Power grid operation scene extraction method and device, electronic equipment and storage medium

Abstract

The invention discloses a power grid operation scene extraction method and device, electronic equipment and a storage medium, and the method comprises the steps: obtaining the operation data of a to-be-extracted power grid, carrying out the preprocessing of the operation data, obtaining the feature data, determining the initial neighborhood radius of each data point in the feature data, and obtaining the initial neighborhood radius of each data point; adjusting the initial neighborhood radius of each data point to obtain the adjusted neighborhood radius of each data point; clustering the feature data to obtain a plurality of clusters, marking noise points, screening out a plurality of target noise points meeting a preset stability condition and a preset scene condition, marking a small probability scene of the plurality of target noise points, optimizing the plurality of clusters, marking an operation scene of each optimized cluster, and obtaining a plurality of target noise points; and determining an actual operation scene of the to-be-extracted power grid according to the small probability scene of the plurality of noise points and the operation scene of each cluster. Therefore, the problem that the accuracy, adaptability and robustness of power grid operation scene extraction are insufficient is solved, the adaptability to a complex power grid is improved, misjudgment of a special small-probability operation mode is reduced, and the accuracy of power grid operation scene extraction is improved.

Description

Technical Field

[0001] The present application relates to the technical field of power system data analysis, and in particular to a method, device, electronic device and storage medium for extracting power grid operation scenarios. Background Art

[0002] With the high proportion of renewable energy sources such as wind power and photovoltaics being connected to the grid, the high volatility and uncertainty of these new energy sources have made it increasingly difficult to balance power on the grid. Furthermore, with the increasing use of ultra-high voltage direct current (UHVDC) large-capacity, long-distance transmission and the expansion of AC / DC hybrid transmission, the complexity and diversity of grid operation scenarios have become prominent, necessitating the integration of heterogeneous data from multiple sources across multiple platforms.

[0003] In related technologies, the extraction of power grid operation scenarios mostly relies on manual experience and static models, and is performed based on offline simulation and fixed rule bases. However, physical models are difficult to characterize the nonlinear characteristics of AC / DC hybrid power grids, and the temporal and spatial correlation mining of massive operating data is limited. It is difficult to adapt to the fluctuations of new energy sources and the real-time operating conditions of cross-regional interconnected power grids, and it is difficult to adapt to dynamic operating modes, resulting in incomplete scenario coverage and delayed response speed. Summary of the Invention

[0004] The present application provides a method, device, electronic device and storage medium for extracting power grid operation scenarios to address the problems that related technologies are difficult to adapt to the complexity and dynamic changes of power grid operation scenarios, and the accuracy, adaptability and robustness of power grid operation scenario extraction are insufficient. The method improves adaptability to complex power grids, reduces misjudgment of special low-probability operating modes, and improves the accuracy of power grid operation scenario extraction.

[0005] The first embodiment of the present application provides a method for extracting a power grid operation scenario, comprising the following steps:

[0006] Obtaining the operating data of the power grid to be extracted;

[0007] Preprocessing the operating data to obtain feature data, determining an initial neighborhood radius of each data point in the feature data, calculating a joint entropy of each data point based on the initial neighborhood radius of each data point, and adjusting the initial neighborhood radius of each data point based on the joint entropy of each data point to obtain an adjusted neighborhood radius of each data point;

[0008] The feature data are clustered using a preset clustering algorithm based on the adjusted neighborhood radius of each data point and a preset minimum number of clusters to obtain a plurality of cluster clusters, the number of points within the adjusted neighborhood radius of each data point is determined, data points with a number of points less than the preset minimum number of clusters are marked as noise points, the neighborhood radius of each noise point is adjusted, and a joint entropy of each noise point is calculated based on the adjusted neighborhood radius of each noise point. Based on the joint entropy of each noise point, a plurality of target noise points that meet a preset stability condition and a preset scenario condition are screened out, and low-probability scenarios of the plurality of target noise points are marked;

[0009] Each cluster is evaluated to obtain an evaluation result of each cluster, and the multiple clusters are optimized according to the evaluation result of each cluster to obtain multiple optimized clusters, and the operation scenario of each optimized cluster is marked, and the actual operation scenario of the power grid to be extracted is determined according to the small probability scenarios of multiple noise points and the operation scenario of each cluster.

[0010] Optionally, in some embodiments, evaluating each cluster to obtain an evaluation result of each cluster, and optimizing the multiple clusters according to the evaluation result of each cluster to obtain multiple optimized clusters, includes:

[0011] Calculate the average entropy of each cluster, and mark the clusters with average entropy greater than the preset segmentation threshold as high entropy clusters;

[0012] Calculate the mutual information between any two clusters, and mark the clusters whose mutual information is greater than the preset merging threshold as cluster pairs to be merged;

[0013] The high entropy clusters are divided to obtain a first optimization result, and the to-be-merged cluster pairs are merged to obtain a second optimization result. The optimized multiple clusters are obtained according to the first optimization result and the second optimization result.

[0014] Optionally, in some embodiments, calculating the joint entropy of each data point according to the initial neighborhood radius of each data point includes:

[0015] Based on a preset first joint entropy calculation formula, the joint entropy of each data point is calculated according to the initial neighborhood radius of each data point, wherein the preset first joint entropy calculation formula is:

[0016]

[0017] Among them, H local (X i ) is the joint entropy of data point i within the initial neighborhood radius, D is the number of features, H(X j |Neighbors(X i)) is the entropy value of the j-th feature within the initial neighborhood radius.

[0018] Optionally, in some embodiments, adjusting the initial neighborhood radius of each data point according to the joint entropy of each data point to obtain the adjusted neighborhood radius of each data point includes:

[0019] Based on a preset first adjustment formula, the initial neighborhood radius of each data point is adjusted according to the joint entropy of each data point to obtain the adjusted neighborhood radius of each data point.

[0020] The preset first adjustment formula is:

[0021]

[0022] Among them, Eps i is the neighborhood radius after adjustment of data point i, Eps init is the initial neighborhood radius, α is the adjustment factor, H max is the maximum normalized entropy value.

[0023] Optionally, in some embodiments, calculating the joint entropy of each noise point according to the adjusted neighborhood radius of each noise point includes:

[0024] Based on a preset second joint entropy calculation formula, the joint entropy of each noise point is calculated according to the adjusted neighborhood radius of each noise point, wherein the preset second joint entropy calculation formula is:

[0025]

[0026] Among them, H noise (x i ) is the joint entropy of the noise point i within the adjusted neighborhood radius, D is the number of features, B j is the number of bins for the jth feature, p jk is the probability of the jth feature in the kth box.

[0027] A second embodiment of the present application provides a device for extracting a power grid operation scenario, including:

[0028] An acquisition module, used to acquire the operation data of the power grid to be extracted;

[0029] a determination module, configured to preprocess the operating data to obtain feature data, determine an initial neighborhood radius of each data point in the feature data, calculate a joint entropy of each data point based on the initial neighborhood radius of each data point, and adjust the initial neighborhood radius of each data point based on the joint entropy of each data point to obtain an adjusted neighborhood radius of each data point;

[0030] a clustering module for clustering the feature data based on the adjusted neighborhood radius of each data point and a preset minimum number of clusters using a preset clustering algorithm to obtain a plurality of cluster clusters, determining the number of points within the adjusted neighborhood radius of each data point, marking data points with a number of points less than the preset minimum number of clusters as noise points, adjusting the neighborhood radius of each noise point, and calculating a joint entropy of each noise point based on the adjusted neighborhood radius of each noise point; based on the joint entropy of each noise point, screening out a plurality of target noise points that meet a preset stability condition and a preset scenario condition, and marking low-probability scenarios for the plurality of target noise points;

[0031] An optimization module is used to evaluate each cluster to obtain an evaluation result of each cluster, optimize the multiple clusters according to the evaluation result of each cluster to obtain multiple optimized clusters, mark the operating scenario of each optimized cluster, and determine the actual operating scenario of the power grid to be extracted based on the low-probability scenarios of multiple noise points and the operating scenario of each cluster.

[0032] Optionally, in some embodiments, the optimization module includes:

[0033] A first calculation unit is used to calculate the average entropy of each cluster, and mark the clusters with an average entropy greater than a preset segmentation threshold as high entropy clusters;

[0034] A second calculation unit is used to calculate the mutual information between any two clusters, and mark the clusters whose mutual information is greater than a preset merging threshold as cluster pairs to be merged;

[0035] An optimization unit is configured to divide the high entropy clusters to obtain a first optimization result, merge the to-be-merged cluster pairs to obtain a second optimization result, and obtain the optimized multiple clusters according to the first optimization result and the second optimization result.

[0036] Optionally, in some embodiments, the determining module includes:

[0037] A third calculation unit is configured to calculate the joint entropy of each data point according to the initial neighborhood radius of each data point based on a preset first joint entropy calculation formula, wherein the preset first joint entropy calculation formula is:

[0038]

[0039] Among them, H local (X i ) is the joint entropy of data point i within the initial neighborhood radius, D is the number of features, H(X j |Neighbors(X i )) is the entropy value of the j-th feature within the initial neighborhood radius.

[0040] Optionally, in some embodiments, the determining module includes:

[0041] The adjustment unit is configured to adjust the initial neighborhood radius of each data point according to the joint entropy of each data point based on a preset first adjustment formula to obtain an adjusted neighborhood radius of each data point.

[0042] The preset first adjustment formula is:

[0043]

[0044] Among them, Eps i is the neighborhood radius after adjustment of data point i, Eps init is the initial neighborhood radius, α is the adjustment factor, H max is the maximum normalized entropy value.

[0045] Optionally, in some embodiments, the clustering module includes:

[0046] The fourth calculation unit is configured to calculate the joint entropy of each noise point according to the adjusted neighborhood radius of each noise point based on a preset second joint entropy calculation formula, wherein the preset second joint entropy calculation formula is:

[0047]

[0048] Among them, H noise (x i ) is the joint entropy of the noise point i within the adjusted neighborhood radius, D is the number of features, B j is the number of bins for the jth feature, p jk is the probability of the jth feature in the kth box.

[0049] The third aspect of the present application provides an electronic device, comprising: 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 method for extracting power grid operation scenarios as described in the above embodiments.

[0050] The fourth aspect of the present application provides a computer-readable storage medium having a computer program stored thereon, which is executed by a processor to implement the method for extracting power grid operation scenarios as described in the above embodiments.

[0051] Therefore, this application has at least the following beneficial effects:

[0052] (1) This application introduces entropy values ​​to dynamically adjust the neighborhood radius parameter in the DBSCAN clustering algorithm. By dynamically adjusting the neighborhood radius, the neighborhood of noise points can be expanded in data-sparse areas, thereby preventing low-density scenes from being misclassified as noise. In data-dense areas, the neighborhood radius can be kept small to avoid over-segmentation of data. This improvement solves the problem that the fixed parameters of the traditional DBSCAN algorithm are sensitive to complex power grid scenarios and significantly improves the algorithm's adaptability to fluctuations in renewable energy output and changes in cross-regional interconnection line power.

[0053] (2) This application adds secondary noise filtering to the scene extraction of the power grid, and adopts a dual verification mechanism of entropy stability verification and variance threshold constraint to ensure the filtering effect of random noise and the high robustness of noise judgment in dynamic scenes, thereby reducing the misjudgment of special low-probability operating modes.

[0054] (3) This application realizes high-entropy cluster segmentation and low-entropy cluster merging by calculating the average entropy evaluation within the cluster, thereby improving the optimization capability of the classification cluster structure and further enhancing the effect of power grid operation scenario extraction.

[0055] Additional aspects and advantages of the present application will be given in part in the description below, and in part will become apparent from the description below, or will be learned through practice of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] The above and / or additional aspects and advantages of the present application will become apparent and easily understood from the following description of the embodiments in conjunction with the accompanying drawings, in which:

[0057] Figure 1 A flowchart of a method for extracting power grid operation scenarios according to an embodiment of the present application;

[0058] Figure 2 A flowchart of a method for extracting power grid operation scenarios according to one embodiment of the present application;

[0059] Figure 3 A schematic diagram of a neighborhood radius adjustment process provided according to an embodiment of the present application;

[0060] Figure 4 Schematic diagram of a block diagram of a device for extracting power grid operation scenarios according to an embodiment of the present application;

[0061] Figure 5 A schematic diagram of the structure of an electronic device provided according to an embodiment of the present application. DETAILED DESCRIPTION

[0062] The following describes in detail embodiments of the present application, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are intended to be used to explain the present application, and should not be construed as limiting the present application.

[0063] The following describes the method, device, electronic device and storage medium for extracting the power grid operation scenario of the embodiment of the present application with reference to the accompanying drawings. In view of the fact that the related technologies mentioned in the above background technology are difficult to adapt to the complexity and dynamic changes of the power grid operation scenario, and the accuracy, adaptability and robustness of the power grid operation scenario extraction are insufficient, the present application provides a method for extracting the power grid operation scenario. In this method, the power grid operation data is first preprocessed and entropy weight feature screening is performed, and then the DBSCAN algorithm is used for clustering. The local neighborhood entropy calculation is introduced to dynamically adjust the neighborhood radius, and then the noise points are filtered twice based on the entropy-density joint criterion. Finally, the clustering results are optimized, and the typical operation mode extraction is completed in combination with the actual power grid operation. As a result, the problem that the related technologies are difficult to adapt to the complexity and dynamic changes of the power grid operation scenario, and the accuracy, adaptability and robustness of the power grid operation scenario extraction are solved, the adaptability to complex power grids is improved, the misjudgment of special low-probability operation modes is reduced, and the accuracy of power grid operation scenario extraction is improved.

[0064] Specifically, Figure 1 A flowchart of a method for extracting power grid operation scenarios provided in an embodiment of the present application.

[0065] like Figure 1 As shown, the method for extracting the power grid operation scenario includes the following steps:

[0066] In step S101 , the operation data of the power grid to be extracted is obtained.

[0067] The operating data may include: the power generation power P of each partition G , load power P L , photovoltaic power generation power P PV , wind power generation power P WD , UHVDC transmission power P DC , Main interconnection line power P Tie At least one of .

[0068] Specifically, combined Figure 2 As shown, the embodiment of the present application can extract the operation data of the power grid through the power company system, for example, the historical time series measurement data of the power grid to be extracted, including: the power generation power P of each partition G , load power P L , photovoltaic power generation power P PV, wind power generation power P WD , UHVDC transmission power P DC , Main interconnection line power P Tie At least one of .

[0069] In step S102, the running data is preprocessed to obtain feature data, and the initial neighborhood radius of each data point in the feature data is determined, and the joint entropy of each data point is calculated based on the initial neighborhood radius of each data point, and the initial neighborhood radius of each data point is adjusted based on the joint entropy of each data point to obtain the adjusted neighborhood radius of each data point.

[0070] Specifically, when obtaining the operation data of the power grid to be extracted, it is necessary to perform data preprocessing on the operation data, wherein the data preprocessing may include the following steps:

[0071] (1) Data standardization processing.

[0072] To avoid the influence of the original feature dimension difference, the running data is normalized by Z-score, and the formula is as follows:

[0073]

[0074] in, is the original eigenvalue, X is the eigenvalue after standardization, μ is the mean of the feature population data, and σ is the standard deviation of the feature population data.

[0075] (2) Feature binning and entropy calculation.

[0076] Converting continuous eigenvalues ​​into discrete variables by using equal-width binning can reduce the interference of noise on the model. Divide the intervals into equal intervals according to the range of each eigenvalue, set the number of bins K, and calculate the interval width:

[0077]

[0078] Among them, X max and X min are the maximum and minimum values ​​of feature X, respectively.

[0079] According to the interval width, the feature data is divided into intervals [X min ,X min +width), [X min +width,X min +2width),…, discretized into K intervals.

[0080] Calculate the information entropy of each feature:

[0081]

[0082] Among them, p(X jk ) is the probability distribution of the j-th feature in the k-th interval.

[0083] (3) Feature screening.

[0084] Set the feature importance threshold H threshold , retain the feature data with entropy values ​​higher than this threshold, wherein the feature importance threshold can be preset by the user, obtained through a limited number of experiments, or obtained through a limited number of computer simulations, and is not specifically limited here. Preferably, H threshold 1.2 is recommended.

[0085] It can be understood that for the power grid operation data after preprocessing and feature screening, the embodiment of the present application can use the DBSCAN algorithm for clustering. DBSCAN (DBSCAN, Density-based Spatial Clustering of Applications with Noise) is a representative density-based clustering algorithm, which can distinguish data into core points, boundary points and noise points according to data density, and can find classes of arbitrary shapes in the data set. The two important parameters of the DBSCAN algorithm are the neighborhood radius Eps and the minimum number of clusters MinPts. The values ​​of these two parameters directly affect the clustering effect. However, a fixed neighborhood radius value is difficult to adapt to the differences in the density of power grid operation data, resulting in low-density operation mode data being easily misjudged as noise data. Therefore, the embodiment of the present application improves the clustering effect and reduces parameter sensitivity by introducing a method of dynamically adjusting the neighborhood radius through neighborhood entropy.

[0086] (1) Calculation of local neighborhood entropy.

[0087] Initialization parameter neighborhood radius Eps init is a random value in the range of [0,1], and the standardized feature data after feature screening is used for each data point X i , calculate its initial neighborhood radius Eps init The joint entropy of all points in:

[0088]

[0089] Among them, D is the number of features after screening, H(X j |Neighbors(X i )) is the entropy value of the jth feature in the neighborhood. The higher the entropy value, the more complex or uneven the data distribution in the neighborhood.

[0090] (2) Dynamically adjust the neighborhood radius Eps i .

[0091] The neighborhood radius is dynamically adjusted based on the joint entropy of all points within the initial neighborhood radius:

[0092]

[0093] Among them, Eps init is the initial neighborhood radius, α is the adjustment factor used to control the sensitivity of the entropy value to the neighborhood radius adjustment, H max is the maximum normalized entropy value, obtained through historical data statistics

[0094] The adjustment process of the neighborhood radius in the embodiment of the present application is as follows: Figure 3 As shown, if H local (X i ) value is high, indicating that the data distribution in this area is complex and the neighborhood radius needs to be expanded to capture more similar points. i Stop adjusting when the rate of change is <1%.

[0095] In step S103, the feature data is clustered based on the adjusted neighborhood radius of each data point and the preset minimum number of clusters through a preset clustering algorithm to obtain multiple cluster clusters, the number of points within the adjusted neighborhood radius of each data point is determined, the data points with a number of points less than the preset minimum number of clusters are marked as noise points, the neighborhood radius of each noise point is adjusted, and the joint entropy of each noise point is calculated based on the adjusted neighborhood radius of each noise point. Based on the joint entropy of each noise point, multiple target noise points that meet the preset stability conditions and preset scene conditions are screened out, and the low-probability scenes of multiple target noise points are marked.

[0096] The preset clustering algorithm is the DBSCAN algorithm, and the preset minimum number of clusters can be pre-set by the user, obtained through a limited number of experiments, or obtained through a limited number of computer simulations, which is not specifically limited here.

[0097] Specifically, the embodiment of the present application performs secondary filtering of noise based on the entropy-density joint criterion.

[0098] (1) DBSCAN algorithm clustering.

[0099] The embodiment of this application uses the dynamically adjusted Eps i The DBSCAN algorithm clustering is performed with a fixed minimum number of clusters MinPts, where the parameter MinPts is generally taken as the feature dimension + 1, that is, D+1.

[0100] If the number of points in the adjusted neighborhood is ≥ MinPts, the point is marked as a core point. Points that do not meet the core point conditions are initially marked as noise points. Noise points may contain real low-probability scenarios of power grid operation mode, so secondary noise filtering is required.

[0101] (2) Dynamic adjustment of the noise point neighborhood radius.

[0102] For each marked noise point x i Dynamically adjust the neighborhood radius based on the density characteristics of power grid operation data:

[0103]

[0104] Dist max =max{Dist(x1,core),Dist(x2,core),…,Dist(x n ,core)};

[0105] Among them, Eps init is the initial neighborhood radius, β is the distance adjustment factor, Dist(x i , nearest core point) is the noise point x i Euclidean distance to the nearest core point, Dist max is the maximum distance from all noise points to the nearest core point in the dataset. The farther the noise point is from the core point, the larger the neighborhood radius of the noise point is to capture possible sparse association patterns.

[0106] (3) Calculation of neighborhood entropy.

[0107] In a dynamic neighborhood, the joint entropy of multidimensional features is calculated to quantify the complexity of the data distribution:

[0108]

[0109] Where D is the number of features, B j is the number of bins for the jth feature, p jk is the probability of the jth feature in the kth box.

[0110] (4) Entropy stability judgment.

[0111] Divide the time series data of the noise points into sliding windows (e.g., 30 minutes per window), calculate the mean and variance of the entropy value within the window, and it is stable if the following conditions are met:

[0112] Stability conditions:

[0113] Among them, μ H is the mean entropy value, is the entropy variance, H′ threshold is the entropy stability threshold, which can be taken as 2.0 according to experience, corresponding to high disorder.

[0114] (5) Filtering rules and output.

[0115] The noise points include real low-probability scenarios of power grid operation, such as sudden changes in renewable energy output, and random noise. Real low-probability scenarios usually have high entropy values ​​and small fluctuations, while random noise entropy values ​​fluctuate violently.

[0116] For noise points marked by the DBSCAN algorithm, if any of the following conditions are met, it is determined to be a low-probability scenario:

[0117] Condition 1: H noise (x i )>H″ threshold And entropy stability is achieved through, H″ threshold The empirical value can be 2.0;

[0118] Condition 2: Match with an existing historical low-probability scenario library, and its DTW (Dynamic Time Warping) similarity must be greater than a preset threshold, which can be set as 0.7 based on experience.

[0119] Noise points with stable high entropy values ​​are marked as low-probability scenes and added to the clustering scene library, while noise points with low entropy values ​​or fluctuations are marked as invalid noise and discarded.

[0120] In step S104, each cluster is evaluated to obtain an evaluation result of each cluster, and multiple clusters are optimized according to the evaluation result of each cluster to obtain multiple optimized clusters, and the operating scenario of each optimized cluster is marked. The actual operating scenario of the power grid to be extracted is determined according to the low-probability scenarios of multiple noise points and the operating scenario of each cluster.

[0121] Optionally, in some embodiments, each cluster is evaluated to obtain an evaluation result of each cluster, and multiple clusters are optimized based on the evaluation result of each cluster to obtain multiple optimized clusters, including: calculating the average entropy of each cluster, marking the clusters with an average entropy greater than a preset segmentation threshold as high entropy clusters; calculating the mutual information of any two clusters, marking the clusters with a mutual information greater than a preset merging threshold as cluster pairs to be merged; splitting the high entropy clusters to obtain a first optimization result, and merging the cluster pairs to be merged to obtain a second optimization result, and obtaining multiple optimized clusters based on the first optimization result and the second optimization result.

[0122] Specifically, the embodiment of the present application performs intra-cluster entropy evaluation on the clustered results. Assuming that m clustering results C are obtained after secondary filtering, m , calculate each cluster C m The average entropy of:

[0123]

[0124] The clustering result C is calculated based on the average entropy of each cluster.m The optimization includes two aspects: splitting high entropy clusters and merging low entropy clusters.

[0125] Split high entropy clusters: If the average entropy of a cluster H cluster >θ split ,θ split is the segmentation threshold, which can be taken as θ according to experience split =2.0, and the K-means algorithm is used to perform secondary segmentation on the cluster result.

[0126] Merge low entropy clusters: If the mutual information MI (C a ,C b )>θ merge ,θ merge is the merging threshold, which can be taken as θ according to experience merge =0.6, the two clusters can be merged into one cluster.

[0127] The embodiment of this application marks the optimized final clusters as typical grid operation scenarios and low-probability scenarios, and combines grid dispatch logs and historical data to confirm the actual operation scenarios corresponding to each cluster. For example, according to the load level of each partition, it can be divided into load peak, load flat, and load valley scenarios; according to the different DC operating powers, it can be divided into DC minimum operation scenarios and DC full power scenarios; according to the wind and solar power generation power, it can be divided into new energy high-probability scenarios and different power generation simultaneous rate scenarios.

[0128] Thus, the grid operation scenario extraction method of the embodiment of the present application mines information from grid time-series data in a data-driven manner, capturing the dynamic correlation between renewable energy output, inter-regional DC, and grid operation status. Based on the DBSCAN clustering algorithm, it adaptively partitions scenarios under different grid operation modes and extracts typical grid operation scenarios. This method can improve the intelligence and refinement of grid operation modes and adapt to the complex operation scenarios of new power systems.

[0129] According to the method for extracting power grid operation scenarios proposed in the embodiment of the present application, the power grid operation data is first preprocessed and entropy weight features are screened. Then, the DBSCAN algorithm is used for clustering. Local neighborhood entropy calculation is introduced to dynamically adjust the neighborhood radius. Then, noise points are filtered twice based on the entropy-density joint criterion. Finally, the clustering results are optimized, and typical operation mode extraction is completed in combination with the actual power grid operation. This solves the problem that related technologies are difficult to adapt to the complexity and dynamic changes of power grid operation scenarios, and the accuracy, adaptability, and robustness of power grid operation scenario extraction are insufficient. This improves adaptability to complex power grids, reduces misjudgments of special low-probability operation modes, and improves the accuracy of power grid operation scenario extraction.

[0130] Next, a device for extracting power grid operation scenarios proposed in an embodiment of the present application will be described with reference to the accompanying drawings.

[0131] Figure 4 It is a block diagram of a device for extracting power grid operation scenarios according to an embodiment of the present application.

[0132] like Figure 4 As shown, the power grid operation scenario extraction device 10 includes: an acquisition module 100, a determination module 200, a clustering module 300 and an optimization module 400.

[0133] The acquisition module 100 is used to acquire the operation data of the power grid to be extracted.

[0134] Determination module 200 is used to preprocess the operating data to obtain feature data, and determine the initial neighborhood radius of each data point in the feature data, and calculate the joint entropy of each data point based on the initial neighborhood radius of each data point, and adjust the initial neighborhood radius of each data point based on the joint entropy of each data point to obtain the adjusted neighborhood radius of each data point.

[0135] The clustering module 300 is used to cluster the feature data based on the adjusted neighborhood radius of each data point and the preset minimum number of clusters using a preset clustering algorithm to obtain multiple cluster clusters, determine the number of points within the adjusted neighborhood radius of each data point, mark the data points with a number of points less than the preset minimum number of clusters as noise points, adjust the neighborhood radius of each noise point, and calculate the joint entropy of each noise point based on the adjusted neighborhood radius of each noise point. Based on the joint entropy of each noise point, screen out multiple target noise points that meet preset stability conditions and preset scenario conditions, and mark low-probability scenarios of multiple target noise points.

[0136] The optimization module 400 is used to evaluate each cluster to obtain an evaluation result of each cluster, optimize multiple clusters according to the evaluation result of each cluster to obtain multiple optimized clusters, mark the operating scenario of each optimized cluster, and determine the actual operating scenario of the power grid to be extracted based on the low-probability scenarios of multiple noise points and the operating scenario of each cluster.

[0137] Optionally, in some embodiments, the optimization module 400 includes: a first calculation unit, a second calculation unit and an optimization unit.

[0138] The first calculation unit is used to calculate the average entropy of each cluster, and mark the clusters with an average entropy greater than a preset segmentation threshold as high entropy clusters.

[0139] The second calculation unit is configured to calculate the mutual information between any two clusters, and mark the clusters whose mutual information is greater than a preset merging threshold as cluster pairs to be merged.

[0140] The optimization unit is used to divide the high entropy cluster to obtain a first optimization result, merge the cluster pairs to be merged to obtain a second optimization result, and obtain multiple optimized clusters according to the first optimization result and the second optimization result.

[0141] Optionally, in some embodiments, the determination module 200 includes: a third computing unit.

[0142] The third calculation unit is configured to calculate the joint entropy of each data point according to the initial neighborhood radius of each data point based on a preset first joint entropy calculation formula, wherein the preset first joint entropy calculation formula is:

[0143]

[0144] Among them, H local (X i ) is the joint entropy of data point i within the initial neighborhood radius, D is the number of features, H(X j |Neighbors(X i )) is the entropy value of the jth feature within the initial neighborhood radius.

[0145] Optionally, in some embodiments, the determination module 200 includes: an adjustment unit.

[0146] The adjustment unit is configured to adjust the initial neighborhood radius of each data point according to the joint entropy of each data point based on a preset first adjustment formula to obtain an adjusted neighborhood radius of each data point.

[0147] The preset first adjustment formula is:

[0148]

[0149] Among them, Eps i is the neighborhood radius after adjustment of data point i, Eps init is the initial neighborhood radius, α is the adjustment factor, H max is the maximum normalized entropy value.

[0150] Optionally, in some embodiments, the clustering module 300 includes: a fourth computing unit.

[0151] The fourth calculation unit is configured to calculate the joint entropy of each noise point according to the adjusted neighborhood radius of each noise point based on a preset second joint entropy calculation formula, wherein the preset second joint entropy calculation formula is:

[0152]

[0153] Among them, H noise (x i) is the joint entropy of the noise point i within the adjusted neighborhood radius, D is the number of features, B j is the number of bins for the jth feature, p jk is the probability of the jth feature in the kth box.

[0154] It should be noted that the above explanation of the embodiment of the method for extracting the power grid operation scenario is also applicable to the device for extracting the power grid operation scenario of this embodiment, and will not be repeated here.

[0155] According to the power grid operation scenario extraction device proposed in the embodiment of the present application, the power grid operation data is first preprocessed and entropy weight feature screened. Then, the DBSCAN algorithm is used for clustering. Local neighborhood entropy calculation is introduced to dynamically adjust the neighborhood radius. Then, noise points are secondary filtered based on the entropy-density joint criterion. Finally, the clustering results are optimized, and the typical operation mode extraction is completed in combination with the actual power grid operation. This solves the problem that related technologies are difficult to adapt to the complexity and dynamic changes of power grid operation scenarios, and the accuracy, adaptability and robustness of power grid operation scenario extraction are insufficient. This improves the adaptability to complex power grids, reduces the misjudgment of special low-probability operation modes, and improves the accuracy of power grid operation scenario extraction.

[0156] Figure 5 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. The electronic device may include:

[0157] Memory 501 , processor 502 , and computer programs stored in the memory 501 and executable on the processor 502 .

[0158] When the processor 502 executes the program, the method for extracting the power grid operation scenario provided in the above embodiment is implemented.

[0159] Furthermore, the electronic device further includes:

[0160] The communication interface 503 is used for communication between the memory 501 and the processor 502 .

[0161] The memory 501 is used to store computer programs that can be run on the processor 502 .

[0162] The memory 501 may include a high-speed RAM (Random Access Memory) memory, and may also include a non-volatile memory, such as at least one disk memory.

[0163] If the memory 501, the processor 502, and the communication interface 503 are implemented independently, the communication interface 503, the memory 501, and the processor 502 can be connected to each other via a bus and communicate with each other. The bus can be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, or an EISA (Extended Industry Standard Architecture) bus. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 5 Only one thick line is used in the diagram, but this does not mean that there is only one bus or one type of bus.

[0164] Optionally, in a specific implementation, if the memory 501, the processor 502 and the communication interface 503 are integrated on a chip, the memory 501, the processor 502 and the communication interface 503 can communicate with each other through an internal interface.

[0165] The processor 502 may be a CPU (Central Processing Unit), or an ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of the present application.

[0166] An embodiment of the present application further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-mentioned method for extracting power grid operation scenarios.

[0167] In the description of this specification, the description with reference to the terms "one embodiment", "some embodiments", "example", "specific example", or "some examples" means that the specific features, structures, materials or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in any one or N embodiments or examples in a suitable manner. In addition, those skilled in the art can combine and combine different embodiments or examples described in this specification and features of different embodiments or examples without contradiction.

[0168] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of technical features indicated. Thus, a feature specified as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of this application, "N" means at least two, for example, two, three, etc., unless otherwise specifically defined.

[0169] Any process or method description in a flowchart or otherwise described herein may be understood to represent a module, fragment or portion of code comprising one or more executable instructions for implementing the steps of a custom logical function or process, and the scope of the preferred embodiments of the present application includes alternative implementations in which functions may be performed out of the order shown or discussed, including performing functions in a substantially simultaneous manner or in reverse order depending on the functions involved, which should be understood by those skilled in the art to which the embodiments of the present application belong.

[0170] It should be understood that various parts of the present application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiment, the N steps or methods can be implemented using software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented using hardware, as in another embodiment, any one of the following technologies known in the art or a combination thereof can be used: a discrete logic circuit having a logic gate circuit for implementing a logic function on a data signal, an application-specific integrated circuit having a suitable combination of logic gate circuits, a programmable gate array, a field programmable gate array, etc.

[0171] Those skilled in the art will understand that all or part of the steps in the method of the above embodiment can be completed by instructing related hardware through a program, and the program can be stored in a computer-readable storage medium. When the program is executed, it includes one or a combination of the steps of the method embodiment.

[0172] Although the embodiments of the present application have been shown and described above, it can be understood that the above embodiments are exemplary and cannot be understood as limitations on the present application. Ordinary technicians in this field can change, modify, replace and modify the above embodiments within the scope of the present application.

Claims

1. A method for extracting power grid operation scenarios, characterized in that: The following steps are involved: Obtaining the operating data of the power grid to be extracted; Preprocessing the operating data to obtain feature data, determining an initial neighborhood radius of each data point in the feature data, calculating a joint entropy of each data point based on the initial neighborhood radius of each data point, and adjusting the initial neighborhood radius of each data point based on the joint entropy of each data point to obtain an adjusted neighborhood radius of each data point; The feature data are clustered using a preset clustering algorithm based on the adjusted neighborhood radius of each data point and a preset minimum number of clusters to obtain a plurality of cluster clusters, the number of points within the adjusted neighborhood radius of each data point is determined, data points with a number of points less than the preset minimum number of clusters are marked as noise points, the neighborhood radius of each noise point is adjusted, and a joint entropy of each noise point is calculated based on the adjusted neighborhood radius of each noise point. Based on the joint entropy of each noise point, a plurality of target noise points that meet a preset stability condition and a preset scenario condition are screened out, and low-probability scenarios of the plurality of target noise points are marked; Each cluster is evaluated to obtain an evaluation result of each cluster, and the multiple clusters are optimized according to the evaluation result of each cluster to obtain multiple optimized clusters, and the operation scenario of each optimized cluster is marked, and the actual operation scenario of the power grid to be extracted is determined according to the small probability scenarios of multiple noise points and the operation scenario of each cluster.

2. The method according to claim 1, characterized in that The step of evaluating each cluster to obtain an evaluation result of each cluster, and optimizing the plurality of clusters according to the evaluation result of each cluster to obtain a plurality of optimized clusters, includes: Calculate the average entropy of each cluster, and mark the clusters with average entropy greater than the preset segmentation threshold as high entropy clusters; Calculate the mutual information between any two clusters, and mark the clusters whose mutual information is greater than the preset merging threshold as cluster pairs to be merged; The high entropy clusters are divided to obtain a first optimization result, and the to-be-merged cluster pairs are merged to obtain a second optimization result. The optimized multiple clusters are obtained according to the first optimization result and the second optimization result.

3. The method according to claim 1, characterized in that Calculating the joint entropy of each data point according to the initial neighborhood radius of each data point includes: Based on a preset first joint entropy calculation formula, the joint entropy of each data point is calculated according to the initial neighborhood radius of each data point, wherein the preset first joint entropy calculation formula is: Among them, H local (X i ) is the joint entropy of data point i within the initial neighborhood radius, D is the number of features, H(X j |Neighbors(X i )) is the entropy value of the j-th feature within the initial neighborhood radius.

4. The method according to claim 1, wherein The adjusting the initial neighborhood radius of each data point according to the joint entropy of each data point to obtain the adjusted neighborhood radius of each data point includes: Based on a preset first adjustment formula, the initial neighborhood radius of each data point is adjusted according to the joint entropy of each data point to obtain the adjusted neighborhood radius of each data point, wherein, The preset first adjustment formula is: Among them, Eps i is the neighborhood radius after adjustment of data point i, Eps init is the initial neighborhood radius, α is the adjustment factor, H max is the maximum normalized entropy value.

5. The method according to claim 1, wherein The calculating the joint entropy of each noise point according to the adjusted neighborhood radius of each noise point includes: Based on a preset second joint entropy calculation formula, the joint entropy of each noise point is calculated according to the adjusted neighborhood radius of each noise point, wherein the preset second joint entropy calculation formula is: Among them, H noise (x i ) is the joint entropy of the noise point i within the adjusted neighborhood radius, D is the number of features, B j is the number of bins for the jth feature, p jk is the probability of the jth feature in the kth box.

6. A device for extracting power grid operation scenarios, characterized in that: include: An acquisition module, used to acquire the operation data of the power grid to be extracted; a determination module, configured to preprocess the operating data to obtain feature data, determine an initial neighborhood radius of each data point in the feature data, calculate a joint entropy of each data point based on the initial neighborhood radius of each data point, and adjust the initial neighborhood radius of each data point based on the joint entropy of each data point to obtain an adjusted neighborhood radius of each data point; a clustering module for clustering the feature data based on the adjusted neighborhood radius of each data point and a preset minimum number of clusters using a preset clustering algorithm to obtain a plurality of cluster clusters, determining the number of points within the adjusted neighborhood radius of each data point, marking data points with a number of points less than the preset minimum number of clusters as noise points, adjusting the neighborhood radius of each noise point, and calculating a joint entropy of each noise point based on the adjusted neighborhood radius of each noise point; based on the joint entropy of each noise point, screening out a plurality of target noise points that meet a preset stability condition and a preset scenario condition, and marking low-probability scenarios for the plurality of target noise points; An optimization module is used to evaluate each cluster to obtain an evaluation result of each cluster, optimize the multiple clusters according to the evaluation result of each cluster to obtain multiple optimized clusters, mark the operating scenario of each optimized cluster, and determine the actual operating scenario of the power grid to be extracted based on the low-probability scenarios of multiple noise points and the operating scenario of each cluster.

7. The device according to claim 6, characterized in that The optimization module includes: A first calculation unit is used to calculate the average entropy of each cluster, and mark the clusters with an average entropy greater than a preset segmentation threshold as high entropy clusters; A second calculation unit is used to calculate the mutual information between any two clusters, and mark the clusters whose mutual information is greater than a preset merging threshold as cluster pairs to be merged; An optimization unit is configured to divide the high entropy clusters to obtain a first optimization result, merge the to-be-merged cluster pairs to obtain a second optimization result, and obtain the optimized multiple clusters according to the first optimization result and the second optimization result.

8. The device according to claim 6, characterized in that The determining module includes: A third calculation unit is configured to calculate the joint entropy of each data point according to the initial neighborhood radius of each data point based on a preset first joint entropy calculation formula, wherein the preset first joint entropy calculation formula is: Among them, H local (X i ) is the joint entropy of data point i within the initial neighborhood radius, D is the number of features, H(X j |Neighbors(X i )) is the entropy value of the j-th feature within the initial neighborhood radius.

9. An electronic device, characterized in that: include: 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 method for extracting a power grid operation scenario according to any one of claims 1 to 5.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: The program is executed by a processor to implement the method for extracting a power grid operation scenario as described in any one of claims 1 to 5.

Images