Electricity consumption data analysis operation and maintenance method and system of electric energy meter

Abstract

The invention relates to the technical field of electric energy meter data analysis, and discloses a power consumption data analysis operation and maintenance method and system of an electric energy meter. The method comprises the following steps of: acquiring and preprocessing electricity utilization state data, namely synchronously acquiring voltage, current and power data, and performing timestamp alignment, data cleaning and resampling; classifying according to preset peak, flat and valley time intervals, and extracting time-frequency domain characteristics such as voltage fluctuation amplitude and current change rate in each interval to generate a set; detecting time period switching frequency, and marking an abnormal interval if the time period switching frequency exceeds a threshold value; generating an abnormal feature subset according to similarity decomposition of historical anomalies and current features; correcting the abnormal feature subset based on the electrical and time sequence parameter association relationship; and reversely backtracking a potential fault source along the fault association map, determining an operation and maintenance priority according to the number of abnormal parameters, position depth and the like, and generating an operation and maintenance instruction. The system comprises corresponding modules such as data acquisition and processing. According to the scheme, data can be accurately analyzed, a fault source is positioned, and the operation and maintenance efficiency is improved.

Description

Technical Field

[0001] The present invention relates to the technical field of electric energy meter data analysis, and in particular to a method and system for analyzing and operating electric energy meter power consumption data. Background Art

[0002] With the rapid development of smart grids, electricity meters, as key devices for metering and data exchange between power users and the grid, have a direct impact on the accuracy and stability of their operating status, including metering precision, user experience, and the safe and economical operation of the grid. With the continuous expansion of the grid and the increasing diversification of user electricity usage, the electricity consumption data collected by meters is becoming massive, multi-dimensional, and dynamically changing. Traditional methods for analyzing and maintaining electricity data face numerous challenges.

[0003] Existing energy meter data analysis and maintenance efforts suffer from deficiencies in data collection and preprocessing. The accuracy of synchronously collecting multiple power status data, such as voltage, current, and power, is insufficient. The simple timestamp alignment method struggles with complex power usage scenarios, resulting in an unreliable data foundation for subsequent analysis. Furthermore, the limited data cleaning method fails to effectively remove high-frequency interference components and eliminate outliers, impacting data quality.

[0004] Traditional methods for processing power usage intervals typically use fixed time period divisions, such as simply dividing them into peak, flat, and off-peak periods, without the ability to dynamically adjust based on actual power usage. When the switching frequency of power usage intervals is abnormal, the lack of effective detection and analysis methods makes it difficult to accurately identify the abnormal period, which in turn affects the judgment of the meter's operating status.

[0005] When it comes to extracting abnormal features from electricity meters, existing technologies often fail to fully consider the similarity between historical abnormal data and current data, and are unable to adopt appropriate decomposition patterns based on different degrees of matching. This results in the extracted abnormal features being inaccurate and incomplete, making it difficult to deeply explore the potential faults of electricity meters.

[0006] In the parameter correlation correction link, traditional methods do not fully utilize the correlation between the electrical parameters and timing parameters of the electricity meter. The correction of parameters such as current amplitude and power gradient is not accurate enough, which makes the reliability of the abnormal feature subset low, affecting subsequent fault analysis and location.

[0007] In addition, in terms of fault source tracing and operation and maintenance decision-making, the existing methods lack systematic analysis based on fault correlation maps, making it difficult to accurately determine the potential fault source electricity meter and its operation and maintenance priority. The generated operation and maintenance instructions are not scientific and reasonable enough, and efficient operation and maintenance management cannot be achieved, resulting in increased operation and maintenance costs and reduced grid operation efficiency. Summary of the Invention

[0008] The object of the present invention is to provide a method and system for analyzing and operating the electricity consumption data of an electric energy meter, so as to solve the problems raised in the above-mentioned background technology.

[0009] To achieve the above objectives, the present invention provides a method for analyzing and operating electricity consumption data of an electric energy meter, the method comprising: S1. Collect and pre-process the electricity consumption status data of the energy meter; S2. Classify the power consumption status data according to the preset power consumption time intervals, extract the time-frequency domain features within each power consumption time interval, and generate a time interval feature set; S3. Detect the switching frequency of the power consumption period. If the switching frequency exceeds the switching frequency threshold, start dynamic analysis strategy switching to mark the abnormal period. S4. Decomposing the period feature set of the non-abnormal period interval according to the similarity matching results between the historical abnormal phase of the electric energy meter and the current time-frequency domain features to generate an abnormal feature subset of the electric energy meter; S5. Based on the correlation between the electrical parameters and the timing parameters of the electric energy meter, perform parameter correlation correction on the abnormal feature subset of the electric energy meter; S6. Based on the corrected subset of abnormal features of the electric energy meter, reversely trace back to the electric energy meter of the potential fault source along the electric energy meter fault association map, determine the operation and maintenance priority and generate operation and maintenance instructions.

[0010] Preferably, collecting and preprocessing the electricity consumption status data of the electric energy meter includes: Synchronously collect the voltage data, current data and power data of the electric energy meter, and align the timestamps of the voltage data, current data and power data; Data cleaning is performed on the voltage, current, and power data after timestamp alignment. This involves using a sliding window filter to remove high-frequency interference components from the voltage data and removing abnormal outliers based on the correlation between the current and power data. The cleaned voltage data, current data and power data are resampled according to the preset sampling frequency to generate a standardized power consumption status data set.

[0011] Preferably, the power consumption status data is classified according to the preset power consumption time intervals, and the time-frequency domain features within each power consumption time interval are extracted to generate a time interval feature set, including: The standardized electricity consumption status data set is divided into multiple electricity consumption period intervals according to the electricity consumption period value, and the electricity consumption period intervals include peak period intervals, normal period intervals and valley period intervals; Extract time-frequency domain features from voltage data, current data, and power data within each electricity consumption period; The time-frequency domain features within the same electricity consumption period are merged according to the time window to generate a multi-dimensional period feature set.

[0012] Preferably, the time-frequency domain characteristics of voltage data include voltage fluctuation amplitude, time-domain mean and frequency-domain center frequency; the time-frequency domain characteristics of current data include current change rate and current fluctuation amplitude; and the time-frequency domain characteristics of power data include power fluctuation frequency and power change gradient.

[0013] Preferably, detecting the switching frequency of the power consumption time interval, and if the switching frequency exceeds the switching frequency threshold, initiating dynamic analysis strategy switching to mark the abnormal time interval, includes: Count the number of switching times during the power consumption period based on the preset time window, calculate the switching frequency per unit time, and determine whether the switching frequency exceeds the switching frequency threshold; If the switching frequency exceeds the switching frequency threshold, the time-frequency domain features of adjacent power consumption time intervals are aligned using the dynamic time warping method, and the cross-time continuity index is extracted; If the cross-time continuity index is lower than the continuity threshold, the power consumption period interval corresponding to the current time window is marked as an abnormal period interval, and the re-extraction of the time-frequency domain features within the abnormal period interval is triggered; otherwise, it is marked as a non-abnormal period interval.

[0014] Preferably, extracting the inter-time period continuity index includes: aligning the time-frequency domain features of adjacent power consumption time periods, calculating the Pearson correlation coefficients of the aligned voltage fluctuation amplitude, current change rate and power change gradient respectively, and taking the average value of the Pearson correlation coefficients as the inter-time period continuity index.

[0015] Preferably, the period feature set of the non-abnormal time interval is decomposed according to the similarity matching result between the historical abnormal stage of the electric energy meter and the current time-frequency domain feature to generate the electric energy meter abnormal feature subset, including: For the current time-frequency domain features in the period feature set of the non-abnormal time interval, the matching degree between the current time-frequency domain features and the historical abnormal stage label of the electric energy meter is determined based on the similarity calculation result of the current time-frequency domain features; Among them, the energy meter's historical abnormal stage label is divided by the voltage amplitude trend and power gradient trend in the energy meter's historical electricity consumption data; If the matching degree is greater than or equal to the preset matching threshold, a low-frequency narrowband decomposition mode is used to extract an abnormal feature subset of the electric energy meter from the time period feature set of the non-abnormal time period; If the matching degree is less than the preset matching threshold, based on the correlation between the harmonic component distribution of the power spectrum in the time period feature set of the non-abnormal time period and the reference frequency of the electrical parameters of the electricity meter, the frequency band where the harmonic energy is concentrated is selected for enhanced decomposition to generate the abnormal feature subset of the electricity meter.

[0016] Preferably, based on the correlation between the electrical parameters and the timing parameters of the electric energy meter, parameter correlation correction is performed on the abnormal feature subset of the electric energy meter, including: According to the correlation between the voltage effective value and the current amplitude in the electrical parameters of the electric energy meter, the amplitude-frequency characteristic of the current amplitude in the abnormal characteristic subset of the electric energy meter is corrected; Based on the matching relationship between the time period identifier in the time series parameter and the power gradient, the power gradient in the abnormal feature subset of the electric energy meter is corrected by time series compensation; The corrected current amplitude and power gradient are combined with the voltage fluctuation amplitude features to generate a parameter-correlated corrected electric energy meter abnormality feature subset.

[0017] Preferably, based on the corrected subset of abnormal characteristics of the electric energy meter, tracing back to the electric energy meter of the potential fault source along the electric energy meter fault association map, determining the operation and maintenance priority and generating the operation and maintenance instructions, including: Obtaining a fault correlation map for electric energy meters. The fault correlation map for electric energy meters is constructed based on the connection relationship between electric energy meter lines and the energy transfer path, and includes the fault correlation direction and weight between electric energy meter nodes. Extract abnormal parameters from the abnormal feature subset of the electric energy meter after parameter association correction, and match the corresponding nodes in the electric energy meter fault association map according to the abnormal parameter type; Traverse in reverse along the fault correlation direction of the electric energy meter fault correlation map to locate the electric energy meter node that is the potential fault source of abnormal parameters; Determine the operation and maintenance priority based on the number and location depth of abnormal parameters of the potential fault source electricity meter node; Based on the operation and maintenance priority and the remaining service life of the electric energy meter, an operation and maintenance instruction is generated, which includes the identification of the fault source electric energy meter and the operation and maintenance time.

[0018] Preferably, the present invention further includes a power consumption data analysis and operation and maintenance system for an electric energy meter, for implementing the above-mentioned power consumption data analysis and operation and maintenance method for an electric energy meter, the system comprising: Data acquisition and processing module: collects and pre-processes the electricity consumption status data of the energy meter; Time period classification and extraction module: classifies the power consumption status data according to the preset power consumption period interval, extracts the time-frequency domain features within each power consumption period interval, and generates a time period feature set; Dynamic strategy marking module: detects the switching frequency of the power consumption period. If the switching frequency exceeds the switching frequency threshold, it starts the dynamic analysis strategy switching to mark the abnormal period. Abnormal feature decomposition module: Decomposes the period feature set of the non-abnormal period interval based on the similarity matching results between the historical abnormal stage of the electric energy meter and the current time-frequency domain features, and generates the abnormal feature subset of the electric energy meter; Parameter association correction module: Based on the correlation between the electrical parameters and timing parameters of the energy meter, parameter association correction is performed on the abnormal feature subset of the energy meter; Fault backtracking decision module: Based on the corrected subset of abnormal features of the electric energy meter, it reversely traces back to the electric energy meter of the potential fault source along the electric energy meter fault association map, determines the operation and maintenance priority and generates operation and maintenance instructions.

[0019] Compared with the prior art, the present invention has the following beneficial effects: The method and system for analyzing and maintaining electricity meter data, provided by this invention, feature optimizations and innovations in multiple aspects. During the data acquisition and preprocessing phase, the data's temporal consistency is ensured by synchronously collecting voltage, current, and power data and aligning timestamps. Sliding window filtering is used to remove high-frequency voltage interference, and outliers are removed based on the correlation between current and power data. The system then resamples data at a preset frequency to generate a standardized data set, improving data quality and providing a reliable foundation for subsequent analysis.

[0020] In the processing of electricity consumption period intervals, the data is divided into peak, flat and valley period intervals, and the time-frequency domain characteristics of the voltage, current and power data in each interval are extracted, such as the voltage fluctuation amplitude, the current change rate, etc., and merged according to the time window to generate a multi-dimensional period feature set, which realizes the refined description of the characteristics of different electricity consumption periods and helps to more accurately analyze the operating status of the electricity meter in different periods.

[0021] By detecting the switching frequency of power consumption time periods, when it exceeds the threshold, the dynamic time warping method is used to align the time-frequency domain features of adjacent time periods, and the average value of the Pearson correlation coefficient is calculated as the continuity indicator across time periods. If it is lower than the threshold, the abnormal time period is marked and the features are re-extracted. This can timely detect abnormal time periods, improve the sensitivity to changes in power consumption time periods and the accuracy of anomaly detection.

[0022] Based on the similarity matching results between historical anomaly phases and current time-frequency domain features, the feature set for non-anomalous periods is decomposed. When the match is high, a low-frequency narrowband decomposition mode is used. When the match is low, a frequency band-based enhanced decomposition is performed based on the correlation between the power spectrum harmonic components and the reference frequency, generating anomaly feature subsets. This differentiated decomposition approach improves the pertinence and accuracy of anomaly feature extraction, enabling more comprehensive exploration of potential anomalies.

[0023] Based on the correlation between the effective value of voltage and the current amplitude in the electrical parameters, the amplitude-frequency characteristics of the current amplitude are corrected. Based on the matching relationship between the time period identifier and the power gradient in the timing parameters, the power gradient is corrected by timing compensation. The corrected features are combined to generate an abnormal feature subset, making the abnormal features more accurate and reliable, providing a more precise basis for subsequent fault analysis.

[0024] By using the fault correlation map constructed based on line connections and energy transfer paths, corresponding nodes are matched according to abnormal parameters, and the potential fault source nodes are located in reverse traversal along the fault correlation direction. The operation and maintenance priority is determined according to the number of abnormal parameters and the depth of the location. Combined with the remaining service life, operation and maintenance instructions containing the fault source identification and operation and maintenance time are generated. This achieves accurate positioning of the fault source and scientific operation and maintenance decision-making, improves operation and maintenance efficiency, reduces operation and maintenance costs, and ensures the safe and stable operation of the power grid. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 This is a working principle diagram of the electricity consumption data analysis and operation and maintenance method of the electric energy meter according to the present invention; Figure 2 Flowchart for electricity status data collection and preprocessing; Figure 3 Flowchart for electricity consumption period classification and time-frequency domain feature extraction; Figure 4 Flowcharts generated for fault tracing and operation and maintenance instructions; Figure 5 This is a flowchart of the electricity consumption data analysis and operation and maintenance system module. DETAILED DESCRIPTION

[0026] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0027] See also Figure 1-Figure 5 The present invention provides a method for analyzing and operating the electricity consumption data of an electric energy meter. The specific implementation steps are as follows: Collect and preprocess electricity meter usage data. Simultaneously collect voltage, current, and power data from the meter and align their timestamps to ensure consistency across the time dimension. The timestamp-aligned data is then cleaned using a sliding window filter to remove high-frequency interference from the voltage data. Outliers are removed based on the correlation between current and power data to ensure data accuracy and reliability. Finally, the cleaned data is resampled at the preset sampling frequency to generate a standardized electricity usage data set.

[0028] The power consumption status data is classified according to the preset power consumption time intervals, and the time-frequency domain features within each power consumption time interval are extracted to generate a time period feature set. The standardized power consumption status data set is divided into multiple power consumption time intervals such as peak time intervals, normal time intervals, and valley time intervals according to the power consumption time interval values. The time-frequency domain features of the voltage data, current data, and power data within each power consumption time interval are extracted separately. The time-frequency domain features of the voltage data include the voltage fluctuation amplitude, time domain mean, and frequency domain center frequency; the time-frequency domain features of the current data include the current change rate and current fluctuation amplitude; and the time-frequency domain features of the power data include the power fluctuation frequency and power change gradient. The time-frequency domain features within the same power consumption time interval are merged according to the time window to generate a multi-dimensional time period feature set.

[0029] The switching frequency of the power consumption period is detected. If the switching frequency exceeds the switching frequency threshold, the dynamic analysis strategy switching is initiated to mark the abnormal period. The number of switching times of the power consumption period is counted based on the preset time window, the switching frequency per unit time is calculated, and it is determined whether the switching frequency exceeds the preset switching frequency threshold. If it exceeds the threshold, the time-frequency domain features of adjacent power consumption periods are aligned using the dynamic time warping method, and the cross-period continuity index is extracted. Specifically, the time-frequency domain features of adjacent power consumption periods are aligned, and the Pearson correlation coefficients of the aligned voltage fluctuation amplitude, current change rate, and power change gradient are calculated respectively. The average value of these Pearson correlation coefficients is used as the cross-period continuity index. If the cross-period continuity index is lower than the continuity threshold, the power consumption period corresponding to the current time window is marked as an abnormal period, and the re-extraction of the time-frequency domain features within the abnormal period is triggered; otherwise, it is marked as a non-abnormal period.

[0030] Based on the similarity matching results between the meter's historical abnormal phases and the current time-frequency domain features, the time feature set for the non-abnormal time interval is decomposed to generate a subset of abnormal meter features. The degree of match between the current time-frequency domain features and the meter's historical abnormal phases is determined based on the similarity calculation results between the meter's historical abnormal phase labels and the current time-frequency domain features. The meter's historical abnormal phase labels are classified based on the voltage amplitude and power gradient trends in the meter's historical electricity usage data. If the degree of match is greater than or equal to a preset matching threshold, a low-frequency narrowband decomposition mode is used to extract a subset of abnormal meter features from the time feature set for the non-abnormal time interval. If the degree of match is less than the preset matching threshold, based on the correlation between the harmonic component distribution of the power spectrum in the time feature set for the non-abnormal time interval and the reference frequency of the meter's electrical parameters, frequency bands with concentrated harmonic energy are selected for enhanced decomposition to generate a subset of abnormal meter features.

[0031] Based on the correlation between the electrical parameters and timing parameters of the energy meter, parameter correlation correction is performed on the energy meter abnormal feature subset. Based on the correlation between the voltage RMS value and the current amplitude in the energy meter electrical parameters, the current amplitude in the energy meter abnormal feature subset is corrected for its amplitude-frequency characteristics. Based on the matching relationship between the time period identifier and the power gradient in the timing parameters, the power gradient in the energy meter abnormal feature subset is corrected for timing compensation. The corrected current amplitude and power gradient are combined with the voltage fluctuation amplitude characteristics to generate the energy meter abnormal feature subset with parameter correlation correction.

[0032] Based on the corrected subset of energy meter abnormality features, the potential fault source energy meter is traced back along the energy meter fault association map to determine the operation and maintenance priority and generate operation and maintenance instructions. The energy meter fault association map is obtained. This map is constructed based on the energy meter line connection relationship and energy transfer path, and contains the fault association direction and weight between energy meter nodes. Abnormal parameters are extracted from the energy meter abnormality feature subset after parameter association correction, and the corresponding nodes in the energy meter fault association map are matched according to the abnormal parameter type. The energy meter fault association map is traversed in reverse along the fault association direction to locate the potential fault source energy meter node of the abnormal parameter. The operation and maintenance priority is determined based on the number and location depth of abnormal parameters of the potential fault source energy meter node. Based on the operation and maintenance priority and the remaining service life of the energy meter, an operation and maintenance instruction is generated, which includes the fault source energy meter identifier and operation and maintenance time.

[0033] Example 1: When collecting and preprocessing the meter's power usage data, the meter's voltage, current, and power data are collected simultaneously. Voltage data is collected using a specific voltage sensor, which monitors the potential difference across the meter in real time and converts it into an electrical signal for output. Current data is acquired using a current transformer. This current transformer, mounted on the circuit's conductors, uses electromagnetic induction to convert high currents into low current signals for subsequent collection and processing. Power data is collected based on the product of voltage and current data. During the collection process, synchronization of voltage and current data is crucial to ensure the accuracy of the power data.

[0034] After collecting voltage, current, and power data, these data need to be timestamped. Timestamp alignment is achieved by adding a precise time stamp to each data point. Specifically, during data collection, the system's clock module generates high-precision timestamps with millisecond accuracy or higher, ensuring consistency across the time dimension for different data types. For each collected voltage, current, and power data point, a corresponding timestamp is associated with it, ensuring a clear correspondence between these data points along the time axis.

[0035] After completing the timestamp alignment, the next step is data cleaning. A sliding window filter is used to process the high-frequency interference components in the voltage data. The specific operation of sliding window filtering is to set a fixed-length window on the voltage data sequence. For example, the window can be set to contain N consecutive data points. Then, for the N voltage data points within the window, their average or other statistics are calculated, and this statistic is used to replace the intermediate data points in the window. As the window slides point by point on the data sequence, the above operation is repeated continuously, thereby achieving smoothing of the voltage data and effectively removing the high-frequency interference components. This method can filter out high-frequency noise caused by external electromagnetic interference and other factors while retaining the main change trends of the voltage data, making the voltage data more stable.

[0036] At the same time, abnormal outliers are removed based on the correlation between current data and power data. Since power is equal to voltage multiplied by current, when the circuit parameters are stable, there is a strong linear correlation between current data and power data. In order to identify abnormal outliers, the correlation coefficient of the current data series and the power data series is first calculated. The correlation coefficient can be calculated using the Pearson correlation coefficient method, which can measure the degree of linear correlation between two variables. For each current data point and the corresponding power data point, their correlation with other data points is calculated. When the correlation coefficient of a data point is lower than the preset threshold, it means that the correlation between the data point and other data points is poor, which may be caused by measurement error, equipment failure or other abnormal conditions. Therefore, it is identified as an abnormal outlier and removed. In this way, the accuracy and reliability of current data and power data can be guaranteed.

[0037] Finally, the cleaned voltage data, current data and power data are resampled according to the preset sampling frequency. The preset sampling frequency is determined based on the needs of subsequent data analysis and the working characteristics of the electricity meter. For example, it can be set to a sampling frequency of 100 times per second or higher. The resampling process is to convert the cleaned data from the original sampling rate to the preset sampling frequency. Specifically, interpolation algorithms such as linear interpolation and polynomial interpolation can be used to insert new data points between the original data points, so that the data is more evenly distributed on the time axis, thereby generating a standardized power consumption status data set. The standardized power consumption status data set has a unified time interval and data format, which facilitates various subsequent analysis and processing operations, such as feature extraction, pattern recognition, etc.

[0038] Throughout the data collection and preprocessing process, attention must be paid to the accuracy and reliability of each link. For example, during data collection, ensure the accuracy and stability of sensors and perform regular sensor calibration and maintenance. During timestamp alignment, ensure the accuracy of the system clock. Timestamp accuracy can be improved by synchronizing with an external clock source. During data cleaning and resampling, appropriate parameters must be selected, such as the sliding window length, correlation coefficient threshold, and resampling frequency, to ensure that the processed data truly reflects the meter's power usage.

[0039] Example 2: When classifying electricity usage data according to preset time periods and extracting time-frequency domain features to generate time period feature sets, the standardized electricity usage data set must first be divided according to the time period. The division of electricity usage time periods is based on the conventional load characteristics of the power system and typically includes peak, normal, and valley periods. Peak periods are generally set as the daily peak electricity consumption period, such as 8:00 AM to 10:00 PM, when electricity demand is high and load fluctuations are relatively significant. Normal periods can be set as certain regular time periods outside of peak periods, such as the period from 10:00 PM to 6:00 AM the following day, when electricity demand is relatively stable. Valley periods are periods of low electricity consumption, such as the early morning to early morning period, when load levels are low. During this division, the system assigns each electricity usage data point to the corresponding time period based on the timestamp information in the data, forming data sets for different time periods.

[0040] For voltage data within each power usage period, its unique time-frequency domain characteristics must be extracted. The voltage fluctuation amplitude is extracted by calculating the difference between the maximum and minimum voltage values ​​within that period, thereby reflecting the voltage's range of variation within that period. For example, during peak hours, voltage may fluctuate significantly due to load changes, and the fluctuation amplitude provides a visual representation of the extent of this variation. The time-domain mean is calculated by taking the arithmetic average of all voltage data points within that period to obtain the average voltage level for that period. This indicator reflects the baseline voltage status during that period. Extracting the frequency-domain center frequency requires first performing a frequency-domain conversion on the voltage data. Methods such as the Fast Fourier Transform (FFT) can be used to convert the time-domain voltage signal into a frequency-domain signal. The center frequency of the energy distribution is then calculated. This frequency reflects the frequency band where the primary energy in the voltage signal is concentrated, facilitating analysis of the frequency characteristics of voltage fluctuations.

[0041] For current data, the current change rate and current fluctuation amplitude are extracted. The current change rate is calculated by performing differential processing on the current data within the time period to obtain the current change between adjacent data points. This is then divided by the time interval to obtain the current change rate per unit time. This indicator can be used to measure how quickly the current changes within the time period. For example, when the load starts or stops, the current change rate will change significantly. The current fluctuation amplitude is extracted in a similar way to the voltage. By calculating the difference between the maximum and minimum values ​​of the current data within the time period, the current fluctuation range is reflected, helping to determine whether the current is stable.

[0042] The time-frequency domain characteristics of power data include the power fluctuation frequency and the power change gradient. Extracting the power fluctuation frequency requires first performing a frequency domain analysis on the power data. Similarly, transformation methods such as FFT can be used to convert the power time domain signal into a frequency domain signal. The main frequency component of the power fluctuation can then be determined. This frequency can reflect the speed of the power fluctuation. For example, when certain devices are started, the power may fluctuate at a specific frequency. The power change gradient is calculated by fitting the power data to obtain the slope of the power change over time during the period. This reflects the power change trend during the period, whether it is rising, falling, or stabilizing. This plays an important role in analyzing the changing characteristics of the power load.

[0043] After extracting the time-frequency domain features within each time period, the features within the same electricity usage period need to be merged according to time windows. The time window setting can be determined based on actual analysis requirements, for example, 15 minutes, 30 minutes, or 1 hour. The feature data within each time window is processed using an aggregation method. For example, statistics are collected on features such as the voltage fluctuation amplitude, time domain mean, frequency domain center frequency, current change rate, fluctuation amplitude, power fluctuation frequency, and change gradient within the window, and their average, maximum, minimum, or other statistical quantities are calculated. This allows the multidimensional features within each time window to be integrated into a single feature vector. The feature vectors of multiple time windows are combined to generate a multidimensional time period feature set.

[0044] This time period feature set contains multi-dimensional feature information for different power consumption periods within different time windows, and can comprehensively reflect the power consumption status of the electricity meter in different time periods. For example, in the feature set of the peak period, the voltage fluctuation amplitude may be large, and the power change gradient may also show a specific trend of change, while the characteristics of the valley period may be characterized by relatively stable voltage and low power levels. By analyzing these feature sets, it can provide rich data support for subsequent power consumption period switching frequency detection and abnormal feature identification, helping operation and maintenance personnel to gain a deeper understanding of the operating status of the electricity meter and promptly detect potential abnormalities.

[0045] Example 3: When detecting the switching frequency between power usage periods and marking abnormal periods, the number of switching times within the power usage period is counted based on a preset time window. The length of the preset time window can be set based on the actual application scenario, for example, 1 hour, 6 hours, or 1 day. The specific duration should be determined based on the operating rules of the power system and data analysis requirements. Within this time window, the system monitors changes in the power usage period in real time. Each time the power usage status switches from one period (such as peak period) to another period (such as normal period), the switching count is increased by 1. By accumulating the number of switching times within this time window, a data basis is provided for the subsequent calculation of the switching frequency.

[0046] Calculate the switching frequency of the power consumption period within a unit time. The calculation formula for the switching frequency is: ; in, Indicates the switching frequency, in times / hour; Indicates the number of switching times between power consumption periods within the preset time window; Indicates the duration of the preset time window, in hours. For example, if the preset time window is 1 hour and 5 time period switches are detected within that window, the switching frequency is 5 times / hour. After calculating the switching frequency, it is compared with the preset switching frequency threshold. The switching frequency threshold is pre-set based on the normal operating parameters and historical data of the power system and is used to determine whether time period switching is too frequent. If the switching frequency exceeds this threshold, it indicates that the current time period switching is abnormally frequent, which may indicate unstable power consumption or other potential problems, requiring further analysis.

[0047] When the switching frequency exceeds the threshold, the time-frequency domain features of adjacent power consumption periods need to be aligned using the Dynamic Time Warping (DTW) method. The dynamic time warping method is an algorithm for aligning two time series. It allows the sequences to be stretched or compressed on the time axis to find the best match between the two sequences. When processing the time-frequency domain features of adjacent time periods, the DTW algorithm calculates the mapping relationship between different time points based on the shape and similarity of the feature sequences, thereby achieving feature alignment. For example, for the voltage fluctuation amplitude sequence of the previous period and the voltage fluctuation amplitude sequence of the next period, the DTW algorithm will find a time mapping method that minimizes the difference between the two sequences, thereby completing the alignment process.

[0048] After alignment, it is necessary to extract the inter-time continuity index. Specifically, after aligning the time-frequency domain features of adjacent power consumption time intervals, the Pearson correlation coefficient is calculated for the aligned voltage fluctuation amplitude, current change rate, and power change gradient, and the average of these Pearson correlation coefficients is used as the inter-time continuity index. The calculation formula of the Pearson correlation coefficient is: ; in, represents the Pearson correlation coefficient; Indicates the length of the feature sequence after alignment; and Represents the first The value of the data point; and Represents the average value of two characteristic sequences. The value range of Pearson correlation coefficient is , the larger its absolute value is, the stronger the linear correlation between the two feature sequences is.

[0049] For the voltage fluctuation amplitude, the Pearson correlation coefficient of the voltage fluctuation amplitude sequence after alignment in the previous period and the voltage fluctuation amplitude sequence after alignment in the next period is calculated, which is recorded as ; For the current change rate, calculate the Pearson correlation coefficient of the current change rate sequence after adjacent time periods are aligned, which is recorded as ; For the power change gradient, calculate the Pearson correlation coefficient of the power change gradient sequence after alignment of adjacent time periods, denoted as Then, the inter-period continuity indicator The calculation formula is: ; This indicator is used to measure the continuity of the time-frequency domain characteristics of adjacent electricity consumption time periods.

[0050] After obtaining the cross-time continuity index, it needs to be compared with the continuity threshold. The continuity threshold is pre-set based on historical data and the correlation between features under normal operating conditions. If the cross-time continuity index is lower than the continuity threshold, it means that the time-frequency domain features between adjacent time periods are quite different and the continuity is poor, and there may be an abnormality. At this time, it is necessary to mark the power consumption period interval corresponding to the current time window as an abnormal period interval and trigger the re-extraction of the time-frequency domain features within the abnormal period interval. When re-extracting features, a finer time window or a more complex feature extraction algorithm can be used to more accurately capture the power consumption characteristics within the period. If the cross-time continuity index is not lower than the continuity threshold, it means that the feature continuity of adjacent time periods is good and it is normal. In this case, the power consumption period interval corresponding to the current time window is marked as a non-abnormal period interval.

[0051] Throughout the detection and labeling process, attention should be paid to the reasonable selection of the preset time window, the scientific setting of the switching frequency threshold and the continuity threshold, as well as the accuracy of the dynamic time warping method and the Pearson correlation coefficient calculation. The selection of these parameters and algorithms directly affects the effectiveness and accuracy of abnormal period detection. For example, if the preset time window is too long, it may mask the frequent switching phenomenon in the short term, while if it is too short, the switching number statistics may be unstable. The setting of the switching frequency threshold and the continuity threshold needs to fully consider the actual operation of the power system and historical data to avoid misjudgment or missed judgment. In addition, when performing feature alignment and correlation coefficient calculation, the integrity and accuracy of the data must be ensured. Missing or abnormal data points must be processed in advance, such as interpolation or elimination, to ensure the reliability of the calculation results.

[0052] Example 4: When decomposing the time feature set for non-abnormal periods to generate a subset of meter abnormality features based on the similarity matching results between the meter's historical abnormal phases and the current time-frequency domain features, the current time-frequency domain features in the time feature set for non-abnormal periods must first be analyzed to determine their degree of match with the meter's historical abnormal phases. The meter's historical abnormal phase labels are assigned based on an analysis of voltage amplitude and power gradient trends in the meter's historical electricity usage data.

[0053] For example, if the voltage amplitude shows a continuous downward trend over a period of time in historical electricity consumption data, while the power gradient also shows an abnormal increase, this period may be classified as the "abnormal voltage decline phase." Meanwhile, in another period of historical data, the voltage amplitude fluctuates frequently within a certain range, and the power gradient also shows irregular changes. This period may be marked as the "abnormal power fluctuation phase." The system will classify these historical abnormal phases and assign corresponding labels to form a historical abnormal phase database.

[0054] When processing the time feature set for the current non-abnormal period, the system calculates the similarity between the current time-frequency domain features and the features corresponding to the historical abnormal phase labels. For example, consider features such as voltage fluctuation amplitude, current change rate, and power fluctuation frequency. Assume the voltage fluctuation amplitude for the current period is 20V, while the corresponding average voltage fluctuation amplitude features for the historical abnormal phase "voltage abnormal decrease phase" are 25V, 5A / s, and 10Hz, respectively. The system then uses specific algorithms (such as Euclidean distance and cosine similarity) to calculate the similarity between the current features and the features of the historical abnormal phases.

[0055] If the calculated matching degree is greater than or equal to the preset matching threshold, it means that the current time-frequency domain features are relatively similar to a certain historical abnormal stage, and there may be an abnormal situation. At this time, the system uses the low-frequency narrowband decomposition mode to extract the abnormal feature subset of the electricity meter from the period feature set of the non-abnormal time interval. The low-frequency narrowband decomposition mode mainly decomposes the signal in the low-frequency band. For example, when it is detected that the voltage fluctuation feature has abnormal energy concentration in the low-frequency band (such as below 50Hz), the system will focus on the low-frequency band and perform more detailed decomposition and extraction of features such as the voltage fluctuation amplitude and time domain mean to obtain a feature subset that may contain abnormalities.

[0056] If the match level is less than the preset matching threshold, it indicates that the current feature has a low similarity to the historical abnormal period. In this case, processing is required based on the correlation between the harmonic component distribution of the power spectrum in the period feature set of the non-abnormal period and the reference frequency of the electrical parameters of the energy meter. The reference frequency of the electrical parameters of the energy meter is typically 50Hz or 60Hz. The harmonic components in the power spectrum are frequency components above the reference frequency.

[0057] For example, suppose the power spectrum for the current time period shows concentrated energy at 150 Hz (3rd harmonic) and 250 Hz (5th harmonic). The base frequency is 50 Hz, and the 3rd and 5th harmonics are integer multiples of the base frequency. In this case, the system selects the frequency band where these harmonic energy is concentrated (e.g., 150 Hz to 250 Hz) for enhanced decomposition. During this enhanced decomposition, the power fluctuation frequency, power variation gradient, and corresponding voltage and current characteristics within this frequency band are analyzed and extracted in greater depth, generating a subset of abnormal meter features.

[0058] Taking a specific application scenario as an example, suppose that within the time feature set for a certain electricity meter during a non-abnormal period, the current time-frequency domain features show a voltage fluctuation amplitude of 18V, a time-domain mean of 220V, and a frequency-domain center frequency of 50Hz; a current change rate of 3A / s, a current fluctuation amplitude of 10A; and a power fluctuation frequency of 50Hz, with a power change gradient of 100W / s. The system matches these features with historical abnormal phase labels. The historical abnormal phase "voltage sag" is characterized by a voltage fluctuation amplitude of 25V, a time-domain mean of 210V, a frequency-domain center frequency of 50Hz, a current fluctuation rate of 4A / s, and a power change gradient of 150W / s. Based on a similarity calculation, if the matching degree is 0.8 and the preset matching threshold is 0.7, the system will use low-frequency narrowband decomposition mode to decompose the voltage, current, and power features in the low-frequency band, extracting features such as the voltage fluctuation amplitude and the time-domain mean as the abnormal feature subset.

[0059] In another case, the degree of match between the current time-frequency domain features and the historical abnormal phase is 0.6, which is less than the preset threshold of 0.7. In this case, the system analyzes the harmonic component distribution of the power spectrum and finds that energy is concentrated at 200 Hz (the fourth harmonic). Since the reference frequency is 50 Hz, the fourth harmonic is correlated with the reference frequency. The system selects a frequency band near 200 Hz for enhanced decomposition, extracting features such as the power fluctuation frequency, power change gradient, and corresponding voltage fluctuation amplitude and current change rate within this frequency band to generate an abnormal feature subset.

[0060] Throughout the entire processing process, attention must be paid to the accuracy and comprehensiveness of the labeling of historical anomaly stages, which directly impacts the effectiveness of similarity matching. Furthermore, the preset matching threshold needs to be adjusted based on actual operating data and experience to ensure accurate identification of potential anomalies while avoiding misjudgments. Furthermore, the selection and parameter settings of the low-frequency narrowband decomposition mode and the enhanced decomposition method based on harmonic component distribution also need to be optimized based on factors such as the specific meter model and operating environment to ensure that the extracted anomaly feature subset truly reflects the meter's potential anomalies.

[0061] Example 5: When performing parameter correlation correction on the meter's abnormal feature subset based on the correlation between the meter's electrical parameters and timing parameters, the current amplitude in the meter's abnormal feature subset must be corrected for its amplitude-frequency characteristics based on the correlation between the voltage RMS value and the current amplitude within the meter's electrical parameters. The voltage RMS value refers to the root mean square value of the voltage within a cycle and reflects the voltage's actual work capacity. During normal circuit operation, a specific correlation exists between the voltage RMS value and the current amplitude, which is typically determined by the circuit's load characteristics. For example, when the load is linear, the current amplitude is proportional to the voltage RMS value. When the load is nonlinear, this relationship may become more complex, but it still maintains a certain regularity.

[0062] The system first calculates the current RMS voltage and establishes a mapping model based on the correspondence between voltage RMS and current amplitude in historical data. If the current amplitude in the meter's abnormal feature subset deviates from the expected value in this mapping model, the current amplitude needs to be corrected for its amplitude-frequency characteristics. For example, if the current RMS voltage is 220V, the historical data model indicates a normal current amplitude of approximately 5A, while the current amplitude in the abnormal feature subset is 8A. The system then analyzes whether this deviation is due to a change in the amplitude-frequency characteristics and, based on the analysis, makes appropriate corrections to the current amplitude to better align it with actual operating conditions.

[0063] Based on the matching relationship between the time period identifier and the power gradient in the timing parameters, time series compensation correction is performed on the power gradient in the abnormal feature subset of the electricity meter. The time period identifier is used to distinguish different electricity consumption periods, such as peak period, normal period, and off-peak period. The characteristics of the power load vary in different time periods, and the power gradient will also show different variation patterns. For example, during peak period, due to the large number of electrical devices, the power may fluctuate greatly, and the change in power gradient is relatively large. During off-peak period, there are fewer electrical devices, the power is relatively stable, and the change in power gradient is small.

[0064] The system will first determine the time period identifier corresponding to the current abnormal feature subset, and then obtain the corresponding normal power gradient variation range and pattern from historical data based on the time period identifier, and establish a matching model between the time period and the power gradient. When the power gradient in the abnormal feature subset does not match the expected value of the matching model, timing compensation correction is required. For example, the current time period is a peak period. According to the matching model, the normal variation range of the power gradient should be within ±500W / s, while the power gradient in the abnormal feature subset is 1000W / s. At this time, the system will analyze whether the deviation is caused by timing factors, and based on the analysis results, perform timing compensation correction on the power gradient to make it conform to the normal variation pattern of the time period.

[0065] After correcting the current amplitude and power gradient, these corrected values ​​need to be combined with the voltage fluctuation amplitude characteristics to generate a parameter-correlated and corrected subset of meter anomaly features. The voltage fluctuation amplitude characteristic reflects voltage fluctuations during operation and is a key indicator for meter anomaly detection. Combining the corrected current amplitude, power gradient, and voltage fluctuation amplitude characteristics creates a more comprehensive and accurate set of anomaly features, effectively reflecting the actual operating status of the meter.

[0066] For example, the corrected current amplitude is 6A, the power gradient is 300W / s, and the voltage fluctuation amplitude is 20V. Combining these features creates a subset of abnormal features encompassing current, power, and voltage. This subset can more comprehensively describe potential abnormalities during meter operation, providing a more reliable basis for subsequent fault tracing and maintenance decisions.

[0067] During the entire parameter correlation correction process, the following points need to be noted: First, when establishing the correlation model between the voltage RMS value and the current amplitude, as well as the matching model between the time period identifier and the power gradient, it is necessary to make full use of historical data to ensure the accuracy and reliability of the model. The collection and processing of historical data should be as comprehensive and accurate as possible, covering a variety of different operating conditions and environmental conditions. Second, when making corrections, the influence of various factors should be taken into account, such as harmonic interference in the circuit and sudden changes in the load, to avoid deviations in the correction results due to these factors. Finally, when merging features, attention should be paid to the correlation and complementarity between features to ensure that the generated abnormal feature subset can comprehensively and accurately reflect the abnormal conditions of the electricity meter.

[0068] By performing parameter correlation correction on the abnormal feature subset of the energy meter through the above steps, the accuracy and reliability of the abnormal features can be effectively improved. This lays a solid foundation for subsequently tracing back to the potential fault source energy meter along the energy meter fault correlation map based on the corrected abnormal feature subset, determining operation and maintenance priorities, and generating operation and maintenance instructions. This correction method can fully utilize the correlation between the electrical parameters and timing parameters of the energy meter to perform multi-angle and comprehensive correction of abnormal features, thereby improving the accuracy and reliability of the entire electricity data analysis and operation and maintenance system, helping operation and maintenance personnel to more accurately identify potential faults in the energy meter and take appropriate operation and maintenance measures in a timely manner to ensure the normal operation of the energy meter.

[0069] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus.

[0070] While embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions, and variations may be made to these embodiments without departing from the principles and spirit of the invention, and that the scope of the invention is defined by the appended claims and their equivalents.

Claims

1. A method for analyzing and operating electricity consumption data of an electric energy meter, characterized in that: The steps include: S1. Collect and pre-process the electricity consumption status data of the energy meter; S2. Classify the power consumption status data according to the preset power consumption time intervals, extract the time-frequency domain features within each power consumption time interval, and generate a time interval feature set; S3. Detect the switching frequency of the power consumption period. If the switching frequency exceeds the switching frequency threshold, start dynamic analysis strategy switching to mark the abnormal period. S4. Decomposing the period feature set of the non-abnormal period interval according to the similarity matching results between the historical abnormal phase of the electric energy meter and the current time-frequency domain features to generate an abnormal feature subset of the electric energy meter; S5. Based on the correlation between the electrical parameters and the timing parameters of the electric energy meter, perform parameter correlation correction on the abnormal feature subset of the electric energy meter; S6. Based on the corrected subset of abnormal features of the electric energy meter, reversely trace back to the electric energy meter of the potential fault source along the electric energy meter fault association map, determine the operation and maintenance priority and generate operation and maintenance instructions.

2. The method for analyzing and operating the electricity consumption data of an electric energy meter according to claim 1, characterized in that: Collect and pre-process the electricity consumption status data of the energy meter, including: Synchronously collect the voltage data, current data and power data of the electric energy meter, and align the timestamps of the voltage data, current data and power data; Data cleaning is performed on the voltage, current, and power data after timestamp alignment. This involves using a sliding window filter to remove high-frequency interference components from the voltage data and removing abnormal outliers based on the correlation between the current and power data. The cleaned voltage data, current data and power data are resampled according to the preset sampling frequency to generate a standardized power consumption status data set.

3. The method for analyzing and operating the electricity consumption data of an electric energy meter according to claim 1, characterized in that: The power consumption status data is classified according to the preset power consumption time intervals, and the time-frequency domain features within each power consumption time interval are extracted to generate a time period feature set, including: The standardized electricity consumption status data set is divided into multiple electricity consumption period intervals according to the electricity consumption period value, and the electricity consumption period intervals include peak period intervals, normal period intervals and valley period intervals; Extract time-frequency domain features from voltage data, current data, and power data within each electricity consumption period; The time-frequency domain features within the same electricity consumption period are merged according to the time window to generate a multi-dimensional period feature set.

4. The method for analyzing and operating the electricity consumption data of an electric energy meter according to claim 3, characterized in that: The time-frequency domain characteristics of voltage data include voltage fluctuation amplitude, time-domain mean and frequency-domain center frequency; the time-frequency domain characteristics of current data include current change rate and current fluctuation amplitude; the time-frequency domain characteristics of power data include power fluctuation frequency and power change gradient.

5. The method for analyzing and operating the electricity consumption data of an electric energy meter according to claim 1, characterized in that: Detect the switching frequency of the power consumption period. If the switching frequency exceeds the switching frequency threshold, initiate dynamic analysis strategy switching to mark the abnormal period, including: Count the number of switching times during the power consumption period based on the preset time window, calculate the switching frequency per unit time, and determine whether the switching frequency exceeds the switching frequency threshold; If the switching frequency exceeds the switching frequency threshold, the time-frequency domain features of adjacent power consumption time intervals are aligned using the dynamic time warping method, and the cross-time continuity index is extracted; If the cross-time continuity index is lower than the continuity threshold, the power consumption period interval corresponding to the current time window is marked as an abnormal period interval, and the re-extraction of the time-frequency domain features within the abnormal period interval is triggered; otherwise, it is marked as a non-abnormal period interval.

6. The method for analyzing and operating the electricity consumption data of an electric energy meter according to claim 5, characterized in that: Extracting the inter-time continuity index includes: aligning the time-frequency domain features of adjacent power consumption time intervals, calculating the Pearson correlation coefficients of the aligned voltage fluctuation amplitude, current change rate and power change gradient, and taking the average value of the Pearson correlation coefficients as the inter-time continuity index.

7. The method for analyzing and operating the electricity consumption data of an electric energy meter according to claim 1, characterized in that: Based on the similarity matching results between the historical abnormal phases of the electric energy meter and the current time-frequency domain features, the period feature set of the non-abnormal period is decomposed to generate the electric energy meter abnormal feature subset, including: For the current time-frequency domain features in the period feature set of the non-abnormal time interval, the matching degree between the current time-frequency domain features and the historical abnormal stage label of the electric energy meter is determined based on the similarity calculation result of the current time-frequency domain features; Among them, the energy meter's historical abnormal stage label is divided by the voltage amplitude trend and power gradient trend in the energy meter's historical electricity consumption data; If the matching degree is greater than or equal to the preset matching threshold, a low-frequency narrowband decomposition mode is used to extract an abnormal feature subset of the electric energy meter from the time period feature set of the non-abnormal time period; If the matching degree is less than the preset matching threshold, based on the correlation between the harmonic component distribution of the power spectrum in the time period feature set of the non-abnormal time period and the reference frequency of the electrical parameters of the electricity meter, the frequency band where the harmonic energy is concentrated is selected for enhanced decomposition to generate the abnormal feature subset of the electricity meter.

8. The method for analyzing and operating the electricity consumption data of an electric energy meter according to claim 1, characterized in that: Based on the correlation between the electrical parameters and timing parameters of the energy meter, parameter correlation correction is performed on the abnormal feature subset of the energy meter, including: According to the correlation between the voltage effective value and the current amplitude in the electrical parameters of the electric energy meter, the amplitude-frequency characteristic of the current amplitude in the abnormal characteristic subset of the electric energy meter is corrected; Based on the matching relationship between the time period identifier in the time series parameter and the power gradient, the power gradient in the abnormal feature subset of the electric energy meter is corrected by time series compensation; The corrected current amplitude and power gradient are combined with the voltage fluctuation amplitude features to generate a parameter-correlated corrected electric energy meter abnormality feature subset.

9. The method for analyzing and operating the electricity consumption data of an electric energy meter according to claim 1, characterized in that: Based on the corrected subset of abnormal energy meter features, the system traces back to the potential fault source energy meter along the energy meter fault association map, determines the operation and maintenance priority, and generates operation and maintenance instructions, including: Obtaining a fault correlation map for electric energy meters. The fault correlation map for electric energy meters is constructed based on the connection relationship between electric energy meter lines and the energy transfer path, and includes the fault correlation direction and weight between electric energy meter nodes. Extract abnormal parameters from the abnormal feature subset of the electric energy meter after parameter association correction, and match the corresponding nodes in the electric energy meter fault association map according to the abnormal parameter type; Traverse in reverse along the fault correlation direction of the electric energy meter fault correlation map to locate the electric energy meter node that is the potential fault source of abnormal parameters; Determine the operation and maintenance priority based on the number and location depth of abnormal parameters of the potential fault source electricity meter node; Based on the operation and maintenance priority and the remaining service life of the electric energy meter, an operation and maintenance instruction is generated, which includes the identification of the fault source electric energy meter and the operation and maintenance time.

10. An electric energy meter power consumption data analysis and operation and maintenance system, used to implement the electric energy meter power consumption data analysis and operation and maintenance method according to any one of claims 1 to 9, characterized in that: include: Data acquisition and processing module: collects and pre-processes the electricity consumption status data of the energy meter; Time period classification and extraction module: classifies the power consumption status data according to the preset power consumption period interval, extracts the time-frequency domain features within each power consumption period interval, and generates a time period feature set; Dynamic strategy marking module: detects the switching frequency of the power consumption period. If the switching frequency exceeds the switching frequency threshold, it starts the dynamic analysis strategy switching to mark the abnormal period. Abnormal feature decomposition module: Decomposes the period feature set of the non-abnormal period interval based on the similarity matching results between the historical abnormal stage of the electric energy meter and the current time-frequency domain features, and generates the abnormal feature subset of the electric energy meter; Parameter association correction module: Based on the correlation between the electrical parameters and timing parameters of the energy meter, parameter association correction is performed on the abnormal feature subset of the energy meter; Fault backtracking decision module: Based on the corrected subset of abnormal features of the electric energy meter, it reversely traces back to the electric energy meter of the potential fault source along the electric energy meter fault association map, determines the operation and maintenance priority and generates operation and maintenance instructions.

Images