21101-S21104:
[0085] S21101. Decompose the power generation of the photovoltaic power station in the training sample set using a preset decomposition algorithm to obtain at least two IMFs and corresponding at least two RESs.
[0086] In this embodiment, a preset decomposition algorithm is used to decompose the power generation of the photovoltaic power station in the training sample set to obtain at least two IMFs and corresponding at least two RESs. Specifically, a pair of opposite positive and negative white noises are added to the original power signal of the photovoltaic power generation to generate two groups of mixed signals; the two groups of mixed signals are modally decomposed to obtain two groups of positive and negative IMF components; the steps of adding a pair of opposite positive and negative white noises to generate two groups of mixed signals and performing modal decomposition on the two groups of mixed signals are repeated N times, each time a new normally distributed Gaussian white noise is added until the number of extreme points of the remaining components does not exceed two, and the average of each set of positive and negative IMF components is calculated to obtain multiple IMFs and RESs.
[0087] S21102. Form each IMF and the corresponding RES into training sets respectively, so as to establish SVM regression models corresponding to each IMF and the corresponding RES respectively.
[0088] Specifically, the upper extreme point and the lower extreme point of the new photovoltaic power generation signal are extracted, and the upper envelope and the lower envelope are generated based on the new photovoltaic power generation signal and the upper extreme point and the lower extreme point, and the mean of the envelope is calculated, and the new photovoltaic power generation signal is subtracted from the mean of the envelope to obtain the detail signal of the local oscillation, and it is judged whether the detail signal meets the termination condition of the decomposition, such as whether the number of extreme points is less than 3, or whether the maximum and minimum values appear alternately without repeated extreme points, if the termination condition is met; if the detail signal does not meet the termination condition, it is regarded as a new photovoltaic power generation signal, and the above steps are repeated to obtain an IMF; return to the step of judging whether the detail signal meets the termination condition of the decomposition until the last IMF that meets the termination condition is obtained.
[0089] In this embodiment, by decomposing the historical photovoltaic power data into intrinsic mode components and residual components of different frequencies, each IMF component is input into the prediction model respectively, and finally the prediction results of each frequency component and residual component are sequenced to obtain the final prediction result, which can effectively reduce the large error in the prediction result caused by the volatility of photovoltaic power generation.
[0090] In one embodiment, modal decomposition is performed on a first mixed power signal and a second mixed power signal respectively to obtain corresponding first IMF components and second IMF components, including: extracting upper extreme points and lower extreme points corresponding to the first mixed power signal and the second mixed power signal respectively; generating corresponding upper envelopes and lower envelopes according to the first mixed power signal, the second mixed power signal and the corresponding upper extreme points and lower extreme points respectively, and determining envelope mean values of the upper envelopes and the lower envelopes; subtracting the first mixed power signal and the second mixed power signal from the corresponding envelope mean values respectively to obtain corresponding target power signals having local oscillation characteristics; judging whether the target power signal satisfies a decomposition termination condition; if so, taking the target power signal as the first IMF component and the second IMF component; if not, taking the target signal as the first mixed power signal and the second mixed power signal, and returning to the step of extracting upper extreme points and lower extreme points corresponding to the first mixed power signal and the second mixed power signal respectively, until the decomposition termination condition is met; wherein the decomposition termination condition at least includes: the number of extreme points is greater than a preset threshold, or whether the maximum value and the minimum value appear alternately without repeated extreme points.
[0091] S21103. Use the preset sparrow search algorithm to optimize the penalty parameter c and kernel function parameter g of each SVM regression model to obtain the optimized optimal SVM regression models, and use each optimal SVM regression model as the preset first power prediction model to perform power prediction on the preset test set to obtain the corresponding photovoltaic power generation power value.
[0092] In this embodiment, the preset sparrow search algorithm is used to optimize the penalty parameter c and kernel function parameter g of each SVM regression model to obtain each optimized optimal SVM regression model, and each optimal SVM regression model is used as a preset first power prediction model to perform power prediction on a preset test set to obtain the corresponding photovoltaic power generation value. Among them, the preset sparrow search algorithm is a new intelligent optimization algorithm in the prior art, which is an intelligent optimization algorithm proposed by using the behavior of sparrows foraging and escaping predators. This embodiment will not be introduced in detail here.
[0093] S21104. Superimpose the photovoltaic power values as an initial prediction value.
[0094] In this embodiment, each photovoltaic power generation value is superimposed as an initial prediction value.
[0095] S2120, obtaining actual historical generated power, and determining the difference between the actual historical generated power and the initial power prediction value, and using the difference as correction data to form a correction data set.
[0096] In this embodiment, the actual historical generated power is obtained, and the difference between the actual historical generated power and the initial power prediction value is determined, and the difference is used as correction data to form a correction data set.
[0097] S2130: Input the correction data set into a preset photovoltaic power error correction model to obtain a correction prediction value.
[0098] In this embodiment, the correction data set is input into the preset photovoltaic power error correction model to obtain the correction prediction value. Among them, the photovoltaic power error correction model adopts the following method: initialize the algorithm parameters, including the number of ant colonies, the number of iterations, the initial position and initial speed of the ants, and the spatial dimension; construct a network topology structure, and form corresponding particles with the parameters to be optimized; use the mean absolute error MAE as the fitness function of the particles, calculate the fitness value according to the initial position of the ants, and thus obtain the initial global optimal and individual optimal values of the ants; use the obtained fitness value to update the ant speed and position, and calculate the fitness value again according to the updated ant position to find the individual optimal position of the ants and the global optimal position; when the number of iterations meets the maximum value, the iterative process ends, and the found hyperparameters are brought into the prediction model to predict the output results; if not satisfied, return to the previous step to continue the iteration.
[0099] S2140, superimposing the initial prediction value and the corrected prediction value to obtain a corrected photovoltaic power generation value, and using the corrected photovoltaic power generation value to perform model training on a photovoltaic power generation prediction model to determine an optimal photovoltaic power generation prediction model.
[0100] In this embodiment, the initial prediction value is superimposed with the corrected prediction value to obtain the corrected photovoltaic power value, and the corrected photovoltaic power value is used to train the photovoltaic power prediction model to determine the optimal photovoltaic power prediction model. In this embodiment, by adding the output of the first power prediction model to the result of the corrected prediction model, the average absolute percentage error and root mean square difference obtained are superior to the prediction results of the traditional LSTM model, so that the above-mentioned photovoltaic power prediction based on error compensation of the present invention can effectively improve the prediction accuracy of photovoltaic power.
[0101] S2150, obtaining future meteorological factor data, and inputting the future meteorological factor data into an optimal power generation prediction model to obtain photovoltaic power generation prediction data.
[0102] The technical solution of the embodiment of the present invention, by adding the output of the first power prediction model with the result of the correction prediction model, the average absolute percentage error and the root mean square error are superior to the prediction results of the traditional SVM model and the LSTM model, so that the above-mentioned photovoltaic power prediction based on error compensation of the present invention can effectively improve the prediction accuracy of photovoltaic power; decompose the historical photovoltaic power data into intrinsic mode components and residual components of different frequencies, input each IMF component into the prediction model respectively, and finally reconstruct the prediction results of each frequency component and the residual component to obtain the final prediction result, which can effectively reduce the large error in the prediction result caused by the volatility of photovoltaic power generation.
[0103] In one embodiment, Figure 3 This is a structural block diagram of a photovoltaic power generation prediction device provided in one embodiment of the present invention. The device is suitable for predicting the power generation of photovoltaic power stations in a photovoltaic power station cluster. The device can be implemented by hardware / software. It can be configured in an electronic device to implement a photovoltaic power generation prediction processing method in an embodiment of the present invention.
[0104] like Figure 3 As shown, the device includes: a correlation determination module 310, a grouping module 320, a data regularization module 330, an optimization module 340 and a prediction module 350;
[0105] The correlation determination module 310 is used to obtain the historical power generation and meteorological parameter information of the photovoltaic power stations included in the photovoltaic power station cluster, and perform correlation analysis on the historical power generation and meteorological parameter information corresponding to each photovoltaic power station to obtain a correlation analysis result;
[0106] A grouping module 320 is used to determine the meteorological parameter information at the same time when the correlation with the historical power generation reaches a preset correlation threshold according to the correlation analysis result, and cluster the photovoltaic power stations based on the correlation analysis result to obtain a cluster grouping result;
[0107] The data regularization module 330 is used to regularize the historical power generation of the photovoltaic power station contained in each of the grouping results and the meteorological parameter information at the same time when the correlation with the historical power generation reaches a preset correlation threshold, so as to obtain target data;
[0108] An optimization module 340 is used to use the target data as a training sample set, and optimize the model parameters of the photovoltaic power generation prediction model based on the training sample set to obtain an optimal power generation prediction model;
[0109] The prediction module 350 is used to obtain future meteorological factor data, and input the future meteorological factor data into the optimal power generation prediction model to obtain photovoltaic power generation prediction data.
[0110] According to an embodiment of the present invention, correlation analysis is performed on the historical power generation and meteorological parameter information corresponding to each photovoltaic power station, and each photovoltaic power station is clustered and grouped based on the correlation analysis result to obtain a group grouping result. For each group group in the group grouping result, data regularization is performed on the historical power generation and meteorological parameter information of the photovoltaic power stations contained in each group group to obtain a training sample set. Based on the training sample set, model parameters of a photovoltaic power prediction model are optimized to obtain an optimal power prediction model, and photovoltaic power prediction data is obtained through the optimal power prediction model, thereby effectively reducing the large error in the prediction result caused by the volatility of photovoltaic power, and effectively improving the prediction accuracy of photovoltaic power.
[0111] In one embodiment, the grouping module 320 includes:
[0112] A data matrix forming unit, used for standardizing the historical power generation of each photovoltaic power station to form a data matrix; wherein the data matrix includes the photovoltaic power station and the corresponding historical power generation;
[0113] A distance matrix forming unit, used to calculate the distance between each photovoltaic power station in the data matrix by using a preset clustering algorithm to form a clustering distance matrix;
[0114] A similarity determination unit, configured to determine the similarity between the photovoltaic power stations according to the cluster distance matrix and the correlation analysis result;
[0115] The grouping result determination unit is used to classify the photovoltaic power stations whose similarity reaches a preset similarity threshold to obtain a group grouping result, update the number of categories, and determine the average distance between each photovoltaic power station in the group grouping result according to a preset weighted average method, and update the cluster distance matrix according to the average distance.
[0116] In one embodiment, the data regularization module 330 includes:
[0117] A first data determination unit is used to use the historical power generation and the meteorological parameter information at the same time as feature data, and to count all recorded values of each feature data in a time series as target feature data of each feature data;
[0118] A second data determination unit is used to determine a first limit value, a second limit value and a limit interval from the target characteristic data; wherein the first limit value is a value that reaches a first proportion in the target characteristic data, the second limit value is a value that reaches a second proportion in all data, and the limit interval is a difference between the first limit value and the second limit value; the first proportion is less than the second proportion;
[0119] An abnormal value upper limit and lower limit determination unit, used to obtain an abnormal value upper limit and an abnormal value lower limit according to the first limit value, the second limit value and the limit interval; wherein the abnormal value upper limit is the sum of the first limit value and the limit interval of a preset multiple; the abnormal value lower limit is the difference between the second limit value and the limit interval of a preset multiple;
[0120] An abnormal value determination unit, configured to determine that the target feature data is a normal value when the target feature data is within the range of the abnormal value upper limit and the abnormal value lower limit, and that the target feature data is an abnormal value when the value is above the abnormal value upper limit and below the abnormal value lower limit;
[0121] The target data determination unit is used to treat the detected abnormal data as missing values, and use interpolation method to calculate the filling value to fill in the missing values to obtain the processed target data.
[0122] In one embodiment, the optimization module 340 includes:
[0123] An initial power prediction unit, used for inputting the training sample set into a preset first power prediction model to obtain an initial power prediction value;
[0124] A correction data determination unit, used for obtaining actual generated power, determining a difference between the actual generated power and the initial power prediction value, and using the difference as correction data to form a correction data set;
[0125] A correction prediction unit, used for inputting the correction data set into a preset photovoltaic power error correction model to obtain a correction prediction value;
[0126] The prediction model determination unit is used to superimpose the initial prediction value and the corrected prediction value to obtain the corrected photovoltaic power generation value, and use the corrected photovoltaic power generation value to perform model training on the photovoltaic power generation prediction model to determine the optimal photovoltaic power generation prediction model.
[0127] In one embodiment, the initial power prediction unit includes:
[0128] A decomposition unit, configured to decompose the power generation of the photovoltaic power station in the training sample set by using a preset decomposition algorithm to obtain at least two intrinsic mode components IMF and corresponding at least two residual components RES;
[0129] A regression model building unit, used for forming training sets for each of the intrinsic mode components IMF and the corresponding residual components RES, so as to build SVM regression models corresponding to each of the intrinsic mode components IMF and the corresponding residual components RES;
[0130] A prediction unit, used for initially optimizing the penalty parameter c and the kernel function parameter g of each SVM regression model by a preset sparrow search algorithm, obtaining each optimized optimal SVM regression model, and using each optimal SVM regression model as a preset first power prediction model to perform power prediction on a preset test set to obtain a corresponding photovoltaic power generation value;
[0131] The initial prediction determination unit is used to superimpose the photovoltaic power generation values as an initial prediction value.
[0132] In one embodiment, the decomposition unit comprises:
[0133] A signal generating subunit, used for acquiring an original power signal of photovoltaic power generation, adding a pair of opposite positive and negative white noises to the original power signal to generate a corresponding first mixed power signal and a second mixed power signal;
[0134] a component determination subunit, configured to perform modal decomposition on the first mixed power signal and the second mixed power signal respectively to obtain corresponding first IMF components and second IMF components, and store the first IMF components in a first component set, and store the second IMF components in a second component set;
[0135] An updating unit, used for returning to the step of adding a pair of opposite positive and negative white noises to the original power signal to generate the corresponding first mixed power signal and the second mixed power signal, adding a new normally distributed Gaussian white noise each time to update the first component set and the second component set until the number of extreme value points of the first IMF component and the second IMF component does not exceed a preset number threshold, thereby obtaining an updated target first component set and a target second component set;
[0136] The solving unit is used to solve the IMF component means of the first IMF component included in the target first component set and the second IMF component included in the target second component set, respectively, to obtain at least two intrinsic mode components IMF and at least two residual components RES.
[0137] In one embodiment, the component determination subunit is specifically configured to:
[0138] Extracting upper extreme value points and lower extreme value points corresponding to the first mixed power signal and the second mixed power signal respectively;
[0139] Generate a corresponding upper envelope and a lower envelope according to the first mixed power signal, the second mixed power signal and the upper extreme point and the lower extreme point respectively corresponding to each other, and determine envelope mean values of the upper envelope and the lower envelope;
[0140] Subtracting the first mixed power signal and the second mixed power signal from the corresponding envelope mean values respectively to obtain corresponding target power signals having local oscillation characteristics;
[0141] Determining whether the target power signal satisfies a decomposition termination condition;
[0142] If satisfied, the target power signal is used as the first IMF component and the second IMF component;
[0143] If not satisfied, take the target signal as the first mixed power signal and the second mixed power signal, and return to the step of extracting the upper extreme points and the lower extreme points corresponding to the first mixed power signal and the second mixed power signal respectively until the decomposition termination condition is met; wherein the decomposition termination condition includes at least: the number of extreme points is greater than a preset threshold, or whether the maximum and minimum values appear alternately without repeated extreme points.
[0144] The photovoltaic power prediction device provided in the embodiment of the present invention can execute the photovoltaic power prediction method provided in any embodiment of the present invention, and has the corresponding functional modules and beneficial effects of the execution method.
[0145] In one embodiment, Figure 4 A schematic diagram of the structure of an electronic device provided for an embodiment of the present invention. The electronic device 10 is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or required herein.
[0146] like Figure 4As shown, the electronic device 10 includes at least one processor 11, and a memory connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program that can be executed by at least one processor, and the processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 to the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other through a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.
[0147] A number of components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.
[0148] The processor 11 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various special artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any appropriate processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as a photovoltaic power generation power prediction method.
[0149] In some embodiments, the photovoltaic power generation power prediction method may be implemented as a computer program, which is tangibly contained in a computer-readable storage medium, such as a storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the photovoltaic power generation power prediction method described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to execute the photovoltaic power generation power prediction method in any other appropriate manner (e.g., by means of firmware).
[0150] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.
[0151] The computer program for implementing the method of the present invention can be written in any combination of one or more programming languages. These computer programs can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable photovoltaic power generation prediction device, so that when the computer program is executed by the processor, the functions / operations specified in the flow chart and / or block diagram are implemented. The computer program can be executed entirely on the machine, partially on the machine, partially on the machine as an independent software package and partially on a remote machine, or completely on a remote machine or server.
[0152] In the context of the present invention, a computer-readable storage medium may be a tangible medium that may contain or store a computer program for use by or in combination with an instruction execution system, device or equipment. A computer-readable storage medium may include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. Alternatively, a computer-readable storage medium may be a machine-readable signal medium. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
[0153] To provide interaction with a user, the systems and techniques described herein may be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the electronic device. Other types of devices may also be used to provide interaction with the user; for example, the feedback provided to the user may be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user may be received in any form (including acoustic input, voice input, or tactile input).
[0154] The systems and techniques described herein may be implemented in a computing system that includes backend components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes frontend components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such backend components, middleware components, or frontend components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), a blockchain network, and the Internet.
[0155] A computing system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The client and server relationship is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and VPS services.
[0156] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the present invention can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solution of the present invention can be achieved, and this document does not limit this.
[0157] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims
1. A photovoltaic power generation power prediction method, characterized in that: The method comprises: Acquire historical power generation and meteorological parameter information of photovoltaic power stations included in the photovoltaic power station cluster, and perform correlation analysis on the historical power generation and meteorological parameter information corresponding to each photovoltaic power station to obtain a correlation analysis result; Determine, according to the correlation analysis result, the meteorological parameter information at the moment when the correlation with the historical power generation reaches a preset correlation threshold, and cluster the photovoltaic power stations based on the correlation analysis result to obtain a cluster grouping result; For each of the grouping results, data regularization is performed on the historical power generation of the photovoltaic power station contained in each of the groupings and the meteorological parameter information at the same time when the correlation with the historical power generation reaches a preset correlation threshold, so as to obtain target data; The target data is used as a training sample set, and based on the training sample set, a model parameter tuning is performed on a photovoltaic power generation prediction model to obtain an optimal power generation prediction model; Acquire future meteorological factor data, and input the future meteorological factor data into the optimal power generation prediction model to obtain photovoltaic power generation prediction data.
2. The method according to claim 1, characterized in that The clustering and grouping of the photovoltaic power stations based on the correlation analysis results to obtain a cluster grouping result includes: Standardizing the historical power generation of each photovoltaic power station to form a data matrix; wherein the data matrix includes the photovoltaic power station and the corresponding historical power generation; Calculating the distance between each photovoltaic power station in the data matrix using a preset clustering algorithm to form a clustering distance matrix; Determining the similarity between the photovoltaic power stations according to the cluster distance matrix and the correlation analysis result; The photovoltaic power stations whose similarity reaches a preset similarity threshold are classified to obtain a grouping result, the number of categories is updated, and the average distance between each photovoltaic power station in the grouping result is determined according to a preset weighted average method, and the clustering distance matrix is updated according to the average distance.
3. The method according to claim 1, characterized in that The historical power generation of the photovoltaic power station included in each of the group groups and the meteorological parameter information at the same time when the correlation with the historical power generation reaches a preset correlation threshold are respectively regularized to obtain target data, including: The historical power generation and the meteorological parameter information at the same moment are respectively used as feature data, and all recorded values of each feature data in the time series are counted as target feature data of each feature data; Determine a first limit value, a second limit value, and a limit interval from the target feature data; wherein the first limit value is a value that reaches a first proportion in the target feature data, the second limit value is a value that reaches a second proportion in all data, and the limit interval is a difference between the first limit value and the second limit value; the first proportion is less than the second proportion; The upper limit of the abnormal value and the lower limit of the abnormal value are obtained according to the first limit value, the second limit value and the limit interval; wherein the upper limit of the abnormal value is the sum of the first limit value and the limit interval of a preset multiple; the lower limit of the abnormal value is the difference between the second limit value and the limit interval of a preset multiple; When the target feature data is within the range of the abnormal value upper limit and the abnormal value lower limit, the target feature data is determined to be a normal value, and values of the target feature data above the abnormal value upper limit and below the abnormal value lower limit are all abnormal values; The detected abnormal data are regarded as missing values, and the filling values are calculated using the interpolation method to obtain the processed target data.
4. The method according to claim 1, characterized in that: The step of optimizing the model parameters of the photovoltaic power generation prediction model based on the training sample set to obtain the optimal power generation prediction model includes: Inputting the training sample set into a preset first power prediction model to obtain an initial power prediction value; Acquire actual historical generated power, and determine the difference between the actual historical generated power and the initial power prediction value, and use the difference as correction data to form a correction data set; Inputting the correction data set into a preset photovoltaic power error correction model to obtain a correction prediction value; The initial prediction value is superimposed on the corrected prediction value to obtain a corrected photovoltaic power generation value, and the corrected photovoltaic power generation value is used to perform model training on a photovoltaic power generation prediction model to determine an optimal photovoltaic power generation prediction model.
5. The method according to claim 4, characterized in that The step of inputting the training sample set into a preset first power prediction model to obtain an initial power prediction value includes: Decomposing the power generation of the photovoltaic power station in the training sample set by using a preset decomposition algorithm to obtain at least two intrinsic mode components IMF and corresponding at least two residual components RES; Forming training sets for each of the intrinsic mode components IMF and the corresponding residual components RES respectively, so as to establish SVM regression models corresponding to each of the intrinsic mode components IMF and the corresponding residual components RES respectively; The penalty parameter c and the kernel function parameter g of each SVM regression model are optimized by using a preset sparrow search algorithm to obtain each optimized optimal SVM regression model, and each optimal SVM regression model is used as a preset first power prediction model to perform power prediction on a preset test set to obtain a corresponding photovoltaic power generation value; The photovoltaic power generation values are superimposed as an initial prediction value.
6. The method according to claim 5, characterized in that The method of using a preset decomposition algorithm to decompose the power generation power of the photovoltaic power station in the training sample set to obtain at least two intrinsic mode components IMF and corresponding at least two residual components RES includes: Acquire an original power signal corresponding to the power generated by photovoltaic power generation, and add a pair of opposite positive and negative white noises to the original power signal to generate a corresponding first mixed power signal and a second mixed power signal; Performing modal decomposition on the first mixed power signal and the second mixed power signal respectively to obtain corresponding first IMF components and second IMF components, and storing the first IMF components in a first component set, and storing the second IMF components in a second component set; Returning to the step of adding a pair of opposite positive and negative white noises to the original power signal to generate the corresponding first mixed power signal and the second mixed power signal, each time adding a new normally distributed Gaussian white noise to update the first component set and the second component set until the number of extreme value points of the first IMF component and the second IMF component does not exceed a preset number threshold, to obtain an updated target first component set and a target second component set; The IMF component means are respectively solved for the first IMF component included in the target first component set and the second IMF component included in the target second component set to obtain at least two intrinsic mode components IMF and at least two residual components RES.
7. The method according to claim 6, characterized in that The performing modal decomposition on the first mixed power signal and the second mixed power signal respectively to obtain corresponding first IMF components and second IMF components includes: Extracting upper extreme value points and lower extreme value points corresponding to the first mixed power signal and the second mixed power signal respectively; Generate a corresponding upper envelope and a lower envelope according to the first mixed power signal, the second mixed power signal and the upper extreme point and the lower extreme point respectively corresponding to each other, and determine envelope mean values of the upper envelope and the lower envelope; Subtracting the first mixed power signal and the second mixed power signal from the corresponding envelope mean values respectively to obtain corresponding target power signals having local oscillation characteristics; Determining whether the target power signal satisfies a decomposition termination condition; If satisfied, the target power signal is used as the first IMF component and the second IMF component; If not satisfied, take the target signal as the first mixed power signal and the second mixed power signal, and return to the step of extracting the upper extreme points and the lower extreme points corresponding to the first mixed power signal and the second mixed power signal respectively until the decomposition termination condition is met; wherein the decomposition termination condition includes at least: the number of extreme points is greater than a preset threshold, or whether the maximum and minimum values appear alternately without repeated extreme points.
8. A photovoltaic power generation power prediction device, characterized in that: include: A correlation determination module is used to obtain the power generation and meteorological parameter information of the photovoltaic power stations included in the photovoltaic power station cluster, and to perform correlation analysis on the historical power generation and meteorological parameter information corresponding to each of the photovoltaic power stations to obtain a correlation analysis result; A grouping module, used to determine the meteorological parameter information at the moment when the correlation with the historical power generation reaches a preset correlation threshold according to the correlation analysis result, and cluster the photovoltaic power stations based on the correlation analysis result to obtain a cluster grouping result; A data regularization module is used to regularize the historical power generation of the photovoltaic power station contained in each of the grouping results and the meteorological parameter information at the same time when the correlation with the historical power generation reaches a preset correlation threshold, so as to obtain target data; An optimization module, used to use the target data as a training sample set, and to tune the model parameters of the photovoltaic power generation prediction model based on the training sample set to obtain an optimal power generation prediction model; The prediction module is used to obtain future meteorological factor data, and input the future meteorological factor data into the optimal power generation prediction model to obtain photovoltaic power generation prediction data.
9. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the photovoltaic power generation prediction method according to any one of claims 1 to 7.
10. 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 processor to implement the photovoltaic power prediction method according to any one of claims 1 to 7 when executed.
Citation Information
Cited By
Distributed photovoltaic short-term prediction method based on meteorological data
CN120338198A
Source network load storage system operation optimization method and system based on digital twinning
CN120389441A
Photovoltaic module output power prediction method for photovoltaic power station economic benefit evaluation
CN120956212A
