Wind farm cluster short-term power prediction method and system

Abstract

The application discloses a kind of based on wind farm cluster short-term power prediction method and system, method includes: the similarity distance matrix of the time-varying law of different station key meteorological elements is constructed;The Copula entropy between each station historical power sample is calculated, and the spatial similarity matrix of station power is constructed according to each Copula entropy;Similarity distance matrix and spatial similarity matrix are normalized, and the similarity distance matrix and spatial similarity matrix after normalization are summed, to obtain the space-time similarity distance matrix;Similarity distance matrix is clustered using hierarchical clustering algorithm, and the zoning scheme is determined in combination with the geographical distribution of new energy station, each wind farm in target area is divided into each region;According to the total power prediction result of the region obtained by the output of LSTM time series prediction model. Improve the overall accuracy and robustness of power prediction, provide more reliable scheduling basis for large-scale access of new energy station to power grid.

Description

Technical Field

[0001] The present invention belongs to the technical field of wind energy generation and comprehensive consumption, and in particular relates to a short-term power prediction method and system based on a wind farm cluster. Background Art

[0002] The volatility and uncertainty of wind power are key factors affecting grid dispatch and operational stability. During wind farm power forecasting, wind power output exhibits a high degree of spatiotemporal correlation, as it is influenced by complex meteorological factors, the geographical environment, and the operational characteristics of the grid. Traditional wind power forecasting methods often use individual farms as units, ignoring the collaborative relationships between farms and failing to fully exploit the similarities in power variations exhibited by wind farms within a region driven by similar meteorological backgrounds. Furthermore, existing cluster forecasting methods often suffer from crude regional divisions and overreliance on geographic information or subjective experience, making it difficult to ensure consistent characteristics across farms within a given zone. Summary of the Invention

[0003] The present invention provides a short-term power prediction method and system based on a wind farm cluster, which is used to solve the technical problem of failing to fully exploit the similarity of power changes exhibited by wind farms in a region driven by similar meteorological backgrounds.

[0004] In a first aspect, the present invention provides a method for short-term power prediction based on a wind farm cluster, comprising:

[0005] Get target area wind farms 5-minute power data for each historical sample day , Measured wind speed data at the wind turbine hub and predicted wind speed data at the turbine hub ,in, is a three-dimensional real space, represents the field of real numbers, express sample day, 288 represents the number of 5-minute samples in a day, represents the number of wind farms, For the The power sequence of a wind farm station, For the The measured wind speed series of a wind farm station, For the The predicted wind speed series of each wind farm station;

[0006] Calculate the ACF curve of wind speed at the hub of the wind turbine, calculate the weighted Euclidean distance between the ACF curves of different stations, and construct the similarity distance matrix of the time-varying laws of key meteorological elements between different stations ,in, is a two-dimensional real number space, Indicates the number of wind farms;

[0007] Calculate the Copula entropy between the historical power samples of each station, and construct the spatial similarity matrix of the station power based on each Copula entropy ;

[0008] The similarity distance matrix and spatial similarity matrix Perform normalization and normalize the similarity distance matrix and spatial similarity matrix Sum and get the spatiotemporal similarity distance matrix ;

[0009] The hierarchical clustering algorithm is used to classify the spatiotemporal similarity distance matrix Clustering is performed, and the zoning scheme is determined based on the geographical distribution of new energy stations. A wind farm is divided into K areas;

[0010] An LSTM time series prediction model is established for each region. The power sequence in a certain region, the measured wind speed sequence at the wind turbine hub of a representative station in a certain region, and the predicted wind speed sequence at the wind turbine hub of a representative station in a certain region are input into a certain LSTM time series prediction model. The LSTM time series prediction model outputs the total power prediction result of a certain region, wherein the representative station in a certain region is selected in a certain region based on the maximum correlation-minimum redundancy principle.

[0011] In a second aspect, the present invention provides a short-term power prediction system based on a wind farm cluster, comprising:

[0012] Acquisition module, configured to obtain the target area wind farms 5-minute power data for each historical sample day , Measured wind speed data at the wind turbine hub and predicted wind speed data at the turbine hub ,in, is a three-dimensional real space, represents the field of real numbers, express sample day, 288 represents the number of 5-minute samples in a day, represents the number of wind farms, For the The power sequence of a wind farm station, For the The measured wind speed series of a wind farm station, For the The predicted wind speed series of each wind farm station;

[0013] The calculation module is configured to calculate the ACF curve of the wind speed at the hub of the wind turbine, calculate the weighted Euclidean distance between the ACF curves of different stations, and construct a similarity distance matrix of the time-varying laws of key meteorological elements between different stations ,in, is a two-dimensional real number space, Indicates the number of wind farms;

[0014] A construction module is configured to calculate the Copula entropy between the historical power samples of each station and construct the spatial similarity matrix of the station power based on each Copula entropy ;

[0015] The summation module is configured to sum the similarity distance matrix and spatial similarity matrix Perform normalization and normalize the similarity distance matrix and spatial similarity matrix Sum and get the spatiotemporal similarity distance matrix ;

[0016] A partitioning module is configured to use a hierarchical clustering algorithm to partition the spatiotemporal similarity distance matrix Clustering is performed, and the zoning scheme is determined based on the geographical distribution of new energy stations. A wind farm is divided into K areas;

[0017] The output module is configured to establish an LSTM time series prediction model for each area, and input the power sequence in a certain area, the measured wind speed sequence at the wind turbine hub of a representative station in a certain area, and the predicted wind speed sequence at the wind turbine hub of a representative station in a certain area into a certain LSTM time series prediction model. The output of the certain LSTM time series prediction model obtains the total power prediction result of a certain area, wherein the representative station in a certain area is selected in a certain area based on the maximum correlation-minimum redundancy principle.

[0018] According to a third aspect, an electronic device is provided, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can perform the steps of the short-term power prediction method based on a wind farm cluster according to any embodiment of the present invention.

[0019] In a fourth aspect, the present invention further provides a computer-readable storage medium having a computer program stored thereon, wherein when the program instructions are executed by a processor, the processor executes the steps of the method for short-term power prediction based on a wind farm cluster according to any embodiment of the present invention.

[0020] The present application discloses a short-term power prediction method and system based on wind farm clusters. The method and system utilize the spatiotemporal coupling characteristics between wind speed and power to cluster and partition wind farms, and predict the total power of each sub-region. By optimizing the site partitioning strategy and selecting representative sites for cluster prediction modeling, the method can effectively reduce random errors, improve the overall accuracy and robustness of power prediction, and provide a more reliable scheduling basis for large-scale access of new energy sites to the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following is a brief introduction to the drawings required for use in the description of the embodiments. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0022] Figure 1 A flow chart of a method for short-term power prediction based on a wind farm cluster provided by one embodiment of the present invention;

[0023] Figure 2 A flow chart of wind farm station partition aggregation based on spatiotemporal correlation is provided for an embodiment of the present invention;

[0024] Figure 3 A flowchart of selecting a regional representative station based on maximum correlation and minimum redundancy according to a specific embodiment of an embodiment of the present invention;

[0025] Figure 4 A flowchart of an LSTM prediction model according to a specific embodiment of the present invention is provided;

[0026] Figure 5 A flowchart of LSTM model training and prediction according to a specific embodiment of the present invention is provided;

[0027] Figure 6 A geographical distribution map of wind farms according to a specific embodiment of the present invention;

[0028] Figure 7 A schematic diagram of the zoning results of 68 wind farms and the selection results of regional representative farms in a specific embodiment of an embodiment of the present invention;

[0029] Figure 8 Each partition of a specific embodiment provided by an embodiment of the present invention Histogram of prediction evaluation indicators;

[0030] Figure 9 The models of a specific embodiment provided by an embodiment of the present invention 、 、 Evaluation index bar chart;

[0031] Figure 10 A power prediction curve diagram of a comparative model of a specific embodiment provided by an embodiment of the present invention;

[0032] Figure 11 A structural block diagram of a short-term power prediction system based on a wind farm cluster provided by an embodiment of the present invention;

[0033] Figure 12 It is a structural diagram of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0034] To make the objectives, technical solutions, and advantages of the embodiments of the present invention more clear, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0035] See also Figure 1 , which shows a flow chart of a short-term power prediction method based on a wind farm cluster in the present application.

[0036] like Figure 1 As shown, the short-term power prediction method based on wind farm cluster specifically includes the following steps:

[0037] Step S101: Get the target area wind farms 5-minute power data for each historical sample day , measured data of wind speed at the wind turbine hub and wind speed prediction data at the turbine hub .

[0038] In this step, For the The power sequence of a wind farm station, For the The measured wind speed series of the wind farm station, For the The predicted wind speed series of each wind farm station, is a three-dimensional real space, represents the field of real numbers, express sample day, 288 represents the number of 5-minute samples in a day, Indicates the number of wind farms;

[0039] Step S102: Calculate the ACF curve of the wind speed at the hub of the wind turbine, calculate the weighted Euclidean distance between the ACF curves of different stations, and construct a similarity distance matrix of the time-varying laws of key meteorological elements between different stations. .

[0040] In this step, is a two-dimensional real number space, Represents the number of wind farms; for the time-varying laws of key meteorological elements, the similarity between the ACF curves of key meteorological elements of each station is calculated. First, the autocorrelation function (ACF) curve of the wind speed at the hub of the wind turbine is calculated. Then, considering that the correlation coefficient on the ACF curve increases with the increase of lag time, the difference between different stations gradually increases. Therefore, combined with the characteristic of the influence weight of time series correlation that "near is large and far is small", the weighted Euclidean distance between the ACF curves of key meteorological elements of the stations is calculated, and finally a similarity distance matrix of the time-varying laws of key meteorological elements between different stations is formed. According to Figure 2 Similarity distance matrix The calculation process is explained:

[0041] Calculate the wind speed at the hub of each wind turbine at each wind farm The ACF curve of the i-th station at the lag time t is calculated. The expression is:

[0042] ,

[0043] Where, is the length of the wind speed sequence, is the mean of the measured wind speed sequence of the i-th wind farm station, , is the ACF value of the i-th wind farm at lag time t, is the measured wind speed sequence of the i-th wind farm at time point The wind speed value, is the measured wind speed sequence of the i-th wind farm at time point Wind speed value;

[0044] Construct the ACF matrix, the expression is:

[0045] ,

[0046] Where, is the ACF matrix, is the ACF value of the mth wind farm at lag time n;

[0047] The autocorrelation value of each lag point of the ACF curve represents the correlation of the wind speed data at that time lag. Time points with larger lags indicate weaker correlations, while time points with smaller lags have stronger correlations.

[0048] When multiple wind farms are involved, the ACF curve of a single wind farm may fluctuate greatly, making it difficult to draw a stable conclusion. Therefore, the average ACF curve is used to integrate the autocorrelated information of multiple data sets to obtain a more reliable weight. The average ACF value of all wind farms at lag time t is calculated. , and calculate the Euclidean distance weight of each lag time point based on the average ACF value, where the expression for calculating the average ACF value is:

[0049] ,

[0050] Where, is the ACF value of wind farm k at lag time t, is the number of wind farms;

[0051] The expression for calculating the Euclidean distance weight is:

[0052] ,

[0053] According to the Euclidean distance weight, the weighted Euclidean distance matrix between the ACF curves of each station is calculated, that is, the similarity distance matrix of the time-varying laws of key meteorological elements between different stations is constructed. , the expression is:

[0054] ,

[0055] Where, is the weighted Euclidean distance between the m-th wind farm and the m-th wind farm, and its value is 0. is the weighted Euclidean distance between wind farm h and wind farm j;

[0056] ,

[0057] Where, 、 are the autocorrelation coefficients of wind farm h and wind farm j at lag time t, is the maximum lag time of ACF. The smaller the value, the higher the temporal similarity between the wind speeds at the hubs of the wind turbines at the two stations.

[0058] Step S103: Calculate the Copula entropy between the historical power samples of each station, and construct the spatial similarity matrix of the station power based on each Copula entropy. .

[0059] In this step, according to the randomness of historical power, the output power sample value of each station is calculated. The nonlinear spatial correlation between the power data of different stations is reflected by calculating the Copula entropy between the historical power samples of each station. The smaller the Copula entropy, the stronger the correlation. Finally, the spatial correlation matrix of the station power is formed. .according to Figure 2 Similarity distance matrix The calculation process is explained:

[0060] Calculate the power time series samples of wind farms h and j and Copula entropy between , where T is the number of samples, and the calculation process is:

[0061] ,

[0062] ,

[0063] Where, is the empirical marginal distribution function of the power time series sample of the h-th wind farm, is the number of power time series samples, is an indicator function, if When it is established, the function value is 1. If When it is not true, the function value is 0. is the sth sample in the power time series of the hth wind farm, is the sth sample in the power time series of the jth wind farm, is the power time series of the wind farm;

[0064] The wind farm and wind farms The power time series samples of are mapped to marginal probabilities through the empirical distribution function, and the expression is:

[0065] ,

[0066] ,

[0067] Where, For wind farms h The tth sample in the power sequence of For wind farmsj The tth sample of the power sequence, For wind farms j The empirical marginal distribution function of the power sequence is, For wind farms h The empirical marginal distribution function of the power sequence is, For wind farms h The marginal probability of the t-th sample in the power sequence of For wind farms j The marginal probability of the t-th sample in the power sequence of ;

[0068] Constructing joint sample pairs based on marginal probabilities , and calculate the distance between sample pairs based on Chebyshev distance, the expression is:

[0069] , ,

[0070] Where, is the sth sample pair and the vth sample pair The Chebyshev distance between , For sample pairs The marginal probability of the sth sample in the wind farm h is, For sample pairs The marginal probability of the vth sample in the wind farm h is, For sample pairs The marginal probability of the sth sample in wind farm j is, For sample pairs The marginal probability of the vth sample in wind farm j is, is the sth power sample, is the vth power sample;

[0071] For each sample pair, find the kth nearest neighbor distance of each sample pair, where k is the number of neighbors. In this method, k=3 is set to take into account both the bias and variance of the estimation. The Copula entropy between the historical power samples of each station is estimated based on the k nearest neighbor distances. The expression is:

[0072] ,

[0073] Where, is the volume of the two-dimensional unit sphere, under the Chebyshev distance =1, is the Chebyshev distance of the kth nearest neighbor of sample pair s, is the Copula entropy between station h and station j, is the Digamma function;

[0074] Depend on The spatial correlation matrix representing the nonlinearity between the two stations is constructed, that is, the spatial similarity matrix of the station power is constructed. , the expression is:

[0075] ,

[0076] Where, is the Copula entropy between station h and station j, is the Copula entropy between station m and station m. The smaller the value, the higher the nonlinear similarity between the two stations.

[0077] Step S104: similarity distance matrix and spatial similarity matrix Perform normalization and normalize the similarity distance matrix and spatial similarity matrix Sum and get the spatiotemporal similarity distance matrix .

[0078] In the step, the maximum and minimum normalization method is used, and the matrix elements before and after normalization are set as and The maximum and minimum values ​​of the matrix elements are 、 , the specific normalization formula is: Based on the obtained normalized matrix, and Sum to form the spatiotemporal similarity distance matrix , . The elements Represents the similarity distance between the i-th and j-th stations. The smaller the value, the higher the similarity between the two stations.

[0079] Spatiotemporal similarity distance matrix The expression is:

[0080] ,

[0081] Where, is the normalized Copula entropy matrix, is the normalized weighted Euclidean distance matrix between the ACF curves of each station, is the spatiotemporal similarity distance between stations h and j, is the spatiotemporal similarity distance between the m-th wind farm and the m-th wind farm.

[0082] Step S105: using a hierarchical clustering algorithm to cluster the similarity distance matrix Clustering is performed, and the zoning scheme is determined based on the geographical distribution of new energy stations. A wind farm is divided into K areas.

[0083] In this step, the spatiotemporal similarity distance matrix obtained in step S104 is input Initially, each wind farm station is treated as a separate cluster, that is, there are m clusters in total. .

[0084] The initial distance between the two clusters The clustering process uses Ward linkage to minimize the total square error of the merged clusters. and clusters are any two clusters to be merged, |C| is the number of wind farms in the cluster to be merged, is the spatiotemporal similarity distance between wind farms h and j, then and The Ward distance calculation formula between them is:

[0085] ,

[0086] Find the two clusters with the smallest d distance and ,merge and For new clusters . Then recalculate the distance between the new cluster and other clusters:

[0087] ,

[0088] Repeatedly calculate the distance between the new cluster and other clusters until the number of clusters is reduced to the preset value K, and determine the partition result based on the geographical distribution, which is recorded as ,in is the set of wind farms contained in the kth region, i is the wind farm number, The number of internal stations is .

[0089] Step S106: Establish an LSTM time series prediction model for each region respectively, input the historical total power time series in a certain region, the measured wind speed time series at the wind turbine hub of the representative station in a certain region, and the predicted wind speed time series at the wind turbine hub of the representative station in a certain region into a certain LSTM time series prediction model, and the output of the certain LSTM time series prediction model obtains the total power prediction result of a certain region, wherein the representative station in a certain region is selected in a certain region based on the maximum relevance-minimum redundancy principle.

[0090] In this step, in order to accurately represent the wind speed in the region and improve the regional prediction accuracy, based on the wind farm station partitioning results obtained in step 5, the maximum relevance-minimum redundancy principle (MRMR) is used to select regional representative stations for each partition. The number of representative stations in each region is Not less than 1 / 3 of the total number of regional stations, that is ,in Indicates rounding up. Figure 3 Explanation of the process for selecting representative sites:

[0091] In the kth region, calculate The measured wind speed sequence at the hub of the wind turbine in each wind farm station Total regional power The mutual information between them is expressed as:

[0092] ,

[0093] Where, is the mutual information between the measured wind speed sequence at the hub of the wind turbine at wind farm station b and the total regional power, is the marginal probability function of the measured wind speed at the hub of the wind turbine at wind farm station b, is the marginal probability function of the total wind power, is the joint probability distribution between the marginal probability function of the measured wind speed at the hub of the wind turbine at wind farm station b and the total wind power, is the total power of region k, is the measured wind speed sequence at the hub of the wind turbine at wind farm station b, is a specific observation value of the wind speed measured at the hub of the wind turbine at wind farm station b;

[0094] Store the calculation results as a correlation vector , select the first representative station with the greatest correlation with the total regional power , is the correlation vector, Join the representative station gathering , the remaining station set is , for each candidate station u∈U, calculate the MRMR score , is the wind speed at the hub of the candidate wind farm station in region k, is the wind speed at the hub of the selected representative wind farm station in region k, is the mutual information between the wind speed at the hub of the candidate wind farm in region k and the total power in region k, The mutual information between the wind speed at the hub of the candidate wind farm station in region k and the wind speed at the hub of the selected representative wind farm station in region k is selected. The station with the highest MRMR score is added to the representative set. When , the output represents the station set As the representative station of region k.

[0095] It should be noted that after determining the partitions and regional representative stations, we began to establish an LSTM time series prediction model for each region to predict the total regional wind power generation. For each region, we established an LSTM time series prediction model to output the regional power 288-step prediction results. Figure 4 、 Figure 5 Explain the LSTM prediction model.

[0096] LSTM (Long Short-Term Memory) is a recurrent neural network (RNN) variant designed specifically to address the long-term dependencies of sequential data. Compared to traditional RNNs, LSTM effectively alleviates the vanishing gradient problem and is suitable for modeling complex time series, such as the changing patterns of wind power. An LSTM network consists of multiple LSTM units connected in series, each of which updates the input sequence at each time step t. Given the current input , the hidden state at the previous moment , memory status , output the current hidden state and memory status The specific calculation process of LSTM includes the following steps:

[0097] Forget Gate : Control the proportion of information that needs to be forgotten in the memory state of the previous moment. The formula is:

[0098]

[0099] in, represents the Sigmoid activation function, is the weight matrix of the forget gate, is the bias term of the forget gate.

[0100] Input Gate and Candidate Memory Cells: The input gate determines how much new information is written, and the candidate memory cells are the candidate values ​​of the new information:

[0101] ,

[0102] ,

[0103] in, is the input gate output, is a candidate memory unit, is the hyperbolic tangent activation function, and are the weight matrices of the input gate, and are bias terms of the input gate.

[0104] Memory state update: Integrate the results of the forget gate and the input gate to update the current memory unit state:

[0105] ,

[0106] Where, is the updated memory cell state at the current time t, is the Hadamard product, that is, the corresponding elements are multiplied one by one, is the state of the memory unit at the previous moment t-1;

[0107] Output gate and hidden state update: Determine the output hidden state based on the current memory state:

[0108] ,

[0109] ,

[0110] in, is the output of the output gate, is the weight matrix of the output gate, is the bias term of the output gate, is the hidden state at time t-1.

[0111] Collect the measured wind speed of the representative wind farm in area k , predicted wind speed and regional total power .

[0112] Data preprocessing: All kinds of data used as input variables and output variables of the LSTM prediction model are normalized according to their characteristics. The wind power data is normalized to the interval [0,1] based on the rated capacity of the wind farm, and the wind speed meteorological data adopts the maximum and minimum normalization method. The wind power before and after normalization is set to and , wind speed is and The maximum and minimum values ​​of the wind speed samples are 、 , the rated capacity of wind power is , the specific normalization formula is as follows:

[0113] ,

[0114] ,

[0115] The dataset is divided into training set, validation set and test set, and the input sequence is constructed according to the collected and processed data, which contains:

[0116] representative site measured wind speed : look-back time window representative site normalized value of measured wind speed, wherein is the number of representative sites of region k;

[0117] historical regional total power : look-back time window normalized value of partition total power, .

[0118] representative site predicted wind speed : prediction time window normalized value of wind speed prediction sequence at the hub of the representative site, .

[0119] In the LSTM time series prediction model, the historical wind speed of the representative site of input region k in the look-back window and the regional historical total power are spliced along the feature dimension to form the input sequence . The input sequence is input into the LSTM network, and after extracting the time series features by the LSTM, the hidden state of the last time step is output; for the prediction period, the predicted wind speed of the representative site in the prediction period is flattened into a one-dimensional vector , which is input into a fully connected layer to extract global features and compressed into a feature vector , is the weight matrix of the fully connected layer, is the one-dimensional vector after flattening the predicted wind speed of the representative site in the prediction period, is the bias term of the fully connected layer weight matrix. Then the time series features output by the LSTM and the feature vector mapped by the NWP (numerical weather prediction, numerical weather prediction data) are spliced into a fusion vector , which is input into a fusion fully connected layer network: , is the weight matrix of the fully connected layer, is the bias term of the fully connected layer, is the first output of the fusion fully connected layer network, is the weight matrix of the fully connected layer, is the normalized prediction result of the total power of the region k in the future 288 steps, and the output prediction result is de-normalized to obtain the total wind power prediction result of each partition , is the de-normalized prediction result of the total power of the region k in the future 288 steps, is the normalized prediction result of the total power of the region k in the future 288 steps.

[0120] In summary, the method of the present application collects the historical wind speed and power data of each wind farm, quantifies the spatial and temporal similarity between the farms through the autocorrelation function (ACF) and Copula entropy; secondly, the hierarchical clustering method is used to divide the region based on the calculated spatial and temporal similarity distance matrix, to ensure that the farms in the same partition have high consistency; then, the maximum correlation minimum redundancy (MRMR) principle is used to select the representative farm station, and the representative farm station is selected from each region, which effectively reduces the input data size while ensuring the prediction accuracy; finally, based on the wind speed and historical regional total power data of the selected representative farm station, a time series prediction model is constructed to realize the short-term prediction of the regional scale day-ahead wind power. The present application improves the accuracy and computational efficiency of regional power prediction by fusing wind speed-power spatio-temporal correlation modeling, regional division and representative farm station selection, which provides effective support for the stable operation and dispatching of large-scale wind power clusters.

[0121] In one specific embodiment, in this embodiment, 68 wind farms in a certain region are used, and the 5-minute level wind power data in 2024 is divided into 8:1:1 sample training set, validation set and test set to perform day-ahead wind farm cluster power prediction. The geographical distribution of each wind farm is shown in Figure 6 The input data of partition aggregation and representative farm station selection includes the 5-minute level power data of the 68 wind farms in the region in 2024 and the hub wind speed data; the input data of the regional power prediction stage includes the corresponding partition total power and the measured wind speed of the representative farm station within the lookback window, and the predicted wind speed of the representative farm station in the prediction period. The mean absolute percentage error and the mean absolute error and the root mean square error are used as the evaluation criteria for the prediction accuracy of each model, and the calculation formula of each evaluation index is:

[0122] ,

[0123] ,

[0124] ,

[0125] In the formula, is the number of test samples; and The predicted time The actual and predicted wind power values ​​at each sampling point; is the sum of the rated capacities of all wind farms in the corresponding area.

[0126] Applying the method of this application, 68 stations are partitioned and aggregated. Figure 7 The following is a schematic diagram of the results of the selection of regional representative sites. There are five regions in total. Combined with the dominant wind direction in the region, which is from northeast to southwest, it can be seen from the figure that the wind farms in each region also show a northeast to southwest direction, which proves the rationality of the zoning. Observing the representative sites in each region, they are evenly distributed in the region and are not limited to a certain area, which proves the rationality of the selection of regional representative sites. Table 1 and Figure 8 The prediction performance evaluation results of the method proposed in the present invention on the wind power partitions of each region are demonstrated. By statistically comparing the prediction error indicators of each partition, it can be seen that the proposed method shows high prediction accuracy in each partition, with small differences and low volatility between the evaluation indicators. The average MAPE of each partition is 14.19%, indicating that the present invention has good regional adaptability and stability. At the same time, there were no obvious deviations or local failures in the prediction results of each partition, verifying the versatility and reliability of the method of the present invention under different spatial distributions. In general, the present invention can effectively improve the accuracy and robustness of regional wind power prediction through the collaborative design of representative site selection, feature fusion and modeling methods, and meet the requirements for high efficiency and generalization ability of regional wind power prediction in practical applications.

[0127] Table 1 Power prediction evaluation indicators for each region

[0128] ,

[0129] In order to verify the application effect of the prediction model of the present invention in actual engineering, experimental verification is carried out using partition 2 as an example, and the following comparison methods are selected:

[0130] MRMR-LSTM: This paper proposes a method that uses the maximum relevance-minimum redundancy principle to select regional representative stations and uses the LSTM prediction model to perform regional power prediction.

[0131] All-LSTM: Instead of selecting a representative station for a region, the historical wind speed and power data of all stations in the region are selected as input, and the LSTM prediction model is used for regional power prediction.

[0132] MRMR-TCN: The maximum relevance-minimum redundancy principle is used to select regional representative stations. The predicted and measured wind speed data and power data of the representative stations are used as input, and the TCN prediction model is used to perform regional power prediction.

[0133] MRMR-CNN-LSTM: The maximum relevance-minimum redundancy principle is used to select regional representative stations. The predicted and measured wind speed data and power data of the representative stations are used as input, and the CNN-LSTM prediction model is used to perform regional power prediction.

[0134] MRMR-Bi-LSTM: This method uses the maximum relevance-minimum redundancy principle to select regional representative stations. The predicted and measured wind speed data and power data of the representative stations are used as input, and the Bi-LSTM prediction model is used to perform regional power prediction.

[0135] MRMR-LSTM-OnlyWS: This method uses the maximum relevance-minimum redundancy principle to select regional representative stations, takes only the measured wind speed and power data of the representative stations as input, and uses the LSTM prediction model to perform regional power prediction.

[0136] The comparison between MRMR-LSTM and All-LSTM in the above comparison model is intended to verify the effectiveness of selecting regional representative stations. Figure 9 、 Figure 10 As can be seen, the proposed method performs comparable to the proposed method in terms of MAPE and MAE, but outperforms the proposed method in terms of RMSE. Table 2 also shows that MRMR-LSTM improves prediction time by 14.95, 31.04, 28.35, 25.01, and −4.05 percentage points, respectively, compared to all the comparison models. Since MRMR-LSTM-OnlyWS does not incorporate predicted wind speed information for the prediction period and cannot guarantee prediction accuracy, this demonstrates that the proposed method significantly outperforms all the comparison models in terms of computational time consumption. These results demonstrate that selecting representative wind farms using the maximum relevance minimum redundancy (MRMR) principle not only effectively reduces model input dimensionality, computational resource consumption, and modeling complexity, but also ensures prediction accuracy. Compared to directly using all site data, this method effectively avoids prediction errors caused by redundant information and noise interference, providing a more efficient and robust solution for regional wind power modeling.

[0137] Table 2 Training time of each model in partition 2

[0138] ,

[0139] The comparison between All-LSTM and MRMR-TCN, MRMR-CNN-LSTM, and MRMR-Bi-LSTM in the above comparison models is intended to verify the rationality of the prediction model design. From Table 2, Figure 9 、 Figure 10 It can be seen that the MRMR-LSTM model, which is selected as a representative station based on the maximum relevance minimum redundancy (MRMR) principle, is compared with the MRMR-TCN, MRMR-CNN-LSTM, and MRMR-Bi-LSTM models. Compared with the comparison model, it increased by 3.81 percentage points. increased by 17.23 percentage points. An increase of 18.1 percentage points. In comparison with other methods, the model proposed in this paper shows better prediction performance. This is mainly due to the model's ability to effectively model the complex and causal mapping relationship between wind speed and power, while avoiding the introduction of invalid information in the feature extraction process. Compared with structures such as convolutional neural networks (CNN) or bidirectional recurrent neural networks (Bi-LSTM), the time series modeling strategy adopted by the model in this paper is more in line with the physical characteristics of wind power prediction tasks, and can fully explore the long-term dependency characteristics in time series to ensure the stability and accuracy of the prediction results. In addition, the model structure is relatively simple and the training efficiency is high. While maintaining high prediction accuracy, it has good engineering application potential, verifying its rationality and practicality as a regional power prediction tool.

[0140] The comparison between the All-LSTM and MRMR-LSTM-OnlyWS models described above aims to verify the effectiveness of integrating wind speed data during the forecast period. Table 3 shows that MRMR-LSTM significantly outperforms MRMR-LSTM-OnlyWS on all forecast evaluation metrics. This is because the All-LSTM model incorporates features containing wind speed information for the future forecast time step into its input, effectively improving the model's ability to perceive future power changes and making the forecast results more accurate. The MRMR-LSTM-OnlyWS model, on the other hand, relies solely on measured wind speed data from representative stations and fails to consider future meteorological information, making it difficult to capture wind speed trends and limiting forecast accuracy. This comparison demonstrates the importance of incorporating NWP information for improving regional wind power forecast accuracy.

[0141] Table 3 Evaluation indicators of each model in partition 2

[0142] .

[0143] See also Figure 11 , which shows a structural block diagram of a short-term power prediction system based on a wind farm cluster in the present application.

[0144] like Figure 11 As shown, the wind farm cluster short-term power prediction system 200 includes an acquisition module 210 , a calculation module 220 , a construction module 230 , a summation module 240 , a division module 250 and an output module 260 .

[0145] The acquisition module 210 is configured to acquire the target area wind farms 5-minute power data for each historical sample day , Measured wind speed data at the wind turbine hub ,in, is a three-dimensional real space, represents the field of real numbers, express sample day, 288 represents the number of 5-minute samples in a day, represents the number of wind farms, For the The power sequence of a wind farm station, For the The measured wind speed series of a wind farm station, For the The predicted wind speed series of each wind farm station;

[0146] The calculation module 220 is configured to calculate the ACF curve of the wind speed at the hub of the wind turbine, calculate the weighted Euclidean distance between the ACF curves of different stations, and construct a similarity distance matrix of the time-varying laws of key meteorological elements between different stations. ,in, is a two-dimensional real number space, Indicates the number of wind farms;

[0147] The construction module 230 is configured to calculate the Copula entropy between the historical power samples of each station and construct the spatial similarity matrix of the station power according to each Copula entropy. ;

[0148] The summing module 240 is configured to sum the similarity distance matrix and spatial similarity matrix Perform normalization and normalize the similarity distance matrix and spatial similarity matrix Sum and get the spatiotemporal similarity distance matrix ;

[0149] The partitioning module 250 is configured to use a hierarchical clustering algorithm to partition the spatiotemporal similarity distance matrix Clustering is performed, and the zoning scheme is determined based on the geographical distribution of new energy stations. A wind farm is divided into K areas;

[0150] The output module 260 is configured to establish an LSTM time series prediction model for each area, and input the power sequence in a certain area, the measured wind speed sequence at the wind turbine hub of the representative station in a certain area, and the predicted wind speed sequence at the wind turbine hub of the representative station in a certain area into a certain LSTM time series prediction model. The output of the certain LSTM time series prediction model obtains the total power prediction result of a certain area, wherein the representative station in a certain area is selected in a certain area based on the maximum correlation-minimum redundancy principle.

[0151] It should be understood that Figure 11 Modules and references documented in Figure 1 Therefore, the operations and features described above for the method and the corresponding technical effects also apply to Figure 11 The modules in it will not be described in detail here.

[0152] In other embodiments, embodiments of the present invention further provide a computer-readable storage medium having a computer program stored thereon, wherein when the program instructions are executed by a processor, the processor is caused to execute the method for short-term power prediction based on a wind farm cluster in any of the above method embodiments;

[0153] As an embodiment, the computer-readable storage medium of the present invention stores computer-executable instructions, and the computer-executable instructions are configured as follows:

[0154] Get target area wind farms 5-minute power data for each historical sample day , Measured wind speed data at the wind turbine hub and predicted wind speed data at the turbine hub ,in, is a three-dimensional real space, represents the field of real numbers, express sample day, 288 represents the number of 5-minute samples in a day, represents the number of wind farms, For the The power sequence of a wind farm station, For the The measured wind speed series of the wind farm station, For the The predicted wind speed series of each wind farm station;

[0155] An ACF curve of the wind speed at the fan hub is calculated, a weighted Euclidean distance between ACF curves of different stations is calculated, and a similarity distance matrix of time-varying rules of key meteorological elements between different stations is constructed wherein, is a two-dimensional real number space, represents the number of wind power stations;

[0156] A Copula entropy between historical power samples of each station is calculated, and a spatial similarity matrix of the power of the stations is constructed according to the Copula entropies ;

[0157] The similarity distance matrix and the spatial similarity matrix are normalized, and the normalized similarity distance matrix and the spatial similarity matrix are summed to obtain a spatio-temporal similarity distance matrix ;

[0158] The hierarchical clustering algorithm is used to cluster the spatio-temporal similarity distance matrix , and a partition scheme is determined in combination with the geographical distribution of the new energy stations to divide K wind power stations in the target region into K regions;

[0159] An LSTM time series prediction model is established for each region, a power sequence in a region, a measured wind speed sequence at a fan hub of a representative station in the region, and a predicted wind speed sequence at the fan hub of the representative station in the region are input into the LSTM time series prediction model, and a total power prediction result of the region is output by the LSTM time series prediction model, wherein the representative station in the region is selected in the region based on the maximum correlation-minimum redundancy principle.

[0160] The computer-readable storage medium can include a program storage area and a data storage area, wherein the program storage area can store an operating system and at least one application required by a function; the data storage area can store data created according to the use of the wind power plant cluster short-term power prediction system, etc. In addition, the computer-readable storage medium can include a high-speed random access memory, and can also include a memory such as at least one disk storage device, a flash memory device, or other non-volatile solid-state memory device. In some embodiments, the computer-readable storage medium can optionally include a memory remotely arranged relative to the processor, and these remote memories can be connected to the wind power plant cluster short-term power prediction system through a network. Examples of the above network include but are not limited to the Internet, an intranet, a local area network, a mobile communication network, and a combination thereof.

[0161] Figure 12 ​Schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. Figure 12 As shown, the device includes: a processor 310 and a memory 320. The electronic device may also include: an input device 330 and an output device 340. The processor 310, the memory 320, the input device 330 and the output device 340 may be connected via a bus or other means. Figure 12 The example of a bus connection is used. Memory 320 is the aforementioned computer-readable storage medium. Processor 310 executes various server functional applications and data processing by running the non-volatile software programs, instructions, and modules stored in memory 320, thereby implementing the aforementioned method embodiment for short-term power forecasting of a wind farm cluster. Input device 330 can receive input digital or character information and generate key signal input related to user settings and function control of the wind farm cluster-based short-term power forecasting system. Output device 340 can include a display device such as a display screen.

[0162] The electronic device can execute the method provided by the embodiment of the present invention, and has the functional modules and beneficial effects corresponding to the execution method. For technical details not fully described in this embodiment, please refer to the method provided by the embodiment of the present invention.

[0163] As an embodiment, the electronic device is applied to a short-term power forecasting system based on a wind farm cluster and is used for a client, and includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to:

[0164] Get target area wind farms 5-minute power data for each historical sample day , Measured wind speed data at the wind turbine hub and predicted wind speed data at the turbine hub ,in, is a three-dimensional real space, represents the field of real numbers, express sample day, 288 represents the number of 5-minute samples in a day, represents the number of wind farms, For the The power sequence of a wind farm station, For the The measured wind speed series of the wind farm station, For the The predicted wind speed series of each wind farm station;

[0165] Calculate the ACF curve of wind speed at the hub of the wind turbine, calculate the weighted Euclidean distance between the ACF curves of different stations, and construct the similarity distance matrix of the time-varying laws of key meteorological elements between different stations ,in, is a two-dimensional real number space, Indicates the number of wind farms;

[0166] Calculate the Copula entropy between the historical power samples of each station, and construct the spatial similarity matrix of the station power based on each Copula entropy ;

[0167] The similarity distance matrix and spatial similarity matrix Perform normalization and normalize the similarity distance matrix and spatial similarity matrix Sum and get the spatiotemporal similarity distance matrix ;

[0168] The hierarchical clustering algorithm is used to classify the spatiotemporal similarity distance matrix Clustering is performed, and the zoning scheme is determined based on the geographical distribution of new energy stations. A wind farm is divided into K areas;

[0169] An LSTM time series prediction model is established for each region. The power sequence in a certain region, the measured wind speed sequence at the wind turbine hub of a representative station in a certain region, and the predicted wind speed sequence at the wind turbine hub of a representative station in a certain region are input into a certain LSTM time series prediction model. The LSTM time series prediction model outputs the total power prediction result of a certain region, wherein the representative station in a certain region is selected in a certain region based on the maximum correlation-minimum redundancy principle.

[0170] Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented using software plus a necessary general-purpose hardware platform, or of course, hardware. Based on this understanding, the essence of the above technical solution, or the portion that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, or an optical disk, and includes a number of instructions for causing a computer device (such as a personal computer, server, or network device) to execute the methods of each embodiment or certain portions of the embodiments.

[0171] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims

1. A short-term power prediction method based on a wind farm cluster, characterized in that: include: Get target area wind farms 5-minute power data for each historical sample day , Measured wind speed data at the wind turbine hub and predicted wind speed data at the turbine hub ,in, is a three-dimensional real space, represents the field of real numbers, express sample day, 288 represents the number of 5-minute samples in a day, represents the number of wind farms, For the The power sequence of a wind farm station, For the The measured wind speed series of a wind farm station, For the The predicted wind speed series of each wind farm station; Calculate the ACF curve of wind speed at the hub of the wind turbine, calculate the weighted Euclidean distance between the ACF curves of different stations, and construct the similarity distance matrix of the time-varying laws of key meteorological elements between different stations ,in, is a two-dimensional real number space, Indicates the number of wind farms; Calculate the Copula entropy between the historical power samples of each station, and construct the spatial similarity matrix of the station power based on each Copula entropy ; The similarity distance matrix and spatial similarity matrix Perform normalization and normalize the similarity distance matrix and spatial similarity matrix Sum and get the spatiotemporal similarity distance matrix ; The hierarchical clustering algorithm is used to classify the spatiotemporal similarity distance matrix Clustering is performed, and the zoning scheme is determined based on the geographical distribution of new energy stations. A wind farm is divided into K areas; An LSTM time series prediction model is established for each region. The power sequence in a certain region, the measured wind speed sequence at the wind turbine hub of a representative station in a certain region, and the predicted wind speed sequence at the wind turbine hub of a representative station in a certain region are input into a certain LSTM time series prediction model. The LSTM time series prediction model outputs the total power prediction result of a certain region, wherein the representative station in a certain region is selected in a certain region based on the maximum correlation-minimum redundancy principle.

2. The method for short-term power prediction based on a wind farm cluster according to claim 1, characterized in that: The ACF curve of wind speed at the hub of the wind turbine is calculated, and the weighted Euclidean distance between the ACF curves of different stations is calculated, and the similarity distance matrix of the time-varying laws of key meteorological elements between different stations is constructed. include: Calculate the wind speed at the hub of each wind turbine at each wind farm The ACF curve of the i-th station at the lag time t is calculated. The expression is: ， Where, is the length of the wind speed sequence, is the mean of the measured wind speed sequence of the i-th wind farm station, , is the ACF value of the i-th wind farm at lag time t, is the measured wind speed sequence of the i-th wind farm at time point The wind speed value, is the measured wind speed sequence of the i-th wind farm at time point Wind speed value; Construct the ACF matrix, the expression is: ， Where, is the ACF matrix, is the ACF value of the mth wind farm at lag time n; Calculate the average ACF value of all wind farms at lag time t , and calculate the Euclidean distance weight of each lag time point based on the average ACF value, where the expression for calculating the average ACF value is: ， Where, is the ACF value of wind farm k at lag time t, is the number of wind farms; The expression for calculating the Euclidean distance weight is: ， According to the Euclidean distance weight, the weighted Euclidean distance matrix between the ACF curves of each station is calculated, that is, the similarity distance matrix of the time-varying laws of key meteorological elements between different stations is constructed. , the expression is: ， Where, is the weighted Euclidean distance between the m-th wind farm and the m-th wind farm, and its value is 0. is the weighted Euclidean distance between wind farm h and wind farm j; ， Where, 、 are the autocorrelation coefficients of wind farm h and wind farm j at lag time t, is the maximum lag time of ACF.

3. The method for short-term power prediction based on a wind farm cluster according to claim 1, characterized in that: The Copula entropy between the historical power samples of each station is calculated, and the spatial similarity matrix of the station power is constructed based on each Copula entropy. include: The empirical marginal distribution functions of wind farm h and wind farm j are estimated respectively by using order statistics. The order statistics is a statistic that describes the relative position of each data point in the sequence, and its expression is: ， ， Where, is the empirical marginal distribution function of the power time series sample of the h-th wind farm, is the number of power time series samples, is an indicator function, if When it is established, the function value is 1. When it is not true, the function value is 0. is the sth sample in the power time series of the hth wind farm, is the sth sample in the power time series of the jth wind farm, is the power time series of the wind farm; The wind farm and wind farms The power time series samples of are mapped to marginal probabilities through the empirical distribution function, and the expression is: ， ， Where, For wind farms h The tth sample in the power sequence of For wind farms j The tth sample of the power sequence, For wind farms j The empirical marginal distribution function of the power sequence is, For wind farms h The empirical marginal distribution function of the power sequence is, For wind farms h The marginal probability of the t-th sample in the power sequence of For wind farms j The marginal probability of the t-th sample in the power sequence of ; Constructing joint sample pairs based on marginal probabilities , and calculate the distance between sample pairs based on Chebyshev distance, the expression is: ， ， Where, is the sth sample pair and the vth sample pair The Chebyshev distance between , For sample pairs The marginal probability of the sth sample in the wind field h is, For sample pairs The marginal probability of the vth sample in the wind farm h is, For sample pairs The marginal probability of the sth sample in wind farm j is, For sample pairs The marginal probability of the vth sample in wind farm j is, is the sth power sample, is the vth power sample; For each sample pair, find the kth nearest neighbor distance of each sample pair, and estimate the Copula entropy between the historical power samples of each station based on the k nearest neighbor distance. The expression is: ， Where, is the volume of the two-dimensional unit sphere, under the Chebyshev distance =1, is the Chebyshev distance of the kth nearest neighbor of sample pair s, is the Copula entropy between station h and station j, is the Digamma function; Depend on The spatial correlation matrix representing the nonlinearity between the two stations is constructed, that is, the spatial similarity matrix of the station power is constructed. , the expression is: ， Where, is the Copula entropy between station h and station j, is the Copula entropy between station m and station m.

4. The method for short-term power prediction based on a wind farm cluster according to claim 1, characterized in that: The spatiotemporal similarity distance matrix The expression is: ， Where, is the normalized Copula entropy matrix, is the normalized weighted Euclidean distance matrix between the ACF curves of each station, is the spatiotemporal similarity distance between stations h and j, is the spatiotemporal similarity distance between the m-th wind farm and the m-th wind farm.

5. The method for short-term power prediction based on a wind farm cluster according to claim 1, characterized in that: Before establishing the LSTM time series prediction model for each region, the method further includes: In the kth region, calculate The measured wind speed sequence at the hub of the wind turbine in each wind farm station Total regional power The mutual information between them is expressed as: ， Where, is the mutual information between the measured wind speed sequence at the hub of the wind turbine at wind farm station b and the total regional power, is the marginal probability function of the measured wind speed at the hub of the wind turbine at wind farm station b, is the marginal probability function of the total wind power, is the joint probability distribution between the marginal probability function of the measured wind speed at the hub of the wind turbine at wind farm station b and the total wind power, is the total power of region k, is the measured wind speed sequence at the hub of the wind turbine at wind farm station b, is a specific observation value of the wind speed measured at the hub of the wind turbine at wind farm station b; Store the calculation results as a correlation vector , select the first representative station with the greatest correlation with the total regional power , is the correlation vector, Join the representative station gathering , the remaining station set is , for each candidate station u∈U, calculate the MRMR score , is the wind speed at the hub of the candidate wind farm station in region k, is the wind speed at the hub of the selected representative wind farm station in region k, is the mutual information between the wind speed at the hub of the candidate wind farm in region k and the total power in region k, The mutual information between the wind speed at the hub of the candidate wind farm station in region k and the wind speed at the hub of the selected representative wind farm station in region k is selected. The station with the highest MRMR score is added to the representative set. When , the output represents the station set As the representative station of region k.

6. A short-term power prediction system based on wind farm clusters, characterized in that: include: Acquisition module, configured to obtain the target area wind farms 5-minute power data for each historical sample day , Measured wind speed data at the wind turbine hub and predicted wind speed data at the turbine hub ,in, is a three-dimensional real space, represents the field of real numbers, express sample day, 288 represents the number of 5-minute samples in a day, represents the number of wind farms, For the The power sequence of a wind farm station, For the The measured wind speed series of a wind farm station, For the The predicted wind speed series of each wind farm station; The calculation module is configured to calculate the ACF curve of the wind speed at the hub of the wind turbine, calculate the weighted Euclidean distance between the ACF curves of different stations, and construct a similarity distance matrix of the time-varying laws of key meteorological elements between different stations ,in, is a two-dimensional real number space, Indicates the number of wind farms; A construction module is configured to calculate the Copula entropy between the historical power samples of each station and construct the spatial similarity matrix of the station power based on each Copula entropy ; The summation module is configured to sum the similarity distance matrix and spatial similarity matrix Perform normalization and normalize the similarity distance matrix and spatial similarity matrix Sum and get the spatiotemporal similarity distance matrix ; A partitioning module is configured to use a hierarchical clustering algorithm to partition the spatiotemporal similarity distance matrix Clustering is performed, and the zoning scheme is determined based on the geographical distribution of new energy stations. A wind farm is divided into K areas; The output module is configured to establish an LSTM time series prediction model for each area, and input the power sequence in a certain area, the measured wind speed sequence at the wind turbine hub of a representative station in a certain area, and the predicted wind speed sequence at the wind turbine hub of a representative station in a certain area into a certain LSTM time series prediction model. The output of the certain LSTM time series prediction model obtains the total power prediction result of a certain area, wherein the representative station in a certain area is selected in a certain area based on the maximum correlation-minimum redundancy principle.

7. An electronic device, characterized in that: include: At least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 5 is implemented.

Images