Method and device for extracting fast climbing and switching scene points of electric power system

Abstract

The invention provides a rapid climbing and switching scene point extraction method and device for a power system, and belongs to the field of data processing. The method comprises the following steps: acquiring load data of a target power system within a preset time; determining a daily periodic feature and a local difference of the load data by adopting an adaptive periodic local feature anomaly detection algorithm APLAD, and removing abnormal data in the load data based on the daily periodic feature and the local difference to obtain preprocessed load data; reconstructing the preprocessed load data to obtain a plurality of groups of reconstructed data; obtaining multiple groups of fusion feature data according to the multiple groups of reconstruction data; and based on the multiple groups of fusion feature data, obtaining a rapid climbing and switching scene point of the target power system. According to the method, the problem of low accuracy of rapid climbing and switching scene point extraction caused by poor quality of used data in a traditional method can be solved.

Description

Technical Field

[0001] The present invention relates to the field of data processing technology, and in particular to a method and device for extracting fast ramping and switching scene points of an electric power system. Background Art

[0002] As an important part of the new energy system, the Energy Internet integrates a large number of distributed renewable energy generation. It has the advantages of immediate use, environmental protection and high efficiency, and is considered to be the development trend of the future energy system. However, with the high proportion of renewable energy access, the Energy Internet faces many challenges, especially the randomness, volatility and intermittency of renewable energy, which has brought great pressure to the safe and stable operation of the energy system. In this context, effective measures need to be taken to face these challenges.

[0003] To solve this problem, the power system needs to have the ability to quickly climb, which requires accurate determination of the fast climbing and switching scene points. However, due to the existence of low-quality and erroneous data, the accuracy of the extraction of fast climbing and switching scene points is often affected, so it is necessary to finely clean the individual data to improve the data quality. However, the inventors found that the current methods usually extract wind power data on an hourly basis when extracting data, ignoring long-term meteorological changes and system operation trends, and insufficient consideration of the data processing process, resulting in poor data quality, which in turn affects the accuracy of the fast climbing and switching scene points. Summary of the invention

[0004] The embodiment of the present invention provides a method and device for extracting fast ramping and switching scene points in an electric power system, so as to solve the problem of low accuracy of extracting fast ramping and switching scene points in an electric power system at the current stage.

[0005] In a first aspect, an embodiment of the present invention provides a method for extracting fast ramping and switching scene points of a power system, comprising:

[0006] Obtaining load data of the target power system within a preset time;

[0007] Adopting the adaptive periodic local feature anomaly detection algorithm APLAD, the daily periodic characteristics and local differences of the load data are determined, and based on the daily periodic characteristics and local differences, the abnormal data in the load data is removed to obtain the preprocessed load data;

[0008] Reconstructing the preprocessed load data to obtain multiple groups of reconstructed data; and obtaining multiple groups of fusion feature data based on the multiple groups of reconstructed data;

[0009] Based on multiple sets of fused feature data, the rapid ramping and switching scenario points of the target power system are obtained.

[0010] In a possible implementation, based on daily periodic characteristics and local differences, removing abnormal data from load data includes:

[0011] For any data in the load data, a comprehensive score of the load data is obtained according to the daily periodic characteristics and local differences corresponding to the load data;

[0012] When the comprehensive score of the load data is greater than the preset abnormal point threshold, the load data is removed as abnormal data.

[0013] In a possible implementation, the calculation formula of the comprehensive score is:

[0014]

[0015] Among them, ξ and δ are adjustment coefficients respectively; α is the local difference; β is the daily periodicity characteristic.

[0016] In a possible implementation, multiple sets of fusion feature data are obtained according to multiple sets of reconstructed data, including:

[0017] For any set of reconstructed data, calculate the time domain characteristics, statistical characteristics and periodic characteristics of the set of reconstructed data;

[0018] Extracting short-time local features and long-time trend features corresponding to the group of reconstructed data according to the time domain features, statistical features and periodic features of the group of reconstructed data;

[0019] The short-term local features and the long-term trend features corresponding to the group of reconstructed data are weightedly fused to obtain fused feature data corresponding to the group of reconstructed data.

[0020] In a possible implementation, based on multiple sets of fused feature data, the rapid ramping and switching scenario points of the target power system are obtained, including:

[0021] An improved density-based clustering algorithm DBSCAN is used to cluster multiple groups of fusion features to obtain the rapid climbing and switching scenario points of the target power system; among them, the improved DBSCAN introduces the giant water lily optimization algorithm into the traditional DBSCAN to adaptively adjust the numerical differences and changing trends of the data during the calculation process.

[0022] In a possible implementation, an improved density-based clustering algorithm DBSCAN is used to cluster multiple groups of fusion features to obtain the rapid ramping and switching scenario points of the target power system, including:

[0023] The improved density-based clustering algorithm DBSCAN is used to cluster multiple groups of fusion features to obtain multiple clustering points;

[0024] For any cluster point, determine all data points of the cluster point within the preset distance threshold range; if the number of data points is greater than the preset number, establish the cluster point and all data points within the preset distance threshold range as new cluster labels; if the number of data points is not greater than the preset number, establish the cluster point and all data points within the preset distance threshold range as fast climbing and switching scene point labels;

[0025] Traverse all cluster points and visit all data points within the preset distance threshold range corresponding to these cluster points; if there are data points without assigned labels, determine whether the data point is a cluster point, if not, assign the data point the label corresponding to its cluster point; if so, explore the objects that can be reached by its density until no further exploration is possible, and obtain a complete cluster, and return to the step of determining all data points within the preset distance threshold range for any cluster point, until there are no data points without assigned labels, and output the clustering result;

[0026] Based on the clustering results, the rapid ramping and switching scenario points of the target power system are obtained.

[0027] In a possible implementation, after obtaining the rapid ramping and switching scene points of the target power system based on the fused feature data, the method further includes:

[0028] Reversely map the rapid ramping and switching scene points to the corresponding time series or observation intervals in the load data;

[0029] According to the index operation, the position of each fast ramping and switching scene point in the load data is located, and the multi-dimensional features and time series information of each fast ramping and switching scene point are restored.

[0030] In a possible implementation, after removing abnormal data from the load data based on the daily periodic characteristics and the local differences, the method further includes:

[0031] The load data after removing abnormal data is used to train the data filling model to fill in the missing values ​​in the load data to obtain the preprocessed load data.

[0032] In a possible implementation, the daily periodicity characteristic is calculated as follows:

[0033]

[0034] Where n is the number of load data at the same time of day; x i (h) is the load data corresponding to time point h; x t is the standardized value of the load data at time point t;

[0035] The local difference is calculated as:

[0036]

[0037] Where k is the length of the local change window; x c (i) is the load data at time point i.

[0038] In a second aspect, an embodiment of the present invention provides a device for extracting fast ramping and switching scene points of a power system, comprising:

[0039] A collection module, used to obtain load data of the target power system within a preset time;

[0040] A processing module is used to determine the daily periodic characteristics and local differences of the load data by using an adaptive periodic local feature anomaly detection algorithm APLAD, and remove abnormal data in the load data based on the daily periodic characteristics and local differences to obtain pre-processed load data;

[0041] An extraction module is used to reconstruct the preprocessed load data to obtain multiple groups of reconstructed data; and obtain multiple groups of fusion feature data based on the multiple groups of reconstructed data;

[0042] The clustering module is used to obtain the rapid ramping and switching scenario points of the target power system based on multiple sets of fused feature data.

[0043] The embodiment of the present invention provides a method and device for extracting fast ramping and switching scene points of an electric power system. By analyzing the daily periodic characteristics and local differences of load data through an adaptive periodic local feature anomaly detection algorithm (Adaptive Periodic-Local Anomaly Detection, APLAD), it is possible to analyze the short-term volatility and periodic deviation in the load data, accurately capture the fluctuation law of the load data, remove the abnormal data in the load data, and preliminarily ensure the accuracy of fast ramping and switching scene point extraction. Then, considering that the collected load data at this stage is mostly wind power data extracted on an hourly basis, the embodiment of the present invention reconstructs the preprocessed load data, and extracts fused feature data from the reconstructed data that can reflect long-term meteorological changes and system operation trends, and uses the fused feature data to determine the corresponding fast ramping and switching scene points, which can overcome the problem of low accuracy of fast ramping and switching scene point extraction caused by poor data quality in traditional methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0045] Figure 1 is a flowchart of a method for extracting fast ramping and switching scene points of a power system provided by an embodiment of the present invention;

[0046] Figure 2 is the k-distance graph for finding eps parameters;

[0047] Figure 3 is a schematic diagram of clustering results of a method for extracting fast ramping and switching scene points of a power system provided by an embodiment of the present invention;

[0048] Figure 4 is a flowchart of a method for extracting fast ramping and switching scene points of a power system provided by another embodiment of the present invention;

[0049] Figure 5 It is a structural schematic diagram of a device for extracting fast ramping and switching scene points of a power system provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0050] In the following description, specific details such as specific system structures, technologies, etc. are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present invention. However, it should be clear to those skilled in the art that the present invention may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present invention.

[0051] In order to make the purpose, technical solutions and advantages of the present invention more clear, specific embodiments will be described below in conjunction with the accompanying drawings.

[0052] Figure 1 The flowchart of the method for extracting the rapid ramping and switching scene points of the power system provided by the embodiment of the present invention is as follows. Figure 1 As shown, the method includes:

[0053] Step 110: Obtain load data of the target power system within a preset time.

[0054] In this embodiment, the load data within the preset time period may be wind power load data in a medium or long time scale. The load data within the preset time period may be time series data.

[0055] Step 120: Adopting the adaptive periodic local feature anomaly detection algorithm APLAD, determine the daily periodic features and local differences of the load data, and remove the abnormal data in the load data based on the daily periodic features and local differences to obtain the preprocessed load data.

[0056] In this embodiment, APLAD can extract the daily periodic characteristics of load data to capture the periodic fluctuations in the load data. In order to capture local differences more accurately, the load data on the time series can be smoothed using a sliding filter with a preset sliding window to calculate the difference between each load data and its neighborhood, thereby obtaining the local difference of each load data. Then, based on the analyzed daily periodic characteristics and local differences, the regular fluctuations of the load data are analyzed, and the degree of difference of the load data collected at each collection point is quantified.

[0057] Accordingly, in an optional embodiment, for any load data, its daily periodicity characteristics can be calculated by the following formula:

[0058]

[0059] Where n is the number of load data at the same time of day; x i (h) is the load data corresponding to time point h; x t is the standardized value of the load data at time point t;

[0060] For any load data, its local difference can be calculated by the following formula:

[0061]

[0062] Where k is the length of the local change window; x c (i) is the load data at time point i.

[0063] After determining the daily periodic characteristics and local differences corresponding to each load data, the abnormal data can be removed by the following steps:

[0064] For any load data, a comprehensive score of the load data is obtained according to the daily periodic characteristics and local differences corresponding to the load data.

[0065] When the comprehensive score of the load data is greater than the preset abnormal point threshold, the load data is removed as abnormal data.

[0066] In this embodiment, the comprehensive score of any load data can be calculated by the following formula:

[0067]

[0068] Among them, ξ and δ are adjustment coefficients, which can be dynamically adjusted according to local differences and daily periodic characteristics; α is the local difference; β is the daily periodic characteristics.

[0069] The threshold of the preset outlier point can be determined by the following formula:

[0070] θ=P(ρ,100×(1-p))

[0071] Among them, p is the proportion of outliers.

[0072] When the comprehensive score of any load data is greater than the preset abnormal point threshold, the load data can be considered as abnormal data and the load data is removed.

[0073] In this embodiment, after removing the abnormal data from the load data according to the method provided in step 120, the load data after the abnormal data is removed can be obtained. Considering that in addition to the abnormal data, there may also be missing data in the acquired data, therefore, after removing the abnormal data, the following steps may also be included:

[0074] The load data after removing abnormal data is used to train the data filling model to fill in the missing values ​​in the load data to obtain the preprocessed load data.

[0075] In this embodiment, the data filling model may be a Light Gradient Boosting Machine (LightGBM). The LightGBM model can be used to predict missing values ​​in load data and fill missing load data with predicted data to improve the quality of load data.

[0076] Step 130: reconstruct the preprocessed load data to obtain multiple groups of reconstructed data; and obtain multiple groups of fusion feature data based on the multiple groups of reconstructed data.

[0077] In this embodiment, considering that the current method usually extracts wind power data on an hourly basis and ignores long-term meteorological changes and system operation trends when extracting data, this embodiment reconstructs the load data after mathematical processing. Specifically, the pre-processed load data in the time series can be divided and reconstructed in units of weeks to obtain multiple groups of reconstructed data.

[0078] Then, since the use of APLAD to pre-process the load data will bring short-term noise impact to a single load data, in order to eliminate this impact, this embodiment further processes multiple groups of data to obtain multiple groups of fused feature data.

[0079] Accordingly, in an optional embodiment, obtaining multiple groups of fusion feature data according to multiple groups of reconstruction data in step 130 may include:

[0080] For any set of reconstructed data, the time domain features, statistical features and periodic features of the set of reconstructed data are calculated.

[0081] According to the time domain characteristics, statistical characteristics and periodic characteristics of the group of reconstructed data, the short-term local characteristics and long-term trend characteristics corresponding to the group of reconstructed data are extracted.

[0082] The short-term local features and the long-term trend features corresponding to the group of reconstructed data are weightedly fused to obtain fused feature data corresponding to the group of reconstructed data.

[0083] In this embodiment, for each set of reconstructed data, its maximum value, minimum value, change, fluctuation range and peak value can be calculated as statistical features; its amplitude, frequency, period and waveform can be calculated as statistical features; its main frequency cost and its corresponding amplitude can be extracted through Fourier transform as periodic features. Based on these features, the structure of the load data can be simplified, and short-term fluctuations and noise can be smoothed, while important time series and frequency domain information can be retained. It can not only reduce the short-term noise impact of the APLAD algorithm on the load data, but also highlight the long-term trend and rapid climbing and switching scene points, and fully reflect the characteristics of weekly load data.

[0084] In this embodiment, in order to obtain the short-time local features and long-time trend features of each group of reconstructed data, the time domain features, statistical features and periodic features corresponding to each group of reconstructed data can be input into two convolution modules with different sizes of convolution kernels for convolution to extract the short-time local features and long-time trend features of each group of reconstructed data, and then different weights are assigned to the short-time local features and long-time trend features of each group of reconstructed data, and weighted fusion is performed to obtain multiple groups of fused features. Among them, the obtained fused features can reveal the potential patterns and structures in the load data.

[0085] Step 140: Based on multiple sets of fused feature data, rapid ramping and switching scenario points of the target power system are obtained.

[0086] In this embodiment, a clustering method may be used to cluster multiple groups of fused feature data to obtain rapid ramping and switching scene points of the target power system.

[0087] Accordingly, in an optional embodiment, an improved density-based clustering algorithm (Density-Based Spatial Clustering of Applications with Noise, DBSCAN) may be selected for clustering.

[0088] In this embodiment, the improved DBSCAN algorithm introduces the giant water lily optimization algorithm into the traditional DBSCAN. The giant water lily optimization algorithm can dynamically adjust the numerical difference and change trend of the data in the calculation process.

[0089] Accordingly, in a specific embodiment, obtaining the rapid ramping and switching scene points of the target power system based on multiple sets of fused feature data in step 130 may include:

[0090] The improved density-based clustering algorithm DBSCAN is used to cluster multiple groups of fusion features to obtain multiple clustering points.

[0091] For any cluster point, determine all data points of the cluster point within the preset distance threshold range; if the number of data points is greater than the preset number, establish the cluster point and all data points within the preset distance threshold range as new cluster labels; if the number of data points is not greater than the preset number, establish the cluster point and all data points within the preset distance threshold range as fast climbing and switching scene point labels.

[0092] Traverse all cluster points and visit all data points within the preset distance threshold corresponding to these cluster points; if there are data points without assigned labels, determine whether the data point is a cluster point. If not, assign the data point the label corresponding to its cluster point; if so, explore the objects whose density can be reached until no further exploration is possible, and obtain a complete cluster. Return to the step of determining all data points within the preset distance threshold for any cluster point until there are no data points without assigned labels, and then output the clustering result.

[0093] Based on the clustering results, the rapid ramping and switching scenario points of the target power system are obtained.

[0094] In this embodiment, the improved DBSCAN algorithm can be used to select a central point to divide multiple groups of fused feature data, obtain active sets and candidate sets, and cluster only the active sets to avoid global neighborhood search for all fused feature data, thereby reducing computational complexity. The preset distance threshold is eps, and the preset number is MinPts parameter.

[0095] Among them, the activity set can be expressed as:

[0096] κ={q|q∈D,|qp|≤n·eps}

[0097] Wherein, D represents any set of fused feature data, P represents the center point, q represents any data in the set of fused feature data; n represents the distance between the data and the center point, and eps represents the maximum distance.

[0098] The candidate set is the data points that are far away from the current operation area and do not need to be calculated temporarily, which can be expressed as:

[0099] τ=D-κ.

[0100] The giant water lily algorithm can balance global search and local optimization by simulating the growth and competition mechanism of water lily leaves while searching extensively, avoiding falling into local optimality and accelerating the convergence process, so that the model can flexibly adapt to data characteristics, thereby improving clustering effect and overall stability. Therefore, combined with the position difference of data points in space and their directional similarity, the relative importance of the position difference and directional similarity corresponding to the data points in the candidate set is dynamically adjusted through the giant water lily optimization algorithm, and the parameters in the algorithm are optimized to make it adaptive to data with different density distributions. The weight is determined by calculating the numerical difference and change trend corresponding to each data point. The larger the numerical difference, the more the clustering algorithm tends to emphasize the position difference of the data; and the more similar the change trend is, the more the clustering algorithm pays attention to the directional consistency of the data. On this basis, the global exploration and local development capabilities of the giant water lily optimization algorithm are combined to enhance this dynamic adjustment process.

[0101] Specifically, in the clustering process, the corresponding maximum distance parameter eps and threshold parameter MinPts may be obtained based on the features after each clustering.

[0102] in, Figure 2 is the k-distance graph for finding the eps parameters; Figure 2 As shown, the horizontal axis represents the sample data volume (that is, the number of data points), and the vertical axis represents the corresponding distance. As the number of samples increases, the corresponding distance also increases.

[0103] Through clustering, multiple cluster points are obtained. For any cluster point, all data points of the cluster point within the EPS range are determined, and the number of all data points found is compared with the MinPts parameter. If the number of data points is greater than the MinPts parameter, the cluster point and all data points within the EPS range are established as new cluster labels; if the number of data points is not greater than the MinPts parameter, the cluster point and all data points within the EPS range are established as fast climbing and switching scene point labels.

[0104] Traverse all cluster points and visit all data points within the eps range of these cluster points; determine whether these data points are assigned corresponding labels. If not, it is necessary to determine whether the data points without assigned labels are cluster points. If it is not a cluster point, assign the label of the cluster point corresponding to the data point to the data. If this point is a cluster point, use a recursive exploration method to explore the objects that can be reached by its density until it can no longer be explored, and obtain a complete cluster.

[0105] Repeat the above steps and continue to select data points that have not been visited until there are no data points without assigned labels. Then output the clustering results and use the clusters obtained by clustering as fast climbing and switching scene points.

[0106] In an optional embodiment, after obtaining the rapid ramping and switching scene points of the target power system based on the fused feature data, the following may also be included:

[0107] Reversely map the rapid ramping and switching scene points to the corresponding time series or observation intervals in the load data.

[0108] According to the index operation, the position of each fast ramping and switching scene point in the load data is located, and the multi-dimensional features and time series information of each fast ramping and switching scene point are restored.

[0109] In this embodiment, according to the fast ramping and switching scene points obtained by clustering, the data corresponding to these points are reversely mapped to the original data, that is, the corresponding time series or observation interval in the load data obtained in step 110. Through precise indexing operations, the specific position of each fast ramping and switching scene point in the original data is located, and its original multi-dimensional features and time series information are restored, ensuring that the tracing of the fast ramping and switching scenes is not limited to the summary data level, but traced back to the complete fine-grained data, which is convenient for further analysis and verification.

[0110] In order to verify the effectiveness of the method provided by the embodiment of the present invention, the present embodiment conducted the following experiment: 78,888 load data from January 1, 2006 to December 31, 2014 in the GEFCom2014 dataset were used for simulation test. The operations in step 120 to step 140 were performed on the load data, and the clustering results obtained were as follows: Figure 3 As shown, each graphic represents a data point, and graphics of different shapes correspond to different labels:

[0111] After clustering, 6 clusters and 48 fast climbing and switching scenes far away from the clusters appeared. Finally, the scenes were extracted and analyzed. The fast climbing and switching scenes showed temporal and seasonal patterns. Data analysis found that they were mainly concentrated in January, February, June, July and August. At the same time, the extracted fast climbing and switching scenes had obvious differences from the normal load, indicating that the model achieved good results. The present invention can realize the extraction of fast climbing and switching scenes in medium and long time scales, and can be used in practical engineering applications.

[0112] Figure 4 FIG. 1 is a flowchart of a method for extracting fast ramping and switching scene points of a power system provided by another embodiment of the present invention. Figure 4 As shown, the method may include:

[0113] Collect and input historical load data of the power system within a preset time period.

[0114] APLAD is used to analyze the daily periodic characteristics and short-term trend deviations of load data, dynamically model its regular fluctuations, and quantify the degree of abnormality at each time point. Through sensitivity control, short-term fluctuations and periodic deviations in load data are analyzed, and characteristic changes under different time scales are combined to accurately capture and locate abnormal points, ensure the adaptability and flexibility of anomaly detection, and ultimately remove abnormal data points.

[0115] Use the LightGBM model to predict and fill missing values ​​to complete data cleaning.

[0116] The preprocessed data are summarized by week, and the time domain characteristics and statistical characteristics of the weekly data are extracted to describe the overall trend of load changes. At the same time, the weekly data is analyzed in the frequency domain to reveal the periodic laws and implicit frequency characteristics in the data. These features are input into two convolution modules with different convolution kernel sizes for weighted fusion, and local features and long-term trend features are extracted respectively. Further clustering analysis is then performed, and the k-distance graph is drawn to find the eps parameter of the DBSCAN model algorithm.

[0117] The improved DBSCAN is used to divide the data set into an active set and a candidate set to reduce the computational complexity. The relative importance of the data points in the candidate set is dynamically adjusted through the giant water lily optimization algorithm, combining the position differences of the data points in space and their directional similarities. The parameters in the algorithm are optimized to make it adaptive to data with different density distributions.

[0118] When the label of the clustered data point is -1, that is, all data points have corresponding labels established, the set of fast climbing and switching scene points is restored according to the data point labels.

[0119] For the contents that are not described in detail in this embodiment, reference may be made to other related embodiments mentioned above.

[0120] It should be understood that the order of execution of the steps in the above embodiment does not necessarily mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present invention.

[0121] The following is an embodiment of the device of the present invention. For details not described in detail therein, reference may be made to the corresponding method embodiment described above.

[0122] Figure 5 The schematic diagram of the structure of the fast climbing and switching scene point extraction device of the power system provided by the embodiment of the present invention is shown. For the convenience of explanation, only the part related to the embodiment of the present invention is shown, which is described in detail as follows:

[0123] like Figure 5 As shown, the fast climbing and switching scene point extraction device 5 of the power system includes:

[0124] The acquisition module 51 is used to obtain the load data of the target power system within a preset time;

[0125] The processing module 52 is used to use an adaptive periodic local feature anomaly detection algorithm APLAD to determine the daily periodic characteristics and local differences of the load data, and remove abnormal data in the load data based on the daily periodic characteristics and local differences to obtain pre-processed load data;

[0126] The extraction module 53 is used to reconstruct the preprocessed load data to obtain multiple groups of reconstructed data; and obtain multiple groups of fusion feature data according to the multiple groups of reconstructed data;

[0127] The clustering module 54 is used to obtain the rapid ramping and switching scene points of the target power system based on multiple groups of fused feature data.

[0128] In a possible implementation, the processing module 52 is specifically configured to:

[0129] For any data in the load data, a comprehensive score of the load data is obtained according to the daily periodic characteristics and local differences corresponding to the load data;

[0130] When the comprehensive score of the load data is greater than the preset abnormal point threshold, the load data is removed as abnormal data.

[0131] In a possible implementation, the calculation formula of the comprehensive score is:

[0132]

[0133] Among them, ξ and δ are adjustment coefficients respectively; α is the local difference; β is the daily periodicity characteristic.

[0134] In a possible implementation, the extraction module 53 is specifically configured to:

[0135] For any set of reconstructed data, calculate the time domain characteristics, statistical characteristics and periodic characteristics of the set of reconstructed data;

[0136] Extracting short-time local features and long-time trend features corresponding to the group of reconstructed data according to the time domain features, statistical features and periodic features of the group of reconstructed data;

[0137] The short-term local features and the long-term trend features corresponding to the group of reconstructed data are weightedly fused to obtain fused feature data corresponding to the group of reconstructed data.

[0138] In a possible implementation, the clustering module 54 is specifically configured to:

[0139] An improved density-based clustering algorithm DBSCAN is used to cluster multiple groups of fusion features to obtain the rapid climbing and switching scenario points of the target power system; among them, the improved DBSCAN introduces the giant water lily optimization algorithm into the traditional DBSCAN to adaptively adjust the numerical differences and changing trends of the data during the calculation process.

[0140] In a possible implementation, the clustering module 54 is specifically configured to:

[0141] The improved density-based clustering algorithm DBSCAN is used to cluster multiple groups of fusion features to obtain multiple clustering points;

[0142] For any cluster point, determine all data points of the cluster point within the preset distance threshold range; if the number of data points is greater than the preset number, establish the cluster point and all data points within the preset distance threshold range as new cluster labels; if the number of data points is not greater than the preset number, establish the cluster point and all data points within the preset distance threshold range as fast climbing and switching scene point labels;

[0143] Traverse all cluster points and visit all data points within the preset distance threshold range corresponding to these cluster points; if there are data points without assigned labels, determine whether the data point is a cluster point, if not, assign the data point the label corresponding to its cluster point; if so, explore the objects that can be reached by its density until no further exploration is possible, and obtain a complete cluster, and return to the step of determining all data points within the preset distance threshold range for any cluster point, until there are no data points without assigned labels, and output the clustering result;

[0144] Based on the clustering results, the rapid ramping and switching scenario points of the target power system are obtained.

[0145] In a possible implementation, the clustering module 54 is further configured to:

[0146] Reversely map the rapid ramping and switching scene points to the corresponding time series or observation intervals in the load data;

[0147] According to the index operation, the position of each fast ramping and switching scene point in the load data is located, and the multi-dimensional features and time series information of each fast ramping and switching scene point are restored.

[0148] In a possible implementation, the processing module 52 is further configured to:

[0149] The load data after removing abnormal data is used to train the data filling model to fill in the missing values ​​in the load data to obtain the preprocessed load data.

[0150] In a possible implementation, the daily periodicity characteristic is calculated as follows:

[0151]

[0152] Where n is the number of load data at the same time of day; x i (h) is the load data corresponding to time point h; x t is the standardized value of the load data at time point t;

[0153] The local difference is calculated as:

[0154]

[0155] Where k is the length of the local change window; x c (i) is the load data at time point i.

[0156] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described or recorded in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0157] Those of ordinary skill in the art will appreciate that the templates, units, and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0158] If the module / unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium. When the computer program is executed by the processor, the steps of the above-mentioned embodiments of the method for extracting the scene point of the rapid climbing and switching of each power system can be implemented. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device that can carry the computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory, random access memory, electrical carrier signal, telecommunication signal and software distribution medium.

[0159] The embodiments described above are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features may be replaced by equivalents. Such modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. A method for extracting rapid ramping and switching scene points of a power system, characterized in that: include: Obtaining load data of the target power system within a preset time; Adopting an adaptive periodic local feature anomaly detection algorithm APLAD to determine the daily periodic characteristics and local differences of the load data, and based on the daily periodic characteristics and the local differences, removing abnormal data in the load data to obtain preprocessed load data; Reconstructing the preprocessed load data to obtain multiple groups of reconstructed data; and obtaining multiple groups of fusion feature data based on the multiple groups of reconstructed data; Based on the multiple groups of fused feature data, rapid ramping and switching scenario points of the target power system are obtained.

2. The method for extracting fast ramping and switching scene points of a power system according to claim 1, characterized in that: The removing abnormal data in the load data based on the daily periodic characteristics and the local differences includes: For any one of the load data, a comprehensive score of the load data is obtained according to the daily periodic characteristics and local differences corresponding to the load data; When the comprehensive score of the load data is greater than the preset abnormal point threshold, the load data is removed as abnormal data.

3. The method for extracting fast ramping and switching scene points of a power system according to claim 2, characterized in that: The calculation formula of the comprehensive score is: Among them, ξ and δ are adjustment coefficients respectively; α is the local difference; β is the daily periodic feature.

4. The method for extracting fast ramping and switching scene points of a power system according to claim 1, characterized in that: The step of obtaining multiple sets of fused feature data according to the multiple sets of reconstructed data includes: For any set of reconstructed data, calculate the time domain characteristics, statistical characteristics and periodic characteristics of the set of reconstructed data; Extracting short-term local features and long-term trend features corresponding to the group of reconstructed data according to the time domain features, statistical features and periodic features of the group of reconstructed data; The short-term local features and the long-term trend features corresponding to the group of reconstructed data are weightedly fused to obtain fused feature data corresponding to the group of reconstructed data.

5. The method for extracting fast ramping and switching scene points of a power system according to claim 1, characterized in that: The obtaining of the rapid ramping and switching scene points of the target power system based on the multiple sets of fused feature data includes: An improved density-based clustering algorithm DBSCAN is used to cluster the multiple groups of fusion features to obtain the rapid climbing and switching scenario points of the target power system; wherein the improved DBSCAN introduces a giant water lily optimization algorithm into the traditional DBSCAN to adaptively adjust the numerical differences and changing trends of the data during the calculation process.

6. The method for extracting fast ramping and switching scene points of a power system according to claim 5, characterized in that: The improved density-based clustering algorithm DBSCAN is used to cluster the multiple groups of fusion features to obtain the rapid ramping and switching scene points of the target power system, including: Using an improved density-based clustering algorithm DBSCAN, clustering the multiple groups of fusion features to obtain multiple clustering points; For any cluster point, determine all data points of the cluster point within a preset distance threshold range; if the number of data points is greater than the preset number, establish the cluster point and all data points within the preset distance threshold range as a new cluster label; if the number of data points is not greater than the preset number, establish the cluster point and all data points within the preset distance threshold range as fast climbing and switching scene point labels; Traverse all cluster points and visit all data points within the preset distance threshold corresponding to these cluster points; if there are data points without assigned labels, determine whether the data point is a cluster point, if not, assign the data point a label corresponding to its cluster point; if so, explore objects whose density can be reached until no further exploration is possible, obtain a complete cluster, and return to the step of determining all data points within the preset distance threshold for any cluster point, until there are no data points without assigned labels, and output the clustering result; Based on the clustering result, fast ramping and switching scenario points of the target power system are obtained.

7. The method for extracting fast ramping and switching scene points of a power system according to claim 1, characterized in that: After obtaining the rapid ramping and switching scene points of the target power system based on the fused feature data, the method further includes: Reverse mapping the rapid ramping and switching scene points to corresponding time series or observation intervals in the load data; According to the indexing operation, the position of each fast ramping and switching scene point in the load data is located, and the multi-dimensional features and time series information of each fast ramping and switching scene point are restored.

8. The method for extracting fast ramping and switching scene points of a power system according to claim 1, characterized in that: After removing abnormal data from the load data based on the daily periodic characteristics and the local differences, the method further includes: The load data after removing abnormal data is used to train the data filling model to fill in the missing values ​​in the load data to obtain the preprocessed load data.

9. The method for extracting fast ramping and switching scene points of a power system according to claim 1, characterized in that: The calculation formula of the daily periodicity characteristic is: Where n is the number of load data at the same time of day; x i (h) is the load data corresponding to time point h; x t is the standardized value of the load data at time point t; The calculation formula of the local difference is: Where k is the length of the local change window; x c (i) is the load data at time point i.

10. A device for extracting fast ramping and switching scene points of a power system, characterized in that: include: A collection module, used to obtain load data of the target power system within a preset time; A processing module, used for adopting an adaptive periodic local feature anomaly detection algorithm APLAD to determine the daily periodic characteristics and local differences of the load data, and based on the daily periodic characteristics and the local differences, remove abnormal data in the load data to obtain pre-processed load data; An extraction module, used to reconstruct the preprocessed load data to obtain multiple groups of reconstructed data; and obtain multiple groups of fusion feature data based on the multiple groups of reconstructed data; A clustering module is used to obtain the rapid ramping and switching scene points of the target power system based on the multiple groups of fused feature data.

Images