Power generation prediction method and device

Abstract

The application discloses a power generation prediction method and device, and belongs to the technical field of data processing. The method comprises the following steps: acquiring first meteorological forecast data of each power generation station in a power generation station cluster in a first time period; determining a cluster to which each power generation station belongs based on the first meteorological forecast data and first power generation data; determining a prediction model corresponding to each cluster based on second power generation data of the power generation stations included in each cluster in a second time period and second meteorological forecast data of the power generation stations in a historical prediction time period; and calling the prediction model corresponding to each cluster to predict power generation data of the power generation stations included in each cluster. The application clusters each power generation station in the power generation station cluster, respectively acquires the prediction model corresponding to the power generation stations included in each cluster, and predicts the power generation data of the power generation stations included in each cluster based on the prediction model, thereby reducing the consumption of computing resources and solving the bottleneck problem existing in the computing power and storage.

Description

Technical Field

[0001] The present application belongs to the field of data processing technology, and in particular relates to a power generation prediction method and device. Background Art

[0002] The deployment of various clean power generation sources, including photovoltaics, wind power, and hydropower, is increasing. Renewable energy sources, such as photovoltaics and wind power, are uncontrollable and weather-dependent. Accurate power generation forecasts for these uncontrollable energy sources facilitate advance planning of power dispatch. However, market-based electricity prices also pose significant challenges to the profitability of wind and solar power plants.

[0003] Existing technologies often focus on mapping key weather parameters and power forecasts for a single power station, using a single model for each station. However, with the growing demand for wind and solar power plants, especially for household clean energy, single-station models require significant data storage, model training, and model inference time. As companies serve hundreds of thousands or even millions of households using clean energy, single-station models face significant computing and storage bottlenecks in model updates and computation. Summary of the Invention

[0004] The present application aims to solve at least one of the technical problems existing in the prior art. To this end, the present application proposes a power generation prediction method, which clusters each power station in a power station cluster, obtains a prediction model corresponding to the power station included in each cluster, and predicts the power generation data of the power station included in each cluster based on the prediction model. This solves the problem of repeated data extraction in obtaining the prediction model corresponding to each power station in the prior art, which consumes computing resources and has bottlenecks in computing power and storage.

[0005] In a first aspect, the present application provides a power generation prediction method, the method comprising:

[0006] Obtaining first weather forecast data for each power station in the power station cluster within a first time period;

[0007] determining, based on the first weather forecast data and first power generation data of each power station in the first time period, the cluster to which each power station belongs;

[0008] Determining a prediction model corresponding to each cluster based on second power generation data of the power plants included in each cluster within a second time period and second meteorological forecast data of the power plants included in each cluster within a historical prediction time period;

[0009] The prediction model corresponding to each cluster is called to predict the power generation data of the power stations included in each cluster.

[0010] According to the power generation prediction method of the present application, each power station in the power station cluster is clustered, and the prediction model corresponding to the power station included in each cluster is obtained respectively. The power generation data of the power station included in each cluster is predicted based on the prediction model, which solves the problem of repeated data pulling when obtaining the prediction model corresponding to each power station in the existing technology, resulting in consumption of computing resources, and bottlenecks in computing power and storage.

[0011] According to one embodiment of the present application, determining the cluster to which each power station belongs based on the first weather forecast data and the first power generation data of each power station in the first time period includes:

[0012] Extracting the irradiance characteristics in the first meteorological forecast data to obtain a first statistic corresponding to each power station;

[0013] Performing feature extraction on the irradiation features in the first meteorological forecast data to obtain a first preset number of largest first feature values ​​corresponding to each power station;

[0014] Using the first statistic and the first eigenvalue as derived features, generating a derived feature group corresponding to each power station;

[0015] obtaining a target derived feature in the derived feature group having the highest correlation with the first power generation data;

[0016] Based on the target derived characteristics and the first power generation data, obtaining statistical characteristics corresponding to each power station;

[0017] Based on the Gaussian mixture clustering algorithm, the statistical features are clustered to determine the cluster to which each power station belongs.

[0018] According to the power generation prediction method of the present application, the first statistic and the first eigenvalue are obtained by extracting the irradiation characteristics in the first meteorological forecast data, and are used as derivative features to generate a derivative feature group corresponding to each power station. The target derivative feature with the highest correlation with the first power generation data is screened out from the derivative feature group. Based on the target derivative feature, the statistical feature corresponding to each power station is obtained. By clustering the statistical feature, the cluster to which each power station belongs is obtained. Since the statistical feature is used when clustering each power station in the power station cluster, the deviation caused by clustering using a single feature is reduced, and the clustering accuracy is improved.

[0019] According to one embodiment of the present application, obtaining the statistical features corresponding to each power station based on the target-derived features and the first power generation data includes:

[0020] Normalizing the first power generation data to obtain normalized first power generation data;

[0021] Grouping the target derived features corresponding to each power station;

[0022] performing statistics on the normalized first power generation data of the power stations included in each group to obtain a second statistic;

[0023] Based on the second statistic, the statistical feature is obtained.

[0024] According to the power generation prediction method of the present application, the first power generation data of each power station is normalized to obtain the normalized first power generation data, and the target derived features are grouped and the normalized first power generation data of the power stations included in each group are statistically analyzed to obtain statistical features. Since the statistical features are obtained by standardizing the power generation data of different power stations, the clustering results obtained based on the statistical features can reduce the impact of the order of magnitude difference of power generation data between different power stations on the clustering results, thereby improving the clustering accuracy.

[0025] According to one embodiment of the present application, extracting the irradiation features in the first meteorological forecast data to obtain the first statistic corresponding to each power station includes:

[0026] Statistics are performed on the quantile, mean, and variance corresponding to the radiation characteristics to obtain the first statistic.

[0027] According to one embodiment of the present application, extracting the irradiation characteristics in the first meteorological forecast data to obtain the first preset number of largest first characteristic values ​​corresponding to each power station includes:

[0028] Based on a principal component analysis algorithm, feature extraction is performed on the radiation feature to obtain the first eigenvalue.

[0029] According to one embodiment of the present application, determining a prediction model corresponding to each cluster based on second power generation data of the power plants included in each cluster within a second time period and second meteorological forecast data of the power plants included in each cluster within a historical prediction time period includes:

[0030] performing feature extraction on the irradiation features in the second weather forecast data to obtain a third statistic corresponding to the power stations included in each cluster;

[0031] Performing feature extraction on the irradiation features in the second meteorological forecast data to obtain the first second preset number of second feature values ​​corresponding to the power stations included in each cluster;

[0032] Based on a preset time series decomposition algorithm, obtaining a trend component and a period component of the irradiance characteristics in the second meteorological forecast data;

[0033] The third statistic, the second eigenvalue, the trend component, the periodic component and the second power generation data are input into a preset prediction model for training to obtain a prediction model corresponding to each cluster.

[0034] According to the power generation prediction method of the present application, by uniformly training prediction models for the power plants included in each cluster, the problem of long training cycles caused by the need to train prediction models separately for each power plant in the prior art is reduced, the training cycle is shortened, and the prediction efficiency is improved.

[0035] According to one embodiment of the present application, the preset prediction model includes a distributed gradient boosting model, a lightweight gradient boosting machine learning model and a deep learning model.

[0036] According to one embodiment of the present application, obtaining first weather forecast data for each power station in the power station cluster within a first time period includes:

[0037] Based on the geographical location of each power station, first weather forecast data from different sources within the first time period is obtained for each power station.

[0038] According to one embodiment of the present application, calling the prediction model corresponding to each cluster to predict the power generation data of the power stations included in each cluster includes:

[0039] According to the meteorological forecast data of the power plants included in each cluster within the target time period and the power generation data of the power plants included in each cluster within the historical time period, the prediction model corresponding to each cluster is called to predict the power generation data of the power plants included in each cluster within the target time period.

[0040] According to one embodiment of the present application, the power generation data includes:

[0041] Power generation or electricity production.

[0042] According to the power generation prediction method of the present application, the prediction of the generated power and generated amount of the power stations included in each cluster can be achieved through the prediction model of the cluster to which each power station in the power station cluster belongs.

[0043] In a second aspect, the present application provides a power generation prediction device, the device comprising:

[0044] An acquisition module, configured to acquire first weather forecast data for each power station in the power station cluster within a first time period;

[0045] a clustering module, configured to determine the cluster to which each power station belongs based on the first weather forecast data and the first power generation data of each power station within the first time period;

[0046] a model determination module, configured to determine a prediction model corresponding to each cluster based on second power generation data of the power plants included in each cluster within a second time period and second meteorological forecast data of the power plants included in each cluster within a historical prediction time period;

[0047] The prediction module is used to call the prediction model corresponding to each cluster to predict the power generation data of the power stations included in each cluster.

[0048] According to the power generation prediction device of the present application, each power station in the power station cluster is clustered, and the prediction model corresponding to the power station included in each cluster is obtained respectively. The power generation data of the power station included in each cluster is predicted based on the prediction model, which solves the problem of consuming computing resources due to the need to repeatedly pull data when obtaining the prediction model corresponding to each power station in the existing technology, and the bottleneck problems in computing power and storage.

[0049] According to one embodiment of the present application, the clustering module is further configured to:

[0050] Extracting the irradiance characteristics in the first meteorological forecast data to obtain a first statistic corresponding to each power station;

[0051] Performing feature extraction on the irradiation features in the first meteorological forecast data to obtain a first preset number of largest first feature values ​​corresponding to each power station;

[0052] Using the first statistic and the first eigenvalue as derived features, generating a derived feature group corresponding to each power station;

[0053] obtaining a target derived feature in the derived feature group having the highest correlation with the first power generation data;

[0054] Based on the target derived characteristics and the first power generation data, obtaining statistical characteristics corresponding to each power station;

[0055] Based on the Gaussian mixture clustering algorithm, the statistical features are clustered to determine the cluster to which each power station belongs.

[0056] According to the power generation prediction device of the present application, the first statistic and the first eigenvalue are obtained by extracting the irradiation characteristics in the first meteorological forecast data, and are used as derivative features to generate a derivative feature group corresponding to each power station. The target derivative feature with the highest correlation with the first power generation data is screened out from the derivative feature group. Based on the target derivative feature, the statistical feature corresponding to each power station is obtained. By clustering the statistical feature, the cluster to which each power station belongs is obtained. Since the statistical feature is used when clustering each power station in the power station cluster, the deviation caused by clustering using a single feature is reduced, and the clustering accuracy is improved.

[0057] According to one embodiment of the present application, the clustering module is further configured to:

[0058] Normalizing the first power generation data to obtain normalized first power generation data;

[0059] Grouping the target derived features corresponding to each power station;

[0060] performing statistics on the normalized first power generation data of the power stations included in each group to obtain a second statistic;

[0061] Based on the second statistic, the statistical feature is obtained.

[0062] According to the power generation prediction device of the present application, the first power generation data of each power station is normalized to obtain the normalized first power generation data, and the target derived features are grouped and the normalized first power generation data of the power stations included in each group are statistically analyzed to obtain statistical features. Since the statistical features are obtained by standardizing the power generation data of different power stations, the clustering results obtained based on the statistical features can reduce the impact of the order of magnitude difference of power generation data between different power stations on the clustering results, thereby improving the clustering accuracy.

[0063] According to one embodiment of the present application, the clustering module is further configured to:

[0064] Statistics are performed on the quantile, mean, and variance corresponding to the radiation characteristics to obtain the first statistic.

[0065] According to one embodiment of the present application, the clustering module is further configured to:

[0066] Based on a principal component analysis algorithm, feature extraction is performed on the radiation feature to obtain the first eigenvalue.

[0067] According to one embodiment of the present application, the model determination module is further configured to:

[0068] performing feature extraction on the irradiation features in the second weather forecast data to obtain a third statistic corresponding to the power stations included in each cluster;

[0069] Performing feature extraction on the irradiation features in the second meteorological forecast data to obtain the first second preset number of second feature values ​​corresponding to the power stations included in each cluster;

[0070] Based on a preset time series decomposition algorithm, obtaining a trend component and a period component of the irradiance characteristics in the second meteorological forecast data;

[0071] The third statistic, the second eigenvalue, the trend component, the periodic component and the second power generation data are input into a preset prediction model for training to obtain a prediction model corresponding to each cluster.

[0072] According to the power generation prediction device of the present application, by uniformly training prediction models for the power plants included in each cluster, the problem of long training cycle caused by the need to train prediction models separately for each power plant in the prior art is reduced, the training cycle is shortened, and the prediction efficiency is improved.

[0073] According to one embodiment of the present application, the preset prediction model includes any one of a distributed gradient boosting model, a lightweight gradient boosting machine learning model and a deep learning model.

[0074] According to one embodiment of the present application, the acquisition module is further configured to:

[0075] Based on the geographical location of each power station, first weather forecast data from different sources within the first time period is obtained for each power station.

[0076] According to one embodiment of the present application, the prediction module is further configured to:

[0077] According to the meteorological forecast data of the power plants included in each cluster within the target time period and the power generation data of the power plants included in each cluster within the historical time period, the prediction model corresponding to each cluster is called to predict the power generation data of the power plants included in each cluster within the target time period.

[0078] According to one embodiment of the present application, the power generation data includes:

[0079] Power generation or electricity production.

[0080] According to the power generation prediction device of the present application, the prediction of the generated power and generated amount of the power stations included in each cluster can be achieved through the prediction model of the cluster to which each power station in the power station cluster belongs.

[0081] In a third aspect, the present application provides an electronic device comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein when the processor executes the computer program, the power generation prediction method as described in the first aspect above is implemented.

[0082] In a fourth aspect, the present application provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the power generation prediction method as described in the first aspect above.

[0083] In a fifth aspect, the present application provides a chip, which includes a processor and a communication interface, wherein the communication interface is coupled to the processor, and the processor is used to run programs or instructions to implement the power generation prediction method as described in the first aspect.

[0084] In a sixth aspect, the present application provides a computer program product, comprising a computer program, which, when executed by a processor, implements the power generation prediction method as described in the first aspect above.

[0085] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become obvious from the description below, or will be learned through practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0086] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which:

[0087] Figure 1 This is one of the flow charts of the power generation prediction method provided in the embodiment of the present application;

[0088] Figure 2 This is a schematic diagram of a highly correlated cluster combination obtained by clustering the statistical characteristics of a single power station using a Gaussian mixture clustering algorithm provided in an embodiment of the present application;

[0089] Figure 3 is a schematic diagram of a result of normalizing first power generation data of a power station provided by an embodiment of the present application;

[0090] Figure 4 is a schematic diagram of statistical characteristics of a power generation station provided in an embodiment of the present application;

[0091] Figure 5 This is a schematic diagram of prediction results of calling the prediction model corresponding to the power stations of the same cluster provided by an embodiment of the present application;

[0092] Figure 6 This is the second flow chart of the power generation prediction method provided in the embodiment of the present application;

[0093] Figure 7 It is a structural diagram of a power generation prediction device provided in an embodiment of the present application;

[0094] Figure 8 It is a structural diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0095] The following will be combined with the accompanying drawings in the embodiments of the present application to clearly describe the technical solutions in the embodiments of the present application. Obviously, the embodiments described are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field are within the scope of protection of this application.

[0096] The terms "first," "second," and the like in the specification and claims of this application are used to distinguish similar objects, and are not used to describe a specific order or precedence. It should be understood that the terms used in this manner are interchangeable where appropriate, so that the embodiments of this application can be implemented in an order other than that illustrated or described herein, and that the objects distinguished by "first," "second," and the like are generally of the same type, and do not limit the number of objects; for example, the first object can be one or more. In addition, the term "and / or" in the specification and claims refers to at least one of the connected objects, and the character " / " generally indicates that the objects connected are in an "or" relationship.

[0097] The power generation prediction method, power generation prediction device, electronic device and readable storage medium provided in the embodiments of the present application are described in detail below with reference to the accompanying drawings through specific embodiments and their application scenarios.

[0098] The power generation prediction method may be applied to a terminal, and may be specifically executed by hardware or software in the terminal.

[0099] The terminal includes, but is not limited to, a portable communication device such as a mobile phone or tablet computer having a touch-sensitive surface (e.g., a touch screen display and / or a touch pad). It should also be understood that, in some embodiments, the terminal may not be a portable communication device, but a desktop computer having a touch-sensitive surface (e.g., a touch screen display and / or a touch pad).

[0100] In the following embodiments, a terminal including a display and a touch-sensitive surface is described. However, it should be understood that the terminal may include one or more other physical user interface devices such as a physical keyboard, a mouse, and a joystick.

[0101] The power generation prediction method provided in the embodiment of the present application can be executed by an electronic device or a functional module or functional entity in the electronic device that can implement the power generation prediction method. The electronic devices mentioned in the embodiment of the present application include but are not limited to mobile phones, tablets, computers, cameras and wearable devices. The power generation prediction method provided in the embodiment of the present application is described below using an electronic device as an example of the execution subject.

[0102] like Figure 1 As shown, the power generation prediction method includes: step 110, step 120, step 130 and step 140.

[0103] Step 110: Obtain first weather forecast data for each power station in the power station cluster within a first time period.

[0104] Specifically, the power station cluster includes multiple power stations, which can be clean energy power stations such as photovoltaic, wind power, and hydropower.

[0105] The first weather forecast data may specifically be weather forecast data for each power station in the power station cluster within a first time period, which may be a past time period or a pre-set time period. The weather forecast data may be multimodal weather forecasts from different sources, obtained based on the geographical location grid points corresponding to the longitude and latitude of each power station, and may specifically include key characteristics such as irradiance, temperature, and wind speed.

[0106] In a specific implementation, a weather forecast data acquisition module can be constructed to pull weather forecast data using the geographical location grid points corresponding to the longitude and latitude of each power station in the power station cluster as units to obtain weather forecast data for each power station in a first time period, i.e., first weather forecast data.

[0107] Step 120: Determine the cluster to which each power plant belongs based on the first weather forecast data and the first power generation data of each power plant in the first time period.

[0108] Specifically, the first power generation data may be power generation data of each power station in the power station cluster during a first time period, and the power generation data may include power generation power or power generation amount. Both the power generation power and power generation amount can be obtained by querying the power generation data of each power station during the first time period.

[0109] By clustering each power station in the power station cluster based on the first weather forecast data and first power generation data obtained for each power station in a first time period, the cluster to which each power station belongs in the first time period is obtained.

[0110] Step 130: Determine a prediction model corresponding to each cluster based on the second power generation data of the power plants included in each cluster in the second time period and the second meteorological forecast data of the power plants included in each cluster in the historical prediction time period.

[0111] Specifically, the second power generation data may be power generation data for each power plant included in the cluster during a second time period. The power generation data may also be power generation power or power generation amount. The second time period may be a time period within the first time period, or a time period including the first time period. The historical prediction time period may specifically be a time period within the second time period.

[0112] The second weather forecast data can be obtained through the weather forecast data acquisition module constructed above. Specifically, the weather forecast data is pulled using the geographical location grid points corresponding to the longitude and latitude of each power station in the power station cluster as units to obtain the weather forecast data of each power station within the historical forecast time period, i.e., the second weather forecast data.

[0113] By inputting the power generation data (i.e., the second power generation data) of the power stations included in each cluster within the second time period and the meteorological forecast data (i.e., the second meteorological forecast data) of the power stations included in each cluster within the historical prediction time period into the preset prediction module for training, a prediction model corresponding to the power stations included in each cluster is obtained.

[0114] It should be noted that when the second time period is a time period that includes the first time period, the second power generation data and the second weather forecast data can be used to obtain a prediction model for the cluster to which each power station belongs within the first time period. To further improve the accuracy of clustering, the weather forecast data for each power station obtained from the first time period to the second time period can be added to the first weather data, and the power generation data for each power station obtained from the first time period to the second time period can be added to the first power generation data. The power generation data and weather forecast data for the entire second time period can be used to update the cluster to which each power station belongs within the first time period. Based on the power generation data of the power stations included in each updated cluster during the second time period and the weather forecast data of the power stations included in each updated cluster during the historical prediction time period, the prediction model corresponding to the power stations included in each updated cluster is determined.

[0115] Step 140: Call the prediction model corresponding to each cluster to predict the power generation data of the power stations included in each cluster.

[0116] Specifically, by calling the prediction model corresponding to each cluster, the power generation data of the power stations included in each cluster are predicted.

[0117] Similarly, when the second time period is a time period including the first time period, in order to further improve the prediction accuracy of the power generation data, the power generation data of the power stations included in each updated cluster can be predicted by calling the prediction model corresponding to the power stations included in each updated cluster.

[0118] According to the power generation prediction method provided in the embodiment of the present application, each power station in the power station cluster is clustered, and the prediction model corresponding to the power station included in each cluster is obtained respectively. The power generation data of the power station included in each cluster is predicted based on the prediction model. This solves the problem of repeated data pulling when obtaining the prediction model corresponding to each power station in the prior art, which consumes computing resources and has bottlenecks in computing power and storage.

[0119] In some embodiments, step 120, determining the cluster to which each power plant belongs based on the first weather forecast data and the first power generation data of each power plant in the first time period, may specifically include:

[0120] Extracting the irradiance characteristics in the first meteorological forecast data to obtain a first statistic corresponding to each power station;

[0121] Performing feature extraction on the irradiation features in the first meteorological forecast data to obtain a first preset number of largest first feature values ​​corresponding to each power station;

[0122] Using the first statistic and the first eigenvalue as derived features, generating a derived feature group corresponding to each power station;

[0123] obtaining a target derived feature in the derived feature group having the highest correlation with the first power generation data;

[0124] Based on the target derived characteristics and the first power generation data, obtaining statistical characteristics corresponding to each power station;

[0125] Based on the Gaussian mixture clustering algorithm, the statistical features are clustered to determine the cluster to which each power station belongs.

[0126] Specifically, a multimodal data feature engineering unit is constructed. Based on the first weather forecast data obtained from different sources, derived features are constructed based on multiple statistical methods to characterize derived feature groups with different dimensions and smaller deviations.

[0127] Specifically, feature extraction is performed on the irradiation features in the first weather forecast data obtained above to obtain a statistic corresponding to each power station, ie, a first statistic.

[0128] By extracting the radiation characteristics such as short wave radiation and direct solar radiation in the first meteorological forecast data from different sources, the first preset number of largest eigenvalues ​​corresponding to each power station can be obtained, that is, the first eigenvalue.

[0129] The first statistic and the first eigenvalue obtained above are used as derived features to generate a derived feature group corresponding to each power station.

[0130] According to the association relationship between each derived feature in the derived feature group and the actual power generation data of the power station, by calculating the correlation between each derived feature in the derived feature group and the above-mentioned first power generation data, the derived feature with the highest correlation with the first power generation data in the derived feature group is obtained, that is, the target derived feature.

[0131] The number of the target derived features may be one or more, and this application does not impose any specific limitation on this.

[0132] Based on the obtained target derived features and the first power generation data of each power station, the statistical features corresponding to each power station are obtained. The obtained statistical features are then clustered using the Gaussian mixture clustering algorithm to obtain the cluster to which each power station belongs.

[0133] like Figure 2 As shown in Figure 1, the cluster combination with high correlation is obtained by clustering the statistical characteristics of a single power station using the Gaussian mixture clustering algorithm.

[0134] According to the cluster to which each power station belongs, a cluster-power station scheduling module is constructed, and prediction model training and prediction model scheduling are carried out on a cluster basis to predict the power generation data of the power stations included in each cluster.

[0135] In some embodiments, in the above steps, extracting the irradiation features in the first weather forecast data to obtain the first statistic corresponding to each power station may specifically include:

[0136] Statistics are performed on the quantile, mean, and variance corresponding to the radiation characteristics to obtain the first statistic.

[0137] Specifically, the first statistic may be obtained by extracting quantile, mean, and variance features of the irradiance features in the first meteorological forecast data from different sources.

[0138] In a specific implementation, feature extraction can be performed by extracting radiation features such as short wave radiation and direct solar radiation from the first meteorological forecast data from different sources, and obtaining statistics such as the 25% quantile, 50% quantile, 75% quantile, mean, variance, etc. of the same radiation feature at the same moment as the first statistic.

[0139] In some embodiments, in the above steps, extracting the irradiance characteristics in the first meteorological forecast data to obtain the first preset number of largest first characteristic values ​​corresponding to each power station includes:

[0140] Based on a principal component analysis algorithm, feature extraction is performed on the radiation feature to obtain the first eigenvalue.

[0141] In a specific implementation, a principal component analysis algorithm may be used to extract the irradiance features in the first meteorological forecast data from different sources, and extract the top k (for example, k may be 5, 10, etc.) largest first eigenvalues.

[0142] According to the power generation prediction method provided in the embodiment of the present application, the first statistic and the first eigenvalue are obtained by extracting the irradiation characteristics in the first meteorological forecast data, and are used as derivative features to generate a derivative feature group corresponding to each power station. The target derivative feature with the highest correlation with the first power generation data is screened out from the derivative feature group. Based on the target derivative feature, the statistical feature corresponding to each power station is obtained. By clustering the statistical feature, the cluster to which each power station belongs is obtained. Since the statistical feature is used when clustering each power station in the power station cluster, the deviation caused by clustering using a single feature is reduced, and the clustering accuracy is improved.

[0143] In some embodiments, in the above steps, obtaining the statistical features corresponding to each power station based on the target-derived features and the first power generation data may specifically include:

[0144] Normalizing the first power generation data to obtain normalized first power generation data;

[0145] Grouping the target derived features corresponding to each power station;

[0146] performing statistics on the normalized first power generation data of the power stations included in each group to obtain a second statistic;

[0147] Based on the second statistic, the statistical feature is obtained.

[0148] Specifically, the first power generation data obtained above is normalized to obtain the normalized first power generation data (such as Figure 3 As shown in FIG, the first power generation data of 13 power stations are normalized), which can be achieved by normalizing the first power generation data. The formula is as follows:

[0149]

[0150] Where X represents the first power generation data, X′ represents the normalized first power generation data, and X min Represents the minimum value of the first power generation data, X max Represents the maximum value of the first power generation data.

[0151] The target derived features corresponding to each power station are grouped. In the specific implementation, the grouping rule used in this application is grouped at intervals of 50, and the upper limit of the grouping can be set to 500W / m 2 The target derived features corresponding to each power station can be divided into 10 groups by grouping, and statistics can be performed on the normalized first power generation data of the power stations included in each group. For example, statistics can be performed on the 10%, 25%, 50%, 75%, and 90% quantiles of the normalized first power generation data of the power stations included in each group to obtain a statistic, i.e., a second statistic.

[0152] The final statistical features are obtained based on the second statistic obtained for each group. For example, if there are 10 groups, the 10%, 25%, 50%, 75%, and 90% quantiles are obtained for each group, and 50 statistical features can be obtained for each power station.

[0153] like Figure 4 As shown, 13 power stations are selected as test stations, the first power generation data of the 13 power stations are normalized respectively, and the target derived features corresponding to the 13 power stations are grouped to obtain the statistical features of each power station.

[0154] According to the power generation prediction method provided in the embodiment of the present application, the first power generation data of each power station is normalized to obtain the normalized first power generation data, and the target derived features are grouped and the normalized first power generation data of the power stations included in each group are statistically analyzed to obtain statistical features. Since the statistical features are obtained by standardizing the power generation data of different power stations, the clustering results obtained based on the statistical features can reduce the impact of the order of magnitude difference of power generation data between different power stations on the clustering results, thereby improving the clustering accuracy.

[0155] In some embodiments, step 130, determining a prediction model corresponding to each cluster based on the second power generation data of the power plants included in each cluster within the second time period and the second meteorological forecast data of the power plants included in each cluster within the historical prediction time period, may specifically include:

[0156] performing feature extraction on the irradiation features in the second weather forecast data to obtain a third statistic corresponding to the power stations included in each cluster;

[0157] Performing feature extraction on the irradiation features in the second meteorological forecast data to obtain the first second preset number of second feature values ​​corresponding to the power stations included in each cluster;

[0158] Based on a preset time series decomposition algorithm, obtaining a trend component and a period component of the irradiance characteristics in the second meteorological forecast data;

[0159] The third statistic, the second eigenvalue, the trend component, the periodic component and the second power generation data are input into a preset prediction model for training to obtain a prediction model corresponding to each cluster.

[0160] Specifically, feature engineering is performed on the obtained second meteorological forecast data, and the results of the feature engineering extraction of the second meteorological forecast data (including the third statistic, the second eigenvalue, the trend component and the periodic component of the irradiation characteristics in the second meteorological forecast data) and the second power generation data are input into the preset prediction model for training to obtain the prediction model corresponding to the power station included in each cluster.

[0161] The feature engineering extraction of the second weather forecast data may specifically include:

[0162] Feature extraction is performed on the irradiation features in the second weather forecast data obtained above to obtain a statistic corresponding to each power station, ie, a third statistic.

[0163] By extracting the radiation characteristics such as short wave radiation and direct solar radiation in the second meteorological forecast data from different sources, the second preset number of largest eigenvalues ​​corresponding to each power station can be obtained, that is, the second eigenvalues.

[0164] The third statistic can be obtained by extracting quantile, mean and variance features of the irradiance features in the second meteorological forecast data from different sources.

[0165] In a specific implementation, feature extraction can be performed by extracting irradiation features such as short wave radiation and direct solar radiation from the second meteorological forecast data from different sources, and obtaining statistics such as the 25% quantile, 50% quantile, 75% quantile, mean, and variance of the same irradiation feature at the same moment as the third statistic.

[0166] By performing feature extraction on the irradiance features in the obtained second weather forecasts from different sources, the first second preset number of second feature values ​​corresponding to the power stations included in each cluster are obtained.

[0167] In a specific implementation, a principal component analysis algorithm may be used to extract the irradiance features in the second weather forecast data from different sources, and extract the top k1 (for example, k1 may be 5, 10, etc.) largest second eigenvalues.

[0168] The trend component and periodic component of the radiation characteristics such as short wave radiation and direct solar radiation in the second weather forecast data are obtained by a preset time series decomposition algorithm, such as the STL time series decomposition algorithm.

[0169] The obtained third statistic, second eigenvalue, trend component, periodic component and second power generation data are input into a preset prediction model for training to obtain a prediction model corresponding to the power generation station included in each cluster.

[0170] According to the power generation prediction method provided in the embodiment of the present application, by uniformly training prediction models for the power plants included in each cluster, the problem of long training cycles caused by the need to train prediction models separately for each power plant in the prior art is reduced, the training cycle is shortened, and the prediction efficiency is improved.

[0171] In some embodiments, the preset prediction model includes a distributed gradient boosting model, a lightweight gradient boosting machine learning model, and a deep learning model.

[0172] Specifically, the preset prediction model used in this application can be an ensemble learning model group, such as an ensemble learning model group composed of a distributed gradient boosting model, a lightweight gradient boosting machine learning model, and a deep learning model. The result output by the ensemble learning model group is the prediction result of the prediction model. During the training process, a grid search can be used to obtain the parameter range. Based on the searched parameter range, the key parameter optimization method is further used to optimize the parameters during each prediction model training until the preset prediction model converges.

[0173] A timing program may also be designed to extract the second weather forecast data and the second power generation data included in a single cluster and perform model training at certain date intervals.

[0174] In some embodiments, step 110, obtaining first weather forecast data for each power station in the power station cluster within a first time period, includes:

[0175] Based on the geographical location of each power station, first weather forecast data from different sources within the first time period is obtained for each power station.

[0176] Specifically, the first weather forecast data may be weather forecast data from different sources obtained based on the geographical location grid points corresponding to the longitude and latitude of each power station. The weather forecast data may specifically include key characteristics such as irradiation and temperature.

[0177] In some embodiments, step 140, calling the prediction model corresponding to each cluster to predict the power generation data of the power stations included in each cluster, may specifically include:

[0178] According to the meteorological forecast data of the power plants included in each cluster within the target time period and the power generation data of the power plants included in each cluster within the historical time period, the prediction model corresponding to each cluster is called to predict the power generation data of the power plants included in each cluster within the target time period.

[0179] Specifically, when using the trained prediction model to predict the power generation data of the power stations included in each cluster, the same feature engineering extraction can be performed on the meteorological forecast data of the power stations included in each cluster within the target time period, and then the results of the feature engineering extraction of the meteorological forecast data within the target time period and the power generation data of the power stations included in each cluster within the historical time period are input into the prediction model corresponding to the cluster to realize the prediction of the power generation data of the power stations included in each cluster within the target time period.

[0180] The target time period may be a future time period or a pre-set time period. The result of feature engineering extraction of the weather forecast data within the target time period may include a statistic obtained by feature extraction of the irradiance characteristics in the weather forecast data within the target time period, a preset number of top-ranked feature values ​​obtained by feature extraction of the irradiance characteristics in the weather forecast data within the target time period, and a trend component and a periodic component of the irradiance characteristics in the weather forecast data within the target time period.

[0181] In specific implementation, according to the required prediction period, such as once a day, a corresponding timing program can be designed to regularly pull the weather forecast data and historical power generation data of the power plant on the target day, call the prediction model of the cluster to which the power plant belongs, and predict the power generation data of the power plant on the target day.

[0182] like Figure 5 As shown, the prediction results of the power generation data of the four power stations in the same cluster are predicted by calling the prediction model corresponding to the same cluster.

[0183] like Figure 6 As shown, the power generation prediction method of the present application may further specifically include:

[0184] Set up a scheduled task for each cluster's prediction model scheduling to predict the power generation data of the power stations included in each cluster. Specifically:

[0185] The scheduled task is divided into a training program and a prediction program.

[0186] The specific process of the training program is as follows:

[0187] According to the above method, the cluster to which each power station in the power station cluster belongs is obtained.

[0188] Obtain the second power generation data of the power plant included in each cluster within the second time period (which can be obtained by querying the historical power generation data of the power plant, and the historical power generation data can be stored in the power generation database) and the second weather forecast data within the historical prediction time period (the second weather forecast data can be stored in the weather forecast database).

[0189] Feature engineering extraction is performed on the second weather forecast data, including feature extraction of irradiance features in the second weather forecast data to obtain a third statistic and a second eigenvalue; and based on a preset time series decomposition algorithm, a trend component and a periodic component in the second weather forecast data are obtained.

[0190] The obtained third statistic, second eigenvalue, trend component, periodic component and second power generation data are input into a preset prediction model for training, and a grid search is used to obtain the parameter range of the preset prediction model. According to the searched parameter range, the key parameter optimization method is used to optimize the parameters during each prediction model training until the preset prediction model converges to obtain a prediction model.

[0191] The specific process of the prediction program is as follows:

[0192] Obtain the power generation stations included in each cluster;

[0193] Obtaining weather forecast data of the power plants included in each cluster within a target time period from a weather forecast database;

[0194] The same feature engineering extraction is performed on the weather forecast data within the target time period, and the results of the feature engineering extraction of the weather forecast data within the target time period and the weather forecast data of the power stations included in each cluster within the historical time period are input into the prediction model to obtain the prediction results of the power generation data of the power stations included in each cluster within the target time period.

[0195] In some embodiments, the power generation data includes:

[0196] Power generation or electricity production.

[0197] Specifically, the power generation data in this application may specifically include power generation power or power generation amount.

[0198] According to the power generation prediction method provided in the embodiment of the present application, the power generation power and power generation amount of the power stations included in each cluster can be predicted through the prediction model of the cluster to which each power station in the power station cluster belongs.

[0199] The power generation prediction method provided in the embodiment of the present application can be executed by a power generation prediction device. In the embodiment of the present application, the power generation prediction device is used as an example to illustrate the power generation prediction method provided in the embodiment of the present application.

[0200] An embodiment of the present application also provides a power generation prediction device.

[0201] like Figure 7 As shown, the power generation prediction device includes: an acquisition module 710 , a clustering module 720 , a model determination module 730 and a prediction module 740 .

[0202] An acquisition module 710 is configured to acquire first weather forecast data for each power station in the power station cluster within a first time period;

[0203] A clustering module 720 is configured to determine the cluster to which each power station belongs based on the first weather forecast data and the first power generation data of each power station in the first time period;

[0204] A model determination module 730 is configured to determine a prediction model corresponding to each cluster based on second power generation data of the power plants included in each cluster within a second time period and second meteorological forecast data of the power plants included in each cluster within a historical prediction time period;

[0205] The prediction module 740 is used to call the prediction model corresponding to each cluster to predict the power generation data of the power stations included in each cluster.

[0206] According to the power generation prediction device provided in the embodiment of the present application, each power station in the power station cluster is clustered, and the prediction model corresponding to the power station included in each cluster is obtained respectively. The power generation data of the power station included in each cluster is predicted based on the prediction model, which solves the problem of repeated data pulling when obtaining the prediction model corresponding to each power station in the prior art, resulting in consumption of computing resources, and bottlenecks in computing power and storage.

[0207] In some embodiments, the clustering module 720 may also be used to:

[0208] Extracting the irradiance characteristics in the first meteorological forecast data to obtain a first statistic corresponding to each power station;

[0209] Performing feature extraction on the irradiation features in the first meteorological forecast data to obtain a first preset number of largest first feature values ​​corresponding to each power station;

[0210] Using the first statistic and the first eigenvalue as derived features, generating a derived feature group corresponding to each power station;

[0211] obtaining a target derived feature in the derived feature group having the highest correlation with the first power generation data;

[0212] Based on the target derived characteristics and the first power generation data, obtaining statistical characteristics corresponding to each power station;

[0213] Based on the Gaussian mixture clustering algorithm, the statistical features are clustered to determine the cluster to which each power station belongs.

[0214] According to the power generation prediction device provided in the embodiment of the present application, the first statistic and the first eigenvalue are obtained by extracting the irradiation characteristics in the first meteorological forecast data, and the first statistic and the first eigenvalue are used as derivative features to generate a derivative feature group corresponding to each power station. The target derivative feature with the highest correlation with the first power generation data is screened out from the derivative feature group. Based on the target derivative feature, the statistical feature corresponding to each power station is obtained. By clustering the statistical feature, the cluster to which each power station belongs is obtained. Since the statistical feature is used when clustering each power station in the power station cluster, the deviation caused by clustering using a single feature is reduced, and the clustering accuracy is improved.

[0215] In some embodiments, the clustering module 720 may also be used to:

[0216] Normalizing the first power generation data to obtain normalized first power generation data;

[0217] Grouping the target derived features corresponding to each power station;

[0218] performing statistics on the normalized first power generation data of the power stations included in each group to obtain a second statistic;

[0219] Based on the second statistic, the statistical feature is obtained.

[0220] According to the power generation prediction device provided in the embodiment of the present application, the first power generation data of each power station is normalized to obtain the normalized first power generation data, and the target derived features are grouped and the normalized first power generation data of the power stations included in each group are statistically analyzed to obtain statistical features. Since the statistical features are obtained by standardizing the power generation data of different power stations, the clustering results obtained based on the statistical features can reduce the impact of the difference in the order of magnitude of power generation data between different power stations on the clustering results, thereby improving the clustering accuracy.

[0221] In some embodiments, the clustering module 720 may also be used to:

[0222] Statistics are performed on the quantile, mean, and variance corresponding to the radiation characteristics to obtain the first statistic.

[0223] In some embodiments, the clustering module 720 may also be used to:

[0224] Based on a principal component analysis algorithm, feature extraction is performed on the radiation feature to obtain the first eigenvalue.

[0225] In some embodiments, the model determination module 730 may also be used to:

[0226] performing feature extraction on the irradiation features in the second weather forecast data to obtain a third statistic corresponding to the power stations included in each cluster;

[0227] Performing feature extraction on the irradiation features in the second meteorological forecast data to obtain the first second preset number of second feature values ​​corresponding to the power stations included in each cluster;

[0228] Based on a preset time series decomposition algorithm, obtaining a trend component and a period component of the irradiance characteristics in the second meteorological forecast data;

[0229] The third statistic, the second eigenvalue, the trend component, the periodic component and the second power generation data are input into a preset prediction model for training to obtain a prediction model corresponding to each cluster.

[0230] According to the power generation prediction device provided in the embodiment of the present application, by uniformly training prediction models for the power plants included in each cluster, the problem of long training cycles caused by the need to train prediction models separately for each power plant in the prior art is reduced, the training cycle is shortened, and the prediction efficiency is improved.

[0231] In some embodiments, the preset prediction model includes any one of a distributed gradient boosting model, a lightweight gradient boosting machine learning model, and a deep learning model.

[0232] In some embodiments, the acquisition module 710 may also be used to:

[0233] Based on the geographical location of each power station, first weather forecast data from different sources within the first time period is obtained for each power station.

[0234] In some embodiments, the prediction module 740 may also be used to:

[0235] The weather forecast data of the power plants included in each cluster within the target time period and the power generation data of the power plants included in each cluster within the historical time period are input into the prediction model, and the power generation data of the power plants included in each cluster within the target time period are predicted.

[0236] In some embodiments, the power generation data includes:

[0237] Power generation or electricity production.

[0238] According to the power generation prediction device provided in the embodiment of the present application, the power generation power and power generation amount of the power stations included in each cluster can be predicted through the prediction model of the cluster to which each power station in the power station cluster belongs.

[0239] The power generation prediction device in the embodiment of the present application can be an electronic device or a component in an electronic device, such as an integrated circuit or chip. The electronic device can be a terminal or other devices other than a terminal. For example, the electronic device can be a mobile phone, a tablet computer, a laptop computer, a PDA, an in-vehicle electronic device, a mobile Internet device (MID), an augmented reality (AR) / virtual reality (VR) device, a robot, a wearable device, an ultra-mobile personal computer (UMPC), a netbook or a personal digital assistant (PDA), etc. It can also be a server, a network attached storage (NAS), a personal computer (PC), a television (TV), a teller machine or a self-service machine, etc., and the embodiment of the present application does not specifically limit it.

[0240] The XXXX device in the embodiments of the present application may be a device having an operating system. The operating system may be a Microsoft (Windows) operating system, an Android operating system, an iOS operating system, or other possible operating systems, which are not specifically limited in the embodiments of the present application.

[0241] The power generation prediction device provided in the embodiment of the present application can achieve Figures 1 to 6 To avoid repetition, the various processes implemented in the method embodiment are not described here.

[0242] In some embodiments, as Figure 8 As shown, an embodiment of the present application also provides an electronic device 800, including a processor 801, a memory 802, and a computer program stored in the memory 802 and executable on the processor 801. When the program is executed by the processor 801, each process of the above-mentioned power generation prediction method embodiment is implemented, and the same technical effect can be achieved. To avoid repetition, it will not be described here.

[0243] It should be noted that the electronic devices in the embodiments of the present application include the mobile electronic devices and non-mobile electronic devices mentioned above.

[0244] An embodiment of the present application also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the various processes of the above-mentioned power generation prediction method embodiment are implemented, and the same technical effect can be achieved. To avoid repetition, it will not be repeated here.

[0245] The processor is the processor in the electronic device described in the above embodiment. The readable storage medium includes a computer-readable storage medium, such as a computer read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0246] An embodiment of the present application also provides a computer program product, including a computer program, which implements the above-mentioned power generation prediction method when executed by a processor.

[0247] The processor is the processor in the electronic device described in the above embodiment. The readable storage medium includes a computer-readable storage medium, such as a computer read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0248] An embodiment of the present application further provides a chip, which includes a processor and a communication interface, wherein the communication interface is coupled to the processor, and the processor is used to run programs or instructions to implement the various processes of the above-mentioned power generation prediction method embodiment, and can achieve the same technical effect. To avoid repetition, it will not be repeated here.

[0249] It should be understood that the chip mentioned in the embodiments of the present application can also be called a system-level chip, a system chip, a chip system or a system-on-chip chip, etc.

[0250] It should be noted that, in this article, the terms "comprise", "include" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, an element defined by the statement "comprises a ..." does not exclude the presence of other identical elements in the process, method, article or device comprising the element. In addition, it should be noted that the scope of the methods and devices in the embodiments of the present application is not limited to performing functions in the order shown or discussed, and may also include performing functions in a substantially simultaneous manner or in the opposite order according to the functions involved. For example, the described method may be performed in an order different from that described, and various steps may also be added, omitted, or combined. In addition, the features described with reference to certain examples may be combined in other examples.

[0251] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus the necessary general hardware platform, and of course can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art can be embodied in the form of a computer software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), including a number of instructions for enabling a terminal (which can be a mobile phone, computer, server, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0252] The embodiments of the present application are described above in conjunction with the accompanying drawings, but the present application is not limited to the above-mentioned specific implementation methods. The above-mentioned specific implementation methods are merely illustrative and not restrictive. Under the guidance of this application, ordinary technicians in this field can also make many forms without departing from the purpose of this application and the scope of protection of the claims, all of which are within the protection of this application.

[0253] Throughout this specification, reference to terms such as "one embodiment," "some embodiments," "illustrative embodiments," "examples," "specific examples," or "some examples" means that the specific features, structures, materials, or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.

[0254] Although the embodiments of the present application have been shown and described, those skilled in the art will appreciate that various changes, modifications, substitutions, and variations may be made to the embodiments without departing from the principles and purpose of the present application, and that the scope of the present application is defined by the claims and their equivalents.

Claims

1. A power generation prediction method, characterized in that: include: Obtaining first weather forecast data for each power station in the power station cluster within a first time period; determining, based on the first weather forecast data and first power generation data of each power station in the first time period, the cluster to which each power station belongs; Determining a prediction model corresponding to each cluster based on second power generation data of the power plants included in each cluster within a second time period and second meteorological forecast data of the power plants included in each cluster within a historical prediction time period; Call the prediction model corresponding to each cluster to predict the power generation data of the power stations included in each cluster; The determining, based on the first weather forecast data and the first power generation data of each power station in the first time period, the cluster to which each power station belongs includes: Extracting the irradiance characteristics in the first meteorological forecast data to obtain a first statistic corresponding to each power station; Performing feature extraction on the irradiation features in the first meteorological forecast data to obtain a first preset number of largest first feature values ​​corresponding to each power station; Using the first statistic and the first eigenvalue as derived features, generating a derived feature group corresponding to each power station; obtaining a target derived feature in the derived feature group having the highest correlation with the first power generation data; Based on the target derived characteristics and the first power generation data, obtaining statistical characteristics corresponding to each power station; Clustering the statistical features based on a Gaussian mixture clustering algorithm to determine the cluster to which each power station belongs; The determining of the prediction model corresponding to each cluster based on the second power generation data of the power plants included in each cluster within the second time period and the second meteorological forecast data of the power plants included in each cluster within the historical prediction time period includes: performing feature extraction on the irradiation features in the second weather forecast data to obtain a third statistic corresponding to the power stations included in each cluster; Performing feature extraction on the irradiation features in the second meteorological forecast data to obtain the first second preset number of second feature values ​​corresponding to the power stations included in each cluster; Based on a preset time series decomposition algorithm, obtaining a trend component and a period component of the irradiance characteristics in the second meteorological forecast data; The third statistic, the second eigenvalue, the trend component, the periodic component and the second power generation data are input into a preset prediction model for training to obtain a prediction model corresponding to each cluster.

2. The power generation prediction method according to claim 1, characterized in that: The acquiring, based on the target derived feature and the first power generation data, statistical features corresponding to each power generation site includes: Normalizing the first power generation data to obtain normalized first power generation data; Grouping the target derived features corresponding to each power station; performing statistics on the normalized first power generation data of the power stations included in each group to obtain a second statistic; Based on the second statistic, the statistical feature is obtained.

3. The power generation prediction method according to claim 1, characterized in that: The extracting the irradiance feature from the first meteorological forecast data to obtain the first statistic corresponding to each power station includes: Statistics are performed on the quantile, mean, and variance corresponding to the radiation characteristics to obtain the first statistic.

4. The power generation prediction method according to claim 1, characterized in that: The extracting the irradiation features in the first meteorological forecast data to obtain the first preset number of largest first feature values ​​corresponding to each power station includes: Based on a principal component analysis algorithm, feature extraction is performed on the radiation feature to obtain the first eigenvalue.

5. The power generation prediction method according to claim 1, characterized in that: The preset prediction models include a distributed gradient boosting model, a lightweight gradient boosting machine learning model and a deep learning model.

6. The power generation prediction method according to claim 1, characterized in that: The obtaining of first weather forecast data for each power station in the power station cluster within a first time period includes: Based on the geographical location of each power station, first weather forecast data from different sources within the first time period is obtained for each power station.

7. The power generation prediction method according to claim 1, characterized in that: The calling of the prediction model corresponding to each cluster to predict the power generation data of the power stations included in each cluster includes: According to the meteorological forecast data of the power plants included in each cluster within the target time period and the power generation data of the power plants included in each cluster within the historical time period, the prediction model corresponding to each cluster is called to predict the power generation data of the power plants included in each cluster within the target time period.

8. The power generation prediction method according to any one of claims 1 to 7, characterized in that: The power generation data includes: Power generation or electricity production.

9. A power generation prediction device, characterized in that: include: An acquisition module, configured to acquire first weather forecast data for each power station in the power station cluster within a first time period; a clustering module, configured to determine the cluster to which each power station belongs based on the first weather forecast data and the first power generation data of each power station within the first time period; a model determination module, configured to determine a prediction model corresponding to each cluster based on second power generation data of the power plants included in each cluster within a second time period and second meteorological forecast data of the power plants included in each cluster within a historical prediction time period; The prediction module is used to call the prediction model corresponding to each cluster to predict the power generation data of the power stations included in each cluster; The clustering module is further configured to extract the radiation characteristics in the first meteorological forecast data to obtain a first statistic corresponding to each power station; Performing feature extraction on the irradiation features in the first meteorological forecast data to obtain a first preset number of largest first feature values ​​corresponding to each power station; Using the first statistic and the first eigenvalue as derived features, generating a derived feature group corresponding to each power station; obtaining a target derived feature in the derived feature group having the highest correlation with the first power generation data; Based on the target derived characteristics and the first power generation data, obtaining statistical characteristics corresponding to each power station; Clustering the statistical features based on a Gaussian mixture clustering algorithm to determine the cluster to which each power station belongs; The model determination module is further configured to extract the irradiation characteristics in the second meteorological forecast data to obtain a third statistic corresponding to the power generation station included in each cluster; Performing feature extraction on the irradiation features in the second meteorological forecast data to obtain the first second preset number of second feature values ​​corresponding to the power stations included in each cluster; Based on a preset time series decomposition algorithm, obtaining a trend component and a period component of the irradiance characteristics in the second meteorological forecast data; The third statistic, the second eigenvalue, the trend component, the periodic component and the second power generation data are input into a preset prediction model for training to obtain a prediction model corresponding to each cluster.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the power generation prediction method according to any one of claims 1 to 8 is implemented.

11. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the power generation prediction method according to any one of claims 1 to 8 is implemented.

Images