New energy extreme output recognition method and device

Abstract

The invention relates to the technical field of new energy power generation, in particular to a new energy extreme output recognition method and device, and the method comprises the steps: collecting the output data of a new energy station; obtaining clusters of the output data according to the daily output peak value and the volatility characteristic of the output data, and extracting at least one dimension characteristic of the output data to calculate a weighted characteristic value of the at least one dimension characteristic; and performing synchronous retraction iterative screening on the clusters according to the weighted feature values to obtain extreme samples of the clusters meeting a preset screening condition, and identifying an extreme output scene of the new energy station according to the extreme samples. Therefore, the problems that the accuracy and robustness of an identification result are insufficient, and the accuracy and robustness of the identification result are poor due to the fact that the related technology usually depends on a single feature for judgment and lacks effective description of an incidence relation among multiple features and an influence weight, and an extreme output behavior is easily misjudged or missed to be judged are solved. And early warning, control, risk prevention and the like are difficult to support in time in actual scheduling and operation.

Description

Technical Field

[0001] This application relates to the field of new energy power generation technology, and in particular to a method and device for identifying extreme power output of new energy sources. Background Technology

[0002] The output of new energy sources is affected by many factors, including natural conditions and equipment status, and is prone to extreme scenarios such as abnormal peaks, drastic fluctuations, or prolonged troughs. If these extreme scenarios are not identified in time, they may lead to grid dispatch imbalances, equipment overloads, or even outages, thereby affecting the safe and stable operation of the power grid.

[0003] In related technologies, the identification of extreme power output in new energy sources typically uses statistical measures such as peak value, standard deviation, and root mean square deviation to determine whether there is an extreme power output, which can reflect the degree of abnormality in power changes to a certain extent.

[0004] However, related technologies often rely on single features, which cannot fully characterize the differences between various extreme power output modes, making the identification results susceptible to short-term noise interference and misjudgment. At the same time, related technologies do not distinguish the value of different types of features (such as rapid changes in ramp rate, continuous troughs, transient spikes, etc.) in extreme signal identification, and lack effective characterization of the correlation and influence weight between multiple features. They are difficult to adapt to the strong fluctuations, non-stationarity and diversity of new energy power output, resulting in insufficient identification accuracy in extreme scenarios, which makes it difficult to meet the needs of scheduling control and safe operation, and urgently needs to be solved. Summary of the Invention

[0005] This application provides a method for identifying extreme power output of new energy sources, in order to solve the problem that in related technologies, the identification of extreme power output of new energy sources usually relies on a single feature for judgment and lacks an effective characterization of the correlation and influence weight between multiple features, which easily leads to misjudgment or omission of extreme power output behavior, resulting in insufficient accuracy and robustness of the identification results, and difficulty in timely supporting early warning, control and risk prevention in actual scheduling and operation.

[0006] The first aspect of this application provides a method for identifying extreme power output of new energy sources, comprising the following steps: collecting power output data of new energy power plants; obtaining clusters of the power output data based on the daily peak power output and fluctuation characteristics of the power output data, and extracting at least one dimension feature of the power output data to calculate a weighted feature value of the at least one dimension feature; performing synchronous back-shrinking iterative screening on the clusters based on the weighted feature value to obtain extreme samples of the clusters that meet preset screening conditions, and identifying extreme power output scenarios of the new energy power plants based on the extreme samples.

[0007] Through the above technical means, the embodiments of this application can perform synchronous back-shrinking iterative screening of clusters based on the weighted feature values ​​of multi-dimensional features of power output data to identify extreme power output scenarios of new energy power plants. By utilizing the comprehensive information of multi-dimensional features, the output differences under different modes can be uniformly quantified, avoiding deviations caused by single feature judgments, improving identification accuracy and robustness, thereby achieving rapid and accurate identification of extreme power output scenarios. This helps reduce grid dispatch risks, prevent potential hazards such as equipment overload in advance, and provide more targeted decision-making basis for the operation and control of new energy power plants.

[0008] Optionally, in one embodiment of this application, obtaining the clustering of the power output data based on the daily peak power output and fluctuation characteristics of the power output data includes: extracting the daily peak power output and fluctuation characteristics of the power output data to obtain a feature matrix of the power output data; processing the feature matrix to obtain a feature matrix that meets preset standardization conditions; and clustering and dividing the feature matrix to obtain the clustering.

[0009] Through the above technical means, the embodiments of this application can cluster and divide the feature matrix to obtain clusters, which can automatically classify data with similar output patterns or feature distributions into the same category, thereby reducing the interference caused by feature differences and improving the accuracy of subsequent extreme output identification. At the same time, it can effectively reveal the potential structural features of the output data, distinguish different operating modes, and provide a clear pattern basis for subsequent weighted feature calculation and synchronous shrinkage iterative screening.

[0010] Optionally, in one embodiment of this application, the step of clustering and dividing the feature matrix to obtain the clusters includes: setting the number of clusters for each cluster to determine the initial cluster center vector of the feature matrix based on the number of clusters; calculating the distance between the samples of each cluster and the initial cluster center vector to allocate the samples to the clusters based on the distance to obtain an allocation result; calculating the new cluster centers of each cluster based on the allocation result until the change in the cluster centers meets a preset iteration stop condition to obtain the clusters.

[0011] Through the above technical means, the embodiments of this application can allocate samples to corresponding clusters according to the distance between the samples and the initial cluster center vectors of the feature matrix. It can achieve fast and effective clustering based on the similarity between the samples and the center vectors, thereby improving the stability and accuracy of the clustering initialization stage. At the same time, it can enable data with similar features to naturally aggregate, avoiding interference from abnormal samples or noisy data on the overall clustering results, thus providing a high-quality initial clustering structure for subsequent cluster center updates, weighted feature calculations, and synchronous shrinkage iterative screening.

[0012] Optionally, in one embodiment of this application, calculating the weighted feature value of the at least one dimension feature includes: calculating the variance of the at least one dimension feature; performing weight normalization on the at least one dimension feature based on the variance to obtain features that satisfy a preset normalization condition; and calculating the weighted feature value based on the features.

[0013] Through the above technical means, the embodiments of this application can normalize the variance by weight, and then calculate the weighted feature value of the dimensional features. This can ensure that the weights of different dimensional features are within a uniform scale range, avoiding the weight bias problem caused by differences in dimensions or excessively large numerical spans. Through the normalized variance weights, the influence of each dimensional feature on the output fluctuation and extremes can be more accurately reflected, making the weighted feature value more stable and comparable when representing extreme samples, thereby improving the objectivity and accuracy of extreme output identification.

[0014] Optionally, in one embodiment of this application, the step of synchronously shrinking and iteratively filtering the clusters according to the weighted feature values ​​to obtain extreme samples of the clusters that meet preset filtering conditions includes: initializing the candidate set of the clusters; calculating the extreme threshold of the samples in the candidate set according to the weighted feature values; and iteratively filtering the samples with the smallest extreme threshold based on the extreme threshold and a preset multi-feature weighted synchronous shrinking strategy until the size of the candidate set meets the preset filtering conditions to obtain the extreme samples.

[0015] Through the above technical means, the embodiments of this application can iteratively select the sample with the smallest extreme threshold based on an extreme threshold and a preset multi-feature weighted synchronous shrinkage strategy. Under the comprehensive constraints of multi-dimensional features, it can gradually approach the most representative extreme output sample, making the identification of extreme scenarios more accurate and convergent. Through the iterative method of synchronous shrinkage, the feature weights and screening thresholds can be dynamically adjusted to reduce the interference of noisy samples and ensure that the finally selected extreme sample has the most prominent anomaly in the feature space, thereby significantly improving the accuracy and robustness of extreme output scenario identification.

[0016] A second aspect of this application provides a new energy extreme output identification device, comprising: a data acquisition module for acquiring output data of new energy power plants; a calculation module for obtaining clusters of the output data based on the daily output peak value and fluctuation characteristics of the output data, and extracting at least one dimension feature of the output data to calculate a weighted feature value of the at least one dimension feature; and an identification module for performing synchronous back-shrinking iterative filtering on the clusters based on the weighted feature value to obtain extreme samples of the clusters that meet preset filtering conditions, and identifying extreme output scenarios of the new energy power plants based on the extreme samples.

[0017] Optionally, in one embodiment of this application, the calculation module includes: an extraction unit, used to extract the daily peak output and fluctuation characteristics of the output data to obtain a feature matrix of the output data; a processing unit, used to process the feature matrix to obtain a feature matrix that meets preset standardization conditions; and a partitioning unit, used to cluster and partition the feature matrix to obtain the clusters.

[0018] Optionally, in one embodiment of this application, the partitioning unit includes: a setting subunit, configured to set the number of clusters in the sub-clusters to determine the initial cluster center vector of the feature matrix based on the number of clusters; an allocation subunit, configured to calculate the distance between the samples in the sub-clusters and the initial cluster center vector, and allocate the samples to the sub-clusters based on the distance to obtain an allocation result; and a calculation subunit, configured to calculate the new cluster centers of the sub-clusters based on the allocation result, until the change in the cluster centers satisfies a preset iteration stop condition to obtain the sub-clusters.

[0019] Optionally, in one embodiment of this application, the calculation module includes: a first calculation unit for calculating the variance of the at least one dimension feature; a normalization unit for performing weight normalization on the at least one dimension feature according to the variance to obtain a feature that satisfies a preset normalization condition; and a second calculation unit for calculating the weighted feature value according to the feature.

[0020] Optionally, in one embodiment of this application, the identification module includes: an initialization unit for initializing the candidate set of the cluster; a third calculation unit for calculating the extreme threshold of the samples in the candidate set based on the weighted feature values; and a filtering unit for iteratively filtering the samples with the smallest extreme threshold based on the extreme threshold and a preset multi-feature weighted synchronous shrinkage strategy until the size of the candidate set meets the preset filtering conditions to obtain the extreme samples.

[0021] A third aspect of this application provides an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the new energy extreme output identification method as described in the above embodiments.

[0022] A fourth aspect of this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for identifying extreme power output of new energy sources.

[0023] A fifth aspect of this application provides a computer program product, including a computer program that, when executed, is used to implement the above-described method for identifying extreme power output of new energy sources.

[0024] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0025] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 This is a flowchart of a new energy extreme output identification method provided according to an embodiment of this application; Figure 2 This is a flowchart of a new energy extreme output identification method according to an embodiment of this application; Figure 3 This is a block diagram of a new energy extreme output identification device according to an embodiment of this application; Figure 4 This is a schematic diagram of the structure of an electronic device according to an embodiment of this application.

[0026] Figure label: 10-New energy extreme output identification device; 100-Acquisition module, 200-Computation module, 300-Identification module; 401-Memory, 402-Processor, 403-Communication interface. Detailed Implementation

[0027] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain this application, and should not be construed as limiting this application.

[0028] The following describes a method and apparatus for identifying extreme power output of new energy sources according to embodiments of this application, with reference to the accompanying drawings. To address the technical problem mentioned in the background that the identification of extreme power output from renewable energy sources typically relies on a single feature for judgment and lacks an effective characterization of the correlation and influence weights between multiple features, which easily leads to misjudgment or omission of extreme power output behavior, resulting in insufficient accuracy and robustness of the identification results and difficulty in timely supporting early warning, control, and risk prevention in actual scheduling and operation, this application provides a method for identifying extreme power output from renewable energy sources. In this method, the power output data of renewable energy power plants are clustered, and then the clusters are synchronously reduced and iteratively filtered based on the weighted feature values ​​of at least one dimension feature to identify extreme power output scenarios of renewable energy power plants. This method can ensure consistency of extreme standards within the same mode, effectively avoid conflicts in extreme definitions under different modes, and make full use of multi-dimensional feature information to highlight the characteristics of extreme samples, avoiding misjudgment or omission caused by a single feature. It can effectively solve the one-sidedness problem of a single indicator, improve the accuracy and robustness of extreme power output identification, thereby reducing the risk of grid scheduling imbalance and equipment overload, reducing failure losses caused by extreme scenarios, and providing targeted basis for the operation and maintenance of renewable energy power plants, thus reducing operation and maintenance costs. This solves the problem that the identification of extreme power output in new energy sources usually relies on a single feature for judgment and lacks an effective characterization of the correlation and influence weight between multiple features, which easily leads to misjudgment or omission of extreme power output behavior. As a result, the accuracy and robustness of the identification results are insufficient, and it is difficult to support early warning, control and risk prevention in actual scheduling and operation.

[0029] Specifically, Figure 1 This is a flowchart illustrating a method for identifying extreme power output of a new energy source, as provided in an embodiment of this application.

[0030] like Figure 1 As shown, the method for identifying extreme power output of new energy sources includes the following steps: In step S101, the power output data of the new energy power station is collected.

[0031] It is understandable that the data acquisition methods may include, but are not limited to, obtaining real-time operating parameters such as active power, reactive power, voltage, and current of wind turbines / photovoltaic inverters through a site monitoring system, power acquisition terminal, or energy management system, and storing and uploading them at a set sampling period to form a raw output sequence for analysis.

[0032] Furthermore, after obtaining the output data of the new energy power station, this application embodiment can clean the output data and remove redundant parts. As one possible implementation, this application embodiment can extract daily output data to obtain a sample set D, which can be represented as: , Where n is the total number of days.x i Indicates the first i The output data vector of the day.

[0033] In step S102, the power output data is clustered based on the daily power output peak and fluctuation characteristics, and at least one dimension feature of the power output data is extracted to calculate the weighted feature value of at least one dimension feature.

[0034] Among these parameters, daily peak power output and volatility characteristics are important statistical features characterizing the operational status of renewable energy power plants. They can be used to depict the distribution of extreme power values, changes in ramp-up speed, and power output stability at different time scales, thus providing fundamental feature support for identifying potential extreme power output behavior. Clustering of power output data can group power sequences with similar variation patterns into the same category, thereby more fully revealing the differences in power output behavior under different operating scenarios. This allows the model to learn the characteristic patterns of various typical patterns, improving the refinement and accuracy of extreme power output identification.

[0035] It can be noted that the methods for extracting dimensional features may include, but are not limited to, sliding window, moving average, segmented statistics, etc.; dimensional features may include, but are not limited to, mean, variance, peak value, skewness, kurtosis, etc.; the extraction method and dimensional features can be set by those skilled in the art according to the actual situation, and no specific restrictions are imposed here.

[0036] In the embodiments of this application, the weighted feature values ​​can be weighted and fused according to the contribution of different features in extreme scene recognition, thereby improving the sensitivity of comprehensive features to abnormal patterns, enabling the recognition model to more accurately reflect the influence of multiple types of features, thereby improving the reliability and accuracy of extreme output recognition.

[0037] Optionally, in one embodiment of this application, obtaining clusters of power output data based on the daily peak power output and fluctuation characteristics of the power output data includes: extracting the daily peak power output and fluctuation characteristics of the power output data to obtain a feature matrix of the power output data; processing the feature matrix to obtain a feature matrix that meets preset standardization conditions; and clustering and dividing the feature matrix to obtain clusters.

[0038] In the embodiments of this application, the preset standardization conditions can be achieved by normalizing each feature, using zero-mean unit variance standardization, or interval scaling, so that features of different dimensions and ranges in the feature matrix are unified to a comparable scale. This avoids the dominant effect of large-scale features on small-scale features, and improves the stability and accuracy of subsequent feature fusion and recognition model training. The clustering and partitioning methods can include, but are not limited to, distance-based partitioning methods, such as K-means clustering and hierarchical clustering; they can also be combined with feature dimensionality reduction or embedding representation methods to map high-dimensional features to a low-dimensional space before clustering, so as to more accurately partition data subsets with similar output patterns, thereby facilitating the subsequent identification and analysis of extreme output behaviors.

[0039] As a specific example, embodiments of this application can extract two key features from daily output data to construct a feature matrix, which may include, but is not limited to: 1) Peak characteristics p i That is, the first i The maximum output value at all points in time throughout the day p i =max ( x i ); 2) Volatility characteristics v i That is, the first i Standard deviation of output values ​​at all points in time throughout the day: , in, m The number of times per day. x i,t For the first i Heavenly t Output value at each point in time For the first i The average daily output value.

[0040] Finally, the feature matrix can be represented as X =[ f 1, f 2,..., f n ] T ,in fi =[ p i , v i ] T For the first i Eigenvectors of the sky.

[0041] Furthermore, to eliminate the influence of dimensions, embodiments of this application can standardize the feature matrix, which can be expressed as: , in, For the standardized first i The first sample j Features ( j =1 corresponds to the peak value. j =2 corresponds to volatility); μ j For the first j The mean of each feature; σ j For the first j Standard deviation of each feature; =10 8 To avoid the minimum value where the denominator is 0.

[0042] Optionally, in one embodiment of this application, clustering and dividing the feature matrix to obtain clusters includes: setting the number of clusters to determine the initial cluster center vector of the feature matrix based on the number of clusters; calculating the distance between the samples of the clusters and the initial cluster center vector to allocate the samples to the clusters based on the distance to obtain the allocation result; calculating the new cluster centers of the clusters based on the allocation result until the change in the cluster centers meets the preset iteration stop condition to obtain the clusters.

[0043] It can be explained that the preset iteration stopping condition can be that the change in cluster centers is lower than a set threshold, or that the set maximum number of iterations is reached, or that the clustering result remains stable and no longer changes in multiple iterations, so as to ensure that the clustering process achieves a balance between convergence and computational cost; it can be set by those skilled in the art according to the actual situation, and no specific restrictions are made here.

[0044] For example, the present application embodiment clusters and partitions the feature matrix based on K-means clustering, which may include the following steps: 1) Initialize cluster centers: First, the number of clusters can be set in the embodiments of this application. k From the standardized feature matrix Random selection k 1 sample was used as the initial cluster center. C ={ μ 1, μ 2,..., μ k};in, μ j =[ μ j,1 , μj,2 ] T For the first j The center vector of each cluster.

[0045] 2) Sample allocation: Secondly, the embodiments of this application can calculate each sample With each cluster center μ j Euclidean distance: , Sample Assign to the nearest cluster, i.e. , in, c i For the first i The cluster label to which each sample belongs (1≤ c i ≤ k ).

[0046] 3) Update cluster centers: Based on the allocation results, this embodiment of the application can calculate the new center for each cluster: , in, n j For the first j The number of samples in each cluster For the first j The sum of the feature vectors of all samples within a cluster.

[0047] 4) Convergence judgment: Repeat steps 2) and 3) until the change in cluster centers is less than the threshold. δ =10 6 ,Right now , in, t The iteration number represents the number of iterations. At this point, the clustering is stable, and the final cluster partitioning result is output.

[0048] Through the above steps, the embodiments of this application can divide the data into... k Each cluster represents a typical power output mode, laying the foundation for subsequent clustering to identify extreme scenarios and avoiding misjudgments caused by confusion in the definitions of extreme modes.

[0049] Optionally, in one embodiment of this application, calculating the weighted feature value of at least one dimension feature includes: calculating the variance of at least one dimension feature; normalizing the weight of at least one dimension feature based on the variance to obtain features that satisfy a preset normalization condition; and calculating the weighted feature value based on the features.

[0050] The preset normalization condition can be to map each weighted feature value to a fixed interval, or to standardize it with zero mean and unit variance, so that different features can be directly compared and fused on the same scale, thereby ensuring that the comprehensive features have consistent contribution weights in extreme output identification and improving the stability and accuracy of the identification model; it can be set by those skilled in the art according to the actual situation, and no specific restrictions are made here.

[0051] The following is a specific example illustrating the calculation of weighted feature values ​​for at least one dimension of a feature according to an embodiment of this application, which may include the following steps: (1) Input data preparation: First, the embodiments of this application can be obtained using K-means clustering. k Cluster data, taking a single cluster of data as input, denoted as S ={ s 1, s 2,..., s m};in m The number of samples in this cluster. s i Indicates the first i The daily output data. This application embodiment can filter the valid output data and delete invalid values.

[0052] (2) Multi-dimensional feature extraction: This application embodiment can extract five core features from daily output data and construct a feature matrix. It can include: 1) Peak characteristics F i,1 =max( s i ), that is, the first i The maximum output value at all points in time throughout the day; 2) Volatility characteristics F i,2 =std( s i ), that is, the first i Standard deviation of output at all points in time: , in, t The number of times per day. s i,kFor the first i Heavenly k Output value at each point in time For the first i Average daily output.

[0053] 3) Mean characteristics F i,3 = That is, the first i The average daily output; 4) Valley characteristics F i,4 =min( s i ), that is, the first i The minimum output at all points in time throughout the day; 5) Peak-valley difference characteristics F i,5 = F i,1 F i,4 That is, the first i The difference between the peak value and the trough value.

[0054] (3) Feature standardization: Furthermore, embodiments of this application can be applied to the feature matrix. F Standardization, eliminating the influence of dimensions, can be expressed as: , In the formula, For the standardized first i The first sample j Features ( j =1,2,3,4,5). μ j For the first j The mean of each feature; σ j For the first j Standard deviation of each feature; =10 6. To avoid the minimum value where the denominator is 0.

[0055] (4) Feature weight calculation: The embodiments of this application can calculate weights based on feature variance to highlight the extreme signal value of stable features, which may include: 1) Calculate the variance of each feature: , in, For the standardized first j The mean of each feature.

[0056] 2) Weight normalization: , Where wj is the first j The weights of each feature.

[0057] (5) Calculation of weighted eigenvalues: Furthermore, in embodiments of this application, the standardized features can be summed according to their weights to obtain the weighted feature value for each sample: , In the formula, S i For the first i The weighted feature values ​​of each sample. w j For the first j The weights of each feature S i The larger the absolute value, the stronger the extremity.

[0058] In step S103, the clusters are synchronously shrunk and iterated to obtain extreme samples of the clusters that meet the preset screening conditions, and extreme power output scenarios of the new energy power station are identified based on the extreme samples.

[0059] In the embodiments of this application, the synchronous backsliding iterative screening may include, but is not limited to, MV-SBR (Multi-feature Weighted Synchronous Backward Reduction) algorithm, multivariate linear regression, neural network regression, etc., which gradually identify and converge extreme samples by iteratively adjusting feature weights and screening thresholds; it can be set by those skilled in the art according to the actual situation, and no specific restrictions are made here.

[0060] It can be explained that the preset screening conditions include, but are not limited to, output data points that exceed the set threshold or quantile in dimensions such as peak power, fluctuation amplitude, ramp rate, and continuous trough; it can combine time continuity, spatial correlation and historical operation mode to make joint judgment on candidate samples under multiple conditions to ensure that the selected samples truly reflect the extreme output behavior of new energy power plants.

[0061] Extreme samples refer to data points or time periods that show abnormal peaks, violent fluctuations, or continuous lows in renewable energy output data, which are significantly deviating from the normal operating mode. By screening and analyzing these samples, the corresponding extreme output scenarios can be obtained, which can reflect the output characteristics of renewable energy power plants under specific meteorological conditions, abnormal equipment status, or abnormal control strategies, and provide data support and decision-making basis for extreme output identification, dispatching early warning, and safe operation of the power grid.

[0062] Optionally, in one embodiment of this application, the clusters are synchronously shrunk and iteratively screened according to weighted feature values ​​to obtain extreme samples of the clusters that meet preset screening conditions, including: initializing the candidate set of the clusters; calculating the extreme threshold of the samples in the candidate set according to the weighted feature values; and iteratively screening the samples with the smallest extreme threshold based on the extreme threshold and a preset multi-feature weighted synchronous shrunk strategy until the size of the candidate set meets the preset screening conditions to obtain extreme samples.

[0063] Among them, the pre-set multi-feature weighted synchronous shrinkage strategy refers to the process of extracting multi-dimensional features from new energy power output data, forming weighted feature values ​​by assigning different feature weights, and synchronously updating feature weights and screening thresholds (shrinkage operation) during the iterative screening process to gradually highlight the feature information of extreme power output samples. This strategy can make full use of the complementarity of multi-dimensional features, improve the accuracy and robustness of extreme sample identification, and at the same time take into account iterative convergence and computational efficiency, thereby providing reliable data for the identification of extreme power output of new energy.

[0064] Specifically, the embodiments of this application utilize the MV-SBR strategy to screen extreme samples, which may include the following steps; 1) Initialize the candidate set C ={1,2,..., m}, contains all sample indices, and sets extreme sample ratios. α =5%, the number of extreme samples that should be retained K =max(1,[ α m ]); 2) Calculate the extreme thresholds for each sample in the candidate set. T i =∣ S i |; 3) Iteratively remove the least extreme samples to find the current candidate set. Ti Remove the corresponding sample when the minimum value is found, until the candidate set size is reached. K ; 4) Output the samples in the candidate set as the extreme scenarios for this cluster, which can be denoted as... E ={ e 1, e 2,..., e K}

[0065] By applying the MV-SBR strategy in a clustered manner, the embodiments of this application can accurately identify extreme scenarios under different operating modes, avoid misjudgment based on a single standard, and provide targeted basis for risk warning of new energy power plants.

[0066] like Figure 2The following is a specific example to further illustrate the new energy extreme output identification method of this application embodiment, which may include the following steps: In step S201, the raw data is acquired.

[0067] First, the embodiments of this application can collect the output data of new energy power plants, including but not limited to real-time active power, reactive power, voltage, current and other operating parameters, and at the same time collect equipment status information to form a complete dataset.

[0068] In step S202, data cleaning and processing are performed.

[0069] Furthermore, embodiments of this application can perform missing value imputation, outlier removal, time sequence alignment, and standardization on the collected data to ensure data integrity and consistency, facilitating subsequent analysis.

[0070] In step S203, clustering features are extracted.

[0071] Secondly, the embodiments of this application can extract multi-dimensional features from the processed output data, which may include, but are not limited to, peak values, fluctuations, ramp rates, and continuous troughs, and calculate the weighted values ​​of each feature to form a feature matrix.

[0072] In step S204, K-meana clustering is performed.

[0073] The embodiments of this application can use the K-means clustering method to cluster the feature matrix, dividing data with similar output patterns into different clusters to reveal typical operating patterns and potential extreme behaviors.

[0074] In step S205, clustering is performed using MV-SBR.

[0075] Within each cluster, the embodiments of this application can apply the MV-SBR strategy to iteratively screen and identify extreme samples, gradually highlighting extreme output characteristics.

[0076] In step S206, the results are integrated and output.

[0077] Finally, the embodiments of this application can integrate the extreme sample identification results of each cluster to generate a comprehensive extreme power output scenario, and output identification results and analysis reports that can be used for power grid dispatching, early warning strategy formulation and equipment management.

[0078] According to the new energy extreme output identification method proposed in this application, the output data of new energy power stations are first clustered. Then, the clusters are synchronously shrunk and iteratively filtered based on the weighted feature values ​​of at least one dimension feature to identify the extreme output scenarios of new energy power stations. This ensures that the extreme standards are consistent within the same mode, effectively avoiding conflicts in the definition of extremes under different modes. At the same time, it can make full use of multi-dimensional feature information, highlight the characteristics of extreme samples, and avoid misjudgment or omission caused by a single feature. It can effectively solve the one-sidedness problem of a single indicator, improve the accuracy and robustness of extreme output identification, thereby reducing the risk of grid dispatch imbalance and equipment overload, reducing the failure losses caused by extreme scenarios, and providing targeted basis for the operation and maintenance of new energy power stations, thus reducing operation and maintenance costs.

[0079] Next, referring to the accompanying drawings, the new energy extreme output identification device proposed according to the embodiments of this application is described.

[0080] Figure 3 This is a block diagram of a new energy extreme output identification device according to an embodiment of this application.

[0081] like Figure 3 As shown, the new energy extreme output identification device 10 includes: a data acquisition module 100, a calculation module 200, and an identification module 300.

[0082] Among them, the acquisition module 100 is used to collect the output data of the new energy power station.

[0083] The calculation module 200 is used to obtain the clustering of the power output data based on the daily power output peak and fluctuation characteristics of the power output data, and to extract at least one dimension feature of the power output data to calculate the weighted feature value of at least one dimension feature.

[0084] The identification module 300 is used to perform synchronous back-shrinking iterative screening of clusters based on weighted feature values ​​to obtain extreme samples of clusters that meet preset screening conditions, and to identify extreme power output scenarios of new energy power plants based on the extreme samples.

[0085] Optionally, in one embodiment of this application, the calculation module 200 includes: an extraction unit, a processing unit, and a partitioning unit.

[0086] The extraction unit is used to extract the daily peak output and fluctuation characteristics of the output data to obtain the feature matrix of the output data.

[0087] The processing unit is used to process the feature matrix to obtain a feature matrix that meets the preset standardization conditions.

[0088] The partitioning unit is used to cluster and divide the feature matrix to obtain clusters.

[0089] Optionally, in one embodiment of this application, the partitioning unit includes: a setting subunit, an allocation subunit, and a calculation subunit.

[0090] The sub-unit is used to set the number of clusters in the clustering, so as to determine the initial cluster center vector of the feature matrix based on the number of clusters.

[0091] The allocation subunit is used to calculate the distance between the clustered samples and the initial cluster center vector, so as to allocate the samples to the clusters according to the distance to obtain the allocation result.

[0092] The calculation subunit is used to calculate the new cluster centers of the clusters based on the allocation results, until the change in the cluster centers meets the preset iteration stop condition, so as to obtain the clusters.

[0093] Optionally, in one embodiment of this application, the calculation module 200 includes: a first calculation unit, a normalization unit, and a second calculation unit.

[0094] The first calculation unit is used to calculate the variance of at least one dimension feature.

[0095] The normalization unit is used to normalize the weights of at least one dimension of features based on the variance, so as to obtain features that meet the preset normalization conditions.

[0096] The second calculation unit is used to calculate the weighted eigenvalues ​​based on the features.

[0097] Optionally, in one embodiment of this application, the identification module 300 includes: an initialization unit, a third calculation unit, and a filtering unit.

[0098] The initialization unit is used to initialize the candidate set for clustering.

[0099] The third calculation unit is used to calculate the extreme thresholds of samples in the candidate set based on the weighted feature values.

[0100] The filtering unit is used to iteratively filter out the sample with the smallest extreme threshold based on the extreme threshold and the preset multi-feature weighted synchronous shrinkage strategy until the size of the candidate set meets the preset filtering conditions to obtain the extreme sample.

[0101] It should be noted that the foregoing explanation of the embodiment of the new energy extreme output identification method also applies to the new energy extreme output identification device of this embodiment, and will not be repeated here.

[0102] According to the new energy extreme output identification device proposed in this application, the output data of new energy power stations are first clustered. Then, the clusters are synchronously shrunk and iteratively filtered based on the weighted feature values ​​of at least one dimension feature to identify the extreme output scenarios of new energy power stations. This ensures that the extreme standards are consistent within the same mode, effectively avoiding conflicts in the definition of extremes under different modes. At the same time, it can make full use of multi-dimensional feature information, highlight the characteristics of extreme samples, and avoid misjudgment or omission caused by a single feature. It can effectively solve the problem of the one-sidedness of a single indicator, improve the accuracy and robustness of extreme output identification, thereby reducing the risk of grid dispatch imbalance and equipment overload, reducing the failure losses caused by extreme scenarios, and providing targeted basis for the operation and maintenance of new energy power stations, thus reducing operation and maintenance costs.

[0103] Figure 4 A schematic diagram of the structure of an electronic device provided in an embodiment of this application. The electronic device may include: The memory 401, the processor 402, and the computer program stored on the memory 401 and capable of running on the processor 402.

[0104] When the processor 402 executes the program, it implements the new energy extreme output identification method provided in the above embodiments.

[0105] Furthermore, electronic devices also include: Communication interface 403 is used for communication between memory 401 and processor 402.

[0106] The memory 401 is used to store computer programs that can run on the processor 402.

[0107] Memory 401 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0108] If the memory 401, processor 402, and communication interface 403 are implemented independently, then the communication interface 403, memory 401, and processor 402 can be interconnected via a bus to complete communication between them. The bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of representation, Figure 4 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0109] Optionally, in a specific implementation, if the memory 401, processor 402, and communication interface 403 are integrated on a single chip, then the memory 401, processor 402, and communication interface 403 can communicate with each other through an internal interface.

[0110] Processor 402 may be a central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of this application.

[0111] This embodiment also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method for identifying extreme power output of new energy sources.

[0112] This application also provides a computer program product, including a computer program that can run computer instructions. When the computer instructions are executed by a processor, they implement the new energy extreme output identification method provided in this application.

[0113] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0114] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "N" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0115] Any process or method described in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or N executable instructions for implementing custom logic functions or processes, and the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functions involved, as should be understood by those skilled in the art to which embodiments of this application pertain.

[0116] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0117] It should be understood that the various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, the N steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. If implemented in hardware, as in another embodiment, it can be implemented using any one or more of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0118] Those skilled in the art will understand that all or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage medium, and when executed, the program includes one or a combination of the steps of the method embodiments.

[0119] Furthermore, the functional units in the various embodiments of this application can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium.

[0120] The storage medium mentioned above can be a read-only memory, a disk, or an optical disk, etc. Although embodiments of this application have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting this application. Those skilled in the art can make changes, modifications, substitutions, and variations to the above embodiments within the scope of this application.

Claims

1. A method for identifying extreme power output of new energy sources, characterized in that, Includes the following steps: Collect power output data from new energy power plants; Based on the daily peak output and fluctuation characteristics of the output data, the output data is clustered, and at least one dimension feature of the output data is extracted to calculate the weighted feature value of the at least one dimension feature. The clusters are synchronously shrunk and iterated to filter based on the weighted feature values ​​to obtain extreme samples of the clusters that meet the preset filtering conditions, and the extreme power output scenarios of the new energy power station are identified based on the extreme samples.

2. The method according to claim 1, characterized in that, The step of obtaining the clustering of the power output data based on the daily power output peak and fluctuation characteristics of the power output data includes: Extract the daily peak output and fluctuation characteristics of the output data to obtain the feature matrix of the output data; The feature matrix is ​​processed to obtain a feature matrix that meets the preset standardization conditions; The feature matrix is ​​clustered and divided to obtain the clusters.

3. The method according to claim 2, characterized in that, The process of clustering and partitioning the feature matrix to obtain the clusters includes: Set the number of clusters for the clustering, and determine the initial cluster center vector of the feature matrix based on the number of clusters; Calculate the distance between the samples in the cluster and the initial cluster center vector, and assign the samples to the clusters according to the distance to obtain the assignment results; The new cluster centers of the clusters are calculated based on the allocation results until the change in the cluster centers meets the preset iteration stop condition, so as to obtain the clusters.

4. The method according to claim 1, characterized in that, The calculation of the weighted feature value of the at least one dimension feature includes: Calculate the variance of the at least one dimension feature; The weights of the at least one dimension feature are normalized according to the variance to obtain features that satisfy the preset normalization conditions. The weighted feature value is calculated based on the features.

5. The method according to claim 1, characterized in that, The step of synchronously shrinking and iteratively filtering the clusters based on the weighted feature values ​​to obtain extreme samples of the clusters that meet preset filtering conditions includes: Initialize the candidate set of the cluster; Calculate the extreme threshold of the samples in the candidate set based on the weighted feature values; Based on the extreme threshold and the preset multi-feature weighted synchronous shrinkage strategy, the sample with the smallest extreme threshold is iteratively screened until the size of the candidate set meets the preset screening conditions to obtain the extreme sample.

6. A new energy extreme output identification device, characterized in that, include: The data acquisition module is used to collect power output data from renewable energy power plants. The calculation module is used to obtain the clustering of the output data based on the daily output peak and fluctuation characteristics of the output data, and extract at least one dimension feature of the output data to calculate the weighted feature value of the at least one dimension feature; The identification module is used to perform synchronous back-shrinking iterative filtering on the clusters according to the weighted feature values ​​to obtain extreme samples of the clusters that meet the preset filtering conditions, and to identify the extreme output scenarios of the new energy power station based on the extreme samples.

7. The apparatus according to claim 6, characterized in that, The computing module includes: An extraction unit is used to extract the daily peak output and fluctuation characteristics of the output data to obtain the feature matrix of the output data; The processing unit is used to process the feature matrix to obtain a feature matrix that meets preset standardization conditions. A partitioning unit is used to cluster and partition the feature matrix to obtain the clusters.

8. An electronic device, characterized in that, include: The memory, the processor, and the computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the new energy extreme output identification method as described in any one of claims 1-5.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, The program is executed by the processor to implement the new energy extreme output identification method as described in any one of claims 1-5.

10. A computer program product, comprising a computer program, characterized in that, The computer program is executed to implement the new energy extreme output identification method as described in any one of claims 1-5.

Images