Characteristic analysis method, device, equipment and storage medium for distribution transformer load data

Abstract

The present application discloses a method, apparatus, device, and storage medium for feature analysis of distribution transformer load data, which relates to power grid data analysis technology. The method comprises: obtaining power time series data of distribution transformer loads in multiple distribution areas within the same time period, wherein the power time series data includes original time series data and target time series data corresponding to each time point; performing differential processing on the original time series data and target time series data corresponding to each time point to obtain multiple two-dimensional point cloud data; processing each two-dimensional point cloud data based on a preset time weight function to obtain time-tilted point cloud data corresponding to each two-dimensional point cloud data; constructing a VR complex based on the multiple time-tilted point cloud data, and determining the topological features of the distribution transformer load based on the continuous coherence results output by the VR complex. The present application can extract topological features with significant duration, accurately reflect the dynamic change law of the distribution transformer load, and improve the robustness of feature extraction.

Description

Technical Field

[0001] The present application relates to the technical field of power grid data analysis, and in particular to a method, apparatus, device, and storage medium for characteristic analysis of distribution transformer load data. Background Art

[0002] Distribution transformer load data (referring to distribution transformers) records various power parameters during operation, such as active power, reactive power, voltage, and current. Characteristic analysis of distribution transformer load data is a key step in power system operation and management. Its purpose is to achieve load forecasting, fault diagnosis, and grid optimization by analyzing the characteristic changes in distribution transformer load data.

[0003] Existing methods for feature analysis of distribution transformer load data are mainly based on deep learning methods. Although deep learning methods can extract deep features of data, they have problems such as large computing resource requirements and long training time during model training, and have high requirements on data quality and integrity.

[0004] Therefore, how to accurately perform characteristic analysis on distribution transformer load data is a key issue that needs to be solved urgently. Summary of the Invention

[0005] The present application provides a method, apparatus, device and storage medium for characterizing distribution transformer load data, which can improve the problems existing in existing solutions.

[0006] In a first aspect, the present application provides a method for characterizing distribution transformer load data, comprising:

[0007] Acquire power time series data of distribution transformer loads in multiple distribution areas within the same time period, the power time series data including original time series data and target time series data corresponding to each time point, the target time series data being obtained by preprocessing the original time series data; perform differential processing on the original time series data and the target time series data corresponding to each time point to obtain multiple two-dimensional point cloud data; process each of the two-dimensional point cloud data based on a preset time weight function to obtain time-tilted point cloud data corresponding to each of the two-dimensional point cloud data; construct a VR complex based on the multiple time-tilted point cloud data, and determine the topological characteristics of the distribution transformer load according to the continuous coherence result output by the VR complex.

[0008] Optionally, the differential processing is performed on the original time series data and the target time series data corresponding to each time point to obtain multiple two-dimensional point cloud data, including: differential calculation is performed on the original time series data and the target time series data corresponding to each time point to obtain the power change rate corresponding to each time point; and multiple two-dimensional point cloud data are obtained based on the original time series data corresponding to each time point and the corresponding power change rate.

[0009] Optionally, the processing of each of the two-dimensional point cloud data based on a preset time weight function to obtain time-tilted point cloud data corresponding to each of the two-dimensional point cloud data includes: performing weighted calculations on the two-dimensional point cloud data corresponding to each of the time points in turn using the preset time weight function to obtain time-tilted point cloud data corresponding to each of the two-dimensional point cloud data.

[0010] Optionally, constructing a VR complex based on multiple time-tilted point cloud data, and determining the topological characteristics of the distribution transformer load according to the continuous coherence result output by the VR complex, includes: determining the threshold parameter range required for constructing the VR complex based on the multiple time-tilted point cloud data; performing continuous coherence analysis on the multiple time-tilted point cloud data according to the threshold parameter range to obtain a feature life cycle diagram, the feature life cycle diagram including the appearance time and disappearance time of the topological feature corresponding to each time-tilted point cloud data; performing feature analysis on the appearance time and disappearance time of the topological feature corresponding to each time-tilted point cloud data to obtain the topological characteristics of the distribution transformer load.

[0011] Optionally, determining the threshold parameter range required for constructing the VR complex based on multiple time-tilted point cloud data includes: obtaining the coordinate point corresponding to each time-tilted point cloud data, and calculating the Euclidean distance between each time-tilted point cloud data according to the coordinate point corresponding to each time-tilted point cloud data; obtaining the minimum distance value and the maximum distance value in the Euclidean distance; determining the minimum distance value as the threshold lower limit value, and determining a preset multiple of the maximum distance value as the threshold upper limit value; and determining the threshold parameter range according to the threshold lower limit value and the threshold upper limit value.

[0012] Optionally, the continuous coherent analysis of the multiple time-tilted point cloud data is performed according to the threshold parameter range to obtain a feature life cycle diagram, including: obtaining a distance threshold parameter, connecting the time-tilted point cloud data whose distance is not greater than the distance threshold parameter in the multiple time-tilted point cloud data, and constructing an initial complex; setting a distance step parameter, gradually increasing the distance threshold parameter according to the distance step parameter, so as to update the initial complex according to the gradually increased distance threshold parameter, and recording the feature information of the updated initial complex; and determining the feature life cycle diagram according to the feature information of the updated initial complex.

[0013] Optionally, the feature analysis of the appearance time and disappearance time of the topological feature corresponding to each of the time-tilted point cloud data to obtain the topological feature of the distribution transformer load includes: reconstructing the appearance time and disappearance time of the topological feature corresponding to each of the time-tilted point cloud data based on a piecewise linear function method to generate a feature continuous peak graph; the feature continuous peak graph includes density information corresponding to multiple time-tilted point cloud data; vectorizing the density information corresponding to multiple time-tilted point cloud data to obtain a feature vector, and determining the topological feature of the distribution transformer load based on the feature vector.

[0014] In a second aspect, the present application provides a device for characterizing distribution transformer load data, the device comprising:

[0015] A data acquisition module is used to acquire power time series data of distribution transformer loads in multiple distribution areas within the same time period. The power time series data includes original time series data and target time series data corresponding to each time point. The target time series data is obtained by preprocessing the original time series data.

[0016] a data processing module, configured to perform differential processing on the original time series data and the target time series data corresponding to each time point to obtain a two-dimensional point cloud data set, wherein the two-dimensional point cloud data set includes a plurality of two-dimensional point cloud data;

[0017] The data processing module is further configured to process each of the two-dimensional point cloud data based on a preset time weight function to obtain time-inclined point cloud data corresponding to each of the two-dimensional point cloud data;

[0018] A feature determination module is used to construct a VR complex based on the multiple time-tilted point cloud data, and determine the topological features of the distribution transformer load according to the continuous synchronization result output by the VR complex.

[0019] In a third aspect, the present application further provides an electronic device, comprising:

[0020] at least one processor; and a memory communicatively coupled to the at least one processor;

[0021] The memory stores a computer program that can be executed by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the characteristic analysis method of distribution transformer load data described in any embodiment of the present application.

[0022] In a fourth aspect, the present application further provides a computer-readable storage medium, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the characteristic analysis method of distribution transformer load data described in any embodiment of the present application when executed.

[0023] In a fifth aspect, the present application further provides a computer program product, comprising a computer program, which, when executed by a processor, implements the characteristic analysis method for distribution transformer load data described in any embodiment of the present application.

[0024] The characteristic analysis scheme for distribution transformer load data provided in the embodiment of the present application can directly capture the dynamic change characteristics of load data by differentially processing the original time series data and the target time series data corresponding to each time point to obtain multiple two-dimensional point cloud data, which is convenient for quickly discovering abnormal data; then, by introducing a preset time weight function to process each two-dimensional point cloud data, a method of obtaining time-tilted point cloud data is used to integrate the time factor into the point cloud data to facilitate capturing the time dependency of the time series data; finally, a VR complex is constructed by multiple time-tilted point cloud data and continuous coherence analysis is performed to extract the topological characteristics of the distribution transformer load, which is able to capture the global structure and persistence characteristics in the data. The scheme provided in this embodiment can extract topological features with significant duration, achieve the beneficial effect of accurately reflecting the dynamic change law of the distribution transformer load and improving the robustness of feature extraction.

[0025] It should be noted that the aforementioned computer instructions may be stored in whole or in part on a computer-readable storage medium. The computer-readable storage medium may be packaged together with the processor of the distribution transformer load data characteristic analysis device, or may be packaged separately from the processor of the distribution transformer load data characteristic analysis device, and this application does not limit this.

[0026] The descriptions of the second, third, fourth and fifth aspects of this application can refer to the detailed description of the first aspect; and the beneficial effects of the descriptions of the second, fourth and fifth aspects can refer to the analysis of the beneficial effects of the first aspect, which will not be repeated here.

[0027] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present application, nor is it intended to limit the scope of the present application. Other features of the present application will become easily understood through the following description.

[0028] It is understandable that before using the technical solutions disclosed in the embodiments of this application, the type, scope of use, and usage scenarios of the personal information involved in this application should be informed to the user and the user's authorization should be obtained in an appropriate manner in accordance with relevant laws and regulations. BRIEF DESCRIPTION OF THE DRAWINGS

[0029] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following is a brief introduction to the drawings required for use in the embodiments. It should be understood that the following drawings only show certain embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other relevant drawings can be obtained based on these drawings without creative work.

[0030] Figure 1 This is a flow chart of a characteristic analysis method for distribution transformer load data provided by an embodiment of the present application;

[0031] Figure 2 This is another flow chart of the characteristic analysis method of distribution transformer load data provided by an embodiment of the present application;

[0032] Figure 3 This is a schematic diagram of a feature life cycle diagram provided by an embodiment of the present application;

[0033] Figure 4 This is a structural diagram of a characteristic analysis device for distribution transformer load data provided by an embodiment of the present application;

[0034] Figure 5 It is a structural diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0035] In order to enable those skilled in the art to better understand the present invention, the technical solutions in this application will be clearly and completely described below in conjunction with the drawings in this embodiment. Obviously, the described embodiments are only part of the embodiments of this application, not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts should fall within the scope of protection of this application.

[0036] It should be noted that the terms "original", "target" etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequential order. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in a sequence other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0037] The present application will be further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are intended only to illustrate the present application and are not intended to limit the present application. It should also be noted that, for ease of description, the accompanying drawings only illustrate portions of the present application, not all of the structures.

[0038] Figure 1 This is a flow chart of a method for analyzing distribution transformer load data characteristics, provided in an embodiment of the present application. This embodiment is applicable to analyzing distribution transformer load data characteristics in a power grid. The method can be performed by a distribution transformer load data characteristic analysis device, which can be implemented in hardware and / or software and integrated into an electronic device that performs the method. Preferably, the electronic device in the embodiment of the present application can be a server, a computer, or the like.

[0039] refer to Figure 1 The characteristic analysis method of distribution transformer load data in this embodiment includes but is not limited to the following steps:

[0040] S110: Obtain power time series data of distribution transformer loads in multiple distribution areas within the same time period.

[0041] The solution provided in this embodiment primarily uses power data, such as active power and reactive power, when analyzing distribution transformer load data. This data allows users to understand the grid's operating status. Distribution transformer load data directly reflects user electricity usage (such as residential and industrial load characteristics) and is influenced by factors such as time of day (day and night, season), weather, and electricity policies, resulting in strong temporal and cyclical characteristics.

[0042] In order to obtain the real-time changes in the user's electricity demand, the power data in this embodiment may specifically refer to active power data, that is, the power actually output by the distribution transformer.

[0043] A distribution area refers to the scope of distribution transformer load data analysis. This includes, for example, analyzing distribution data for multiple subdistricts (townships) within a county, multiple districts (counties) within a city, or multiple districts (counties) within a province. It can also encompass residential communities (or villages), industrial parks, and other areas. The specific scope of multiple distribution areas is not limited.

[0044] The power time series data represents a power data sequence indexed by time. In this embodiment, each time point corresponds to an original time series data and a target time series data. The target time series data is obtained by preprocessing the original time series data.

[0045] Among them, the original time series data is the original power data of the distribution and transformation equipment directly collected (such as the load value uploaded by the smart meter in real time), which may contain noise, outliers or missing values; the target time series data is the data obtained after preprocessing the original time series data (such as denoising, outlier processing, missing value filling and normalization, etc.), which is used to reflect the actual trend or expected state of the load (such as the load curve during normal operation).

[0046] Specifically, raw time series data can be obtained by using IoT devices such as distribution transformer terminals and smart meters to collect real-time and historical load data from multiple distribution areas (such as residential communities and industrial parks) at a preset sampling frequency, covering the same time period. The preset sampling frequency can be 10 minutes or 15 minutes, and the same time period can be one day or seven consecutive days.

[0047] S120 , performing differential processing on the original time series data and the target time series data corresponding to each time point to obtain a plurality of two-dimensional point cloud data.

[0048] For each time point , T represents the number of time points in the same time period. For the original time series data and target time series data corresponding to any time point, the 2D point cloud data corresponding to the current time point is generated. This process continues until all time points are processed, resulting in multiple 2D point cloud data.

[0049] In this embodiment, the two-dimensional point cloud data includes the original time series data of each time point and the difference data between the original time series data corresponding to the current time period and the target time series data. The specific method of obtaining multiple two-dimensional point cloud data is as follows:

[0050] The original time series data and the target time series data corresponding to each time point are differentially calculated to obtain the power change rate corresponding to each time point; and multiple two-dimensional point cloud data are obtained based on the original time series data and the corresponding power change rate corresponding to each time point.

[0051] The power change rate indicates the absolute change between the target data value and the initial data value at each time point.

[0052] The specific implementation method can be: the input data is the original time series data sequence of a certain distribution area in the same time period and target time series data sequence ; and for each time point , determine the power change rate corresponding to each time point, and obtain , and the original time series data The horizontal axis is the power change rate As the vertical coordinate, construct the two-point cloud data at the current time point ( , ),in, The multiple two-dimensional point cloud data obtained can be expressed as When the power distribution area includes multiple areas, each area generates independent point cloud data which can be expressed as .

[0053] This embodiment combines time series data and differential data to obtain two-dimensional point cloud data, which can intuitively display the dynamic change trend and fluctuation amplitude of load data in the time dimension. For example, when the load surges or drops suddenly for a short time, the differential result will form a point that deviates significantly from the mean, which appears as an abnormal longitudinal (variation) offset in the two-dimensional point cloud data, and intuitively displays dynamic characteristics (such as peak values, valley values, and oscillation frequencies). In two-dimensional point cloud data, abnormal data (such as equipment failure, sudden increase / decrease in load, and data collection errors) appear as outliers or isolated points that deviate from the main cluster. For example, when the power change at a certain point in time far exceeds the historical statistical range (such as 3 times the standard deviation), the point will be significantly away from the main distribution of the point cloud data, and the abnormal data can be quickly located through geometric analysis (such as distance measurement and density estimation).

[0054] S130 , processing each two-dimensional point cloud data based on a preset time weight function to obtain time-inclined point cloud data corresponding to each two-dimensional point cloud data.

[0055] In this embodiment, a preset time weight function is used to analyze the influence of the time dimension using a linear function method. This method incorporates temporal sequence information into the point cloud data to capture the temporal dependencies of the time series data. In this embodiment, the linear function can be a time-varying weight, such as multiplying the time index t by a certain coefficient, or a time-dependent function that adjusts the distribution of the point cloud, resulting in a certain skew in the time direction of the point cloud data to amplify the influence of the time dimension.

[0056] Preferably, in this embodiment, the above step S130 can be implemented as follows: the two-dimensional point cloud data corresponding to each time point are weighted calculated in turn using a preset time weight function to obtain the time-tilted point cloud data corresponding to each two-dimensional point cloud data.

[0057] Specifically, the preset time weight function improved in this embodiment is: , for each time point , process the corresponding point cloud data in turn ,in, Used to control the tilt rate, is the offset; and each point cloud data is multiplied by , obtain time-tilted point cloud data ,specific It can be expressed as follows:

[0058]

[0059] The solution for obtaining time-tilted point cloud data provided in this embodiment incorporates a time-weighted mechanism. This allows the time-tilted point cloud data to retain the characteristics of the original data while incorporating the importance of the time dimension. This allows current analysis to focus more on recent data and quickly respond to sudden changes in load (such as equipment failures and abnormal user behavior). For example, when a distribution transformer is overloaded, the point cloud of recent time points is enhanced, making abnormal features more prominent and easier to identify in the time-tilted point cloud. This significantly improves the ability to analyze dynamic load characteristics and the reliability of anomaly detection.

[0060] S140 , constructing a VR complex based on the multiple time-tilted point cloud data, and determining the topological characteristics of the distribution transformer load according to the continuous synchronization result output by the VR complex.

[0061] The Vietoris-Rips Complex (VR complex) is used to construct complexes of varying thickness based on given point cloud data by adjusting threshold parameters. This allows the data to capture its topological evolution from discrete points to fully connected structures. In this embodiment, the VR complex is constructed to convert point cloud data into a topological structure, capturing the connectivity between points and void characteristics. This allows the identification of features such as holes and loops in load fluctuations, which can be used to reflect the dynamic stability of the power grid.

[0062] Specifically, when constructing a VR complex based on multiple time-tilted point cloud data, the distance between each time-tilted point cloud data point can be pre-determined to gradually generate the VR complex based on a distance threshold. When the distance threshold is very small, the complex consists only of isolated points (a 0-dimensional simplex). When the distance threshold increases to exceed the distance between two points, the two points are connected to form an edge (a 1-dimensional simplex). When the distance threshold is further increased, when the distance between three time-tilted point cloud data points is no greater than the distance threshold, a triangle is formed (a 2-dimensional simplex), and so on.

[0063] Furthermore, by analyzing the data features under each simplicial complex, the topological characteristics of the distribution transformer load are obtained. Topological features are geometric and structural properties extracted from time series data through topological data analysis (TDA). These include connectivity and holes. Specific examples include: using topological features (such as holes and connectivity changes) to capture underlying patterns in load changes (such as holiday patterns and industrial production cycles), improving the accuracy of power load forecasts; if a topological mutation occurs (such as the sudden disappearance of a hole or a breakdown in connectivity), it may indicate an equipment failure (such as a short circuit or overload), which can be used to identify early fault signals. For example, a sudden increase in the load on a transformer may cause a "hole" in the point cloud distribution, which may indicate poor line contact.

[0064] Preferably, please refer to Figure 2 , Figure 2 This is another flow chart of the characteristic analysis method of distribution transformer load data provided by the embodiment of the present application. In the technical solution provided by this embodiment, the above step S140 can be implemented based on the following steps S141 to S143:

[0065] S141. Determine a threshold parameter range required for constructing a VR complex based on multiple time-tilted point cloud data.

[0066] Constructing a VR complex primarily relies on a threshold parameter, requiring analysis of point cloud distribution characteristics based on a determined threshold parameter range. Specifically, the threshold parameter range is obtained by analyzing coordinate points corresponding to multiple time-tilted point cloud data. In a preferred embodiment, step S141 can be implemented through the following steps a) to d):

[0067] a) Obtain the coordinate points corresponding to each time-tilted point cloud data, and calculate the Euclidean distance between each time-tilted point cloud data according to the coordinate points corresponding to each time-tilted point cloud data.

[0068] For 2D point cloud data Each point cloud data corresponds to a coordinate value. For any two points in the set, the Euclidean distance between the time-tilted point cloud data must be calculated. By calculating the Euclidean distance between all points, the spatial position association of each point in the time-tilted point cloud is completely preserved, and the storage provides a full data basis for subsequent threshold determination.

[0069] b) Get the minimum and maximum distance values ​​in the Euclidean distance.

[0070] For the obtained Euclidean distances between the pairwise time-tilted point cloud data, a distance set may be obtained for storage, and each distance value in the distance set may be traversed to obtain a minimum distance value and a maximum distance value from the distance set.

[0071] c) determining the minimum distance value as the lower threshold value, and determining a preset multiple of the maximum distance value as the upper threshold value.

[0072] In this embodiment, the minimum distance value obtained in step b) is used As the lower threshold value , that is, when When , only the closest point pairs are connected, corresponding to the most basic edges in the VR complex. Optionally, the minimum distance value Can be 0.

[0073] And by presetting the multiple k, by presetting the multiple k and the maximum distance value The product of determines the upper threshold value , wherein the preset multiple k can be 1.5 or 2. In this embodiment, the maximum distance value is not directly selected. As the upper threshold value The reason for this is to avoid the upper limit being too small due to individual abnormal long-distance points (such as noise), and to cover the potential reasonable connection range by amplifying the multiple.

[0074] d) determining a threshold parameter range according to the lower threshold value and the upper threshold value.

[0075] The determined threshold parameter range is used to construct the VR complex in subsequent steps and calculate the continuous coordination.

[0076] S142. Perform continuous coherence analysis on multiple time-tilted point cloud data according to a threshold parameter range to obtain a feature life cycle diagram.

[0077] For a given threshold parameter range, the distance threshold can be determined within this range in a variety of ways, such as using equal spacing or gradually increasing values. For each threshold, a corresponding VR complex can be constructed based on the distance between each point in the point cloud data. Specifically, when the distance between two points is less than or equal to the distance, an edge is established between the two points, forming a simple complex structure. For each constructed VR complex, a persistence coherence analysis can be performed to reflect the topological features of the complex, such as connected components and holes. By comparing the persistence coherence under different thresholds, the birth, continuation, and disappearance of topological features can be determined. For example, when a new connected component appears under one threshold, the birth time of this connected component is recorded; when this connected component merges with other connected components and disappears under another threshold, its disappearance time is recorded. In this way, a feature life cycle diagram can be constructed based on the appearance and disappearance times of the corresponding topological features in each time-tilted point cloud data.

[0078] In another preferred embodiment, the above step S142 can be implemented through the following steps e) to g):

[0079] e) obtaining a distance threshold parameter, connecting the time-tilted point cloud data whose distance is not greater than the distance threshold parameter among the plurality of time-tilted point cloud data, and constructing an initial complex.

[0080] In this embodiment, the distance threshold parameter can be the minimum distance value determined in the above steps. , through the minimum distance value Construct an initial complex. That is, for each point cloud data, calculate its distance to all points in other point cloud data. If the current distance is not greater than the set distance threshold parameter, establish a connection between the time-tilted point cloud data where these two points are located, that is, regard them as a whole, and thus construct an initial complex. This initial complex is a structure formed by connecting point cloud data that meet the threshold conditions based on the distance relationship. It preliminarily describes the topological relationship between point cloud data. By setting a distance threshold to construct the initial complex, it is possible to quickly connect closely related point cloud data based on the spatial distance relationship between point cloud data, providing a basic structure for subsequent more in-depth topological analysis. This method can highlight the local structural features within a certain distance range in the point cloud data, help capture the closely related parts of the distribution transformer load in space, and provide a preliminary framework for analyzing its topological characteristics.

[0081] f) setting a distance step parameter, gradually increasing a distance threshold parameter according to the distance step parameter, updating the initial complex according to the gradually increased distance threshold parameter, and recording feature information of the updated initial complex.

[0082] Determine the distance step parameter. Starting from the initial distance threshold parameter, gradually increase the distance threshold according to the step parameter. That is, each time the distance threshold is updated, recheck the connectivity between all time-tilted point cloud data for each updated distance threshold. If two originally unconnected point cloud data satisfy the requirement that the current distance between some point pairs is no greater than the updated distance threshold parameter due to the increase in the distance threshold, a connection is established between the two point cloud data, thereby updating the initial complex. At the same time, record the feature information of the updated initial complex. This feature information may include parameters that can describe the topological properties of the complex, such as the number of connected components, the number of holes, and the dimension.

[0083] g) Determine a feature lifecycle diagram based on the updated feature information of the initial complex.

[0084] A feature lifecycle diagram is drawn based on the recorded feature information of the updated initial complex. The distance threshold parameter is used as the horizontal axis, and the various feature information of the complex is used as the vertical axis. For example, the number of connected components is plotted as the distance threshold increases. If a new connected component appears at a certain distance threshold, its birth is marked at the corresponding location on the diagram. If a connected component disappears due to merging with other connected components during subsequent changes in the distance threshold, the location where it disappears is also marked on the diagram. A similar method is used for other feature information, such as the number of holes, using different lines or colors to distinguish different features, thus forming a complete feature lifecycle diagram. For multiple time-skewed point cloud data sets, their feature lifecycle diagrams can be plotted on the same diagram for comparison and analysis. By observing the feature lifecycle diagram, one can intuitively understand the changes in topological features in the point cloud data at different time points, such as which topological features are stable and which dynamically appear and disappear with time or threshold changes. This provides powerful support for topological feature analysis of distribution transformer loads.

[0085] For details, please refer to Figure 3 , Figure 3 This is a schematic diagram of the feature life cycle diagram provided by the embodiment of the present application. In this embodiment, the distance threshold parameter ε=0 is selected to construct the VR complex. The distance threshold parameter in the VR complex flow is expanded to 10, and the zero-dimensional and one-dimensional topological features are continuously recorded to achieve continuous homology. Figure 3 As shown in the figure, taking an elliptical data point set as an example, by gradually increasing ε, we can see that when ε = 1.5, three connected components are formed. When ε = 2.0, the shape is relatively complete. When ε = 2.3, a relatively good simplicial complex is generated, which now captures the hole in the point cloud data. When ε = 5.0, the hole is still found. Therefore, it can be said that this hole persists from ε = 2.3 to ε = 5.0, indicating that this hole is a topological invariant with a relatively long life cycle in the data, that is, a relatively obvious topological feature contained in this dataset.

[0086] S143 , performing feature analysis on the appearance time and disappearance time of the topological feature corresponding to each time-tilted point cloud data to obtain the topological feature of the distribution transformer load.

[0087] As the distance threshold is gradually increased during the complex construction process, the distance threshold at which each topological feature first appears is recorded as the appearance time of the topological feature. For example, when the distance threshold reaches a certain value, a new connected component appears in the complex for the first time. The time corresponding to this threshold is the appearance time of this connected component and this topological feature. If a topological feature disappears due to changes (such as connected components merging or holes being filled) as the distance threshold continues to increase, the distance threshold at that time is recorded as the disappearance time of the topological feature. Statistical analysis is performed on the appearance and disappearance times of each topological feature in all time-skewed point cloud data. For example, the average appearance and disappearance time and temporal distribution range of different topological features are calculated to understand the general behavior patterns of these features in the overall data. The appearance and disappearance times of topological features are then correlated with actual operating data of the distribution transformer load (such as load level and change trends). For example, the appearance or disappearance of certain topological features is observed to see if they correspond to peaks or troughs in the distribution transformer load or are associated with sudden changes in load. By analyzing large amounts of time-skewed point cloud data, attempts are made to identify common patterns of topological feature change. For example, do certain topological features appear and disappear repeatedly within a specific time period, or do certain features appear and disappear in a certain order? These patterns may reflect the inherent topological structure changes of the distribution transformer load under different operating conditions. By analyzing the appearance and disappearance time of each topological feature in the time-tilted point cloud data, we can fully understand the topological characteristics of the distribution transformer load and provide a strong basis for monitoring, prediction, and optimized management of the distribution transformer load.

[0088] In another preferred embodiment, the above step S143 can be implemented through the following steps h) to j):

[0089] h) Reconstruct the appearance time and disappearance time of the topological feature corresponding to each time-tilted point cloud data based on a piecewise linear function method to generate a feature persistence peak graph.

[0090] The appearance time of the topological feature corresponding to each time-tilted point cloud data is recorded as , the disappearance time is recorded as , which can form the existence time interval corresponding to the current topological feature For example, a certain topological feature appears at time t=3 and disappears at time t=8, so the corresponding existence time interval of the current topological feature can be recorded as (3,8).

[0091] Further for each feature point Perform a linear transformation to map the y=x line (diagonal line) of the original coordinate system to the new horizontal axis. The mapped point represents the "intensity" of a certain feature. For example, the new coordinates of point (3,8) after mapping are (5.5√2, 2.5√2). A piecewise linear function is further used to fit the appearance and disappearance of each topological feature. The piecewise linear function is:

[0092] By plotting piecewise linear functions of different topological features in the same coordinate system, we generate a feature persistence peak graph, which visually displays the appearance, persistence, and disappearance of each topological feature. Finally, we map the density information corresponding to the time-skewed point cloud data onto the feature persistence peak graph. Typically, the vertical axis represents density value, and the horizontal axis represents time. Alternatively, we can use visualization methods such as color depth and graph height to show density differences at different time points and regions.

[0093] Optionally, the density information corresponding to the time-tilted point cloud data can be obtained by clustering, which directly reflects the aggregation of the point cloud in space.

[0094] i) Vectorize the density information corresponding to multiple time-tilted point cloud data to obtain eigenvectors, and determine the topological characteristics of the distribution transformer load based on the eigenvectors.

[0095] The density information corresponding to the multiple time-inclined point cloud data in the characteristic continuous peak graph is vectorized. That is, each density information is arranged into a vector in a certain order. This vector represents the density information of the characteristic continuous peak graph, that is, the characteristic vector. In this embodiment, various machine learning or data analysis methods can be used to further analyze and process the characteristic vector. For example, a clustering algorithm can be used to cluster different characteristic vectors to find similar topological feature patterns; or a feature selection method can be used to extract key features that are representative of the topological features of the distribution transformer load. Based on the results of the analysis and processing, the topological features of the distribution transformer load are determined. For example, the clustering results can be associated with the actual operating status of the distribution transformer load, and different clustering categories can be defined to correspond to different topological features of the distribution transformer load; or the topological features of the distribution transformer load can be described based on the value range and change of the key features.

[0096] This embodiment converts visualized density information into quantifiable numerical vectors through vectorization processing, which helps to explore the intrinsic relationship between complex topological features hidden in large amounts of data and distribution transformer loads, improves the accuracy and scientificity of the description of distribution transformer load topological features, and provides more powerful support for distribution transformer load prediction, fault diagnosis, and optimized operation.

[0097] The characteristic analysis method of distribution transformer load data provided in the embodiment of the present application can directly capture the dynamic change characteristics of load data by differentially processing the original time series data and the target time series data corresponding to each time point to obtain multiple two-dimensional point cloud data, which is convenient for quickly discovering abnormal data; then, by introducing a preset time weight function to process each two-dimensional point cloud data, a method of obtaining time-tilted point cloud data is used to integrate the time factor into the point cloud data so as to capture the time dependency of the time series data; finally, a VR complex is constructed by multiple time-tilted point cloud data and continuous coherence analysis is performed to extract the topological characteristics of the distribution transformer load, which is able to capture the global structure and persistence characteristics in the data. The solution provided in this embodiment can extract topological features with significant duration, achieve the beneficial effect of accurately reflecting the dynamic change law of the distribution transformer load and improving the robustness of feature extraction.

[0098] Figure 4 This is a structural diagram of a characteristic device for distribution transformer load data provided in an embodiment of the present application, which is suitable for executing the characteristic method for distribution transformer load data provided in an embodiment of the present application. Figure 4 As shown, the device may specifically include: a data acquisition module 410, a data processing module 420 and a feature determination module 430.

[0099] The data acquisition module 410 is used to obtain power time series data of distribution transformer loads in multiple distribution areas within the same time period. The power time series data includes original time series data and target time series data corresponding to each time point. The target time series data is obtained by preprocessing the original time series data.

[0100] The data processing module 420 is configured to perform differential processing on the original time series data and the target time series data corresponding to each time point to obtain a two-dimensional point cloud data set, wherein the two-dimensional point cloud data set includes a plurality of two-dimensional point cloud data;

[0101] The data processing module 420 is further configured to process each of the two-dimensional point cloud data based on a preset time weight function to obtain time-inclined point cloud data corresponding to each of the two-dimensional point cloud data;

[0102] The feature determination module 430 is configured to construct a VR complex based on the plurality of time-tilted point cloud data, and determine the topological features of the distribution transformer load according to a continuous coherence result output by the VR complex.

[0103] The characteristic analysis device for distribution transformer load data provided in the embodiment of the present application can directly capture the dynamic change characteristics of load data by differentially processing the original time series data and the target time series data corresponding to each time point to obtain multiple two-dimensional point cloud data, which is convenient for quickly discovering abnormal data; then, by introducing a preset time weight function to process each two-dimensional point cloud data to obtain time-tilted point cloud data, the time factor is integrated into the point cloud data to facilitate capturing the time dependency of the time series data; finally, a VR complex is constructed by multiple time-tilted point cloud data and continuous coherence analysis is performed to extract the topological characteristics of the distribution transformer load, which is able to capture the global structure and persistence characteristics in the data. The solution provided in this embodiment can extract topological features with significant duration, achieve the beneficial effect of accurately reflecting the dynamic change law of the distribution transformer load and improving the robustness of feature extraction.

[0104] In one embodiment, the data processing module 420 is specifically used to perform differential calculation on the original time series data and the target time series data corresponding to each time point to obtain the power change rate corresponding to each time point; and obtain multiple two-dimensional point cloud data based on the original time series data corresponding to each time point and the corresponding power change rate.

[0105] In one embodiment, the data processing module 420 is further configured to perform weighted calculations on the two-dimensional point cloud data corresponding to each of the time points using the preset time weight function in sequence to obtain time-inclined point cloud data corresponding to each of the two-dimensional point cloud data.

[0106] In one embodiment, the feature determination module 430 includes a parameter determination unit, a data analysis unit, and a feature analysis unit.

[0107] wherein the parameter determination unit is configured to determine a threshold parameter range required for constructing the VR complex based on the plurality of time-tilted point cloud data;

[0108] a data analysis unit, configured to perform continuous coherence analysis on the plurality of time-tilted point cloud data according to the threshold parameter range to obtain a feature life cycle diagram, wherein the feature life cycle diagram includes an appearance time and a disappearance time of a topological feature corresponding to each time-tilted point cloud data;

[0109] The feature analysis unit is used to perform feature analysis on the appearance time and disappearance time of the topological feature corresponding to each of the time-tilted point cloud data to obtain the topological feature of the distribution transformer load.

[0110] In one embodiment, the parameter determination unit is specifically used to obtain the coordinate point corresponding to each of the time-tilted point cloud data, calculate the Euclidean distance between each of the time-tilted point cloud data based on the coordinate point corresponding to each of the time-tilted point cloud data; obtain the minimum distance value and the maximum distance value in the Euclidean distance; determine the minimum distance value as the lower threshold value, and determine the preset multiple of the maximum distance value as the upper threshold value; determine the threshold parameter range according to the lower threshold value and the upper threshold value.

[0111] In one embodiment, the data analysis unit is specifically used to obtain a distance threshold parameter, connect the time-tilted point cloud data whose distance is not greater than the distance threshold parameter among the multiple time-tilted point cloud data, and construct an initial complex; set a distance step parameter, gradually increase the distance threshold parameter according to the distance step parameter, so as to update the initial complex according to the gradually increased distance threshold parameter, and record the feature information of the updated initial complex; determine the feature life cycle diagram according to the feature information of the updated initial complex.

[0112] In one embodiment, the feature analysis unit reconstructs the appearance time and disappearance time of the topological feature corresponding to each of the time-tilted point cloud data based on a piecewise linear function method to generate a feature continuous peak diagram; the feature continuous peak diagram includes density information corresponding to multiple time-tilted point cloud data; the density information corresponding to the multiple time-tilted point cloud data is vectorized to obtain a feature vector, and the topological feature of the distribution transformer load is determined according to the feature vector.

[0113] Those skilled in the art will clearly understand that for the sake of convenience and brevity of description, only the division of the above-mentioned functional modules is used as an example for illustration. In actual applications, the above-mentioned functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. The specific working process of the functional modules described above can refer to the corresponding process in the aforementioned method embodiment and will not be repeated here.

[0114] An embodiment of the present application also provides an electronic device, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the characterization method of distribution transformer load data described in any embodiment of the present application.

[0115] An embodiment of the present application further provides a computer-readable medium, wherein the computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the characteristic method of distribution transformer load data described in any embodiment of the present application when executed.

[0116] Reference below Figure 5 , Figure 5 FIG1 is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application, which shows a schematic diagram of the structure of a computer system 500 suitable for implementing the electronic device in an embodiment of the present application. Figure 5 The electronic device shown is merely an example and should not limit the functions and scope of use of the embodiments of the present application.

[0117] like Figure 5 As shown, the computer system 500 includes a central processing unit 501, which can perform various appropriate actions and processes according to a program stored in a read-only memory 502 or a program loaded from a storage unit 508 into a random access memory 503. Various programs and data required for the operation of the computer system 500 are also stored in the random access memory 503. The central processing unit 501, the read-only memory 502, and the random access memory 503 are connected to each other via a bus 504. An input / output interface 505 is also connected to the bus 504.

[0118] The following components are connected to the input / output interface 505: an input section 506 including a keyboard, a mouse, and the like; an output section 507 including devices such as a cathode ray tube (CRT), a liquid crystal display (LCD), and a speaker; a storage section 508 including a hard disk; and a communication section 509 including a network interface card such as a LAN card or a modem. The communication section 509 performs communication processing via a network such as the Internet. A drive 510 is also connected to the input / output interface 505 as needed. A removable medium 511, such as a magnetic disk, an optical disk, a magneto-optical disk, or a semiconductor memory, is installed in the drive 510 as needed, so that computer programs read therefrom can be installed into the storage section 508 as needed.

[0119] In particular, according to the embodiments disclosed in the present application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, the embodiments disclosed in the present application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via the communication section 509 and / or installed from a removable medium 511. When the computer program is executed by the central processing unit 501, the above-mentioned functions defined in the system of the present application are performed.

[0120] It should be noted that the computer-readable medium shown in this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. The computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or any combination of the above. More specific examples of computer-readable storage media can include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this application, a computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, device, or device. In this application, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, which carries computer-readable program code. This propagated data signal can take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device. Program code embodied on a computer-readable medium may be transmitted using any suitable medium, including but not limited to wireless, wired, and optical cables, or any suitable combination thereof.

[0121] The flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present application. In this regard, each box in the flowchart or block diagram can represent a module, program segment, or a part of code, and the above-mentioned module, program segment, or a part of code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flowchart, and the combination of the boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0122] The modules and / or units described in the embodiments of this application may be implemented in software or hardware. The modules and / or units described may also be provided in a processor. For example, a processor may be described as comprising a data acquisition module, a data processing module, and a feature determination module. The names of these modules do not, in some cases, limit the modules themselves.

[0123] As another aspect, the present application also provides a computer-readable medium, which may be included in the device described in the above embodiment; or it may exist independently and not be assembled into the device. The above computer-readable medium carries one or more programs. When the above one or more programs are executed by one of the devices, the device includes: obtaining power time series data of distribution transformer loads in multiple distribution areas in the same time period, the power time series data including original time series data and target time series data corresponding to each time point, the target time series data being obtained by preprocessing the original time series data; performing differential processing on the original time series data and the target time series data corresponding to each time point to obtain multiple two-dimensional point cloud data; processing each of the two-dimensional point cloud data based on a preset time weight function to obtain time-tilted point cloud data corresponding to each of the two-dimensional point cloud data; constructing a VR complex based on the multiple time-tilted point cloud data, and determining the topological characteristics of the distribution transformer load according to the continuous coherence result output by the VR complex.

[0124] According to the technical solution of this embodiment, by performing differential processing on the original time series data and the target time series data corresponding to each time point to obtain multiple two-dimensional point cloud data, the dynamic change characteristics of the load data can be directly captured, facilitating the rapid discovery of abnormal data; then, by introducing a preset time weight function to process each two-dimensional point cloud data to obtain time-tilted point cloud data, the time factor is integrated into the point cloud data to facilitate the capture of the time dependency of the time series data; finally, by constructing a VR complex through multiple time-tilted point cloud data and performing continuous coherence analysis, the topological features of the distribution transformer load are extracted, which can capture the global structure and persistence characteristics in the data. The solution provided by this embodiment can extract topological features with significant duration, achieve the beneficial effect of accurately reflecting the dynamic change law of the distribution transformer load and improving the robustness of feature extraction.

[0125] The above specific embodiments do not constitute a limitation on the scope of protection of this application. Those skilled in the art will appreciate that various modifications, combinations, sub-combinations, and substitutions may occur depending on design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application shall be included within the scope of protection of this application.

Claims

1. A characteristic analysis method for distribution transformer load data, characterized in that: include: Obtain power time series data of distribution transformer loads in multiple distribution areas within the same time period, wherein the power time series data includes original time series data and target time series data corresponding to each time point, and the target time series data is obtained by preprocessing the original time series data; Performing differential processing on the original time series data and the target time series data corresponding to each time point to obtain a plurality of two-dimensional point cloud data; Processing each of the two-dimensional point cloud data based on a preset time weight function to obtain time-inclined point cloud data corresponding to each of the two-dimensional point cloud data; constructing a VR complex based on the plurality of time-tilted point cloud data, and determining the topological characteristics of the distribution transformer load according to a continuous coherence result output by the VR complex; The step of performing differential processing on the original time series data and the target time series data corresponding to each time point to obtain a plurality of two-dimensional point cloud data includes: Performing differential calculation on the original time series data and the target time series data corresponding to each time point to obtain the power change rate corresponding to each time point; Obtain a plurality of two-dimensional point cloud data according to the original time series data corresponding to each time point and the corresponding power change rate; The step of performing differential processing on the original time series data and the target time series data corresponding to each time point to obtain a plurality of two-dimensional point cloud data includes: The two-dimensional point cloud data corresponding to each of the time points is weightedly calculated using the preset time weight function in sequence to obtain time-inclined point cloud data corresponding to each of the two-dimensional point cloud data.

2. The characteristic analysis method of distribution transformer load data according to claim 1 is characterized in that: The step of constructing a VR complex based on the plurality of time-tilted point cloud data and determining the topological characteristics of the distribution transformer load according to a continuous coherence result output by the VR complex includes: determining a threshold parameter range required for constructing the VR complex based on the plurality of time-tilted point cloud data; Performing continuous coherence analysis on the plurality of time-tilted point cloud data according to the threshold parameter range to obtain a feature life cycle diagram, wherein the feature life cycle diagram includes an appearance time and a disappearance time of a topological feature corresponding to each time-tilted point cloud data; A feature analysis is performed on the appearance time and disappearance time of the topological feature corresponding to each of the time-tilted point cloud data to obtain the topological feature of the distribution transformer load.

3. The characteristic analysis method of distribution transformer load data according to claim 2 is characterized in that: The determining of a threshold parameter range required for constructing the VR complex based on the plurality of time-tilted point cloud data includes: Obtaining a coordinate point corresponding to each of the time-tilted point cloud data, and calculating a Euclidean distance between each of the time-tilted point cloud data according to the coordinate point corresponding to each of the time-tilted point cloud data; Obtaining the minimum distance value and the maximum distance value in the Euclidean distance; Determine the minimum distance value as a lower threshold value, and determine a preset multiple of the maximum distance value as an upper threshold value; The threshold parameter range is determined according to the lower threshold value and the upper threshold value.

4. The characteristic analysis method of distribution transformer load data according to claim 2 is characterized in that: The continuously performing coherent analysis on the plurality of time-tilted point cloud data according to the threshold parameter range to obtain a feature life cycle diagram includes: Obtaining a distance threshold parameter, connecting time-tilted point cloud data whose distance is not greater than the distance threshold parameter among the plurality of time-tilted point cloud data, and constructing an initial complex; Setting a distance step parameter, gradually increasing the distance threshold parameter according to the distance step parameter, updating the initial complex according to the gradually increased distance threshold parameter, and recording feature information of the updated initial complex; The feature lifecycle graph is determined according to the updated feature information of the initial complex.

5. The characteristic analysis method of distribution transformer load data according to claim 2 is characterized in that: The performing feature analysis on the appearance time and disappearance time of the topological feature corresponding to each of the time-tilted point cloud data to obtain the topological feature of the distribution transformer load includes: Reconstructing the appearance time and disappearance time of the topological feature corresponding to each of the time-tilted point cloud data based on a piecewise linear function method to generate a feature persistence peak graph; the feature persistence peak graph includes density information corresponding to a plurality of the time-tilted point cloud data; Vectorization is performed on density information corresponding to a plurality of the time-tilted point cloud data to obtain a feature vector, and a topological feature of the distribution transformer load is determined based on the feature vector.

6. A characteristic analysis device for distribution transformer load data, characterized in that: include: A data acquisition module is used to acquire power time series data of distribution transformer loads in multiple distribution areas within the same time period. The power time series data includes original time series data and target time series data corresponding to each time point. The target time series data is obtained by preprocessing the original time series data. a data processing module, configured to perform differential processing on the original time series data and the target time series data corresponding to each time point to obtain a two-dimensional point cloud data set, wherein the two-dimensional point cloud data set includes a plurality of two-dimensional point cloud data; The data processing module is further configured to process each of the two-dimensional point cloud data based on a preset time weight function to obtain time-inclined point cloud data corresponding to each of the two-dimensional point cloud data; a feature determination module, configured to construct a VR complex based on a plurality of the time-tilted point cloud data, and determine the topological features of the distribution transformer load according to a continuous coherence result output by the VR complex; The data processing module is specifically configured to perform differential calculation on the original time series data and the target time series data corresponding to each time point to obtain the power change rate corresponding to each time point; and obtain a plurality of two-dimensional point cloud data according to the original time series data and the corresponding power change rate corresponding to each time point; Among them, the data processing module is specifically used to perform weighted calculation on the two-dimensional point cloud data corresponding to each of the time points using the preset time weight function in sequence to obtain the time-inclined point cloud data corresponding to each of the two-dimensional point cloud data.

7. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively coupled to the at least one processor; The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can perform the characteristic analysis method for distribution transformer load data according to any one of claims 1 to 5.

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

Images