A bidirectional metering correction method and system for an electric energy meter

Abstract

The present application belongs to the technical field of electric energy metering correction, and provides a two-way metering correction method and system for an electric energy meter, comprising the following steps: obtaining the power generation of distributed energy and the power consumption of users in historical data, identifying a sustained period with significant difference between the power generation and the power consumption as a mismatch high-occurrence period, performing selective analysis on the mismatch high-occurrence period, and determining a time period division rule for prediction analysis; and performing time period division based on the determined time period division rule to obtain an analysis time period. By analyzing the metering abnormality proportion of different types of key mismatch time period groups, a high-proportion group is screened out, a neural network model is used to predict the occurrence trend and time thereof, and electric energy metering correction compensation is performed, which can predict the occurrence of key mismatch time periods in advance, correct and compensate based on the historical metering deviation rate mean, and effectively reduce the settlement error caused by inaccurate electric energy metering.

Description

Technical Field

[0001] This invention belongs to the field of electricity meter metering correction technology, specifically a bidirectional metering correction method and system for electricity meters. Background Technology

[0002] The bidirectional metering function of the electricity meter allows the meter to simultaneously track and record the electrical energy consumed by the user from the grid (positive electrical energy) and the electrical energy supplied by the user to the grid (reverse electrical energy). This metering method has a clear distinction between power and the direction of electrical energy flow. Electricity consumption is regarded as positive power or positive electrical energy, while electricity generation is regarded as negative power or negative electrical energy.

[0003] When distributed power sources such as photovoltaic and wind power are connected to the grid, there may be a time mismatch between power generation and electricity load. For example, during the day, photovoltaic power generation is large, but the user's electricity load is small, resulting in excess electricity that needs to be fed back to the grid. At night, photovoltaic power generation is small, and users may need to draw electricity from the grid. When distributed power sources are connected to the grid, using bidirectional metering meters can accurately measure the electricity fed back to the grid by users, optimizing the problems of "under-counting of electricity" or "settlement disputes" caused by traditional unidirectional meters.

[0004] However, if the electricity meter is inaccurate, it may affect the settlement results between the user and the power grid. In this case, the electricity meter needs to be corrected. However, the lack of accurate identification of the period of mismatch between distributed energy and load leads to the delay in the correction action.

[0005] Therefore, the present invention provides a bidirectional metering correction method and system for electricity meters. Summary of the Invention

[0006] In order to overcome the shortcomings of the prior art, at least one technical problem raised in the background art is solved.

[0007] The technical solution adopted by this invention to solve its technical problem is: a bidirectional metering correction method for electricity meters, comprising the following steps:

[0008] The system acquires historical data on the generation of distributed energy resources and the electricity consumption of users, identifies periods of significant discrepancy between generation and consumption as high-risk periods for mismatch, performs selection analysis on these high-risk periods, and determines the period division rules for predictive analysis.

[0009] The analysis period is divided according to the established time period division rules. The difference between the total power generation and the total power consumption in each analysis period is calculated. Periods with high degree of difference are extracted as key mismatch periods. Cluster analysis on the range of difference values ​​is performed on the key mismatch periods to obtain different types of key mismatch period groups.

[0010] The proportion of abnormal electricity metering is analyzed for different types of critical mismatch time periods. The critical mismatch time periods with a high proportion of metering anomalies are extracted as high proportion groups. The trend of critical mismatch time periods in the high proportion groups is analyzed, the time of the next occurrence is predicted, and electricity metering correction and compensation are performed.

[0011] As a further aspect of the present invention: the process of obtaining the time period division rules is as follows:

[0012] Extract all the periods with high mismatch occurrences, and arrange these periods according to their corresponding durations to obtain an ascending sequence of high mismatch occurrence periods.

[0013] Calculate the mean and standard deviation of the time period length in the ascending sequence of high mismatch periods to obtain the mean and standard deviation of the time length; set the retention range based on the mean and standard deviation of the time length.

[0014] Traverse the ascending sequence of high-incidence mismatch periods, retain the high-incidence mismatch periods within the retention range, and extract the minimum value among the retained high-incidence mismatch periods as the period division rule for predictive analysis.

[0015] As a further aspect of the present invention: the process for obtaining the high-incidence period of mismatch is as follows:

[0016] Calculate the absolute value of the difference between power generation and power consumption at each time point as the power difference value. Mark the time points that exceed the preset limit as significant difference time points. Merge adjacent significant difference points by time deviation limit to form an integration group. Calculate the time period corresponding to the integration group as the high mismatch period.

[0017] As a further aspect of the present invention: the process of obtaining the integrated group is as follows:

[0018] The time deviation value is calculated by performing a difference calculation on the time points corresponding to the significant difference values ​​of adjacent electrical quantity differences in the significant difference value sequence; if it is less than or equal to the time deviation limit, it is marked as a group of values ​​that can be merged.

[0019] Extract all mergeable value groups, iterate through all mergeable value groups, and integrate the same power difference in different mergeable value groups to obtain integrated groups.

[0020] As a further aspect of the present invention: the process of obtaining the key mismatch period is as follows:

[0021] Calculate the absolute value of the difference between the total power generation and the total power consumption in each analysis period, and use it as the total power deviation value for that period. If the deviation value is greater than the threshold for judging the degree of deviation, it is marked as a period with a high degree of deviation, and the period with a high degree of deviation is regarded as a critical mismatch period.

[0022] As a further aspect of the present invention: the process of obtaining the high-proportion category group is as follows:

[0023] Obtain the proportion of measurement anomalies corresponding to the key mismatch time period group for each type; use K-means clustering to perform cluster analysis on the proportion of measurement anomalies, setting the number of clusters to 2, i.e., high proportion of measurement anomalies and low proportion of measurement anomalies; the cluster with the higher average proportion of anomalies among the two obtained clusters is the high proportion group.

[0024] As a further aspect of the present invention: the process for obtaining the percentage of measurement anomalies is as follows:

[0025] For each type of critical mismatch period group, calculate the sum of the forward and reverse electricity within the critical mismatch period as the total interactive electricity for the period; take the absolute value of the difference between the meter reading and the actual value as the absolute deviation value for the period; and take the ratio of the absolute deviation for the period to the total interactive electricity for the period as the electricity metering deviation rate; if it is greater than the standard value of the electricity metering deviation rate, it is marked as a metering anomaly.

[0026] Extract the key mismatch periods of measurement anomalies and count their total number as the number of measurement anomalies. The ratio of the number of measurement anomalies to the total number of key mismatch periods of this type is used as the percentage of measurement anomalies.

[0027] As a further aspect of the present invention: the process of predicting the time of the next occurrence is as follows:

[0028] Identify key mismatch periods in high-proportion groups and build a prediction model by training a neural network model.

[0029] Based on the prediction model, the probability of a critical mismatch period occurring within the future analysis period is output. If the output probability of a critical mismatch period is greater than the probability limit, then the analysis period is predicted to be a critical mismatch period, which is the time when it will occur next.

[0030] As a further aspect of the present invention: the process of performing power metering correction and compensation is as follows:

[0031] For the analysis period predicted as a critical mismatch period, the average electricity metering deviation rate corresponding to the critical mismatch period in the high proportion group is calculated as the benchmark deviation compensation rate.

[0032] A bidirectional metering correction system for an electricity meter, the system comprising:

[0033] The time period segmentation rule determination module obtains the power generation of distributed energy sources and the electricity consumption of users from historical data, identifies the continuous periods with significant differences between power generation and electricity consumption as high-incidence periods of mismatch, performs selection analysis on high-incidence periods of mismatch, and determines the time period segmentation rules used for predictive analysis.

[0034] Key mismatch period analysis module: Based on the determined period division rules, the analysis period is divided into periods. The difference between the total power generation and the total power consumption in each analysis period is calculated. Periods with high degree of difference are extracted as key mismatch periods. Cluster analysis on the range of difference values ​​is performed on the key mismatch periods to obtain different types of key mismatch period groups.

[0035] Time Period Prediction and Correction Module: Analyzes the proportion of different types of key mismatch time period groups with abnormal electricity metering, extracts key mismatch time period groups with high proportion of metering abnormalities as high proportion groups, analyzes the trend of key mismatch time periods in the high proportion groups, predicts the time of the next occurrence, and performs electricity metering correction and compensation.

[0036] The beneficial effects of this invention are as follows:

[0037] 1. By analyzing historical data on distributed energy generation and user electricity consumption, this method identifies periods of high frequency of mismatches with significant power disparities. Based on this, it determines period segmentation rules, enabling the segmentation of periods based on actual data characteristics rather than subjective experience. This captures tight mismatch occurrence cycles, reduces subjective bias, and closely matches the actual mismatch time distribution characteristics. Furthermore, it can automatically adapt to the operating characteristics and mismatch patterns of different power systems, forming personalized adaptation schemes. This provides a more scientific and accurate basis for subsequent metering corrections, improving the adaptability of the entire correction method to actual power systems and the accuracy of data analysis.

[0038] 2. Based on time period segmentation rules, the analysis period is divided. By calculating the total electricity deviation value of each time period, key mismatch periods are extracted and clustered. Using the K-means clustering method, key mismatch periods are grouped according to the deviation value range, which can effectively identify different degrees of electricity mismatch types. This clustering analysis can quantify the distribution characteristics of different mismatch situations, providing a classification basis for subsequent targeted analysis of metering anomalies in various mismatch periods. This makes subsequent metering anomaly analysis and correction compensation more targeted and efficient, helps to more accurately locate mismatch periods that need to be focused on, and improves the efficiency and accuracy of metering correction.

[0039] 3. By analyzing the proportion of metering anomalies in different types of critical mismatch period groups, high-proportion groups are selected, and neural network models are used to predict their occurrence trends and times. Electricity metering correction and compensation are then performed, which can predict the occurrence of critical mismatch periods in advance. Based on the historical average metering deviation rate, correction and compensation can be performed, which can effectively reduce settlement errors caused by inaccurate electricity metering. Attached Figure Description

[0040] The invention will now be further described with reference to the accompanying drawings.

[0041] Figure 1This is a flowchart of the steps of a bidirectional metering correction method for an electricity meter according to the present invention.

[0042] Figure 2 This is a partial logic judgment flowchart of step S1 in a bidirectional metering correction method for electricity meters according to the present invention.

[0043] Figure 3 This is an architectural diagram of a bidirectional metering correction system for electricity meters according to the present invention. Detailed Implementation

[0044] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0045] Example 1

[0046] Please see Figure 1 and Figure 2 As shown in the embodiment of the present invention, a bidirectional metering correction method for an electricity meter includes the following steps:

[0047] Step S1: Obtain the power generation of distributed energy sources and the electricity consumption of users from historical data, identify the continuous periods with significant differences between power generation and electricity consumption as high-incidence periods of mismatch, perform selection analysis on high-incidence periods of mismatch, and determine the period division rules for predictive analysis.

[0048] In some embodiments, the power generation data of distributed energy sources and the electricity consumption data of users are obtained from historical data. The power generation data of distributed energy sources and the electricity consumption data of users are the basic data for the entire correction method. By analyzing the spatiotemporal distribution of the differences between the two, the entire logical closed loop of mismatch period identification, period rule division, period type clustering, and deviation prediction correction can be completed.

[0049] Distributed energy sources include, but are not limited to: solar photovoltaic systems and wind turbines.

[0050] Specifically, historical data on distributed energy generation and user electricity consumption are extracted from the historical data storage system of electricity meters and the monitoring system of distributed energy.

[0051] The extraction period can be one month, one quarter, or one year, and the historical period to be extracted is determined by those skilled in the art based on experience and the characteristics of data changes.

[0052] It should be noted that the integrity and accuracy of the data must be ensured during data acquisition, including but not limited to: data timestamps, power generation values, and power consumption values. For different types of distributed energy, their corresponding power generation data should be acquired separately.

[0053] Based on the extracted historical time period range, the acquired distributed energy generation data and user electricity consumption data are aligned by timestamp to ensure that the two sets of data correspond one-to-one at the same time.

[0054] Calculate the absolute value of the difference between the power generation and the power consumption at each time point, and use it as the power difference value.

[0055] The power difference value is compared with the preset power difference limit. If the power difference value is less than or equal to the power difference limit, it is marked as a non-significant difference time point. If the power difference value is greater than the power difference limit, it is marked as a significant difference time point.

[0056] It should be explained that the power difference limit is used to determine whether the power difference exceeds the limit at the corresponding time point. This value is set by those skilled in the art according to industry standards.

[0057] Extract all time points with significant differences and serialize them according to the time dimension according to the corresponding timestamps to obtain the sequence of electricity difference values ​​corresponding to the time points with significant differences, which is used as the significant difference value sequence.

[0058] The time deviation value is obtained by calculating the difference between the time points corresponding to the significant difference values ​​of adjacent electrical quantity differences in the significant difference value sequence.

[0059] The time deviation value is compared with the time deviation limit. If the time deviation value is less than or equal to the time deviation limit, the two corresponding power difference values ​​are marked as a mergeable value group. If the time deviation value is greater than the time deviation limit, the two corresponding power difference values ​​are marked as an unmergeable value group.

[0060] It should be explained that the time deviation limit is used to determine whether two adjacent values ​​in a significantly different value sequence are approximately continuous at a given time point. This value is set by those skilled in the art by referring to the requirements for load fluctuation tolerance in industry standards and by summarizing experience.

[0061] Extract all mergeable value groups, iterate through all mergeable value groups, and integrate the same power difference in different mergeable value groups to obtain integrated groups.

[0062] For the integration process, for example, given mergeable value group A (power difference value a, power difference value b), mergeable value group B (power difference value b, power difference value c), mergeable value group C (power difference value c, power difference value d), and mergeable value group D (power difference value e, power difference value f), the corresponding mergeable value groups A, B, and C can be integrated to obtain an integrated group.

[0063] Extract all integration groups, and calculate the time period corresponding to the integration group based on the timestamps corresponding to the power difference values ​​in the integration group, which is taken as the high-incidence period of mismatch.

[0064] A significant characteristic of mismatched periods is the large difference between power generation and power consumption. In this case, the impact of metering deviation is more pronounced, creating a high-risk scenario for metering errors.

[0065] Extract all the periods with high mismatch occurrences, and arrange these periods according to their corresponding durations to obtain an ascending sequence of high mismatch occurrence periods.

[0066] By using statistical distribution methods to filter out extreme values ​​in ascending sequences during periods of high mismatch incidence, the accuracy and practicality of data analysis can be improved, and the interference of extreme values ​​can be eliminated.

[0067] Calculate the mean and standard deviation of the time period length in the ascending sequence of high mismatch periods to obtain the mean and standard deviation of the time length.

[0068] By setting an extreme value filtering range based on the mean and standard deviation of the time period, the retention range can be set to... ,in, This is a proportionality coefficient, set according to the data characteristics of the ascending sequence during periods of high mismatch incidence. It can be set to 1.5 or 2. The average of the time lengths, This represents the standard deviation of the time duration.

[0069] Traverse the ascending sequence of high-incidence mismatch periods, retain the high-incidence mismatch periods within the retention range, and extract the minimum value among the retained high-incidence mismatch periods as the period division rule for predictive analysis.

[0070] The minimum time length corresponding to the period with the highest mismatch incidence will be used as the standard for dividing the period size.

[0071] The reason for using the minimum time length corresponding to the period with high mismatch incidence as the standard for dividing the time period is that the time period division standard corresponding to the minimum value can capture mismatch more accurately. This time period length reflects the most compact mismatch occurrence cycle in the system. Using the minimum value as the division standard can ensure that the granularity of the time period division is fine enough. Using this as the time period division for predictive analysis can reduce the subjective bias of experience setting and make the time period division more in line with the actual time distribution characteristics of mismatch.

[0072] Different power systems have different operating characteristics and mismatch modes. By using the minimum value classification standard obtained through analysis of actual historical data, it is possible to automatically adapt to the characteristics of various power systems. Different classification standards can be calculated for different power systems, thus forming a personalized adaptation scheme with one classification for each system.

[0073] Step S1 has at least the following effects: By analyzing the distributed energy generation and user electricity consumption in historical data, it identifies high-incidence periods of significant electricity mismatch and determines the period division rules accordingly. It can divide the period based on actual data characteristics rather than subjective experience, capture the compact mismatch occurrence cycle, reduce subjective bias, and conform to the actual mismatch time distribution characteristics. At the same time, it can automatically adapt to the operating characteristics and mismatch modes of different power systems, form a personalized adaptation scheme, and provide a more scientific and accurate period division basis for subsequent metering correction. This improves the adaptability of the entire correction method to the actual power system and the accuracy of data analysis.

[0074] Step S2: Divide the time periods according to the determined time period division rules to obtain the analysis time periods, calculate the difference between the total power generation and the total power consumption in each analysis time period, extract the time periods with high degree of difference as key mismatch time periods, and perform cluster analysis on the key mismatch time periods about the range of difference values ​​to obtain different types of key mismatch time period groups.

[0075] In some embodiments, the extracted historical time period is divided into several analysis time periods according to the time period division rules obtained in step S1 above.

[0076] For each analysis period, the absolute value of the difference between the total power generation and the total power consumption within the analysis period is calculated as the total power deviation value for that period.

[0077] The total power consumption deviation value of a time period is compared with the deviation degree judgment threshold. If the total power consumption deviation value of a time period is less than or equal to the deviation degree judgment threshold, it is marked as a low-degree deviation time period. If the total power consumption deviation value of a time period is greater than the deviation degree judgment threshold, it is marked as a high-degree deviation time period.

[0078] It should be explained that the deviation judgment threshold is used to judge the degree of difference between the total power generation and the total power consumption in the corresponding analysis period. This value is set by those skilled in the art based on industry standards and the actual characteristics of the power system.

[0079] The periods with high degree of difference are identified as key mismatch periods. The total power deviation value of the corresponding period is obtained, and cluster analysis is performed based on the total power deviation value to obtain different types of key mismatch period groups.

[0080] Specifically, the process of choosing K-means clustering as the clustering method for cluster analysis is as follows:

[0081] S201: Randomly select K different deviation values ​​from the deviation value sequence as the initial cluster centers;

[0082] For example, when K=3, the initial center may be 50kWh, 80kWh, or 120kWh.

[0083] S202: Calculate the distance from each deviation value to all cluster centers and assign it to the nearest cluster. The distance value can be represented by absolute distance or Euclidean distance.

[0084] For example, the total power deviation for a given period is 75 kWh. Its distance to 50 kWh is 25, its distance to 80 kWh is 5, and its distance to 120 kWh is 45. Therefore, it is assigned to the cluster containing 80 kWh.

[0085] S203: Calculate the mean of the total power deviation value for all time periods within each cluster, and use it as the new cluster center;

[0086] For example, a cluster contains deviation values ​​of 50 kWh and 65 kWh, and the new center has a value of 57.5 kWh.

[0087] S204: Repeat steps S202 and S203 until the cluster center no longer changes significantly or the maximum number of iterations is reached; where the standard for no longer changing significantly can be that the difference between the centers of two iterations is less than 0.1 kWh, and the maximum number of iterations is set to 50.

[0088] S205: Verify the results for each cluster. For each cluster, calculate the variance of the total electricity deviation value during the time period and evaluate the dispersion of the data within the group. The smaller the variance, the closer the deviation values ​​within the same cluster are, and the better the clustering effect.

[0089] Retrieve each cluster, which corresponds to a type, and obtain the key mismatch time period groups of different types.

[0090] In this embodiment, the type refers to the range of the total electricity deviation value for the time period.

[0091] Step S2 has at least the following effects: Based on time period segmentation rules, the analysis divides the time periods, calculates the total electricity deviation value for each time period, extracts key mismatch periods, and performs cluster analysis on them. Using the K-means clustering method, the key mismatch periods are grouped according to the deviation value range, effectively identifying different degrees of electricity mismatch types. This cluster analysis quantifies the distribution characteristics of different mismatch situations, providing a classification basis for subsequent targeted analysis of metering anomalies in various mismatch periods. This makes subsequent metering anomaly analysis and correction compensation more targeted and efficient, helping to more accurately locate mismatch periods that require special attention, and improving the efficiency and accuracy of metering correction.

[0092] Step S3: Analyze the proportion of different types of critical mismatch time periods with abnormal electricity metering, extract the critical mismatch time periods with high proportion of metering abnormalities as high proportion groups, analyze the trend of critical mismatch time periods in the high proportion groups, predict the time of the next occurrence, and perform electricity metering correction and compensation.

[0093] In some embodiments, for each type of critical mismatch period group, the corresponding electricity metering deviation rate is calculated.

[0094] The calculation process for the electricity calculation deviation value is as follows: calculate the sum of the forward electricity and the reverse electricity during the key mismatch period as the total interactive electricity of the period; take the absolute value of the difference between the electricity meter reading and the actual value as the absolute deviation value of the period; and take the ratio of the absolute deviation of the period to the total interactive electricity of the period as the electricity metering deviation rate.

[0095] It needs to be explained that the meter reading is the statistical data of the electrical energy exchanged between the user and the power grid during the specified time period, recorded by the meter itself. The actual value is the real data of the electrical energy exchanged between the user and the power grid during the specified time period, obtained through verification methods, including but not limited to: standard power sources, third-party verification devices, and energy conservation verification methods. Forward electricity is the amount of electricity the user draws from the power grid during the specified time period (current flows from the power grid to the user's side), recorded by the meter. Reverse electricity is the amount of electricity the user sends to the power grid during the specified time period (current flows from the user's side to the power grid), recorded by the meter.

[0096] For example, the electricity meter measures: the forward electricity (household electricity) is 20kWh, the reverse electricity (photovoltaic grid connection) is 15kWh, and the electricity meter reading is 20+15=35kWh; after verification, the actual forward electricity is 19.5kWh, the actual reverse electricity is 15.2kWh, and the actual value is 34.7kWh.

[0097] Based on the electricity metering deviation rate corresponding to each critical mismatch period, it is determined whether there is an electricity metering deviation in the corresponding period. If the electricity metering deviation rate is less than or equal to the standard value of the electricity metering deviation rate, it is marked as normal metering. If the electricity metering deviation rate is greater than the standard value of the electricity metering deviation rate, it is marked as abnormal metering.

[0098] Extract the key mismatch periods of measurement anomalies and count their total number as the number of measurement anomalies. The ratio of the number of measurement anomalies to the total number of key mismatch periods of this type is used as the percentage of measurement anomalies.

[0099] The process of clustering the proportion of measurement anomalies using the K-means clustering method is the same as the method in step S2 above. The difference is that the number of clusters needs to be set to 2, that is, high proportion of measurement anomalies and low proportion of measurement anomalies.

[0100] The group with the higher average anomaly percentage among the two obtained clusters is designated as the high-percentage group, and the group with the lower average anomaly percentage is designated as the low-percentage group.

[0101] Identify the key mismatch periods in high-proportion groups, and analyze the occurrence trends of these key mismatch periods by building a predictive model to predict when they will occur again.

[0102] Specifically, a neural network model is used to train the prediction model. The input layer consists of the timestamps corresponding to the critical mismatch periods and the corresponding electricity metering deviation rates. The hidden layer consists of 1-2 layers of LSTM units to capture the long-term dependencies of the time series. The output layer outputs the critical mismatch probability (between 0 and 1) through a fully connected layer.

[0103] Training parameters: Batch size: 32-64, Number of iterations: 50-100, Learning rate: 0.001; Loss function: Binary cross-entropy.

[0104] The prediction model outputs the probability of a critical mismatch period within the future analysis period. If the output probability of a critical mismatch period is greater than the probability limit, the analysis period is predicted to be a critical mismatch period. If the output probability of a critical mismatch period is less than or equal to the probability limit, the analysis period is predicted to be a non-critical mismatch period.

[0105] For the analysis period that is predicted to be a critical mismatch period, power metering correction compensation is performed.

[0106] Specifically, the average electricity metering deviation rate corresponding to the key mismatch period in the high-proportion category is calculated as the benchmark deviation compensation rate.

[0107] By combining the positive and negative values ​​of the difference between the electricity meter readings and the actual values, the electricity meter readings are corrected and compensated for the analysis period when the prediction is the critical mismatch period, using the benchmark deviation compensation rate.

[0108] If the difference between the meter reading and the actual value is positive, the meter reading is large and should be reduced during correction. If the difference is negative, the meter reading is small and should be increased during correction.

[0109] For example, if the average deviation rate of the high-proportion group is 3.5%, then the benchmark deviation compensation rate is 3.5%. For the predicted critical mismatch period, the electricity meter reading is adjusted according to this ratio (e.g., if the original reading is 100kWh, the corrected value is 100×(1-3.5%)=96.5kWh).

[0110] Optionally, to allow sufficient time for data processing and system response, adjustments can be made in advance. The advance adjustment time can be set to half or one-third of the analysis period; for example, if the analysis period is one hour, the adjustment can be completed 30 minutes or 20 minutes in advance.

[0111] Step S3 has at least the following effects: by analyzing the proportion of metering anomalies in different types of key mismatch period groups, high-proportion groups are selected, and their occurrence trends and times are predicted using a neural network model. Electricity metering correction and compensation are then performed, which can predict the occurrence of key mismatch periods in advance. Based on the average historical metering deviation rate, correction and compensation can be performed, which can effectively reduce settlement errors caused by inaccurate electricity metering. At the same time, targeted adjustments are made according to the positive or negative difference between the metered value and the actual value, making the correction more scientific and reasonable.

[0112] In this embodiment, step S1 identifies high-incidence periods of mismatch through historical data and determines time period division rules based on actual data characteristics, providing a time period division framework for step S2. Step S2 divides historical periods into analysis periods based on these rules, enabling subsequent deviation calculations and cluster analyses to be based on time units that closely match the actual mismatch cycle, thus optimizing the analysis errors caused by subjective division. The analysis results of steps S1 and S2 are the core inputs to the prediction model in step S3.

[0113] Example 2

[0114] Based on the same inventive concept as the bidirectional metering correction method for an electricity meter in the foregoing embodiments, such as Figure 3 As shown, this application provides a bidirectional metering correction system for electricity meters, wherein the system specifically includes:

[0115] The time period segmentation rule determination module obtains the power generation of distributed energy sources and the electricity consumption of users from historical data, identifies the continuous periods with significant differences between power generation and electricity consumption as high-incidence periods of mismatch, performs selection analysis on high-incidence periods of mismatch, and determines the time period segmentation rules used for predictive analysis.

[0116] The execution process is as follows: Obtain the power generation and user electricity consumption of distributed energy sources from historical data, calculate the electricity difference value at each time point after aligning with the timestamp, mark the points exceeding the preset limit as significant difference points, merge adjacent significant difference points through the time deviation limit to form an integration group, and then determine the high-incidence period of mismatch. Sort these periods in ascending order of length, calculate the mean and standard deviation to filter out extreme values, and finally take the minimum value within the retention range as the period division rule for predictive analysis, so that the period division fits the actual mismatch cycle.

[0117] Key mismatch period analysis module: Based on the determined period division rules, the analysis period is divided into periods. The difference between the total power generation and the total power consumption in each analysis period is calculated. Periods with high degree of difference are extracted as key mismatch periods. Cluster analysis on the range of difference values ​​is performed on the key mismatch periods to obtain different types of key mismatch period groups.

[0118] The execution process is as follows: Based on the determined time period division rules, the historical time period is divided into several analysis periods. The total deviation value of power generation and power consumption in each period is calculated. Periods that exceed the deviation degree judgment threshold are marked as high-degree difference periods, i.e., critical mismatch periods.

[0119] K-means clustering was used to analyze the deviation values ​​of key mismatch periods: initial cluster centers were randomly selected, and deviation values ​​were assigned to the nearest cluster according to distance. The cluster centers were iteratively updated until convergence or the maximum number of iterations was reached. The clustering effect was verified by variance to obtain key mismatch period groups with different deviation value ranges.

[0120] Time Period Prediction and Correction Module: Analyzes the proportion of different types of key mismatch time period groups with abnormal electricity metering, extracts key mismatch time period groups with high proportion of metering abnormalities as high proportion groups, analyzes the trend of key mismatch time periods in the high proportion groups, predicts the time of the next occurrence, and performs electricity metering correction and compensation.

[0121] The execution process is as follows: calculate the electricity metering deviation rate for each key mismatch period group, mark those exceeding the standard value as metering anomalies, count the percentage of anomalies, and distinguish high and low percentage groups by K-means clustering.

[0122] For high-proportion groups, an LSTM neural network model (input timestamp and deviation rate, hidden layer captures time series dependence) is used to predict the time of the next critical mismatch period. If the predicted probability exceeds the limit, the benchmark compensation rate is calculated based on the average deviation rate of the high-proportion groups. Combining the positive and negative difference between the metered value and the actual value, the metered value of the predicted period is corrected and compensated.

[0123] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for bidirectional metering correction for an electric energy meter, characterized in that: The method comprises the following steps: obtaining the power generation of the distributed energy and the power consumption of the user in historical data, identifying a time period with a significant difference between the power generation and the power consumption as a mismatch high-occurrence time period, performing selection analysis on the mismatch high-occurrence time period, and determining a time period division rule for prediction analysis; calculating the absolute value of the difference between the power generation and the power consumption at each time point as an electric quantity difference value, marking a value exceeding a preset limit value as a significant difference time point, merging adjacent significant difference points through a time deviation limit value to form an integrated group, and calculating the time period corresponding to the integrated group as a mismatch high-occurrence time period; extracting all the mismatch high-occurrence time periods, arranging the time periods according to their corresponding time period lengths to obtain an ascending sequence of the mismatch high-occurrence time periods; calculating the average value and the standard deviation of the time period length in the ascending sequence of the mismatch high-occurrence time periods to obtain a time length average value and a time length standard deviation, and setting a retention range based on the time length average value and the time length standard deviation; traversing the ascending sequence of the mismatch high-occurrence time periods, retaining the mismatch high-occurrence time periods within the retention range, and extracting the minimum value in the retained mismatch high-occurrence time periods as a time period division rule for prediction analysis; dividing the time period based on the determined time period division rule to obtain an analysis time period, calculating the difference value between the total power generation value and the total power consumption value of each analysis time period, extracting a high-degree difference time period as a key mismatch time period, and performing clustering analysis on the key mismatch time period with respect to the difference value range to obtain different types of key mismatch time period groups; the key mismatch time period is obtained by: calculating the absolute value of the difference between the total power generation value and the total power consumption value in each analysis time period as a total electric quantity deviation value; if the value is greater than a deviation degree judgment threshold, the value is marked as a high-degree difference time period, and the high-degree difference time period is taken as a key mismatch time period; analyzing the proportion of the key mismatch time period groups of different types in the occurrence of electric energy metering abnormalities, extracting a key mismatch time period group with a high metering abnormality proportion as a high-proportion type group, analyzing the trend of the key mismatch time period in the high-proportion type group, predicting the time of the next occurrence, and performing electric energy metering correction compensation.

2. The bidirectional metering correction method for an electric energy meter according to claim 1, characterized in that: the integrated group is obtained by: calculating the difference value of the significant difference time points corresponding to adjacent electric quantity difference values in the significant difference value sequence to obtain a time deviation value; if the value is less than or equal to a time deviation limit value, the value is marked as a mergable value group; extracting all the mergable value groups, and traversing all the mergable value groups to integrate the same electric quantity difference values in different mergable value groups to obtain an integrated group.

3. The bidirectional metering correction method for an electric energy meter according to claim 1, characterized in that: the high-proportion type group is obtained by: obtaining the metering abnormality proportion corresponding to each type of key mismatch time period group; performing clustering analysis on the metering abnormality proportion by using a K-means clustering method, setting the number of clusters to 2, i.e., a metering abnormality proportion high and a metering abnormality proportion low; and taking the high-proportion type group as the average value of the metering abnormality proportion in the two clusters.

4. The bidirectional metering correction method for an electric energy meter according to claim 3, characterized in that: the metering abnormality proportion is obtained by: For each type of key mismatch period in the group, the sum of the forward and reverse electric quantities in the key mismatch period is calculated as the total interaction electric quantity of the period; the absolute value of the difference between the metering value and the actual value is taken as the absolute deviation value of the period, and the ratio of the absolute deviation to the total interaction electric quantity of the period is taken as the electric energy metering deviation rate; if it is greater than the standard value of the electric energy metering deviation rate, it is marked as a metering anomaly; The key mismatch periods of the metering anomaly are extracted, and the total number thereof is counted as the metering anomaly number, and the ratio of the metering anomaly number to the total number of key mismatch periods in the type is taken as the metering anomaly proportion.

5. The bidirectional metering correction method for an electric energy meter according to claim 1, characterized in that: The process of predicting the next occurrence time is: The key mismatch periods in the high-proportion class group are obtained, and a prediction model is constructed by training a neural network model; Based on the prediction model, the probability of being a key mismatch period in the future analysis period is output, and if the probability of the key mismatch period is greater than the probability limit value, the analysis period is predicted to be a key mismatch period, which is the next occurrence time.

6. The bidirectional metering correction method for an electric energy meter according to claim 5, characterized in that: The process of performing electric energy metering correction compensation is: For the analysis period predicted to be a key mismatch period, the average electric energy metering deviation rate corresponding to the key mismatch period in the high-proportion class group is calculated as the reference deviation compensation rate.

7. A bidirectional metering correction system for an electric energy meter, characterized by, The system is used to execute the method of any one of claims 1-6, and the system comprises: The period division rule determination module: obtains the power generation of the distributed energy and the power consumption of the user in the historical data, identifies the continuous period with significant difference between the power generation and the power consumption as a mismatch high-occurrence period, analyzes the mismatch high-occurrence period, and determines the period division rule for prediction analysis; The key mismatch period analysis module: based on the determined period division rule, the analysis period is obtained by period division, the difference value between the total value of the power generation and the total value of the power consumption of each analysis period is calculated, the high-degree difference period is extracted as the key mismatch period, and the key mismatch period is analyzed in terms of the difference value range to obtain different types of key mismatch period groups; The period prediction and correction module: analyzes the proportion of different types of key mismatch period groups for the occurrence of electric energy metering anomalies, extracts the key mismatch period group with high metering anomaly proportion as the high-proportion class group, analyzes the trend of the occurrence of the key mismatch period in the high-proportion class group, predicts the next occurrence time, and performs electric energy metering correction compensation.

Images