Wind power plant short-term power prediction method and system based on similar days of meteorological elements

Abstract

The invention provides a wind power plant short-term power prediction method and system based on similar days of meteorological elements, and relates to the field of data processing, and the method comprises the steps: carrying out the data processing of basic information, operation data, meteorological data and landform data of a plurality of wind and light hydropower stations in a target region, carrying out the clustering of the plurality of wind and light hydropower stations, and carrying out the clustering of the plurality of wind and light hydropower stations; determining a plurality of wind power generation clusters, a plurality of photovoltaic power generation clusters and a plurality of hydroelectric power generation clusters, and establishing a plurality of candidate wind power generation prediction models, a plurality of candidate photovoltaic power generation prediction models and a plurality of candidate hydroelectric power generation prediction models; generating a plurality of candidate combination prediction models corresponding to the wind-solar hydropower station, and determining an optimal combination prediction model; obtaining weather forecast data and hydrological forecast data of the wind and light hydropower station to be predicted in the target time period; and the wind power generation power, the photovoltaic power generation power and the hydroelectric power generation power are predicted through the corresponding optimal combination prediction model. The method has the advantage of improving the accuracy of water-wind-light power generation power prediction.

Description

Technical Field

[0001] The invention relates to the field of data processing, and in particular to a method and system for predicting short-term power of a wind farm based on days with similar meteorological elements. Background Art

[0002] Vigorously developing clean energy such as water, wind, and light is a major strategic measure to ensure future energy security and respond to global warming. Wind energy and solar energy are the new energy sources with the greatest prospects for large-scale development, but they are easily affected by meteorological factors, and their output power has strong random fluctuations and is difficult to predict. With the continuous advancement of the construction of wind and light power generation bases. Large-scale direct grid connection of wind and photovoltaic power will bring great pressure to the peak, peak-shaving and stable operation of the power system. Hydropower units have the characteristics of rapid start and stop, flexible operation, large output change amplitude, and fast response to load changes. They are ideal peak-shaving power sources. Utilizing the natural complementarity of resources and the flexibility of hydropower, aggregating multiple energy sources such as water, wind, and light to form a multi-energy complementary power generation system is an effective way to reduce the impact of new energy grid connection and improve the utilization rate of basin resources.

[0003] The current power prediction methods all predict the power of hydropower, wind power and photovoltaic power generation separately, without considering the impact of water, wind and light energy. Therefore, how to comprehensively consider the impact of water, wind and light on each other and make a joint power prediction for hydropower, wind and photovoltaic power generation is an issue that needs to be considered to further improve the accuracy of hydropower, wind power and photovoltaic power prediction. Summary of the invention

[0004] The present invention provides a method and system for short-term power prediction of a wind farm based on days with similar meteorological elements, which can improve the accuracy of power prediction for hydropower, wind and solar power generation.

[0005] In order to realize the above technical features, the object of the present invention is achieved as follows: a short-term power forecasting system for a wind farm based on a day with similar meteorological elements comprises: A data processing module, used for processing basic information, operation data, meteorological data and topographic data of multiple wind-solar-hydropower stations in a target area, and generating processed basic information, operation data, meteorological data and topographic data of multiple wind-solar-hydropower stations in the target area; A site clustering module is used to cluster the multiple wind, solar and hydropower stations in the target area based on the basic information, operation data, meteorological data and topographic data of the multiple wind, solar and hydropower stations in the target area, and determine multiple wind power generation clusters, multiple photovoltaic power generation clusters and multiple hydropower generation clusters; A single modeling module is used to establish, for each of the wind power generation clusters, a plurality of candidate wind power generation prediction models corresponding to the wind power generation cluster, and is also used to establish, for each of the photovoltaic power generation clusters, a plurality of candidate photovoltaic power generation prediction models corresponding to the photovoltaic power generation cluster, and is also used to establish, for each of the hydropower generation clusters, a plurality of candidate hydropower generation prediction models corresponding to the hydropower generation cluster; A joint modeling module is used to generate, for each of the wind-solar-hydropower stations, a plurality of candidate wind power generation prediction models corresponding to the wind power generation cluster to which the wind-solar-hydropower station belongs, a plurality of candidate photovoltaic power generation prediction modules corresponding to the photovoltaic power generation cluster to which the wind-solar-hydropower station belongs, and a plurality of candidate hydropower generation prediction models corresponding to the hydropower generation cluster to which the wind-solar-hydropower station belongs, and to screen the plurality of candidate combination prediction models to determine the optimal combination prediction model corresponding to the wind-solar-hydropower station; A data acquisition module is used to obtain the meteorological forecast data and hydrological forecast data of the wind, solar and hydropower stations to be predicted in the target time period; The power generation prediction module is used to predict the wind power generation, photovoltaic power generation and hydropower generation of the wind, photovoltaic and hydropower stations to be predicted in the target time period based on the meteorological forecast data and hydrological forecast data of the wind, photovoltaic and hydropower stations to be predicted in the target time period through the optimal combined prediction model corresponding to the wind, photovoltaic and hydropower stations to be predicted.

[0006] The data processing module processes basic information, operation data, meteorological data and topographic data of multiple wind, solar and hydropower stations in the target area, including: Correcting the operation data and meteorological data of the plurality of wind-solar-hydropower stations in the target area to generate corrected operation data and meteorological data of the plurality of wind-solar-hydropower stations in the target area; Determine a wind-solar hydropower station to be supplemented among the multiple wind-solar hydropower stations in the target area, wherein the wind-solar hydropower station to be supplemented is a wind-solar hydropower station whose operating data volume does not meet a preset operating data volume requirement; For each wind-solar hydropower station to be supplemented, determine a similar wind-solar hydropower station to the wind-solar hydropower station to be supplemented from multiple wind-solar hydropower stations in the target area, and supplement the operating data of the wind-solar hydropower station to be supplemented based on the operating data and meteorological data of the similar wind-solar hydropower stations; The basic information, operation data, meteorological data and topographic data of multiple wind, solar and hydropower stations after data correction and supplementation are analyzed to determine multiple prediction and control factors.

[0007] The data processing module corrects the operation data and meteorological data of the plurality of wind, solar and hydropower stations in the target area, including: According to the wind power curve of the wind-solar-hydropower station, the wind speed in the meteorological data is corrected once: In combination with the historical maintenance records of the wind-solar-hydropower station, the abnormal wind speed segment data in the meteorological data is eliminated; Eliminate invalid meteorological data and invalid wind power generation data in the operating data according to preset screening rules; The random forest method is used to interpolate the invalid wind power data.

[0008] The data processing module performs data characteristic analysis on the basic information, operation data, meteorological data and topographic data of multiple wind, solar and hydropower stations after data correction and supplementation, and determines multiple prediction control factors, including: Based on the wind resource map, meteorological data and wind measurement data of the wind-solar-hydropower station, the characteristics of the wind resources in the region are analyzed and studied to determine the wind resource characteristics of the wind-solar-hydropower station, wherein the wind resource characteristics at least include the spatial distribution characteristics of the wind resources, the interannual variation characteristics, the intra-annual variation characteristics, the wind frequency distribution and the wind shear index; Based on the solar energy resource map and meteorological data of the wind-solar-hydropower station, determine the light resource characteristics of the wind-solar-hydropower station, wherein the light resource characteristics at least include the spatial distribution characteristics, interannual variation characteristics, intra-annual variation characteristics, solar energy resource stability and direct radiation ratio of solar energy resources; Based on the watershed inflow data of the wind-solar-hydropower station, the randomness of the monthly, seasonal and annual runoff of the wind-solar-hydropower station is determined.

[0009] The site clustering module clusters the multiple wind, solar and hydropower stations in the target area based on the processed basic information, operation data, meteorological data and topographic data of the multiple wind, solar and hydropower stations in the target area, and determines multiple wind power generation clusters, multiple photovoltaic power generation clusters and multiple hydropower generation clusters, including: Based on the processed basic information, operation data, meteorological data and topographic data of the multiple wind, solar and hydropower stations in the target area, clustering the multiple wind, solar and hydropower stations in the target area to determine multiple wind power generation clusters, multiple photovoltaic power generation clusters and multiple hydropower generation clusters; For each of the hydropower generation clusters, multiple meteorological conditions corresponding to the hydropower generation cluster are determined based on meteorological data of multiple wind, solar and hydropower stations included in the hydropower generation cluster.

[0010] The single modeling module establishes a plurality of candidate wind power generation prediction models corresponding to the wind power generation cluster, including: For each of the meteorological conditions, a plurality of wind power generation prediction models to be screened are established, and the plurality of wind power generation prediction models to be screened are screened to determine at least one candidate wind power generation prediction model.

[0011] The single modeling module establishes a plurality of candidate photovoltaic power generation prediction models corresponding to the photovoltaic power generation cluster, including: For each of the meteorological conditions, the correlation between different meteorological elements, combined light resource characteristics and photovoltaic power generation power is determined, a plurality of photovoltaic power generation prediction models to be screened are established, and the plurality of photovoltaic power generation prediction models to be screened are screened to determine at least one candidate photovoltaic power generation prediction model.

[0012] The single modeling module establishes a plurality of candidate hydropower generation prediction models corresponding to the hydropower generation cluster, including: For each of the meteorological conditions, a hydropower modeling feature pool is established, wherein the hydropower modeling feature pool at least includes predicted runoff characteristics, precipitation characteristics, cumulative precipitation characteristics, lead-lag characteristics and meteorological factor characteristics, and a plurality of hydropower prediction models to be screened are established based on the hydropower modeling feature pool, and the plurality of hydropower prediction models to be screened are screened to determine at least one candidate hydropower prediction model.

[0013] The joint modeling module screens the multiple candidate combined prediction models to determine the optimal combined prediction model corresponding to the wind-solar-hydropower station, including: For each of the candidate combination prediction models, a test sample is input into the candidate combination prediction model to obtain a hydropower prediction power sequence, a wind power prediction power sequence and a photovoltaic prediction power sequence, and a sum operation is performed on the hydropower prediction power sequence, the wind power prediction power sequence and the photovoltaic prediction power sequence to obtain a total power prediction sequence; For each of the total power prediction sequences, calculating the residual load standard deviation between the total power prediction sequence and the total power sequence of the test sample; For each of the candidate combined prediction models, generating a hydropower evaluation matrix, a wind power evaluation matrix, and a photovoltaic evaluation matrix based on the residual load standard deviation corresponding to the candidate combined prediction model; Based on the hydropower evaluation matrix, wind power evaluation matrix and photovoltaic evaluation matrix corresponding to each of the candidate combined prediction models, an optimal hydropower power prediction model, an optimal wind power prediction model and an optimal photovoltaic power prediction model are determined.

[0014] The short-term power forecasting method of wind farm based on similar meteorological elements on the day includes: Processing the basic information, operation data, meteorological data and topographic data of multiple wind-solar-hydropower stations in the target area to generate processed basic information, operation data, meteorological data and topographic data of the wind-solar-hydropower stations in the target area; Based on the basic information of the wind farm, wind operation data, meteorological data and topographic data of the target area, multiple wind, solar and hydropower stations in the target area are clustered to determine multiple wind power generation clusters, multiple photovoltaic power generation clusters and multiple hydropower generation clusters; For each of the wind power generation clusters, a plurality of candidate wind power generation prediction models corresponding to the wind power generation clusters are established; for each of the photovoltaic power generation clusters, a plurality of candidate photovoltaic power generation prediction models corresponding to the photovoltaic power generation clusters are established; and for each of the hydropower generation clusters, a plurality of candidate hydropower generation prediction models corresponding to the hydropower generation clusters are established; For each of the wind-solar-hydropower stations, based on a plurality of candidate wind power generation prediction models corresponding to the wind power generation cluster to which the wind-solar-hydropower station belongs, a plurality of candidate photovoltaic power generation prediction modules corresponding to the photovoltaic power generation cluster to which the wind-solar-hydropower station belongs, and a plurality of candidate hydropower generation prediction models corresponding to the hydropower generation cluster to which the wind-solar-hydropower station belongs, a plurality of candidate combined prediction models corresponding to the wind-solar-hydropower station are generated, the plurality of candidate combined prediction models are screened, and the optimal combined prediction model corresponding to the wind-solar-hydropower station is determined; Obtain the meteorological forecast data and hydrological forecast data of the wind, solar and hydropower stations to be predicted in the target time period; The optimal combined prediction model corresponding to the wind, solar and hydropower station to be predicted is used to predict the wind power generation power, photovoltaic power generation power and hydropower generation power of the wind, solar and hydropower station to be predicted in the target time period based on the meteorological forecast data and hydrological forecast data of the wind, solar and hydropower station to be predicted in the target time period.

[0015] Beneficial effects of the present invention: 1. The present invention processes the basic information, operation data, meteorological data and topographic data of multiple wind, solar and hydropower stations in the target area to provide more comprehensive and accurate data support for subsequent modeling. On this basis, by clustering the multiple wind, solar and hydropower stations in the target area, multiple wind power generation clusters, multiple photovoltaic power generation clusters and multiple hydropower generation clusters are determined to achieve classification of multiple wind, solar and hydropower stations, reducing the workload of subsequently establishing a single wind power generation prediction model, photovoltaic power generation prediction model and hydropower generation prediction model. In addition, based on multiple candidate wind power generation prediction models corresponding to the wind power generation cluster to which the wind, solar and hydropower stations belong, multiple candidate photovoltaic power generation prediction modules corresponding to the photovoltaic power generation cluster to which the wind, solar and hydropower stations belong, and multiple candidate hydropower generation prediction models corresponding to the hydropower generation cluster to which the wind, solar and hydropower stations belong, multiple candidate combination prediction models corresponding to the wind, solar and hydropower stations are generated, and the multiple candidate combination prediction models are screened to determine the optimal combination prediction model corresponding to the wind, solar and hydropower stations, thereby ensuring the accuracy of wind and solar power generation power prediction.

[0016] 2. Based on hydrological and meteorological element data, the present invention adopts big data and other methods to analyze the main controlling factors affecting the power forecast of various energy forms, and mines the timing characteristics between the power / meteorological information at historical moments and the current power, further improving the accuracy of water, wind and solar power generation prediction.

[0017] 3. For wind-solar hydropower stations whose operating data volume does not meet the preset operating data volume requirement, the present invention uses the real power generation data of other wind-solar hydropower stations, and conducts grey correlation analysis based on the similarity principle to find out similar wind-solar hydropower stations with similar meteorological conditions, terrain distribution, and power generation laws as the wind-solar hydropower stations to be supplemented. The missing data of the wind-solar hydropower stations to be supplemented are supplemented by the sufficient historical data of similar wind-solar hydropower stations. Grey correlation analysis and error verification are performed based on the similarity principle to infer the historical power generation data of the wind-solar hydropower stations to be supplemented, thereby improving the prediction accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0019] Figure 1 It is a module schematic diagram of a wind farm short-term power prediction system based on days with similar meteorological elements according to some embodiments of this specification.

[0020] Figure 2 It is a flow chart of a method for short-term power prediction of a wind farm based on days with similar meteorological elements according to some embodiments of this specification. DETAILED DESCRIPTION

[0021] In order to more clearly illustrate the technical solutions of the embodiments of this specification, the following is a brief introduction to the drawings required for the description of the embodiments. Obviously, the drawings described below are only some examples or embodiments of this specification. For ordinary technicians in this field, this specification can also be applied to other similar scenarios based on these drawings without creative work. Unless it is obvious from the language environment or otherwise explained, the same reference numerals in the figures represent the same structure or operation.

[0022] Figure 1 is a schematic diagram of a module of a wind farm short-term power prediction system based on similar days of meteorological elements according to some embodiments of this specification, such as Figure 1 As shown, the short-term power prediction system for wind farms based on days with similar meteorological elements may include a data processing module, a site clustering module, a single modeling module, a joint modeling module, a data acquisition module and a power generation prediction module.

[0023] The data processing module can be used to process the basic information, operation data, meteorological data and topographic data of multiple wind, solar and hydropower stations in the target area, and generate the processed basic information, operation data, meteorological data and topographic data of multiple wind, solar and hydropower stations in the target area.

[0024] The basic information, operation data, meteorological data and topographic data of multiple wind, solar and hydropower stations in the target area can be shown in Table 1.

[0025] Table 1

[0026] In some embodiments, the data processing module processes basic information, operation data, meteorological data, and topographic data of multiple wind, solar, and hydropower stations in the target area, including: Correct the operation data and meteorological data of multiple wind, solar and hydropower stations in the target area, and generate corrected operation data and meteorological data of multiple wind, solar and hydropower stations in the target area; Determine a wind-solar hydropower station to be supplemented among a plurality of wind-solar hydropower stations in a target area, wherein the wind-solar hydropower station to be supplemented is a wind-solar hydropower station whose operating data volume does not meet a preset operating data volume requirement; For each wind-solar-hydropower station to be supplemented, similar wind-solar-hydropower stations to the wind-solar-hydropower station to be supplemented are determined from multiple wind-solar-hydropower stations in the target area, and the operating data of the wind-solar-hydropower station to be supplemented is supplemented based on the operating data and meteorological data of the similar wind-solar-hydropower stations. For example, the real power generation data of the similar wind-solar-hydropower stations is used, combined with the similarity of wind, light and water, to infer the historical power generation data of the wind-solar-hydropower station to be supplemented; The basic information, operation data, meteorological data and topographic data of multiple wind, solar and hydropower stations after data correction and supplementation are analyzed to determine multiple prediction and control factors.

[0027] Specifically, the data processing module can determine the abnormal data in the meteorological observation data, wind tower data, and wind power generation data, and develop corresponding identification methods and interpolation algorithms to form high-quality data that can be used for wind power generation modeling. For the gridded short-term, short-term, and medium-term numerical forecast results and observed wind speeds (nacelle wind speed, wind tower wind speed), considering the impact of topography on wind speed, downscaling and correction techniques from grid forecasts to wind turbine point forecasts are established. The data processing module can also clarify and interpolate abnormal data for acquired satellite remote sensing data, NWP (Numerical Weather Prediction) meteorological element data, terrain data, ground meteorological station observation data, power station irradiance and power data. Due to the characteristics of photovoltaics not generating electricity at night, the irradiance and power values ​​during the non-power generation period need to be processed to zero values. The data processing module can also clarify and interpolate abnormal data for acquired hydropower power generation, runoff data, nearby meteorological station observation data, and NWP meteorological element data.

[0028] For example, the data processing module corrects the operating data and meteorological data of multiple wind, solar and hydropower stations in the target area, including: According to the wind power curve of wind, solar and hydropower stations, the wind speed in the meteorological data is corrected once: Combined with the historical maintenance records of wind and solar power stations, the abnormal wind speed data in the meteorological data was eliminated. For example, the abnormal wind speed caused by the incorrect setting of the correction coefficient of the nacelle anemometer after the main control program was upgraded was proposed; Eliminate invalid meteorological data and invalid wind power data in operation data according to preset screening rules; The random forest method is used to interpolate the invalid wind power data. For example, the wind speed interval of 0.5m / s is used to divide the power data set into different wind speed intervals, and the first quartile Q in the data set is selected. 1 and the third quartile Q 3 , press I QR =Q 3 -Q 1 The interquartile range and inner limit interval [F1, Fu] in each set were calculated, and data outside the inner limit interval were defined as abnormal data for elimination. The eliminated power data were supplemented with the random forest method.

[0029] As an example only, the preset screening rules may be as shown in Table 2.

[0030] Table 2

[0031] In some embodiments, the data processing module performs data characteristic analysis on the basic information, operation data, meteorological data and topographic data of multiple wind, solar and hydropower stations after data correction and supplementation, and determines multiple prediction control factors, including: Based on the wind resource map, meteorological data and wind measurement data of the wind-solar-hydropower station, the characteristics of the wind resources in the region are analyzed and studied to determine the wind resource characteristics of the wind-solar-hydropower station, where the wind resource characteristics at least include the spatial distribution characteristics of wind resources, interannual variation characteristics, intra-annual variation characteristics, wind frequency distribution and wind shear index; Based on the solar energy resource map and meteorological data of the wind, solar and hydropower stations, determine the light resource characteristics of the wind, solar and hydropower stations, where the light resource characteristics at least include the spatial distribution characteristics of solar energy resources, interannual variation characteristics, intra-annual variation characteristics, solar energy resource stability and direct radiation ratio; Based on the basin inflow data of wind and solar power stations, the randomness of monthly, seasonal and annual runoff of wind and solar power stations is determined.

[0032] The site clustering module can be used to cluster multiple wind, solar and hydropower stations in the target area based on the basic information, operation data, meteorological data and topographic data of multiple wind, solar and hydropower stations in the target area after processing, and determine multiple wind power generation clusters, multiple photovoltaic power generation clusters and multiple hydropower generation clusters.

[0033] In some embodiments, the site clustering module clusters multiple wind, solar and hydropower stations in the target area based on the basic information, operation data, meteorological data and topographic data of the processed multiple wind, solar and hydropower stations in the target area, and determines multiple wind power generation clusters, multiple photovoltaic power generation clusters and multiple hydropower generation clusters, including: Based on the processed basic information, operation data, meteorological data and topographic data of multiple wind, solar and hydropower stations in the target area, multiple wind, solar and hydropower stations in the target area are clustered to determine multiple wind power generation clusters, multiple photovoltaic power generation clusters and multiple hydropower generation clusters; For each hydropower generation cluster, multiple meteorological conditions corresponding to the hydropower generation cluster are determined based on meteorological data of multiple wind, solar and hydropower stations included in the hydropower generation cluster.

[0034] For example, when generating wind power generation clusters, the site clustering module can determine multiple wind power generation clusters based on the basic information of multiple wind, solar and hydropower stations in the target area after processing, the operating data related to wind power generation, meteorological data and topographic data. When generating photovoltaic power generation clusters, the site clustering module can determine multiple photovoltaic power generation clusters based on the basic information of multiple wind, solar and hydropower stations in the target area after processing, the operating data related to photovoltaic power generation, meteorological data and topographic data. When generating hydropower generation clusters, the site clustering module can determine multiple hydropower generation clusters based on the basic information of multiple hydropower stations in the target area after processing, the operating data related to hydropower generation, meteorological data and topographic data.

[0035] As an example only, firstly, the photovoltaic power and irradiance of all photovoltaic power stations in the target area are obtained as the main feature quantities and accumulated as historical data; then the power and irradiance of the historical meteorological data of each photovoltaic power station are horizontally spliced ​​as features, and the data of different photovoltaic power stations are vertically spliced, and clustering based on different sites is performed to divide the sites into k categories with higher similarity.

[0036] A single modeling module can be used to establish multiple candidate wind power prediction models corresponding to each wind power cluster, and to establish multiple candidate photovoltaic power prediction models corresponding to each photovoltaic cluster. It can also be used to establish multiple candidate hydropower prediction models corresponding to each hydropower cluster.

[0037] In some embodiments, a single modeling module establishes multiple candidate wind power generation prediction models corresponding to the wind power generation cluster, including: for each meteorological condition, multiple wind power generation prediction models to be screened are established, and the multiple wind power generation prediction models to be screened are screened to determine at least one candidate wind power generation prediction model. Specifically, the single modeling module can analyze the time variation characteristics of wind power generation under different weather types based on the cleaned meteorological data, wind tower data, wind power generation data, and topographic data, analyze the difference in wind turbine output under different topography, and combine numerical forecast data to establish a wind power generation model using physical models, nonlinear fitting, machine learning and other algorithms. Based on short-term, short-term, and medium-term numerical forecast modes (including ensemble modes) or multi-mode weather forecast results, using downscaling / correction technology and power forecast models, multiple methods are used to predict the wind power generation power for the next 4 hours, 3 days, and 7 days, respectively, and the optimal prediction model or model combination is screened based on the accessed power data and error dynamic analysis.

[0038] In some embodiments, a single modeling module establishes multiple candidate photovoltaic power generation prediction models corresponding to photovoltaic power generation clusters, including: for each meteorological condition, determining the correlation between different meteorological elements, combined light resource characteristics and photovoltaic power generation power, establishing multiple photovoltaic power generation prediction models to be screened, and screening multiple photovoltaic power generation prediction models to be screened to determine at least one candidate photovoltaic power generation prediction model. Specifically, data fusion is performed on satellite remote sensing data, NWP meteorological element data, terrain data, and ground meteorological observation data to extract key features. For example, for meteorological element data, a channel attention mechanism is introduced to assign weights to each channel to determine the importance of each channel in photovoltaic prediction. In the channel attention layer, global maximum pooling and global average pooling operations are performed on each meteorological element to obtain a feature vector of two dimensions. The two feature vectors are then passed to two fully connected layers for processing. Finally, the weighted meteorological element features are obtained. For satellite data with fewer channels, a convolutional network is used to perform a dimensionality increase operation to expand it to the same dimension as the meteorological element channel. Then, the two sets of feature vectors are fused using a cross attention layer. The cross attention mechanism can learn the correlation between different features and further extract more informative feature representations. Finally, the fused feature vector is input into the LSTM network, which uses the fused features for time series modeling and generates solar radiation prediction results. Based on the solar radiation prediction, it is necessary to compare the correlation between different meteorological elements, combined light resource characteristics and photovoltaic power generation, so as to construct the characteristics of the input model and establish multiple photovoltaic power generation prediction models to be screened. Among them, the correlation between the combined light resource characteristics and photovoltaic power generation can be measured by the correlation coefficient.

[0039] In some embodiments, a single modeling module establishes multiple candidate hydropower prediction models corresponding to the hydropower cluster, including: for each meteorological condition, a hydropower modeling feature pool is established, wherein the hydropower modeling feature pool includes at least the forecast runoff feature, precipitation feature, cumulative precipitation feature, lead-lag feature and meteorological factor feature, and multiple hydropower prediction models to be screened are established based on the hydropower modeling feature pool, and the multiple hydropower prediction models to be screened are screened to determine at least one candidate hydropower prediction model. Specifically, in addition to taking the forecast runoff, precipitation and other related meteorological factors in NWP as features, the power generation corresponding to similar days in the historical data set will be found as a new feature to be added to the model. The determination of similar days is based on the similarity of the forecast meteorological data to the historical days. Specifically, the similarity of the time series of runoff, precipitation, maximum temperature, minimum temperature, air pressure, humidity, etc. is used as the judgment standard. Then, the power generation time series corresponding to the similar days is added to the feature pool as a feature. In addition, considering that there is a certain lag between the power generation of the hydropower station and precipitation, that is, the power generation capacity of the hydropower station is the cumulative effect of precipitation, when constructing the features, it is necessary to take into account the cumulative precipitation over a period of time, as well as the lead and lag terms, and then screen the features.

[0040] As an example only, the candidate wind power prediction model, the candidate photovoltaic power prediction model and the candidate hydropower prediction model can all be BP (Back Propagation) neural network, LSTM (Long Short-Term Memory) neural network, bidirectional LSTM (BiLSTM) neural network, GRU (Gate Recurrent Unit) neural network, bidirectional GRU (BiGRU) neural network, CNN (Convolutional Neural Networks) neural network, extreme learning machine (ELM), etc.

[0041] The joint modeling module can be used for each wind, solar and hydropower station to generate multiple candidate combination prediction models corresponding to the wind power generation cluster to which the wind, solar and hydropower station belongs, based on multiple candidate wind power generation prediction models corresponding to the wind power generation cluster to which the wind, solar and hydropower station belongs, multiple candidate photovoltaic power generation prediction modules corresponding to the photovoltaic power generation cluster to which the wind, solar and hydropower station belongs, and multiple candidate hydropower generation prediction models corresponding to the hydropower generation cluster to which the wind, solar and hydropower station belongs, and to screen the multiple candidate combination prediction models to determine the optimal combination prediction model corresponding to the wind, solar and hydropower station.

[0042] In some embodiments, the joint modeling module screens multiple candidate combined prediction models to determine the optimal combined prediction model corresponding to the wind, solar and hydropower stations, including: For each candidate combination prediction model, the test sample is input into the candidate combination prediction model to obtain a hydropower prediction power sequence, a wind power prediction power sequence and a photovoltaic prediction power sequence, and the hydropower prediction power sequence, the wind power prediction power sequence and the photovoltaic prediction power sequence are summed to obtain a total power prediction sequence; For each total power prediction sequence, calculate the residual load standard deviation between the total power prediction sequence and the total power sequence of the test sample; For each candidate combination prediction model, a hydropower evaluation matrix, a wind power evaluation matrix, and a photovoltaic evaluation matrix are generated based on the residual load standard deviation corresponding to the candidate combination prediction model; Based on the hydropower evaluation matrix, wind power evaluation matrix and photovoltaic evaluation matrix corresponding to each candidate combination prediction model, the optimal hydropower power prediction model, the optimal wind power prediction model and the optimal photovoltaic power prediction model are determined.

[0043] Specifically, the joint modeling module can create corresponding evaluation matrices from the perspectives of hydropower, wind power and photovoltaic power using the standard deviations of each residual load, and make fuzzy optimization decisions based on the created hydropower evaluation matrix, wind power evaluation matrix and photovoltaic evaluation matrix to determine the optimal hydropower power prediction model, optimal wind power prediction model and optimal photovoltaic power prediction model with the largest relative superiority. It can be understood that different wind, solar and hydropower stations correspond to an optimal combination prediction model under each corresponding meteorological condition.

[0044] The data acquisition module can be used to obtain the meteorological forecast data and hydrological forecast data of the wind, solar and hydropower stations to be predicted in the target time period.

[0045] The wind-solar-hydropower station to be predicted may be a wind-solar-hydropower station in the target area.

[0046] The power generation prediction module can be used to predict the wind power generation, photovoltaic power generation and hydropower generation of the wind, solar and hydropower stations to be predicted in the target time period based on the meteorological forecast data and hydrological forecast data of the wind, solar and hydropower stations to be predicted in the target time period through the optimal combination prediction model corresponding to the wind, solar and hydropower stations to be predicted.

[0047] Specifically, based on the meteorological conditions in the target time period, the optimal combination prediction model that matches the meteorological conditions in the target time period can be determined from multiple optimal combination prediction models corresponding to the wind, solar and hydropower stations to be predicted. Then, the optimal combination prediction model corresponding to the wind, solar and hydropower stations to be predicted is input into the optimal combination prediction model based on the meteorological forecast data and hydrological forecast data of the wind, solar and hydropower stations to be predicted in the target time period to obtain the wind power generation power, photovoltaic power generation power and hydropower generation power of the wind, solar and hydropower stations to be predicted in the target time period.

[0048] Figure 2is a flow chart of a method for short-term wind farm power prediction based on similar days of meteorological elements according to some embodiments of this specification, such as Figure 2 As shown, the method for short-term power prediction of a wind farm based on days with similar meteorological elements may include the following steps.

[0049] Step 210, performing data processing on basic information, operation data, meteorological data, and topographic data of multiple wind-solar-hydropower stations in the target area to generate processed basic information, operation data, meteorological data, and topographic data of multiple wind-solar-hydropower stations in the target area; Step 220, based on the basic information of the wind farms, wind power operation data, meteorological data and topographic data of the target area, clustering multiple wind, solar and hydropower stations in the target area to determine multiple wind power generation clusters, multiple photovoltaic power generation clusters and multiple hydropower generation clusters; Step 230, for each wind power generation cluster, establish multiple candidate wind power generation prediction models corresponding to the wind power generation cluster, for each photovoltaic power generation cluster, establish multiple candidate photovoltaic power generation prediction models corresponding to the photovoltaic power generation cluster, and for each hydropower generation cluster, establish multiple candidate hydropower generation prediction models corresponding to the hydropower generation cluster; Step 240, for each wind-solar-hydropower station, based on multiple candidate wind power prediction models corresponding to the wind power generation cluster to which the wind-solar-hydropower station belongs, multiple candidate photovoltaic power generation prediction modules corresponding to the photovoltaic power generation cluster to which the wind-solar-hydropower station belongs, and multiple candidate hydropower generation prediction models corresponding to the hydropower generation cluster to which the wind-solar-hydropower station belongs, generate multiple candidate combination prediction models corresponding to the wind-solar-hydropower station, screen the multiple candidate combination prediction models, and determine the optimal combination prediction model corresponding to the wind-solar-hydropower station; Step 250, obtaining meteorological forecast data and hydrological forecast data of the wind, solar and hydropower stations to be predicted in the target time period; Step 260, predict the wind power generation, photovoltaic power generation and hydropower generation of the wind, photovoltaic and hydropower stations to be predicted in the target time period based on the meteorological forecast data and hydrological forecast data of the wind, photovoltaic and hydropower stations to be predicted in the target time period through the optimal combined prediction model corresponding to the wind, photovoltaic and hydropower stations to be predicted.

[0050] The short-term power prediction method for wind farms based on days with similar meteorological elements can be executed by a short-term power prediction system for wind farms based on days with similar meteorological elements. For more descriptions of the short-term power prediction method for wind farms based on days with similar meteorological elements, please refer to the relevant descriptions of the short-term power prediction system for wind farms based on days with similar meteorological elements, which will not be repeated here.

[0051] Finally, it should be understood that the embodiments described in this specification are only used to illustrate the principles of the embodiments of this specification. Other variations may also fall within the scope of this specification. Therefore, as an example and not a limitation, alternative configurations of the embodiments of this specification may be considered consistent with the teachings of this specification. Accordingly, the embodiments of this specification are not limited to the embodiments explicitly introduced and described in this specification.

Claims

1. A short-term wind farm power forecasting system based on similar days of meteorological elements, characterized by: include: A data processing module, used for processing basic information, operation data, meteorological data and topographic data of multiple wind-solar-hydropower stations in a target area, and generating processed basic information, operation data, meteorological data and topographic data of multiple wind-solar-hydropower stations in the target area; A site clustering module is used to cluster the multiple wind, solar and hydropower stations in the target area based on the basic information, operation data, meteorological data and topographic data of the multiple wind, solar and hydropower stations in the target area, and determine multiple wind power generation clusters, multiple photovoltaic power generation clusters and multiple hydropower generation clusters; A single modeling module is used to establish, for each of the wind power generation clusters, a plurality of candidate wind power generation prediction models corresponding to the wind power generation cluster, and is also used to establish, for each of the photovoltaic power generation clusters, a plurality of candidate photovoltaic power generation prediction models corresponding to the photovoltaic power generation cluster, and is also used to establish, for each of the hydropower generation clusters, a plurality of candidate hydropower generation prediction models corresponding to the hydropower generation cluster; A joint modeling module is used to generate, for each of the wind-solar-hydropower stations, a plurality of candidate wind power generation prediction models corresponding to the wind power generation cluster to which the wind-solar-hydropower station belongs, a plurality of candidate photovoltaic power generation prediction modules corresponding to the photovoltaic power generation cluster to which the wind-solar-hydropower station belongs, and a plurality of candidate hydropower generation prediction models corresponding to the hydropower generation cluster to which the wind-solar-hydropower station belongs, and to screen the plurality of candidate combination prediction models to determine the optimal combination prediction model corresponding to the wind-solar-hydropower station; A data acquisition module is used to obtain the meteorological forecast data and hydrological forecast data of the wind, solar and hydropower stations to be predicted in the target time period; The power generation prediction module is used to predict the wind power generation, photovoltaic power generation and hydropower generation of the wind, photovoltaic and hydropower stations to be predicted in the target time period based on the meteorological forecast data and hydrological forecast data of the wind, photovoltaic and hydropower stations to be predicted in the target time period through the optimal combined prediction model corresponding to the wind, photovoltaic and hydropower stations to be predicted.

2. The short-term power forecasting system for wind farms based on similar days of meteorological elements according to claim 1 is characterized in that: The data processing module processes basic information, operation data, meteorological data and topographic data of multiple wind, solar and hydropower stations in the target area, including: Correcting the operation data and meteorological data of the plurality of wind-solar-hydropower stations in the target area to generate corrected operation data and meteorological data of the plurality of wind-solar-hydropower stations in the target area; Determine a wind-solar hydropower station to be supplemented among the multiple wind-solar hydropower stations in the target area, wherein the wind-solar hydropower station to be supplemented is a wind-solar hydropower station whose operating data volume does not meet a preset operating data volume requirement; For each wind-solar hydropower station to be supplemented, determine a similar wind-solar hydropower station to the wind-solar hydropower station to be supplemented from multiple wind-solar hydropower stations in the target area, and supplement the operating data of the wind-solar hydropower station to be supplemented based on the operating data and meteorological data of the similar wind-solar hydropower stations; The basic information, operation data, meteorological data and topographic data of multiple wind, solar and hydropower stations after data correction and supplementation are analyzed to determine multiple prediction and control factors.

3. The short-term power forecasting system for wind farms based on similar days of meteorological elements according to claim 2 is characterized in that: The data processing module corrects the operation data and meteorological data of the plurality of wind, solar and hydropower stations in the target area, including: According to the wind power curve of the wind-solar-hydropower station, the wind speed in the meteorological data is corrected once: In combination with the historical maintenance records of the wind-solar-hydropower station, the abnormal wind speed segment data in the meteorological data is eliminated; Eliminate invalid meteorological data and invalid wind power generation data in the operating data according to preset screening rules; The random forest method is used to interpolate the invalid wind power data.

4. The short-term power forecasting system for wind farms based on similar days of meteorological elements according to claim 2 is characterized in that: The data processing module performs data characteristic analysis on the basic information, operation data, meteorological data and topographic data of multiple wind, solar and hydropower stations after data correction and supplementation, and determines multiple prediction control factors, including: Based on the wind resource map, meteorological data and wind measurement data of the wind-solar-hydropower station, the characteristics of the wind resources in the region are analyzed and studied to determine the wind resource characteristics of the wind-solar-hydropower station, wherein the wind resource characteristics at least include the spatial distribution characteristics of the wind resources, the interannual variation characteristics, the intra-annual variation characteristics, the wind frequency distribution and the wind shear index; Based on the solar energy resource map and meteorological data of the wind-solar-hydropower station, determine the light resource characteristics of the wind-solar-hydropower station, wherein the light resource characteristics at least include the spatial distribution characteristics, interannual variation characteristics, intra-annual variation characteristics, solar energy resource stability and direct radiation ratio of solar energy resources; Based on the watershed inflow data of the wind-solar-hydropower station, the randomness of the monthly, seasonal and annual runoff of the wind-solar-hydropower station is determined.

5. The wind farm short-term power prediction system based on similar days of meteorological elements according to any one of claims 1 to 4, characterized in that: The site clustering module clusters the multiple wind, solar and hydropower stations in the target area based on the processed basic information, operation data, meteorological data and topographic data of the multiple wind, solar and hydropower stations in the target area, and determines multiple wind power generation clusters, multiple photovoltaic power generation clusters and multiple hydropower generation clusters, including: Based on the processed basic information, operation data, meteorological data and topographic data of the multiple wind, solar and hydropower stations in the target area, clustering the multiple wind, solar and hydropower stations in the target area to determine multiple wind power generation clusters, multiple photovoltaic power generation clusters and multiple hydropower generation clusters; For each of the hydropower generation clusters, multiple meteorological conditions corresponding to the hydropower generation cluster are determined based on meteorological data of multiple wind, solar and hydropower stations included in the hydropower generation cluster.

6. The short-term power forecasting system for wind farms based on similar days of meteorological elements according to claim 5 is characterized in that: The single modeling module establishes a plurality of candidate wind power generation prediction models corresponding to the wind power generation cluster, including: For each of the meteorological conditions, a plurality of wind power generation prediction models to be screened are established, and the plurality of wind power generation prediction models to be screened are screened to determine at least one candidate wind power generation prediction model.

7. The short-term power forecasting system for wind farms based on similar days of meteorological elements according to claim 5 is characterized in that: The single modeling module establishes a plurality of candidate photovoltaic power generation prediction models corresponding to the photovoltaic power generation cluster, including: For each of the meteorological conditions, the correlation between different meteorological elements, combined light resource characteristics and photovoltaic power generation power is determined, a plurality of photovoltaic power generation prediction models to be screened are established, and the plurality of photovoltaic power generation prediction models to be screened are screened to determine at least one candidate photovoltaic power generation prediction model.

8. The short-term power forecasting system for wind farms based on similar days of meteorological elements according to claim 5 is characterized in that: The single modeling module establishes a plurality of candidate hydropower generation prediction models corresponding to the hydropower generation cluster, including: For each of the meteorological conditions, a hydropower modeling feature pool is established, wherein the hydropower modeling feature pool at least includes predicted runoff characteristics, precipitation characteristics, cumulative precipitation characteristics, lead-lag characteristics and meteorological factor characteristics, and a plurality of hydropower prediction models to be screened are established based on the hydropower modeling feature pool, and the plurality of hydropower prediction models to be screened are screened to determine at least one candidate hydropower prediction model.

9. The short-term power forecasting system for wind farms based on similar days of meteorological elements according to claim 5 is characterized in that: The joint modeling module screens the multiple candidate combined prediction models to determine the optimal combined prediction model corresponding to the wind-solar-hydropower station, including: For each of the candidate combination prediction models, a test sample is input into the candidate combination prediction model to obtain a hydropower prediction power sequence, a wind power prediction power sequence and a photovoltaic prediction power sequence, and a sum operation is performed on the hydropower prediction power sequence, the wind power prediction power sequence and the photovoltaic prediction power sequence to obtain a total power prediction sequence; For each of the total power prediction sequences, calculating the residual load standard deviation between the total power prediction sequence and the total power sequence of the test sample; For each of the candidate combined prediction models, generating a hydropower evaluation matrix, a wind power evaluation matrix, and a photovoltaic evaluation matrix based on the residual load standard deviation corresponding to the candidate combined prediction model; Based on the hydropower evaluation matrix, wind power evaluation matrix and photovoltaic evaluation matrix corresponding to each of the candidate combined prediction models, an optimal hydropower power prediction model, an optimal wind power prediction model and an optimal photovoltaic power prediction model are determined.

10. A method for short-term wind farm power forecasting based on days with similar meteorological elements, characterized in that: include: Processing the basic information, operation data, meteorological data and topographic data of multiple wind-solar-hydropower stations in the target area to generate processed basic information, operation data, meteorological data and topographic data of the wind-solar-hydropower stations in the target area; Based on the basic information of the wind farm, wind operation data, meteorological data and topographic data of the target area, multiple wind, solar and hydropower stations in the target area are clustered to determine multiple wind power generation clusters, multiple photovoltaic power generation clusters and multiple hydropower generation clusters; For each of the wind power generation clusters, a plurality of candidate wind power generation prediction models corresponding to the wind power generation clusters are established; for each of the photovoltaic power generation clusters, a plurality of candidate photovoltaic power generation prediction models corresponding to the photovoltaic power generation clusters are established; and for each of the hydropower generation clusters, a plurality of candidate hydropower generation prediction models corresponding to the hydropower generation clusters are established; For each of the wind-solar-hydropower stations, based on a plurality of candidate wind power generation prediction models corresponding to the wind power generation cluster to which the wind-solar-hydropower station belongs, a plurality of candidate photovoltaic power generation prediction modules corresponding to the photovoltaic power generation cluster to which the wind-solar-hydropower station belongs, and a plurality of candidate hydropower generation prediction models corresponding to the hydropower generation cluster to which the wind-solar-hydropower station belongs, a plurality of candidate combined prediction models corresponding to the wind-solar-hydropower station are generated, the plurality of candidate combined prediction models are screened, and the optimal combined prediction model corresponding to the wind-solar-hydropower station is determined; Obtain the meteorological forecast data and hydrological forecast data of the wind, solar and hydropower stations to be predicted in the target time period; The optimal combined prediction model corresponding to the wind, solar and hydropower station to be predicted is used to predict the wind power generation power, photovoltaic power generation power and hydropower generation power of the wind, solar and hydropower station to be predicted in the target time period based on the meteorological forecast data and hydrological forecast data of the wind, solar and hydropower station to be predicted in the target time period.

Images