Data partition processing method and system of electric energy meter

Abstract

The invention is suitable for the technical field of data processing, and provides a data partition processing method and system for an electric energy meter, and the method comprises the following steps: collecting the information of the electric energy meter, the information of the electric energy meter comprising power consumption data and archive information; performing time sequence feature extraction on the electric energy meter information, wherein the time sequence features comprise a peak-to-valley ratio, a daily load rate, a periodic feature and an autocorrelation coefficient; spatial features of the electric energy meter information are extracted, wherein the spatial features comprise a transformer load rate, an electric meter topology depth, a spatial lag feature and a spatial autocorrelation index; screening and optimizing the extracted time sequence features and spatial features based on correlation pruning to obtain feature information; and constructing a clustering partition model based on the feature information, and obtaining a plurality of data partitions according to a clustering result. According to the method, the space-time characteristics of the electric energy meter data are fully utilized, the redundant characteristics are effectively removed, and the data processing efficiency and the accuracy of the clustering result are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of data processing, and in particular to a data partition processing method and system for an electric energy meter. Background Art

[0002] With the continuous development of smart grid technology, the energy meter, as an indispensable part of the power system, plays a vital role in the collection and analysis of its data in power dispatching, load forecasting, fault detection and other aspects. The traditional data processing method of the energy meter is often limited to simple data recording and display, lacking in-depth mining and utilization of data, especially failing to fully utilize the temporal and spatial characteristics of the energy meter data. In addition, in the face of massive amounts of energy meter data, how to efficiently and accurately partition the data so as to better manage the load and optimize resource allocation in the future is a major challenge currently faced. Traditional data partitioning methods are often based on simple geographical locations or administrative divisions, ignoring the inherent connections and spatiotemporal characteristics between the energy meter data, resulting in the partitioning results being not detailed and accurate enough. Therefore, it is necessary to provide a data partitioning processing method and system for an energy meter to solve the above problems. Summary of the invention

[0003] In view of the deficiencies in the prior art, the object of the present invention is to provide a data partition processing method and system for an electric energy meter to solve the problems existing in the above-mentioned background technology.

[0004] The present invention is implemented as follows: a data partition processing method for an electric energy meter, the method comprising the following steps: Collecting electric energy meter information, the electric energy meter information including power consumption data and archive information, the archive information including installation location, substation, feeder, transformer and meter code; Extracting time series features from the electric energy meter information, wherein the time series features include peak-to-valley ratio, daily load rate, periodicity features and autocorrelation coefficient; Extracting spatial features from the electric energy meter information, wherein the spatial features include transformer load rate, electric energy meter topology depth, spatial hysteresis feature and spatial autocorrelation index; Based on correlation pruning, the extracted temporal features and spatial features are screened and optimized to obtain feature information; A clustering partition model is constructed based on the feature information, and a plurality of data partitions are obtained according to the clustering results.

[0005] As a further solution of the present invention: the step of extracting time series features from the electric energy meter information specifically includes: Calculate the peak-to-valley ratio and daily load rate of electricity consumption data. Peak-to-valley ratio = (maximum electricity consumption - minimum electricity consumption) ÷ average electricity consumption; daily load rate = average electricity consumption ÷ maximum electricity consumption; Determine the power consumption time series x(t) based on the power consumption data, perform discrete Fourier transform on x(t) to obtain X(k), and take the amplitudes of the first K coefficients as the periodic characteristics; Calculate hysteresis based on x(t) The autocorrelation coefficient AT is AT= , represents the mean of the time series, is the lag order, which indicates the number of time units of the time series shifted backward, and N is the length of the time series.

[0006] As a further solution of the present invention: the step of extracting spatial features from the electric energy meter information specifically includes: Calculate all transformer load factors and meter topology depths, where the meter topology depths are obtained based on topology paths; Calculate the spatial hysteresis characteristics of the energy meter, , is the spatial lag feature of meter i, B represents the neighbor set of meter i, is the spatial weight between meters i and j; Calculate the spatial autocorrelation index I, I= , n represents the total number of electricity meters, and U represents the mean electricity consumption of all electricity meters.

[0007] As a further solution of the present invention: the step of screening and optimizing the extracted temporal features and spatial features based on correlation pruning to obtain feature information specifically includes: Calculate the variance value of each feature in the temporal features and the spatial features, and remove the features whose variance value is less than the first threshold; Perform correlation pruning, pair the retained features in pairs, and calculate the correlation cardinality between each pair of features , , Representation characteristics and Features The covariance of and Respectively represent characteristics and Features The standard deviation of One of the two features corresponding to the correlation cardinality greater than the second threshold is selected, and feature information is obtained according to the retained feature.

[0008] As a further solution of the present invention: the step of constructing a cluster partition model based on the feature information specifically includes: Determine the feature set F based on the feature information: F= , where m is the number of features, and each feature is standardized to have a mean of 0 and a standard deviation of 1; The temporal features and spatial features in the feature information are fused into a feature matrix X, and K-Means++ or GMM is used to cluster and partition the electric energy meter.

[0009] Another object of the present invention is to provide a data partition processing system for an electric energy meter, the system comprising: An electric meter information collection module is used to collect electric energy meter information, wherein the electric energy meter information includes power consumption data and archive information, wherein the archive information includes installation location, substation, feeder, transformer and electric meter code; A time series feature extraction module is used to extract time series features from the electric energy meter information, wherein the time series features include peak-to-valley ratio, daily load rate, periodicity characteristics and autocorrelation coefficient; A spatial feature extraction module is used to extract spatial features from electric energy meter information, wherein the spatial features include transformer load rate, electric energy meter topology depth, spatial hysteresis feature and spatial autocorrelation index; The feature screening and optimization module is used to screen and optimize the extracted temporal features and spatial features based on correlation pruning to obtain feature information; The data clustering and partitioning module is used to construct a clustering and partitioning model based on the feature information and obtain a number of data partitions according to the clustering results.

[0010] As a further solution of the present invention: the temporal feature extraction module includes: Daily load rate calculation unit, used to calculate the peak-to-valley ratio and daily load rate of power consumption data. Peak-to-valley ratio = (maximum power consumption - minimum power consumption) ÷ average power consumption. Daily load rate = average power consumption ÷ maximum power consumption. The periodic feature unit is used to determine the power consumption time series x(t) according to the power consumption data, perform discrete Fourier transform on x(t) to obtain X(k), and take the amplitude of the first K coefficients as the periodic feature; Autocorrelation unit, used to calculate lag based on x(t) The autocorrelation coefficient AT is AT= , represents the mean of the time series, is the lag order, which indicates the number of time units of the time series shifted backward, and N is the length of the time series.

[0011] As a further solution of the present invention: the spatial feature extraction module includes: A topology depth determination unit, used to calculate all transformer load rates and electric meter topology depths, wherein the electric meter topology depths are obtained according to the topology path; The spatial hysteresis characteristic unit is used to calculate the spatial hysteresis characteristics of the electric energy meter. , is the spatial lag feature of meter i, B represents the neighbor set of meter i, is the spatial weight between meters i and j; Spatial autocorrelation index unit, used to calculate the spatial autocorrelation index I, I= , n represents the total number of electricity meters, and U represents the mean electricity consumption of all electricity meters.

[0012] As a further solution of the present invention: the feature screening optimization module includes: A variance value screening unit, used to calculate the variance value of each feature in the time series feature and the spatial feature, and remove features whose variance value is less than a first threshold; The correlation pruning unit is used to perform correlation pruning, pair the retained features in pairs, and calculate the correlation cardinality between each pair of features. , , Representation characteristics and Features The covariance of and Respectively represent characteristics and Features The standard deviation of The feature information determination unit is used to select one of the two features corresponding to the correlation cardinality greater than the second threshold, and obtain feature information according to the retained feature.

[0013] As a further solution of the present invention: the data clustering partitioning module includes: A feature set determination unit is used to determine a feature set F according to feature information: F= , where m is the number of features, and each feature is standardized to have a mean of 0 and a standard deviation of 1; The electric meter clustering and partitioning unit is used to fuse the temporal features and spatial features in the feature information into a feature matrix X, and use K-Means++ or GMM to cluster and partition the electric energy meter.

[0014] Compared with the prior art, the present invention has the following beneficial effects: The present invention makes full use of the spatiotemporal characteristics of the electric energy meter data by extracting the temporal characteristics and spatial characteristics of the electric energy meter information, and provides a more comprehensive and accurate data basis for subsequent clustering and partitioning. The extracted temporal characteristics and spatial characteristics are screened and optimized based on correlation pruning, which effectively removes redundant features, improves data processing efficiency and the accuracy of clustering results, and ensures the differentiation after data partitioning. Finally, a clustering and partitioning model is constructed based on the optimized feature information, which can more finely and accurately reflect the internal connection and spatiotemporal characteristics between the electric energy meter data, thereby obtaining a number of data partitions with more practical significance, which is convenient for subsequent better load management and optimization of resource allocation. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 The present invention is a flow chart of a data partition processing method for an electric energy meter.

[0016] Figure 2 The present invention is a flow chart of extracting time series features of electric energy meter information in a data partition processing method of an electric energy meter.

[0017] Figure 3 The present invention is a flow chart of extracting spatial features of electric energy meter information in a data partitioning processing method of an electric energy meter.

[0018] Figure 4 The present invention is a flowchart for screening and optimizing features in a data partitioning processing method for an electric energy meter.

[0019] Figure 5 The present invention is a flowchart for constructing a cluster partition model in a data partition processing method for an electric energy meter.

[0020] Figure 6 The present invention is a structural schematic diagram of a data partition processing system for an electric energy meter. DETAILED DESCRIPTION

[0021] In order to make the purpose, technical solution and advantages of the present invention clearer, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0022] The specific implementation of the present invention is described in detail below in conjunction with specific embodiments.

[0023] like Figure 1 As shown, an embodiment of the present invention provides a data partition processing method for an electric energy meter, the method comprising the following steps: S100, collecting electric energy meter information, the electric energy meter information including power consumption data and archive information, the archive information including installation location, substation, feeder, transformer and meter code; S200, extracting time series features from the electric energy meter information, where the time series features include peak-to-valley ratio, daily load rate, periodicity features, and autocorrelation coefficient; S300, extracting spatial features from the electric energy meter information, where the spatial features include transformer load rate, electric energy meter topology depth, spatial hysteresis features, and spatial autocorrelation index; S400, screening and optimizing the extracted temporal features and spatial features based on correlation pruning to obtain feature information; S500, constructing a clustering partition model based on the feature information, and obtaining a plurality of data partitions according to the clustering results.

[0024] It should be noted that, at present, facing the massive amount of electricity meter data, how to efficiently and accurately partition the data so as to better manage the load and optimize resource allocation in the future is a major challenge currently faced. Traditional data partitioning methods are often based on simple geographical locations or administrative divisions, ignoring the inherent connections and spatiotemporal characteristics between electricity meter data, resulting in inadequate and inaccurate partitioning results. The embodiments of the present invention are intended to solve the above problems.

[0025] In the embodiment of the present invention, it is first necessary to collect the information of electric energy meters, which includes the data of all electric energy meters in a certain area. Specifically, the electric energy meter information includes electricity consumption data and archive information. Then, the time series features of the electricity consumption data in the electric energy meter information are extracted. The time series features include peak-to-valley ratio, daily load rate, periodic characteristics and autocorrelation coefficient. The peak-to-valley ratio can reflect the volatility of electricity consumption, and the daily load rate can reflect the balance of electricity consumption; the periodic characteristics and the autocorrelation coefficient reveal the inherent laws of electricity consumption data from the perspectives of frequency domain and time domain respectively. The periodic characteristics are used to capture the periodic change laws in electricity consumption data, such as daily cycles, weekly cycles, etc. These laws are usually closely related to the user's work and rest habits, production activities, etc. The autocorrelation coefficient is used to measure the correlation between the time series and its lagged version, reflecting the self-similarity and trend of electricity consumption data. Then, the spatial features of the electric energy meter information are extracted. The spatial features include transformer load rate, meter topology depth, spatial lag features and spatial autocorrelation index. The transformer load rate is used to measure the load of the transformer. The meter topology depth is the number of levels of the meter in the power grid topology structure, which is used to reflect its position in the power grid. The spatial lag feature is used to reflect the impact of neighboring meters on the current meter. The spatial autocorrelation index measures the aggregation of spatial data and is used to reflect the spatial distribution pattern of meter data. Here, many features are extracted, and only features with large differences are retained for more efficient data partitioning. Here, the extracted time series features and spatial features will be screened and optimized according to correlation pruning to obtain feature information. The features in the feature information can better partition the electric energy meter. Finally, a clustering partitioning model is constructed based on the feature information, and several data partitions are obtained according to the clustering results. The partitioning results are accurate, the differences are obvious, and the partitioning process is efficient.

[0026] like Figure 2 As shown, as a preferred embodiment of the present invention, the step of extracting time series features from the electric energy meter information specifically includes: S201, calculating the peak-to-valley ratio and daily load rate of electricity consumption data; S202, determining the power consumption time series x(t) according to the power consumption data, performing a discrete Fourier transform on x(t) to obtain X(k), and taking the amplitudes of the first K coefficients as periodic features; S203, calculate hysteresis based on x(t) The autocorrelation coefficient AT.

[0027] In the embodiment of the present invention, the electricity consumption data is unified to the hourly granularity, and then the peak-to-valley ratio and daily load rate of the electricity consumption data are calculated. The peak-to-valley ratio = (maximum electricity consumption - minimum electricity consumption) ÷ average electricity consumption, and the daily load rate = average electricity consumption ÷ maximum electricity consumption. The maximum electricity consumption is the maximum electricity consumption in a certain hour. It is also necessary to determine the electricity consumption time series x(t) based on the electricity consumption data, perform a discrete Fourier transform on x(t) to obtain X(k), and take the amplitudes of the first K coefficients as periodic features to capture the periodic pattern of the electricity consumption data, where K is a constant. Finally, the lag is calculated based on x(t). The autocorrelation coefficient AT is AT= , represents the mean of the time series, is the lag order, which indicates the number of time units of the time series shifted backward, and N is the length of the time series. The value needs to be set to a fixed range. For hourly data, it is usually , analyze the autocorrelation within 1 day.

[0028] like Figure 3 As shown, as a preferred embodiment of the present invention, the step of extracting spatial features from the electric energy meter information specifically includes: S301, calculating all transformer load rates and electric meter topology depths, where the electric meter topology depths are obtained according to the topology path; S302, calculating the spatial hysteresis characteristics of the electric energy meter; S303, calculating the spatial autocorrelation index I.

[0029] In the embodiment of the present invention, all transformer load rates and meter topology depths are calculated. The transformer load rate is the ratio of the sum of the electricity consumption of all meters under the transformer to the transformer capacity. The meter topology depth reflects the number of levels of the meter in the power grid. The meter topology depth is obtained based on the topological path. Then the spatial hysteresis characteristics of the energy meter are calculated. , is the spatial lag feature of meter i, which is the weighted average of the neighbor meter data of meter i, and is used to reflect the impact of neighbor meters on the current meter; B represents the neighbor set of meter i, and the neighbors can be other meters under the same transformer, or meters within a certain spatial distance range. is the spatial weight between meters i and j, . Represents the spatial distance between meters i and j. Finally, the spatial autocorrelation index I is calculated, I= , n represents the total number of electricity meters, U represents the average power consumption of all electricity meters, Indicates the electricity consumption data of meter i.

[0030] like Figure 4 As shown, as a preferred embodiment of the present invention, the step of screening and optimizing the extracted temporal features and spatial features based on correlation pruning to obtain feature information specifically includes: S401, calculating the variance value of each feature in the temporal feature and the spatial feature, and removing features whose variance value is less than a first threshold; S402, perform correlation pruning, pair the retained features in pairs, and calculate the correlation cardinality between each pair of features ; S403, selecting one of the two features corresponding to the correlation cardinality greater than the second threshold, and obtaining feature information according to the retained feature.

[0031] In the embodiment of the present invention, when screening features, the variance value of each feature in the temporal feature and the spatial feature is first calculated, and the features with variance values ​​less than the first threshold are removed, and the features with significant changes are retained, which are more meaningful for reference. Then, correlation pruning is performed, and the retained features are paired with each other. Any two features need to be paired, and the correlation cardinality between each pair of features is calculated. , , Representation characteristics and Features The covariance of and Respectively represent characteristics and Features The standard deviation of , and finally select one of the two features corresponding to the correlation cardinality greater than the second threshold. If the correlation cardinality is greater than the second threshold (such as 0.85), it means that the two features are highly correlated, so just select one of them. The final feature information can be obtained based on the retained features.

[0032] like Figure 5 As shown, as a preferred embodiment of the present invention, the step of constructing a cluster partition model based on the feature information specifically includes: S501, determine the feature set F according to the feature information: F= , where m is the number of features, and each feature is standardized; S502: The time series features and spatial features in the feature information are integrated into a feature matrix X, and the electric energy meter is clustered and partitioned using K-Means++ or GMM.

[0033] In the embodiment of the present invention, the final feature set F is determined: F= , where m is the number of features. Each feature is standardized to make its mean 0 and standard deviation 1. After standardization, the temporal features and spatial features in the feature information are fused into a feature matrix X, and K-Means++ or GMM is used to cluster and partition the electric energy meter. The following is a brief description of the steps of K-Means++ clustering: randomly select the first cluster center C1; for each sample xi, calculate its minimum distance D(xi) to the selected center; according to the probability D(xi) 2 / (∑ j D(xj) 2 ) Select the next center; repeat until K centers are selected; perform iterative optimization: assign each sample to the nearest cluster center; recalculate the center of each cluster; stop when the cluster center no longer changes or reaches the maximum number of iterations to obtain the clustering result.

[0034] like Figure 6 As shown, an embodiment of the present invention further provides a data partition processing system for an electric energy meter, the system comprising: The electric meter information collection module 100 is used to collect electric energy meter information, wherein the electric energy meter information includes power consumption data and archive information, wherein the archive information includes installation location, substation, feeder, transformer and electric meter code; A time series feature extraction module 200 is used to extract time series features from the electric energy meter information, wherein the time series features include peak-to-valley ratio, daily load rate, periodicity characteristics and autocorrelation coefficient; A spatial feature extraction module 300 is used to extract spatial features from electric energy meter information, wherein the spatial features include transformer load factor, electric energy meter topology depth, spatial hysteresis feature and spatial autocorrelation index; The feature screening and optimization module 400 is used to screen and optimize the extracted temporal features and spatial features based on correlation pruning to obtain feature information; The data clustering and partitioning module 500 is used to construct a clustering and partitioning model based on the feature information and obtain a number of data partitions according to the clustering results.

[0035] As a preferred embodiment of the present invention, the time series feature extraction module 200 includes: Daily load rate calculation unit, used to calculate the peak-to-valley ratio and daily load rate of power consumption data. Peak-to-valley ratio = (maximum power consumption - minimum power consumption) ÷ average power consumption. Daily load rate = average power consumption ÷ maximum power consumption. The periodic feature unit is used to determine the power consumption time series x(t) according to the power consumption data, perform discrete Fourier transform on x(t) to obtain X(k), and take the amplitude of the first K coefficients as the periodic feature; Autocorrelation unit, used to calculate lag based on x(t) The autocorrelation coefficient AT, AT = , represents the mean of the time series, is the lag order, which indicates the number of time units of the time series shifted backward, and N is the length of the time series.

[0036] As a preferred embodiment of the present invention, the spatial feature extraction module 300 includes: A topology depth determination unit, used to calculate all transformer load rates and electric meter topology depths, wherein the electric meter topology depths are obtained according to the topology path; The spatial hysteresis characteristic unit is used to calculate the spatial hysteresis characteristics of the electric energy meter. , is the spatial lag feature of meter i, B represents the neighbor set of meter i, is the spatial weight between meters i and j; Spatial autocorrelation index unit, used to calculate the spatial autocorrelation index I, I= , n represents the total number of electricity meters, and U represents the mean electricity consumption of all electricity meters.

[0037] As a preferred embodiment of the present invention, the feature screening and optimization module 400 includes: A variance value screening unit, used to calculate the variance value of each feature in the time series feature and the spatial feature, and remove features whose variance value is less than a first threshold; The correlation pruning unit is used to perform correlation pruning, pair the retained features in pairs, and calculate the correlation cardinality between each pair of features. , , Representation characteristics and Features The covariance of and Respectively represent characteristics and Features The standard deviation of The feature information determination unit is used to select one of the two features corresponding to the correlation cardinality greater than the second threshold, and obtain feature information according to the retained feature.

[0038] As a preferred embodiment of the present invention, the data clustering and partitioning module 500 includes: A feature set determination unit is used to determine a feature set F according to feature information: F= , where m is the number of features, and each feature is standardized to have a mean of 0 and a standard deviation of 1; The electric meter clustering and partitioning unit is used to fuse the temporal features and spatial features in the feature information into a feature matrix X, and use K-Means++ or GMM to cluster and partition the electric energy meter.

[0039] The above only describes in detail the preferred embodiments of the present invention, which is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

[0040] It should be understood that, although each step in the flow chart of each embodiment of the present invention is shown in sequence according to the indication of the arrow, these steps are not necessarily performed in sequence according to the order indicated by the arrow. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be performed in other orders. Moreover, at least a portion of the steps in each embodiment may include a plurality of sub-steps or a plurality of stages, and these sub-steps or stages are not necessarily performed at the same time, but can be performed at different times, and the execution order of these sub-steps or stages is not necessarily performed in sequence, but can be performed in turn or alternately with at least a portion of other steps or sub-steps or stages of other steps.

[0041] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by instructing the relevant hardware through a computer program, and the program can be stored in a non-volatile computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM is available in many forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0042] Those skilled in the art will readily appreciate other embodiments of the present disclosure after considering the disclosure in the specification and examples. This application is intended to cover any variations, uses or adaptations of the present disclosure, which follow the general principles of the present disclosure and include common knowledge or customary techniques in the art that are not disclosed in the present disclosure. The specification and examples are intended to be exemplary only, and the true scope and spirit of the present disclosure are indicated by the claims.

Claims

1. A data partition processing method for an electric energy meter, characterized in that: The method comprises the following steps: Collecting electric energy meter information, the electric energy meter information including power consumption data and archive information, the archive information including installation location, substation, feeder, transformer and meter code; Extracting time series features from the electric energy meter information, wherein the time series features include peak-to-valley ratio, daily load rate, periodicity features and autocorrelation coefficient; Extracting spatial features from the electric energy meter information, wherein the spatial features include transformer load rate, electric energy meter topology depth, spatial hysteresis feature and spatial autocorrelation index; Based on correlation pruning, the extracted temporal features and spatial features are screened and optimized to obtain feature information; A clustering partition model is constructed based on the feature information, and a plurality of data partitions are obtained according to the clustering results.

2. The data partition processing method of the electric energy meter according to claim 1 is characterized in that: The step of extracting time series features from the electric energy meter information specifically includes: Calculate the peak-to-valley ratio and daily load rate of electricity consumption data. Peak-to-valley ratio = (maximum electricity consumption - minimum electricity consumption) ÷ average electricity consumption; daily load rate = average electricity consumption ÷ maximum electricity consumption; Determine the power consumption time series x(t) based on the power consumption data, perform discrete Fourier transform on x(t) to obtain X(k), and take the amplitudes of the first K coefficients as the periodic characteristics; Calculate hysteresis based on x(t) The autocorrelation coefficient AT is AT= , represents the mean of the time series, is the lag order, which indicates the number of time units of the time series shifted backward, and N is the length of the time series.

3. The data partition processing method of the electric energy meter according to claim 2 is characterized in that: The step of extracting spatial features from the electric energy meter information specifically includes: Calculate all transformer load factors and meter topology depths, where the meter topology depths are obtained based on topology paths; Calculate the spatial hysteresis characteristics of the energy meter, , is the spatial lag feature of meter i, B represents the neighbor set of meter i, is the spatial weight between meters i and j; Calculate the spatial autocorrelation index I, I= , n represents the total number of electricity meters, and U represents the mean electricity consumption of all electricity meters.

4. The data partition processing method of the electric energy meter according to claim 1, characterized in that: The step of screening and optimizing the extracted temporal features and spatial features based on correlation pruning to obtain feature information specifically includes: Calculate the variance value of each feature in the temporal features and the spatial features, and remove the features whose variance value is less than the first threshold; Perform correlation pruning, pair the retained features in pairs, and calculate the correlation cardinality between each pair of features , , Representation characteristics and Features The covariance of and Respectively represent characteristics and Features The standard deviation of One of the two features corresponding to the correlation cardinality greater than the second threshold is selected, and feature information is obtained according to the retained feature.

5. The data partition processing method of the electric energy meter according to claim 1, characterized in that: The step of constructing a cluster partition model based on the feature information specifically includes: Determine the feature set F based on the feature information: F= , where m is the number of features, and each feature is standardized to have a mean of 0 and a standard deviation of 1; The temporal features and spatial features in the feature information are fused into a feature matrix X, and K-Means++ or GMM is used to cluster and partition the electric energy meter.

6. A data partition processing system for an electric energy meter, characterized in that: The system comprises: The electric meter information collection module is used to collect electric meter information, and the electric meter information includes power consumption data and archive information, and the archive information includes installation location, substation, feeder, transformer and electric meter code; the time series feature extraction module is used to extract time series features from the electric meter information, and the time series features include peak-to-valley ratio, daily load rate, periodicity characteristics and autocorrelation coefficient; A spatial feature extraction module is used to extract spatial features from electric energy meter information, wherein the spatial features include transformer load rate, electric energy meter topology depth, spatial hysteresis feature and spatial autocorrelation index; The feature screening and optimization module is used to screen and optimize the extracted temporal features and spatial features based on correlation pruning to obtain feature information; The data clustering and partitioning module is used to construct a clustering and partitioning model based on the feature information and obtain a number of data partitions according to the clustering results.

7. The data partition processing system of the electric energy meter according to claim 6, characterized in that: The temporal feature extraction module comprises: Daily load rate calculation unit, used to calculate the peak-to-valley ratio and daily load rate of power consumption data. Peak-to-valley ratio = (maximum power consumption - minimum power consumption) ÷ average power consumption. Daily load rate = average power consumption ÷ maximum power consumption. The periodic feature unit is used to determine the power consumption time series x(t) according to the power consumption data, perform discrete Fourier transform on x(t) to obtain X(k), and take the amplitude of the first K coefficients as the periodic feature; Autocorrelation unit, used to calculate lag based on x(t) The autocorrelation coefficient AT, AT = , represents the mean of the time series, is the lag order, which indicates the number of time units of the time series shifted backward, and N is the length of the time series.

8. The data partition processing system for electric energy meters according to claim 7, characterized in that: The spatial feature extraction module comprises: A topology depth determination unit, used to calculate all transformer load rates and electric meter topology depths, wherein the electric meter topology depths are obtained according to the topology path; The spatial hysteresis characteristic unit is used to calculate the spatial hysteresis characteristics of the electric energy meter. , is the spatial lag feature of meter i, B represents the neighbor set of meter i, is the spatial weight between meters i and j; Spatial autocorrelation index unit, used to calculate the spatial autocorrelation index I, I= , n represents the total number of electricity meters, and U represents the mean electricity consumption of all electricity meters.

9. The data partition processing system of the electric energy meter according to claim 6, characterized in that: The feature screening optimization module includes: A variance value screening unit, used to calculate the variance value of each feature in the time series feature and the spatial feature, and remove features whose variance value is less than a first threshold; The correlation pruning unit is used to perform correlation pruning, pair the retained features in pairs, and calculate the correlation cardinality between each pair of features. , , Representation characteristics and Features The covariance of and Respectively represent characteristics and Features The standard deviation of The feature information determination unit is used to select one of the two features corresponding to the correlation cardinality greater than the second threshold, and obtain feature information according to the retained feature.

10. The data partition processing system of the electric energy meter according to claim 6, characterized in that: The data clustering partitioning module comprises: A feature set determination unit is used to determine a feature set F according to feature information: F= , where m is the number of features, and each feature is standardized to have a mean of 0 and a standard deviation of 1; The electric meter clustering and partitioning unit is used to fuse the temporal features and spatial features in the feature information into a feature matrix X, and use K-Means++ or GMM to cluster and partition the electric energy meter.

Images