Multi-epitope metering tank branch electricity behavior identification method based on data correlation

Abstract

The application discloses a multi-epitope metering box branch electricity behavior identification method based on data correlation, which comprises the following steps: acquiring total meter, sub-meter frozen data and metering box archives, and establishing original electric parameter sequence and branch user set; calculating common time scale deviation through mutual correlation of total phase power increment sequence and combined power increment sequence of the same phase sub-meter, and further determining individual time scale deviation of each sub-meter; reconfiguring sub-meter electric parameter sequence according to total deviation, and obtaining time scale corrected data; calculating branch current, voltage difference and branch equivalent resistance according to the time scale corrected data, and determining variable technical loss; combining corrected residual error, time scale corrected residual error and loss corrected residual error, identifying collection time scale anomaly, branch resistive anomaly, bypass electricity or metering loop anomaly, and generating records. The method can improve the branch electricity behavior identification accuracy.

Description

Technical Field

[0001] This application relates to the field of low-voltage power distribution metering data processing technology, specifically to a method for identifying the electricity consumption behavior of branch circuits in multi-metering boxes based on data correlation. Background Technology

[0002] Low-voltage distribution area multi-meter boxes typically consist of a main meter, a concentrator, and multiple user sub-meters to collect data such as phase voltage, current, active power, power factor, and time-of-use electricity consumption. Current methods for analyzing line losses and investigating abnormal electricity usage in distribution areas often rely on the difference between the main meter's electricity consumption and the sum of the sub-meters' electricity consumption, combined with a fixed line loss rate or manual experience for screening. In actual operation, the freeze times of the main meter and each sub-meter are often not completely consistent. Different data collectors, meter clocks, or data collection tasks can also cause overall or individual meter time shifts, leading to misaligned data for the same load start / stop event in the main meter and sub-meter data, resulting in false residuals. Simultaneously, branch conductors and terminals within the metering box exhibit technical losses that vary with current. These losses are closely related to load levels, and a fixed line loss rate cannot accurately reflect actual loss changes. The combination of these factors can easily lead to confusion between abnormal data collection timescales, normal fluctuations in technical losses, abnormal branch contact resistance, and abnormal bypass electricity consumption or metering circuits, potentially causing misjudgments or masking genuine anomalies. Summary of the Invention

[0003] This application provides a method for identifying the electricity consumption behavior of multi-meter box branches based on data correlation, so as to at least solve some of the technical problems existing in the related technologies described above.

[0004] According to a first aspect of the embodiments of this application, a method for identifying the electricity consumption behavior of a multi-meter box branch based on data correlation is provided, including: Obtain frozen data from the master table and sub-tables, as well as metering box files, and establish the original electrical parameter sequence and branch user set; In the calculation segment, the total phase power increment sequence and the in-phase sub-meter combined power increment sequence are calculated, and the maximum cross-correlation shift is taken as the common time scale deviation. A reference signal for each table is constructed using the common time scale deviation, and the maximum additional shift of the cross-correlation between the power increment sequence of each table and the corresponding reference signal is taken as the individual time scale deviation. The total deviation of each sub-meter is obtained by adding the common time scale deviation to the individual time scale deviation. The electrical parameter sequence is then reassigned according to the total deviation of each sub-meter to obtain the time scale corrected sub-meter electrical parameter sequence. The branch current and voltage difference are obtained from this sequence, the equivalent resistance of the branch is estimated, and the variable technical loss is calculated using the branch current and the equivalent resistance of the branch. The system generates uncorrected residuals, time-scaled corrected residuals, and loss-corrected residuals, and determines the anomaly type in sequence. Specifically: when the uncorrected residual exceeds the preset residual range and the time-scaled corrected residual falls into the preset zero range, it is recorded as a time-scaled abnormality; when the loss-corrected residual is positively correlated with the square of the branch current and the voltage drop exceeds the voltage drop corresponding to the equivalent resistance, it is recorded as a branch resistivity abnormality; when the loss-corrected residual is synchronized with the load of a single user and the sub-meter current is mismatched, it is recorded as a bypass power consumption or metering circuit abnormality.

[0005] As an optional approach, establishing the original electrical parameter sequence and branch user set includes: retaining the original freezing time of each sampling point in the frozen data of the master meter and sub-meters; marking each sampling point with data source identifier, phase identifier, meter position identifier, acquisition cycle, and data integrity status; and according to the phase relationship and branch affiliation relationship in the metering box file, assigning single-phase users to the corresponding phase branch, and assigning three-phase users with three-phase power or current data to the branch user set according to phases A, B, and C respectively.

[0006] As an optional approach, determining the common timescale deviation includes: sequentially selecting candidate translation amounts within a preset translation range; translating the in-phase sub-table synthesized power increment sequence according to the candidate translation amounts; calculating the sum of the products of the translated in-phase sub-table synthesized power increment sequence and the total in-phase power increment sequence for effective sampling points within the calculation segment; taking the candidate translation amount with the largest product sum as the segment common timescale deviation; taking the median of multiple segment common timescale deviations and using this median as the common timescale deviation.

[0007] As an optional approach, constructing the sub-table reference signal includes: selecting a sub-table as the target sub-table; summing the power increment sequences of the remaining sub-tables (excluding the target sub-table) after shifting them according to a common time scale deviation; subtracting the corrected composite power increment sequences of the remaining sub-tables from the total sub-table power increment sequence to obtain the sub-table reference signal of the target sub-table; and, based on the common time scale deviation shift, continuing to shift the target sub-table power increment sequence within a preset additional shift range, and taking the additional shift amount when the cross-correlation between the target sub-table power increment sequence and the sub-table reference signal of the target sub-table is maximized as the individual time scale deviation of the target sub-table.

[0008] As an optional scheme, the redistribution parameter sequence includes: when the total deviation of each sub-meter corresponds to an integer multiple of the sampling period, shifting the sub-meter parameter sequence forward or backward by the corresponding number of sampling periods, adjusting only the time tag and maintaining the voltage, current, power, and energy values; otherwise, dividing the original sampling interval into a front segment and a back segment based on the total deviation of each sub-meter, linearly interpolating the power at the segment start and end of the interval to obtain the power at the segment point, obtaining the estimated average power of the front segment from the power at the start of the interval and the power at the segment point, obtaining the estimated average power of the back segment from the power at the segment point and the power at the end of the interval, allocating the original energy according to the duration ratio of the front and back segments and the estimated average power, the energy of the back segment is obtained by subtracting the energy allocated in the front segment from the original energy, writing the energy allocated in the front segment and the energy of the back segment into the corresponding total meter sampling period, and interpolating the voltage and current according to adjacent sampling points.

[0009] As an optional approach, the branch current and voltage difference are calculated as follows: for sub-meters with current data, the sub-meter current is adjusted using time-scale correction; for sub-meters without current data but with active power and voltage data, the approximate current is obtained by dividing the time-scale corrected active power by the time-scale corrected voltage; the branch current is obtained by combining the sub-meter current and the approximate current within the same branch; the sub-meter voltage difference is obtained by subtracting the sub-meter voltage within the same branch from the total phase voltage, and the average sub-meter voltage difference within the same branch is taken to obtain the average branch voltage difference.

[0010] As an optional approach, the equivalent resistance of the branch is estimated by collecting branch current and average voltage difference of the branch for multiple sampling periods as data pairs, multiplying the branch equivalent resistance by the branch current to obtain the estimated voltage difference, and using the least squares method to obtain the equivalent resistance of the branch with the condition that the sum of squares of the deviations between the average voltage difference of the branch and the estimated voltage difference is minimized.

[0011] As an optional approach, the data collection includes: retaining sampling periods where the time-stamped residual falls within a preset residual range, the direction of change of user-end voltage drop is consistent with the direction of change of branch current, and the load change amplitude exceeds a preset load change threshold; eliminating sampling periods where the time-stamped residual exceeds the preset residual range; and marking the branch as having insufficient reliability in resistance estimation when the difference in equivalent resistance obtained under multiple load levels exceeds a preset consistency threshold, and reducing the loss correction weight of the branch when calculating variable technical losses.

[0012] As an optional approach, the variable technical loss is calculated according to the sampling period; for a single branch, the active power loss of the branch is obtained by multiplying the square of the branch current by the equivalent resistance of the branch, and the active power loss of the branch is multiplied by the sampling period and processed according to the loss correction weight to obtain the power loss of the branch; for each phase in series between the common incoming line section and the user terminal branch, the power loss of the common incoming line section and the power loss of the user terminal branch are calculated separately, and the two are added together to obtain the variable technical loss of that phase.

[0013] This application calculates the common timescale deviation by correlating the total phase power increment with the combined power increment of the sub-meters in the same phase, and further determines the individual timescale deviation for each sub-meter. This corrects for acquisition time misalignment while preserving the characteristics of the original frozen data, reducing false total and sub-meter residuals caused by asynchronous load start-up and shutdown. Simultaneously, it estimates the equivalent resistance of branches using the timescale-corrected branch current and voltage difference, and calculates variable technical losses based on current changes. This allows residual analysis to no longer rely on a fixed line loss rate, more accurately reflecting the actual loss changes of wires and terminals inside the metering box. Furthermore, this invention generates residuals before correction, residuals after timescale correction, and loss-corrected residuals. It classifies and judges these residuals by combining them with the square of the branch current, single-user load changes, and sub-meter current matching relationships. This allows for the differentiation of situations such as abnormal acquisition timescales, abnormal branch resistance, abnormal bypass power consumption or metering circuits, and insufficient data quality, thereby improving the accuracy of identifying branch power consumption behavior in multi-meter metering boxes and reducing the risk of misjudgment and missed judgment.

[0014] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Furthermore, no embodiment in this disclosure is required to achieve all the effects described above. Attached Figure Description

[0015] The accompanying drawings, which are incorporated in and form a part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure.

[0016] Figure 1 A flowchart illustrating a method for identifying the electricity consumption behavior of a multi-position metering box branch based on data correlation, provided in an embodiment of this disclosure.

[0017] Figure 2 A flowchart for estimating common timescale deviations provided in embodiments of this disclosure.

[0018] Figure 3 A flowchart for calculating variable technology losses provided in this embodiment of the disclosure.

[0019] Figure 4 A flowchart for determining the anomaly type provided in this embodiment of the disclosure.

[0020] Figure 5 This is a schematic block diagram of a multi-meter box branch power consumption behavior identification system based on data correlation, provided as an embodiment of the present disclosure.

[0021] Figure 6 This is a schematic block diagram of an electronic device provided in an embodiment of the present disclosure. Detailed Implementation

[0022] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort should fall within the scope of protection of the present application.

[0023] According to embodiments of this disclosure, a method for identifying branch electricity consumption behavior in multi-meter-position metering boxes is provided, applicable to scenarios in low-voltage distribution areas with multi-meter-position metering boxes configured with concentrators or main meters and multiple user sub-meters. The main meter in the metering box can collect three-phase voltage, three-phase current, active power, reactive power, power factor, and time-of-use electricity; each user sub-meter can collect voltage, current, active power, power factor, and time-of-use electricity. The system simultaneously obtains meter position relationships, phase relationships, branch affiliation relationships, and collection cycles from the metering box archive. In actual low-voltage field operation, the freeze times of the main meter and each sub-meter are usually not completely consistent. Simultaneously, the technical losses on the branch wires and terminals within the metering box vary with current, exhibiting variable characteristics. The superposition of these two factors can lead to distortion of the residual electricity consumption between the main and sub-meters, potentially misjudging misaligned data collection timestamps or normal technical losses as electricity theft, or masking genuine anomalies within fluctuations in line losses. This method sequentially completes the estimation of acquisition time scale deviation and data reconfiguration, the estimation of branch equivalent resistance and the calculation of variable technical losses, the analysis of loss correction residuals and the classification of anomalies, and finally generates verifiable records.

[0024] The implementation process of the method described in this application will be described in detail below with reference to specific embodiments. It should be noted that this embodiment is only used to explain this application and is not intended to limit the scope of protection of this application. Conventional adjustments or substitutions of each step by those skilled in the art without departing from the concept of this application should be included in the scope of protection of this application.

[0025] Please see Figure 1 , Figure 1 This is a flowchart of a method for identifying the electricity consumption behavior of a multi-meter box branch based on data correlation, as provided in an embodiment of this disclosure. Figure 1 As shown, the method includes steps S1-S6: In step S1, the frozen data of the master table and sub-tables and the metering box files are obtained, and the original electrical parameter sequence and branch user set are established.

[0026] In some embodiments, the system acquires the original frozen data of the master table and each sub-table within each acquisition cycle, while simultaneously reading the metering box file. When establishing the original electrical parameter sequence, the original frozen time of each sampling point is retained, and each sampling point is marked with a data source identifier, phase identifier, table position identifier, acquisition cycle, and data integrity status. At this stage, no uniform time interpolation is performed on the data of each table, because subsequent steps require identification and quantification of the frozen time deviation. If all data are directly interpolated to a uniform time point here, the original time-scale deviation information will be erased, and it will be impossible to distinguish whether the residual comes from acquisition misalignment or abnormal power consumption.

[0027] When establishing a branch user set, the system categorizes users based on the phase and branch affiliation relationships in the metering box file: single-phase users are assigned to the corresponding branches A, B, or C according to their phase information; three-phase users with three-phase power or current data are assigned to the respective branch user sets for each phase (A, B, and C); three-phase users providing only total power or total electricity are excluded from subsequent branch equivalent resistance estimation because they cannot be separated by phase, and are only included in the overall electricity balance verification of the metering box. Optionally, when the branch hierarchy information in the file is incomplete and the branch affiliation relationship cannot be determined, the branch equivalent resistance estimation is not performed on these users, and they are only included in the overall electricity balance verification of the same phase; when the file can determine the branch affiliation relationship of the same phase, the subsequent branch current, branch average voltage difference, and branch equivalent resistance are still calculated according to the aforementioned branch user set.

[0028] In step S2, in the calculation segment, the total phase power increment sequence and the in-phase sub-phase power increment sequence are calculated, and the maximum cross-correlation shift is taken as the common time scale deviation.

[0029] In some embodiments, the freeze times of the main meter, concentrator, data collector, and individual sub-meters in a low-voltage distribution area typically deviate. This deviation includes: a systematic time offset between multiple sub-meters and the main meter under the same data collector, i.e., a common timescale deviation; and an additional offset caused by the clock of a specific sub-meter or the data acquisition task, i.e., an individual timescale deviation. This embodiment first estimates the common timescale deviation, and then estimates the individual timescale deviation of each sub-meter based on this, thereby processing the two types of deviations separately and avoiding misinterpreting the start-up and shutdown changes of a high-load user as a time misalignment of the entire distribution box. This hierarchical processing method ensures that subsequent data reconfiguration and impedance estimation are based on accurate time correspondence, eliminating spurious residuals caused by inconsistent freeze times.

[0030] Specifically, please refer to Figure 2 , Figure 2 This is a flowchart of the common timescale deviation estimation provided in the embodiments of this disclosure. Figure 2 As shown in box 201, in the calculation segment, the total phase power increment sequence and the in-phase sub-meter combined power increment sequence are calculated.

[0031] In this context, power increment refers to the difference in active power between two adjacent sampling points from the same data source. For the main meter, the phase power increment sequence is calculated separately for each of phases A, B, and C. For individual meters, the active power increment sequence for each individual meter is calculated. Power increment, rather than the original power value, is used as the basis signal for time-scale estimation because the original power sequence is approximately constant during stable load operation. The curves of the main meter and individual meters have similar shapes and stable values, making it difficult to identify time shifts. Power increment, however, is different. When loads such as electric water heaters, air conditioners, and induction cookers of residential users start or stop, the power increment exhibits significant positive or negative jumps. These jumps constitute characteristic markers on the time axis. If there is a discrepancy between the freeze times of the main meter and individual meters, the positions of the same load start / stop event in the two power increment sequences will be relatively offset. This offset can be used to quantitatively estimate the time shift.

[0032] Specifically, let the general table be a certain phase number The active power in each sampling period is Then the power increment of that phase is ; Let the phase number be The first sub-table The active power in each sampling period is Then the power increment of the sub-meter is The system sums the power increments of all in-phase sub-meters to obtain the composite power increment sequence of the in-phase sub-meters. This sequence serves as the input for subsequent common timescale deviation estimation.

[0033] Not all periods of the aforementioned power increment sequence are suitable for timescale deviation estimation. In some embodiments, the system filters calculation segments from historical data that meet certain conditions, including: retaining segments where the data collection of the master table and sub-tables is complete; retaining segments where the number of sampling points whose absolute power increment exceeds a preset amplitude threshold meets a preset quantity condition to ensure that there are enough load change events within the segment to support cross-correlation calculations; and retaining segments where the phase files have not changed. The preset amplitude threshold can be configured as a certain percentage of the rated power of the master table, which is used to exclude stable periods where the load is almost unchanged. The specific value is calibrated during deployment based on the load characteristics of the transformer area. The preset quantity condition is used to ensure that the calculation segment contains a sufficient number of effective change points, for example, it can be set to no less than a certain number of sampling points.

[0034] Meanwhile, the system will remove segments with more than a preset number of missing sampling points from the calculation segments; segments in the master table and multiple sub-tables where the absolute value of power increment exceeds the preset jump threshold will also be removed; the preset jump threshold is higher than the aforementioned amplitude threshold and is used to identify drastic jumps that may be caused by data anomalies rather than actual load changes; this threshold can be pre-calibrated during deployment based on the statistical distribution of abnormal jumps in the historical data of the transformer area; the calculation segments retained after the above screening have better data quality and contain sufficient load change information.

[0035] In box 202, the median of the common time scale deviations of each segment is selected as the final common time scale deviation. Specifically, the system uses the power increment sequence of a certain phase of the master table as a reference and sequentially selects candidate shift amounts within a preset shift range. The preset shift range is determined based on the maximum freeze time deviation that the acquisition system may experience, and can optionally be set to a positive or negative number of sampling periods. For each candidate shift amount, the system shifts the overall in-phase sub-meter composite power increment sequence by the candidate shift amount, and then calculates the sum of the products of the shifted in-phase sub-meter composite power increment sequence and the master table phase power increment sequence for the effective sampling points within the calculation segment. This product is the cross-correlation value. Cross-correlation is a method to measure the similarity of signals under different time offsets: when the shift amount is exactly equal to the actual freeze time deviation, the load start-stop transitions in the two sequences are aligned in time, and the product sum reaches its maximum value; when deviating from the actual deviation, the transition positions are offset, and the product sum decreases.

[0036] The candidate shift amount when the product sum is at its maximum is used as the common time scale deviation of the calculation segment. After performing the above process on multiple calculation segments, the median of the common time scale deviations of each segment is taken as the final common time scale deviation. The median is used instead of the mean in order to reduce the interference of random factors in individual segments and make the estimation results more robust. The common time scale deviation reflects the time offset direction and offset of the sub-table data as a whole relative to the total table data under the same acquisition path.

[0037] In step S3, a sub-table reference signal is constructed using the common time scale deviation, and the maximum additional shift of the cross-correlation between the sub-table power increment sequence and the corresponding reference signal is taken as the individual time scale deviation.

[0038] In some embodiments, based on the common timescale deviation, the system further estimates the individual timescale deviation of each sub-table. For each target sub-table, the sub-table reference signal is constructed as follows: a sub-table is selected as the target sub-table, and the power increment sequences of the remaining sub-tables (excluding the target sub-table) are shifted according to the common timescale deviation and then summed to obtain the corrected composite power increment sequence of the remaining sub-tables; the corrected composite power increment sequence of the remaining sub-tables is subtracted from the total table power increment sequence to obtain the sub-table reference signal of the target sub-table; wherein, in a physical sense, this reference signal approximately represents the portion of the total table power increment that should belong to the target sub-table, excluding the influence of the remaining sub-tables, so that the subsequent cross-correlation search can focus on reflecting the time offset of the target sub-table itself.

[0039] Based on the common timescale deviation shift, the system continues to shift the target sub-meter power increment sequence within a preset additional shift range. Similarly, the cross-correlation method is used to calculate the sum of the products between the target sub-meter power increment sequence after the continued shift and its sub-meter reference signal. The additional shift amount when the cross-correlation is maximum is taken as the individual timescale deviation of the target sub-meter. The preset additional shift range is usually smaller than the search range of the common timescale deviation because the individual deviation is generally small. It can be optionally set to ±1 to ±2 sampling periods.

[0040] The reason for estimating common deviation and individual deviation separately is that the start-stop changes of a high-load user within the metering box may dominate the increase in total meter power. If each sub-meter is searched individually, the deviation of low-load users is easily masked by high-load users. By first extracting the common deviation to eliminate systematic offsets, and then performing a small-range search on each sub-meter, the estimation results are more accurate. If the individual deviation of a sub-meter deviates significantly from zero while the individual deviations of other sub-meters under the same data collector are close to zero, there may be a single-meter clock anomaly in that sub-meter. This information is used as a reference in subsequent anomaly type judgment. After performing the above process sequentially on each sub-meter, the system obtains the individual time-scale deviations of all sub-meters.

[0041] In step S4, the common time scale deviation and the individual time scale deviation are added together to obtain the total deviation of each sub-meter. The electrical parameter sequence is then reassigned according to the total deviation of each sub-meter to obtain the time scale corrected sub-meter electrical parameter sequence.

[0042] In some embodiments, the system adds the common timescale deviation to the individual timescale deviation to obtain the total deviation of each sub-table, and reconfigures the sub-table electrical parameter sequences according to the total deviation to obtain the timescale-corrected sub-table electrical parameter sequences; the reconfiguration process is divided into two cases according to the relationship between the total deviation and the sampling period.

[0043] Specifically, when the total deviation of a sub-meter corresponds to an integer multiple of the sampling period, the electrical parameter sequence of that sub-meter is shifted forward or backward by the corresponding number of sampling periods. Only the time label is adjusted, keeping the voltage, current, power, and energy values ​​unchanged. In this case, the data of each sampling interval of the sub-meter falls completely into the corresponding sampling period of the main meter, without any splitting or interpolation of the values.

[0044] When the total deviation of a certain sub-meter does not correspond to an integer multiple of the sampling period, a certain sampling interval of the sub-meter spans two adjacent sampling periods of the main meter in time. The original power needs to be reasonably distributed between the two periods. If it is simply divided equally by time, it will introduce obvious errors when the load differs greatly between the first half and the second half of the interval. Therefore, this embodiment performs weighted distribution based on the power change trend within the interval.

[0045] Specifically, the original sampling interval is divided into a front segment and a back segment based on the total deviation; the power at the segmentation point is obtained by linear interpolation of the interval's starting power and ending power, i.e., the starting power plus the proportion of the front segment duration multiplied by the difference between the starting and ending power; the estimated average power of the front segment is the average of the interval's starting power and the power at the segmentation point, and the estimated average power of the back segment is the average of the power at the segmentation point and the power at the interval's ending power.

[0046] Based on this, the original electricity is allocated according to the duration ratio of the first and second segments and the estimated average power. The allocation formula can be expressed as: ; in For the first The electricity consumption of each sub-meter within the original sampling range. It is the ratio of the duration of the first segment to the total duration of the sampling interval. and These are the estimated average power values ​​for the first and second segments, respectively. In the formula, the numerator corresponds to the estimated charge weight for the first segment, and the denominator is the sum of the estimated charge weights for the two segments, ensuring that the total charge after allocation is consistent with the original charge. The charge for the second segment is obtained by subtracting the allocated charge for the first segment from the original charge. The front-end power allocation and the back-end power allocation are written into the corresponding master table sampling period respectively; for voltage and current data, the corrected values ​​are obtained by interpolation according to adjacent sampling points.

[0047] After the above reconfiguration, the sub-table data of each sampling period corresponds to the total table data in time, forming a time-stamped corrected sub-table electrical parameter sequence.

[0048] In step S5, the branch current and voltage difference are obtained from the sequence, the equivalent resistance of the branch is estimated, and the variable technical loss is calculated using the branch current and the equivalent resistance of the branch, forming the residual before correction, the residual after time-scale correction, and the residual after loss correction.

[0049] Specifically, please refer to Figure 3 , Figure 3 This is a flowchart illustrating the calculation of variable technology losses provided in an embodiment of this disclosure. Figure 3 As shown in box 301, based on the time-scaled corrected sub-meter electrical parameter sequence, the branch current and voltage difference are calculated according to phase and branch.

[0050] For submeters with current data, the submeter current after time-scale correction is used directly; for submeters without current data but with active power and voltage data, the approximate current is obtained by dividing the active power after time-scale correction by the voltage after time-scale correction; the submeter currents and approximate currents within the same branch are combined, that is, the branch current of the branch is obtained by summing them one by one.

[0051] Regarding the voltage difference, the voltage difference of each sub-meter in the same branch is obtained by subtracting the phase voltage of the main meter from the voltage of each sub-meter. Since the wire length and terminal of each sub-meter in the same branch may vary slightly from the main meter, the voltage difference of each sub-meter has a certain degree of dispersion. The system averages the voltage difference of each sub-meter in the same branch to obtain the average voltage difference of the branch. This value reflects the overall voltage drop level of the wires and terminals from the main meter to the user end of the branch.

[0052] In box 302, the branch current and voltage difference are obtained from this sequence, and the equivalent resistance of the branch is estimated. The equivalent resistance of the branch refers to the equivalent value of all series resistances from the main metering point to the user terminal of the branch, including the wire resistance and terminal contact resistance. Its magnitude depends on the cross-section and length of the branch wires and the contact condition of the terminals. The system uses branch current and average voltage difference data from multiple sampling periods to estimate the equivalent resistance of the branch using the least squares method.

[0053] Specifically, for a given branch, the average voltage difference of the branch in each sampling period is approximately equal to the equivalent resistance of the branch multiplied by the branch current. The system uses the equivalent resistance of the branch as the parameter to be determined, and multiplies the equivalent resistance of the branch by the branch current to obtain the estimated voltage difference value for each period. The solution is performed under the condition of minimizing the sum of squared deviations between the average voltage difference of the branch and the estimated voltage difference value in all sampling periods. In essence, it fits the best linear coefficient among multiple sets of voltage difference and current data points collected under different load levels, and the fitting result is the equivalent resistance of the branch. The wider the current range covered by the data and the more effective data points, the more accurate the estimation result.

[0054] In some embodiments, the data pairs used for equivalent resistance estimation need to be screened to avoid interference from abnormal power consumption data on impedance parameters. Specifically, the system retains sampling periods that meet the following conditions: the time-scaled residual falls within a preset residual range, meaning there is no significant power imbalance in that period. The preset residual range can be calibrated during the deployment phase based on the statistical distribution of residuals from historical normal data; the direction of voltage drop change at the user end is consistent with the direction of branch current change, i.e., voltage drop increases when current increases and decreases when current decreases; the load change exceeds a preset load change threshold. The preset load change threshold is used to exclude periods where the load is almost constant and the voltage difference changes very little, and can be configured as a certain percentage of the branch's rated current. Sampling periods that do not meet the above conditions are discarded. Among them, sampling periods where the time-scaled residual exceeds the preset residual range may include bypass power consumption or metering circuit anomalies, and including them in impedance estimation will cause the results to deviate from the true value, so they are excluded.

[0055] Optionally, after completing the equivalent resistance estimation, the system calculates the branch equivalent resistance using data from different dates or different load levels and performs cross-validation. If the difference in the branch equivalent resistance obtained under multiple load levels for the same branch exceeds a preset consistency threshold, the system marks the branch as having insufficient resistance estimation reliability and reduces the loss correction weight of the branch when calculating variable technical losses in the future. The loss correction weight ranges from 0 to 1, and is set to 1 when the resistance estimation reliability is sufficient.

[0056] When the maximum equivalent resistance of the same branch is obtained under multiple load levels, The minimum value is The preset consistency threshold is At that time, with Indicates actual differences; when At that time, loss correction weight ,when At that time, loss correction weight The pre-defined consistency threshold is pre-calibrated during the system deployment phase based on the conductor cross-section, terminal type, and typical load range of the metering boxes within the distribution area.

[0057] In box 303, variable technical losses are calculated using branch current and branch equivalent resistance over the sampling period. Specifically, for a single branch segment, the active power loss is obtained by multiplying the square of the branch current by the branch equivalent resistance. This active power loss is then multiplied by the sampling period duration and the corresponding branch loss correction weight. This yields the branch loss power calculated including the loss correction residual; when the resistance estimation is sufficiently reliable... The power loss of this branch is equal to the power loss of the branch before reduction; unlike the fixed line loss rate method, the loss changes with the branch current in this calculation method: when the current increases, the loss increases significantly with a square relationship, and when the current decreases, the loss decreases accordingly, which can reflect the actual physical process.

[0058] For the common incoming line segment and the user terminal branch connected in series, the system calculates the power loss of the common incoming line segment and the power loss of each user terminal branch separately. The current of the common incoming line segment is the sum of the currents of each sub-meter in the same phase, and the current of the user terminal branch is the sum of the currents of each sub-meter in each branch. The estimation method of the equivalent resistance of the common incoming line segment is the same as that of the terminal branch; the sum of the currents of each sub-meter in the same phase is used as the current of the common incoming line segment, and the common change in the average voltage difference of each branch in the same phase within the same sampling period is used as the voltage difference of the common incoming line segment. The current of the common incoming line segment and the voltage difference of the common incoming line segment in multiple sampling periods are used as data pairs, and the equivalent resistance of the common incoming line segment is estimated according to the least squares method described above.

[0059] The variable technical loss for that phase in that sampling period is obtained by adding the power loss of the common incoming line section to the power loss of each user terminal branch. The power loss of the common incoming line section is calculated based on the current of the common incoming line section, the equivalent resistance of the common incoming line section, and the corresponding loss correction weight. The power loss of the user terminal branch is calculated based on the current of each user terminal branch, the equivalent resistance of each user terminal branch, and the corresponding loss correction weight. Calculating the power loss of each phase separately can avoid the mutual masking of the losses of each phase when the three phases are unbalanced.

[0060] In step S6, the following residuals are generated: pre-correction residual, post-timescale correction residual, and loss correction residual. When the pre-correction residual exceeds the preset residual range and the post-timescale correction residual falls into the preset zero range, it is recorded as a timescale abnormality. When the loss correction residual is positively correlated with the square of the branch current and the voltage drop exceeds the voltage drop corresponding to the equivalent resistance, it is recorded as a branch resistivity abnormality. When the loss correction residual is synchronized with the load of a single user and the sub-meter current is mismatched, it is recorded as a bypass power consumption or metering circuit abnormality. The rest are recorded as insufficient data quality, and a record is generated.

[0061] In some embodiments, the system calculates residuals at three levels according to phase and sampling period. Specifically, the residual before correction is the total meter power consumption minus the sum of the original sub-meter power consumption, which reflects the power difference between the total meter and sub-meters without any correction. The residual after time-scale correction is the total meter power consumption minus the sum of the sub-meter power consumption after time-scale correction, which eliminates the false differences caused by misaligned acquisition time scales, but still includes the influence of branch technical losses. The residual after loss correction is the total meter power consumption minus the sum of the sub-meter power consumption after time-scale correction, minus the variable technical losses, where the variable technical losses are the power losses calculated according to the aforementioned steps and included in the loss correction weight. This value further eliminates the fluctuations of normal technical losses with load changes based on the first two corrections.

[0062] The residuals at the three levels are stored separately according to phase and sampling period, and output as a recorded residual sequence. By comparing the changing trends of the residuals at the three levels, the system can preliminarily determine the main source of the residuals: if the residuals before correction are significantly larger and the residuals after time-scale correction decrease significantly, it indicates that the residuals mainly come from the misalignment of the acquisition time; if there are still residuals after time-scale correction but they decrease significantly after deducting variable technical losses, it indicates that the residuals mainly come from the fluctuation of normal technical losses of the branch with the load level; if there are still stable residuals after both corrections, there may be a real power consumption anomaly, and the system will proceed to the subsequent anomaly type identification.

[0063] In some embodiments, please refer to Figure 4 , Figure 4 This is a flowchart of the exception type determination provided in an embodiment of this disclosure. For example... Figure 4 As shown, the system judges the exception type in the following order. Once the condition of the previous type is met, the corresponding record is generated and no further judgment is made.

[0064] First, determine if there is an anomaly in the data acquisition timescale: when the residual before correction exceeds the preset residual range and the residual after timescale correction falls into the preset zero value range, record the data acquisition timescale anomaly and write the common timescale deviation or individual timescale deviation into the record; the preset zero value range can be calibrated during deployment based on the measurement error and data accuracy, for example, taking a certain proportion of the electricity corresponding to the rated capacity of the metering box; this type of anomaly has been eliminated after timescale correction, so it will not proceed to the subsequent judgment.

[0065] If the abnormal conditions of the acquisition time scale are not met, determine whether it is a branch resistance abnormality: The system calculates the correlation between the loss correction residual and the square of the branch current in the same period. If the two show a positive correlation, that is, the larger the branch current, the larger the residual, and the amount of voltage drop at the user end when the branch current increases exceeds the voltage drop corresponding to the branch equivalent resistance, then record the branch resistance abnormality.

[0066] The logic here is that the equivalent resistance estimation uses the filtered normal segment data. When the actual impedance of the branch in the abnormal segment is higher than the estimated value due to poor terminal contact, terminal heating or wire deterioration, the extra loss will be reflected in the loss correction residual in a way that is positively correlated with the square of the current. At the same time, the voltage drop at the user end will also exceed the voltage drop corresponding to the estimated equivalent resistance.

[0067] In anomaly identification, the degree of correlation is represented by the correlation coefficient within the same decision time window; the decision time window consists of multiple consecutive sampling periods and is consistent with the sampling period used to generate the residual sequence; for loss-corrected residual sequences... With the square sequence of branch currents When the correlation coefficient between the two within the judgment time window is greater than zero, it is determined that the loss correction residual is positively correlated with the square of the branch current.

[0068] If the branch resistance abnormality condition is not met, determine whether it is a bypass power consumption or metering circuit abnormality: If the loss correction residual is synchronized with the load change of a single user, that is, the residual changes synchronously when the active power of the user increases or decreases, but the current of the user's submeter does not increase accordingly, and the voltage change at the user end is inconsistent with the voltage change calculated from the branch equivalent resistance and the branch current change, then record the bypass power consumption or metering circuit abnormality.

[0069] Synchronization between loss correction residuals and individual user load changes means that within the same judgment time window, the correlation coefficient between the loss correction residual sequence and the user's time-scaled corrected active power change sequence is the largest and greater than zero among users in the same phase. Inconsistency between user-end voltage changes and voltage changes calculated from branch equivalent resistance and branch current changes means that within the same judgment time window, the actual user-end voltage change direction differs from the estimated voltage change direction obtained from branch equivalent resistance and branch current changes. This judgment reflects a physical scenario where there is a current path bypassing the sub-meter or a sub-meter metering loop fault leading to under-metering.

[0070] If none of the first three conditions are met, the quality of the recorded data is insufficient. This situation is characterized by residuals being scattered among multiple users in the same phase, and the residual patterns of each period lacking consistency, making it impossible to stably point to a single table position or a single branch.

[0071] The above judgments are executed sequentially, and the record type is determined sequentially: For timescale anomalies, the condition is that the residual before correction exceeds the preset residual range and falls into the preset zero range after correction. Once this condition is met, a record is generated and the process does not proceed to subsequent judgments. If the timescale anomaly condition is not met, then branch resistance anomalies are judged. If the branch resistance anomaly condition is not met, then bypass power consumption or metering circuit anomalies are judged, because the residual caused by bypass power consumption is synchronized with the load of a specific user rather than related to the square of the total branch current. Bypass power consumption anomalies are judged based on the residual being synchronized with the load of a single user and the sub-meter current mismatch. Specifically, sub-meter current mismatch means that when the active power of the user increases after timescale correction, the sub-meter current of the user does not increase in the same direction, or when the active power of the user decreases after timescale correction, the sub-meter current of the user does not decrease in the same direction.

[0072] In some embodiments, the system includes an abnormal phase identifier, an abnormal branch identifier, a suspicious meter position identifier, an estimated time scale deviation, a branch equivalent resistance, a variable technical loss value for each sampling period, a residual sequence before correction, a residual sequence after time scale correction, a residual sequence after loss correction, the time period of the abnormality, and the type of abnormality in each record generated by the system. Based on this, on-site personnel can specifically check the data acquisition task configuration, meter clock, terminal block status, branch cables, and suspicious bypass connection points.

[0073] In some embodiments, the system continuously runs the above process in a time sequence and cumulatively updates the equivalent resistance estimation results of the same metering box. When new sampling data arrives, the branch equivalent resistance is updated using the new data increment, and the updated value is compared with the historical value. If the equivalent resistance of a branch continues to rise over a period of time, it can serve as an early warning of the gradual deterioration of poor terminal contact. Optionally, the system summarizes the results of multiple metering boxes, and maintenance personnel can determine whether the anomaly is a problem with an individual meter or a problem with the entire acquisition system based on the distribution of the anomaly type among different metering boxes and different phases.

[0074] Therefore, this method can distinguish between misaligned acquisition time scales, normal variable technical losses, abnormal branch contact, and bypass power consumption, reducing misjudgments caused by inconsistent time scales and the assumption of a fixed line loss rate, and improving the accuracy of anomaly location at the meter level.

[0075] Please see Figure 5 , Figure 5 This is a structural block diagram of a multi-position metering box branch circuit electricity consumption behavior identification system based on data correlation, provided in an embodiment of this application. Figure 5 As shown, the system includes: The data acquisition and modeling module 501 is used to acquire frozen data from the master table and sub-tables, as well as metering box files, and to establish the original electrical parameter sequence and branch user set. The common timescale deviation determination module 502 is used to calculate the total phase power increment sequence and the in-phase sub-phase combined power increment sequence in the calculation segment, and take the maximum cross-correlation shift as the common timescale deviation. The individual timescale deviation determination module 503 is used to construct a sub-table reference signal with the common timescale deviation, and take the maximum additional shift of the cross-correlation between the sub-table power increment sequence and the corresponding reference signal as the individual timescale deviation; The time scale correction module 504 is used to add the common time scale deviation to the individual time scale deviation to obtain the total deviation of each sub-meter, and reassign the electrical parameter sequence according to the total deviation of each sub-meter to obtain the time scale corrected sub-meter electrical parameter sequence; The loss calculation module 505 is used to obtain the branch current and voltage difference based on the time-scaled corrected sub-meter electrical parameter sequence, estimate the branch equivalent resistance, and calculate the variable technical loss based on the branch current and the branch equivalent resistance. The residual identification and recording module 506 is used to generate the pre-correction residual, the time-scaled corrected residual, and the loss-corrected residual. Specifically, when the pre-correction residual exceeds a preset residual range and the time-scaled corrected residual falls within a preset zero range, it is recorded as a time-scaled abnormality. When the loss-corrected residual is positively correlated with the square of the branch current and the voltage drop exceeds the voltage drop corresponding to the equivalent resistance, it is recorded as a branch resistivity abnormality. When the loss-corrected residual is synchronized with the load of a single user and the sub-meter current is mismatched, it is recorded as a bypass power consumption or metering circuit abnormality. If the aforementioned conditions are not met, it is recorded as insufficient data quality, and a record is generated.

[0076] Each processing unit and / or module in the embodiments of this application can be implemented by an analog circuit that implements the functions described in the embodiments of this application, or by software that executes the functions described in the embodiments of this application.

[0077] Based on the same inventive concept, this application also provides an electronic device, the method corresponding to which can be the method in the foregoing embodiments, and its problem-solving principle is similar to that method. For example... Figure 6 As shown, Figure 6 This is a schematic block diagram of an electronic device provided in an embodiment of the present disclosure. The device includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the methods and / or technical solutions of the foregoing embodiments of the present application.

[0078] In particular, the methods and / or embodiments in this application can be implemented as computer software programs. For example, the embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowchart. When the computer program is executed by a processor, it performs the functions defined in the methods of this application.

[0079] Another embodiment of this application provides a storage medium storing computer program instructions thereon, which can be executed by a processor to implement the methods and / or technical solutions of any one or more embodiments of this application.

[0080] In the above embodiments, the descriptions of each embodiment have different focuses. Parts not described in detail in a certain embodiment can be referred to in the relevant descriptions of other embodiments. The above descriptions are merely preferred embodiments of this application and explanations of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in this application is not limited to the technical solutions formed by specific combinations of the above technical features, but should also cover other technical solutions formed by arbitrary combinations of the above technical features or their equivalent features without departing from the inventive concept.

Claims

1. A method for identifying the electricity consumption behavior of branch circuits in multi-position metering boxes based on data correlation, characterized in that, include: Acquire the frozen data of the master meter and sub-meters, as well as the meter box files, and establish the original electrical parameter sequence and branch user set. This includes retaining the original freezing time of each sampling point in the frozen data of the master meter and sub-meters, and marking each sampling point with data source identifier, phase identifier, meter position identifier, acquisition cycle, and data integrity status. Based on the phase relationship and branch affiliation relationship in the meter box files, single-phase users are assigned to the corresponding phase branch, and three-phase users with three-phase power or current data are assigned to the branch user set according to phases A, B, and C respectively. In the calculation segment, the total phase power increment sequence and the in-phase sub-meter combined power increment sequence are calculated, and the maximum cross-correlation shift is taken as the common time scale deviation. A sub-table reference signal is constructed using a common timescale deviation. The maximum additional shift amount of cross-correlation between the sub-table power increment sequence and the corresponding reference signal is taken as the individual timescale deviation. The process includes: selecting a sub-table as the target sub-table; shifting the power increment sequences of the remaining sub-tables (excluding the target sub-table) according to the common timescale deviation and summing them to obtain the corrected composite power increment sequences of the remaining sub-tables; subtracting the corrected composite power increment sequences of the remaining sub-tables from the total phase power increment sequence to obtain the sub-table reference signal of the target sub-table; and, based on the shift of the common timescale deviation, continuing to shift the target sub-table power increment sequence within a preset additional shift range, taking the additional shift amount when the cross-correlation between the target sub-table power increment sequence and the sub-table reference signal of the target sub-table is maximized as the individual timescale deviation of the target sub-table. The total deviation of each sub-meter is obtained by adding the common time scale deviation to the individual time scale deviation. The electrical parameter sequence is then reassigned according to the total deviation of each sub-meter to obtain the time scale corrected sub-meter electrical parameter sequence. The branch current and voltage difference are obtained from this sequence, the equivalent resistance of the branch is estimated, and the variable technical loss is calculated using the branch current and the equivalent resistance of the branch. The system generates the pre-correction residual, the post-timescale correction residual, and the loss correction residual, and determines the anomaly type in sequence. Specifically, when the pre-correction residual exceeds the preset residual range and the post-timescale correction residual falls into the preset zero value range, it is recorded as a timescale acquisition anomaly. When the loss correction residual is positively correlated with the square of the branch current and the voltage drop exceeds the voltage drop corresponding to the equivalent resistance, it is recorded as a branch resistive anomaly. When the loss correction residual is synchronized with the load of a single user and the sub-meter current is mismatched, it is recorded as bypass power consumption or metering circuit abnormality.

2. The method according to claim 1, characterized in that, It also includes filtering calculation segments: retaining segments where the total table and sub-table data collection is complete, the number of sampling points whose absolute power increment exceeds the preset amplitude threshold meets the preset quantity condition, and the phase file has not been changed; removing segments where the number of missing sampling points exceeds the preset value, needs to be supplemented, or where the absolute power increment of the total table and multiple sub-tables simultaneously exceeds the preset jump threshold.

3. The method according to claim 1, characterized in that, Determining the common timescale deviation includes: sequentially selecting candidate translation amounts within a preset translation range; translating the in-phase sub-table synthesized power increment sequence according to the candidate translation amounts; calculating the sum of the products of the translated in-phase sub-table synthesized power increment sequence and the total in-phase power increment sequence for effective sampling points within the calculation segment; taking the candidate translation amount with the largest product sum as the segment common timescale deviation; taking the median of multiple segment common timescale deviations and using this median as the common timescale deviation.

4. The method according to claim 1, characterized in that, The reconfigured power parameter sequence includes: when the total deviation of each sub-meter corresponds to an integer multiple of the sampling period, the sub-meter power parameter sequence is shifted forward or backward by the corresponding number of sampling periods, adjusting only the time tag while maintaining the voltage, current, power, and energy values; otherwise, the original sampling interval is divided into a front segment and a back segment based on the total deviation of each sub-meter, and the power at the dividing point is obtained by linear interpolation based on the starting power and ending power of the interval. The estimated average power of the front segment is obtained from the starting power and the power at the dividing point, and the estimated average power of the back segment is obtained from the power at the dividing point and the power at the ending power of the interval. The original energy is allocated according to the duration ratio of the front segment and the back segment and the estimated average power. The energy of the back segment is obtained by subtracting the allocated energy of the front segment from the original energy. The allocated energy of the front segment and the energy of the back segment are written into the corresponding total meter sampling period, and the voltage and current are obtained by interpolation based on adjacent sampling points.

5. The method according to claim 1, characterized in that, The branch current and voltage difference are calculated as follows: for sub-meters with current data, the sub-meter current is corrected for time scale; for sub-meters without current data but with active power and voltage data, the approximate current is obtained by dividing the time-scale corrected active power by the time-scale corrected voltage; the sub-meter current and the approximate current in the same branch are combined to obtain the branch current. The voltage difference between individual meters is obtained by subtracting the voltage of individual meters within the same branch from the total phase voltage. The average voltage difference between individual meters within the same branch is then calculated by averaging these voltage differences.

6. The method according to claim 5, characterized in that, The equivalent resistance of the branch is estimated by collecting branch current and average voltage difference of the branch for multiple sampling periods as data pairs, multiplying the branch equivalent resistance by the branch current to obtain the estimated voltage difference, and using the least squares method to obtain the equivalent resistance of the branch with the condition of minimizing the sum of squares of the deviations between the average voltage difference of the branch and the estimated voltage difference.

7. The method according to claim 6, characterized in that, The data collection includes: retaining sampling periods where the time-scaled residual falls within a preset residual range, the direction of change of user-end voltage drop is consistent with the direction of change of branch current, and the load change amplitude exceeds a preset load change threshold; eliminating sampling periods where the time-scaled residual exceeds the preset residual range; when the difference in equivalent resistance of a branch obtained under multiple load levels exceeds a preset consistency threshold, marking the branch as having insufficient reliability in resistance estimation, and reducing the loss correction weight of the branch when calculating variable technical losses.

8. The method according to claim 1, characterized in that, The variable technical loss is calculated according to the sampling period; for a single branch, the active power loss of the branch is obtained by multiplying the square of the branch current by the equivalent resistance of the branch, and the active power loss of the branch is multiplied by the sampling period and processed according to the loss correction weight to obtain the power loss of the branch; for the phases in series between the common incoming line section and the user terminal branch, the power loss of the common incoming line section and the power loss of the user terminal branch are calculated separately, and the two are added together to obtain the variable technical loss of the phase.

Images