Method and system for evaluating errors of smart meters and feeder line loss, device, and medium

Abstract

The application discloses an intelligent electric meter error and transformer area line loss evaluation method and system, equipment and a medium, the method constructs a dynamic line loss model of line loss changing with load, considers the dynamic correlation characteristics of line loss and power supply in the actual scene, so that the line loss is dynamically adjusted with the actual load, the estimation accuracy and adaptability are improved, and the joint estimation of the line loss parameters and the electric meter error is realized through the dynamic line loss model. Moreover, a fading memory recursive algorithm based on a dynamic forgetting factor strategy is used for solving, compared with a fixed forgetting factor strategy, data saturation and noise interference can be effectively inhibited, and an effective incremental block which keeps physical continuity on data is constructed, each effective incremental block is taken as an independent iteration unit for solving, a high-quality data basis is provided for the solving process, the noise, missing values and abnormal working conditions such as equipment restart in the measurement data can be well adapted, and therefore the robustness and reliability of the estimation result are improved.

Description

Technical Field

[0001] This invention relates to the field of meter error assessment technology, and in particular to a method and system for assessing smart meter errors and transformer area line losses, an electronic device, and a computer-readable storage medium. Background Technology

[0002] In the scenario of power distribution meter error assessment, traditional verification methods rely on manual on-site sampling, which suffers from low efficiency and limited coverage, making it difficult to meet the full monitoring needs of hundreds of millions of meters in the grid. However, with the widespread application of advanced measurement systems, distribution substations have accumulated massive amounts of power measurement data, providing data support for estimating meter errors. Existing meter error assessment methods rely on the main and sub-meter measurement equipment already deployed in the distribution substation to collect data such as cumulative power readings. By constructing models and solving algorithms, they can monitor the operating errors of smart meters and line losses in the distribution area. No additional dedicated testing equipment is required; error assessment can be completed using only existing measurement data, significantly reducing the manpower and time costs required for on-site verification. Therefore, it has significant application value in fields such as lean operation and maintenance of smart grids. However, existing meter error assessment methods typically employ a fixed line loss assumption, neglecting the dynamic correlation between line loss and power supply in real-world scenarios. This leads to significant deviations in line loss estimation, which in turn affects the accuracy of meter error assessment. Furthermore, the assessment algorithms are prone to data saturation and are insufficiently adaptable to noise, missing values, and abnormal operating conditions such as equipment restarts in the measurement data, resulting in poor robustness and reliability of meter error estimation. Summary of the Invention

[0003] This invention provides a method and system for assessing smart meter errors and transformer area line losses, as well as an electronic device and a computer-readable storage medium. It can achieve joint estimation of line loss parameters and meter errors, and also improves estimation accuracy, robustness and reliability.

[0004] According to one aspect of the present invention, a method for evaluating smart meter errors and transformer area line losses is provided, comprising the following: Collect electricity metering data from the master table and each sub-table to obtain an electricity metering dataset. Extract multiple valid incremental blocks that maintain physical continuity in the data from the electricity metering dataset. Perform outlier cleaning on the data of each valid incremental block. The data of each valid incremental block includes the electricity increment sequence of the master table and each sub-table. Based on the principle of energy conservation and the incremental energy data of the main meter and each sub-meter, a dynamic line loss model is constructed to reflect the changes in line loss with load. Using each effective increment block as an independent iteration unit, the dynamic line loss model is solved by a gradual memory recursion algorithm based on a dynamic forgetting factor strategy, which yields the line loss of the transformer area and the relative error of each sub-meter within the time period corresponding to each effective increment block.

[0005] Furthermore, the process of constructing a dynamic line loss model that varies with load based on the principle of energy conservation and the incremental energy data of the main meter and each sub-meter includes the following: Each effective incremental block is divided into multiple metering cycles, and a power balance model for each metering cycle is constructed based on the principle of energy conservation. The power balance model of the distribution area is calculated by weighted average based on multiple metering periods to obtain the power balance model of the distribution area with weighted average for multiple metering periods. Based on the direct proportional relationship between line loss rate and power supply, the power weighted average balance model of the transformer area is transformed into a dynamic line loss model in the form of linear regression.

[0006] Furthermore, the linear regression form of the dynamic line loss model is as follows: ; in, This represents the difference between the weighted average of the total energy consumption increment from the main meter and the sum of the weighted average of the energy consumption increments from all sub-meters over multiple metering periods. This represents the weighted average of fixed losses. Indicates the number of table partitions. Indicates the first j Only the weighted average of the electricity increments from the individual meters. Indicates the first j Only the relative error weighted average of the table is used. This represents the weighted average of the total electricity consumption increment. Indicates the base line loss rate. This represents the line loss coefficient, used to characterize the degree of fluctuation in the current period load compared to the average load.

[0007] Furthermore, the expression for the gradual fading memory recursive algorithm based on the dynamic forgetting factor strategy is as follows: ; in, Indicates the number of measurement periods. Indicates the first k Forgetting factor per measurement period, Indicates hyperparameters, express( k +1) The difference between the weighted average of the total energy consumption increment of the main meter and the sum of the weighted average of the energy consumption increments of all sub-meters within a metering period. express( k+1) The regression vector at time step (+1), superscript T Indicates transpose. express k The estimation results of the time parameter vector. express( k The gain vector at time +1) express k The covariance matrix at time t, express( k The covariance matrix at time +1), Represents the identity matrix. express( k +1) The estimation result of the parameter vector at time step.

[0008] Furthermore, if the covariance matrix If the condition number is greater than a preset threshold, then use replace , Represents the regularization coefficient. , This represents the maximum absolute value of the correlation coefficient between the electrical energy increment sequences of any two sub-tables.

[0009] Furthermore, the process of extracting multiple effective incremental blocks that maintain physical continuity in the data from the electricity metering dataset includes the following: Detect reset points in the electricity metering dataset; where a reset point represents the moment when the meter restarts or the reading changes. The time axis is divided based on the reset point, and the electricity metering dataset is divided into multiple data blocks that maintain physical continuity of the data, and data blocks with a data size less than the sample size threshold are removed. The logical zero point of each data block is identified, and the data in each data block that is before the logical zero point is removed. The power increment sequence of the master table and each sub-table is calculated based on the logical zero point, thereby transforming each data block into an effective increment block. The logical zero point represents the load stabilization moment of the unified increment benchmark.

[0010] Furthermore, the process of cleaning outliers from the data in each valid incremental block includes the following: For each power increment sequence within the effective increment block, first calculate the 25th percentile, 75th percentile, and interquartile range of the power increment data to set the normal data range, mark the normal data samples, then calculate the mean and standard deviation of the normal data samples, and standardize each normal data sample based on the mean and standard deviation, retaining only the data samples whose standardized values ​​are less than the preset threshold.

[0011] In addition, the present invention also provides an evaluation system for smart meter errors and transformer area line losses, comprising: The data preprocessing module is used to collect the electricity metering data from the master table and each sub-table to obtain the electricity metering dataset. From the electricity metering dataset, multiple effective incremental blocks that maintain physical continuity in the data are extracted, and outlier cleaning is performed on the data of each effective incremental block. The data of each effective incremental block includes the electricity increment sequence of the master table and each sub-table. The model building module is used to build a dynamic line loss model that varies with load based on the principle of energy conservation and the incremental energy data of the main meter and each sub-meter. The iterative solution module is used to solve the dynamic line loss model by using each effective incremental block as an independent iterative unit and employing a gradual memory recursive algorithm based on a dynamic forgetting factor strategy. This yields the line loss of the transformer area and the relative error of each sub-meter within the time period corresponding to each effective incremental block.

[0012] In addition, the present invention also provides an electronic device, including a processor and a memory, wherein the memory stores a computer program, and the processor executes the steps of the method described above by calling the computer program stored in the memory.

[0013] In addition, the present invention provides a computer-readable storage medium for storing a computer program for evaluating smart meter errors and transformer area line losses, wherein the computer program performs the steps of the method described above when run on a computer.

[0014] The present invention has the following beneficial effects: The method for evaluating smart meter errors and transformer area line losses in this invention constructs a dynamic line loss model that adapts to load variations. Compared to existing fixed line loss assumptions, this method considers the dynamic correlation between line loss and power supply in real-world scenarios, allowing line loss to dynamically adjust with the actual load, thus improving estimation accuracy and adaptability. Furthermore, the dynamic line loss model enables joint estimation of line loss parameters and meter errors. A gradual forgetting factor-based recursive algorithm is employed for solving the problem. Compared to a fixed forgetting factor strategy, this effectively suppresses data saturation and noise interference. By constructing effective incremental blocks that maintain physical continuity in the data, and using each effective incremental block as an independent iterative unit, a high-quality data foundation is provided for the solution process. This method effectively adapts to noise, missing values, and abnormal operating conditions such as equipment restarts in the measurement data, thereby improving the robustness and reliability of the estimation results.

[0015] In addition, the smart meter error and transformer area line loss assessment system of the present invention also has the above-mentioned advantages.

[0016] In addition to the objectives, features, and advantages described above, the present invention has other objectives, features, and advantages. The invention will now be described in further detail with reference to the figures. Attached Figure Description

[0017] The accompanying drawings, which form part of this application, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings: Figure 1 This is a flowchart illustrating the method for evaluating smart meter errors and transformer area line losses according to a preferred embodiment of this application. Figure 2 yes Figure 1 A schematic diagram of the sub-process of step S1; Figure 3 yes Figure 1 A schematic diagram of the sub-process of step S2; Figure 4 This is a schematic diagram of the module structure of a smart meter error and transformer area line loss assessment system according to another embodiment of this application. Detailed Implementation

[0018] It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other. This application will now be described in detail with reference to the accompanying drawings and embodiments.

[0019] Reference Figure 1 A preferred embodiment of this application provides a method for evaluating smart meter errors and transformer area line losses, including the following: Step S1: Collect the electricity metering data from the master meter and each sub-meter to obtain the electricity metering dataset. Extract multiple valid incremental blocks that maintain physical continuity in the data from the electricity metering dataset. Perform outlier cleaning on the data of each valid incremental block. The data of each valid incremental block includes the electricity increment sequence of the master meter and each sub-meter. Step S2: Based on the principle of energy conservation and the incremental energy data of the main meter and each sub-meter, construct a dynamic line loss model that varies with load. Step S3: Using each effective incremental block as an independent iteration unit, the dynamic line loss model is solved by the gradual fading memory recursive algorithm based on the dynamic forgetting factor strategy, so as to obtain the transformer area line loss and the relative error of each sub-table within the time period corresponding to each effective incremental block.

[0020] It is understood that the smart meter error and transformer area line loss assessment method in this embodiment constructs a dynamic line loss model that varies with load. Compared to the existing fixed line loss assumption, this method considers the dynamic correlation between line loss and power supply in actual scenarios, allowing line loss to dynamically adjust with the actual load, thus improving estimation accuracy and adaptability. Furthermore, the dynamic line loss model enables joint estimation of line loss parameters and meter error. Moreover, it employs a fading memory recursive algorithm based on a dynamic forgetting factor strategy for solving the problem. Compared to a fixed forgetting factor strategy, this effectively suppresses data saturation and noise interference. It also constructs effective incremental blocks that maintain physical continuity in the data, using each effective incremental block as an independent iterative unit for solving the problem. This provides a high-quality data foundation for the solution process and can well adapt to noise, missing values, and abnormal operating conditions such as equipment restarts in the measurement data, thereby improving the robustness and reliability of the estimation results.

[0021] In step S1, measurement data for the transformer substation is first collected, specifically through a master meter already deployed in the substation. M Only the table is split, collecting cumulative readings of total positive active power, i.e., power metering data. Data acquisition intervals support 1 minute of raw data or 5 minutes, 10 minutes, etc., resampled data. Additionally, to ensure the data length meets the requirements of subsequent models, a larger data acquisition window can be selected, and acquisition cycles with a success rate of less than 100% can be removed. For the first j Table partitioning only t Time reading, For the summary table t Readings at specific times. After collecting the electricity metering data from the master meter and each sub-meter, an electricity metering dataset can be constructed. The electricity metering dataset includes a series of sampling times for the master meter and each sub-meter's electricity metering time series data. Then, multiple effective incremental blocks that maintain physical continuity in the data are extracted from the electricity metering dataset.

[0022] Among them, such as Figure 2 As shown, the process of extracting multiple effective incremental blocks that maintain physical continuity in the data from the electricity metering dataset includes the following: Step S11: Detect the reset point in the electricity metering data set; where the reset point represents the moment when the meter restarts or the reading changes. Step S12: Divide the time axis based on the reset point, divide the electricity metering dataset into multiple data blocks that maintain physical continuity of the data, and remove data blocks whose data volume is less than the sample size threshold; Step S13: Identify the logical zero point of each data block, remove the data in each data block that is before the logical zero point, and calculate the power increment sequence of the master table and each sub-table based on the logical zero point, thereby converting each data block into an effective increment block; where the logical zero point represents the load stabilization moment of the unified increment benchmark.

[0023] Specifically, the reset point in the electricity metering dataset is first detected. The reset point represents the moment when the meter restarts or the reading jumps. For example, when any of the following occurs, the corresponding moment is determined to be a reset point: a single meter drops significantly (e.g., relative change exceeds 30% of the maximum value or absolute change exceeds 50 kWh), multiple meters show negative growth simultaneously (e.g., more than 50% of sub-meters drop in the same direction), or the total meter drops significantly (e.g., more than 100 kWh). After obtaining the reset point list and deduplicating it, the data corresponding to the reset point in the electricity metering dataset is deleted, and the data of the remaining sampling moments are re-sorted according to time.

[0024] Then, the time axis is divided according to the reset point, and the reordered electricity metering dataset is divided into multiple data blocks, each data block Maintain physical continuity of the data and remove data blocks smaller than a sample size threshold (e.g., 50). Additionally, if there is no reset point, the entire electricity metering dataset is treated as a single data block.

[0025] Next, for each data block The system identifies logical zero points, where a logical zero point represents a stable load moment based on a unified incremental benchmark. A logical zero point is defined as a moment in which the total table increment fluctuation is less than a threshold value for multiple consecutive sampling intervals starting from a certain time, or where a preset percentage of sub-table increment fluctuations are less than a threshold value for sub-table increment fluctuations. For example, if a logical zero point is defined as a moment in which the total table increment fluctuation is less than 5% for five consecutive sampling intervals, or where more than 80% of sub-table increment fluctuations are less than 8%, then that moment is considered a logical zero point. The system then further analyzes the data within each data block, prior to the logical zero point segment. Directly exclude, and treat the segment after the logical zero point as a valid incremental block. The power increment sequence of the master meter and each sub-meter is calculated based on the logic zero point. The power increment of the master meter within the effective increment block can be expressed as: , No. j The increase in electrical energy in each sub-meter can be expressed as: The resulting effective increment block is a set of non-zero increments that are physically contiguous and contain more than 50 sampling intervals per block. The data includes the total electricity increment sequence. and the sequence of electrical energy increments for each sub-meter Additionally, data cleaning can be performed, such as deleting completely empty columns and removing outlier samples with negative increments.

[0026] In addition, in step S1, outlier cleaning is also required for the data of each valid incremental block. The specific process is as follows: For each energy increment sequence within the valid increment block, outlier cleaning is first performed using the IQR method, and the 25th percentile of the energy increment data is calculated. 75th percentile and interquartile range To set the normal data range First, label the normal data samples. Then, use the Z-score method to clean up outliers and calculate the mean of the normal data samples. and standard deviation Each normal data sample is standardized based on its mean and standard deviation: Only retain data samples whose standardized values ​​are less than a preset threshold, for example, retain The sample.

[0027] It is understandable that most existing data preprocessing is simple data cleaning, such as missing value filling and outlier removal, without considering the problem of data physical continuity breaks caused by meter restarts and reading jumps. This invention first detects reset points, finds the moments in the power data time series where reset-type load jumps or systemic synchronous rollbacks occur, and regards these moments as data physical continuity breakpoints. Data blocks are segmented at the breakpoints to ensure the physical continuity of each data block. A two-stage outlier filtering is used to improve data purity and data quality. It can well adapt to noise, missing values, and abnormal operating conditions such as equipment restarts in measurement data, thereby improving the robustness and reliability of the estimation results.

[0028] In addition, in step S2, based on the principle of energy conservation and the incremental energy data of the main meter and each sub-meter, a dynamic line loss model is constructed to reflect the change in line loss with load, wherein, for example... Figure 3 As shown, it specifically includes the following: Step S21: Divide each effective incremental block into multiple metering cycles, and construct the power balance model of the transformer area for each metering cycle based on the principle of power conservation; Step S22: Perform a weighted average calculation on the power balance model of the distribution area based on multiple metering periods to obtain the weighted average power balance model of the distribution area for multiple metering periods. Step S23: Based on the direct proportional relationship between line loss rate and power supply, the power weighted average balance model of the transformer area is converted into a dynamic line loss model in the form of linear regression.

[0029] Specifically, each valid increment block Divided into K The measurement cycle is divided into several periods, preferably evenly, with each period lasting [duration].t It can be set according to the actual situation, and recorded. For the first i Within the first metering cycle j Only the actual electricity consumption is measured. For the first i Within the first metering cycle j Only the increment of the electricity meter reading, For the first i The increase in total meter readings within a metering period For the first j If we only consider the approximate relative error of the table, then we have: Since the summary table can be considered to have almost no error, the increment of the summary table can be used to approximate the first... i The total power supply within the first metering cycle, based on the principle of energy conservation, can be used to construct the first... i Power balance model for a single metering cycle in the transformer substation: , M Indicates the number of table partitions. Indicates the first i Line loss rate of the transformer area within a metering cycle. Indicates the first i Total power supply loss of the transformer area within a metering cycle.

[0030] Then, from the first i Starting from each measurement cycle N The power balance model of the distribution area is calculated by weighted averaging over several metering periods to obtain the power balance model of the distribution area with weighted average over multiple metering periods: ,in, N The value can be set as needed, usually between 10 and 30. This represents the weighted average of the total electricity consumption increment. , Indicates the first j Only the weighted average of the electricity increments from the individual meters. , This represents the weighted average of fixed losses. , Indicates the first j Only the relative error weighted average of the table is used. , This represents the weighted average of line loss rates. , Indicates the first k Normalized weighting factors for each measurement period, , Indicates the first k Weights for each measurement period.

[0031] Next, based on circuit principles, the first iThe line loss within each metering cycle meets the following requirements: , t This indicates the duration of each metering cycle, while the total power supply for the distribution area is: Therefore, it can be concluded that the line loss rate is directly proportional to the power supply: This indicates that the greater the load, the higher the line loss rate, accurately reflecting the physical nature of the dynamic change in line loss of a distribution substation with load, and breaking the unreasonable assumption of a fixed line loss in the traditional model. And divide both sides of the expression of the power weighted average balance model of the distribution area by... We can obtain: After weighted averaging, the weighted average of fixed losses and the weighted average of relative errors of each sub-meter can be considered to remain unchanged, i.e. and If it is independent of the metering period, then Represented as ,Will Represented as Meanwhile, based on the direct proportionality between line loss rate and power supply, the above formula can be rewritten as: Then, it is transformed into a linear regression form to facilitate subsequent recursive solutions, thus obtaining a dynamic line loss model in linear regression form: ; in, This represents the difference between the weighted average of the total energy consumption increment from the main meter and the sum of the weighted average of the energy consumption increments from all sub-meters over multiple metering periods. This represents the weighted average of fixed losses. Indicates the number of table partitions. Indicates the first j Only the weighted average of the electricity increments from the individual meters. Indicates the first j Only the relative error weighted average of the table is used. This represents the weighted average of the total electricity consumption increment. Indicates the base line loss rate. This represents the line loss coefficient, used to characterize the fluctuation of the current period's load compared to the average load, reflecting the dynamic correlation of line loss. , This represents the weighted average of the energy increment over K metering periods.

[0032] It is understood that this invention derives the dynamic relationship between line loss rate and power supply based on circuit principles, constructs a dynamic line loss model, which enables line loss to be dynamically adjusted according to the actual load, greatly improving adaptability and estimation accuracy. Furthermore, it reduces noise interference through weighted averaging, and realizes the joint estimation of line loss parameters and meter error.

[0033] It is understandable that the weights for each measurement period are... Existing technologies typically employ a fixed-weight strategy, which leads to poor model generalization ability. Preferably, this invention uses a method that calculates the proportion of high-load samples in the overall table within each measurement period, i.e., the proportion of samples exceeding the 75th percentile. If this proportion is less than a threshold (e.g., 20%), then linear weights are applied. If the proportion is greater than or equal to the preset threshold, then a squared weighting is applied: To highlight high-load effective data, Indicates the first k Total metered energy increment within a metering cycle.

[0034] Furthermore, in step S3, each valid increment block is used as an independent iteration unit, according to... K Based on the dynamic line loss model described above, data from each measurement period can be used to obtain... A system of linear equations consisting of 12 equations: Among them, observation items Regression vector It includes constant terms, sub-table measurement increments, and line loss correlation terms, and the regression matrix. parameter vector Among them, fixed losses Basic line loss rate The line loss parameter to be estimated is the weighted average of the relative errors of each sub-table. Together they form the parameter vector, when If the problem persists, it can be solved using the gradual reduction of memory recursion algorithm.

[0035] The specific solution process is as follows: first, initialize the parameter vector. The dimension of the parameter vector is M +2, including sub-table errors and line loss parameters, initialize the covariance matrix. , Represents the identity matrix. Let represent the initial uncertainty coefficient, typically taken as 1000. Then, based on the diminishing memory recursive algorithm, it can be iterated as follows: ,in, This is the forgetting factor, with a value ranging from 0 to 1. In existing technologies, the forgetting factor... The forgetting factor is usually set to a fixed constant and verified experimentally. The value range of has a significant impact on algorithm performance, and a compromise is usually achieved between 0.97 and 0.99. However, this static setting has inherent technical drawbacks, for example, when When the weight of the model is equal to 1, the algorithm degenerates into the traditional recursive least squares method, where all historical data have the same weight. This prevents new data from effectively correcting the model, resulting in a "data saturation" phenomenon, and the algorithm loses its ability to track the time-varying nature of parameters. However, when the weight of the model is equal to 1, the algorithm degenerates into the traditional recursive least squares method, where all historical data have the same weight. This leads to a situation where new data cannot effectively correct the model, resulting in a "data saturation" phenomenon, and the algorithm loses its ability to track the time-varying nature of parameters. When the value is small (e.g.) <0.97), although it can track parameter changes relatively quickly, it will excessively amplify the interference of measurement noise, leading to instability in the parameter estimation process, a significant increase in mean square error, and a tendency to produce misjudgments and missed judgments; when When the value is large (e.g.) While a value greater than 0.99 provides good smoothing for noise, the algorithm suffers from excessive "inertia." When system parameters (such as meter errors) undergo real, step-like changes, the parameter tracking convergence speed is too slow, resulting in significant dynamic delay. Therefore, this invention addresses these technical shortcomings by abandoning the static trade-off strategy of fixing the forgetting factor and innovatively proposing a dynamic forgetting factor adaptive adjustment strategy. Its core technical idea lies in... The hyperparameter is transformed from a fixed parameter into a state variable that adaptively adjusts according to the inherent characteristics of the data. The adjustment mechanism follows the formula: This allows the dynamic forgetting factor to adaptively adjust with the residuals, better balancing the ability to track new data and model stability, and effectively suppressing data saturation and residual noise interference. Therefore, the iterative solution expression of the gradual fading memory recursive algorithm based on the dynamic forgetting factor strategy of this invention is: ; in, Indicates the number of measurement periods. Indicates the first k Forgetting factor per measurement period, This represents a hyperparameter, typically set to 0.15. express( k +1) The difference between the weighted average of the total energy consumption increment of the main meter and the sum of the weighted average of the energy consumption increments of all sub-meters within a metering period. express( k +1) The regression vector at time step (+1), superscript T Indicates transpose. express k The estimation results of the time parameter vector. express( k The gain vector at time +1) express k The covariance matrix at time t, express( k The covariance matrix at time +1), Represents the identity matrix. express( k+1) Estimation results of the parameter vector at time step. After obtaining the final parameter vector estimation results, the first element of the parameter vector estimation results is extracted as the fixed loss, the last element is the base line loss rate, and the middle element is... M Each element represents the relative error of each sub-meter. Based on this, the system can output the identification results of out-of-tolerance meters with errors exceeding the limits and the line loss analysis report for the transformer area.

[0036] Optionally, if the covariance matrix The condition number is greater than a preset threshold, i.e. , Represents the covariance matrix The condition number of Generally, 10 is taken. 9 Then it is necessary to utilize replace ,Right now , Represents the regularization coefficient. , This represents the maximum absolute value of the correlation coefficient between the energy increment sequences of any two sub-tables. Setting the minimum value to 0.1 can avoid multicollinearity interference, suppress the interference of sub-table incremental collinearity on parameter estimation, and improve the stability of joint estimation. Through the above processing, when When the value is too large, it means that the matrix is ​​ill-conditioned. In this case, regularization can be used to add a positive value to all feature directions, which can raise the small eigenvalues, thereby reducing the condition number, avoiding the matrix from approaching singularity, and helping to further improve the estimation accuracy.

[0037] In addition, such as Figure 4 As shown, another embodiment of the present invention also provides an evaluation system for smart meter errors and transformer area line losses, preferably employing the smart meter error and transformer area line loss evaluation method described above, including: The data preprocessing module is used to collect the electricity metering data from the master table and each sub-table to obtain the electricity metering dataset. From the electricity metering dataset, multiple effective incremental blocks that maintain physical continuity in the data are extracted, and outlier cleaning is performed on the data of each effective incremental block. The data of each effective incremental block includes the electricity increment sequence of the master table and each sub-table. The model building module is used to build a dynamic line loss model that varies with load based on the principle of energy conservation and the incremental energy data of the main meter and each sub-meter. The iterative solution module is used to solve the dynamic line loss model by using each effective incremental block as an independent iterative unit and employing a gradual memory recursive algorithm based on a dynamic forgetting factor strategy. This yields the line loss of the transformer area and the relative error of each sub-meter within the time period corresponding to each effective incremental block.

[0038] It is understood that the smart meter error and transformer area line loss assessment system in this embodiment constructs a dynamic line loss model that varies with load. Compared to existing fixed line loss assumptions, this model considers the dynamic correlation between line loss and power supply in real-world scenarios, allowing line loss to dynamically adjust with the actual load, thus improving estimation accuracy and adaptability. Furthermore, the dynamic line loss model enables joint estimation of line loss parameters and meter errors. Moreover, it employs a fading memory recursive algorithm based on a dynamic forgetting factor strategy for solving the problem. Compared to a fixed forgetting factor strategy, this effectively suppresses data saturation and noise interference. It also constructs effective incremental blocks that maintain physical continuity in the data, using each effective incremental block as an independent iterative unit for solving the problem. This provides a high-quality data foundation for the solution process, effectively adapting to noise, missing values, and abnormal operating conditions such as equipment restarts in the measurement data, thereby improving the robustness and reliability of the estimation results.

[0039] In addition, another embodiment of the present invention provides an electronic device including a processor and a memory, wherein the memory stores a computer program, and the processor executes the steps of the method described above by calling the computer program stored in the memory.

[0040] In addition, another embodiment of the present invention provides a computer-readable storage medium for storing a computer program for evaluating smart meter errors and transformer area line losses, wherein the computer program performs the steps of the method described above when run on a computer.

[0041] Common computer-readable storage media include: floppy disks, flexible disks, hard disks, magnetic tapes, any other magnetic media, CD-ROMs, any other optical media, punch cards, paper tape, any other physical media with perforated patterns, random access memory (RAM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), flash erasable programmable read-only memory (FLASH-EPROM), any other memory chips or cartridges, or any other media readable by a computer. Instructions may further be transmitted or received by a transmission medium. The term transmission medium can include any tangible or intangible medium used to store, encode, or carry instructions for execution by a machine, and includes digital or analog carrier communication signals or intangible media that facilitate communication of such instructions. Transmission media include coaxial cables, copper wires, and optical fibers, which contain conductors for transmitting a bus of computer data signals.

[0042] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of this application can be implemented in various computer languages, such as the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0043] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0044] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0045] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0046] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0047] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

[0048] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for evaluating the error of a smart meter and the line loss of a distribution area, characterized in that, Includes the following: Collect electricity metering data from the master table and each sub-table to obtain an electricity metering dataset. Extract multiple valid incremental blocks that maintain physical continuity in the data from the electricity metering dataset. Perform outlier cleaning on the data of each valid incremental block. The data of each valid incremental block includes the electricity increment sequence of the master table and each sub-table. Based on the principle of energy conservation and the incremental energy data of the main meter and each sub-meter, a dynamic line loss model is constructed to reflect the changes in line loss with load. Using each effective increment block as an independent iteration unit, the dynamic line loss model is solved by a gradual memory recursion algorithm based on a dynamic forgetting factor strategy, which yields the line loss of the transformer area and the relative error of each sub-meter within the time period corresponding to each effective increment block.

2. The method for evaluating smart meter errors and distribution line losses as described in claim 1, characterized in that, The process of constructing a dynamic line loss model that varies with load based on the principle of energy conservation and the incremental energy data of the master meter and each sub-meter includes the following: Each effective incremental block is divided into multiple metering cycles, and a power balance model for each metering cycle is constructed based on the principle of energy conservation. The power balance model of the distribution area is calculated by weighted average based on multiple metering periods to obtain the power balance model of the distribution area with weighted average for multiple metering periods. Based on the direct proportional relationship between line loss rate and power supply, the power weighted average balance model of the transformer area is transformed into a dynamic line loss model in the form of linear regression.

3. The method for evaluating smart meter errors and distribution line losses as described in claim 2, characterized in that, The linear regression form of the dynamic line loss model is as follows: ； in, This represents the difference between the weighted average of the total energy consumption increment from the main meter and the sum of the weighted average of the energy consumption increments from all sub-meters over multiple metering periods. This represents the weighted average of fixed losses. Indicates the number of table partitions. Indicates the first j Only the weighted average of the electricity increments from the individual meters. Indicates the first j Only the relative error weighted average of the table is used. This represents the weighted average of the total electricity consumption increment. Indicates the base line loss rate. This represents the line loss coefficient, used to characterize the degree of fluctuation in the current period load compared to the average load.

4. The method for evaluating smart meter errors and distribution line losses as described in claim 2, characterized in that, The expression for the gradual memory fading recursion algorithm based on the dynamic forgetting factor strategy is as follows: ； in, Indicates the number of measurement periods. Indicates the first k Forgetting factor per measurement period, Indicates hyperparameters, express( k +1) The difference between the weighted average of the total energy consumption increment of the main meter and the sum of the weighted average of the energy consumption increments of all sub-meters within a metering period. express( k +1) The regression vector at time step (+1), superscript T Indicates transpose. express k The estimation results of the time parameter vector. express( k The gain vector at time +1) express k The covariance matrix at time t, express( k The covariance matrix at time +1), Represents the identity matrix. express( k +1) The estimation result of the parameter vector at time step.

5. The method for evaluating smart meter errors and distribution line losses as described in claim 4, characterized in that, If the covariance matrix If the condition number is greater than a preset threshold, then use replace , Represents the regularization coefficient. , This represents the maximum absolute value of the correlation coefficient between the electrical energy increment sequences of any two sub-tables.

6. The method for evaluating smart meter errors and distribution line losses as described in claim 1, characterized in that, The process of extracting multiple effective incremental blocks that maintain physical continuity in the data from the electricity metering dataset includes the following: Detect reset points in the electricity metering dataset; where a reset point represents the moment when the meter restarts or the reading changes. The time axis is divided based on the reset point, and the electricity metering dataset is divided into multiple data blocks that maintain physical continuity of the data, and data blocks with a data size less than the sample size threshold are removed. The logical zero point of each data block is identified, and the data in each data block that is before the logical zero point is removed. The power increment sequence of the master table and each sub-table is calculated based on the logical zero point, thereby transforming each data block into an effective increment block. The logical zero point represents the load stabilization moment of the unified increment benchmark.

7. The method for evaluating smart meter errors and distribution line losses as described in claim 1, characterized in that, The process of cleaning outliers from the data in each valid incremental block includes the following: For each power increment sequence within the effective increment block, first calculate the 25th percentile, 75th percentile, and interquartile range of the power increment data to set the normal data range, mark the normal data samples, then calculate the mean and standard deviation of the normal data samples, and standardize each normal data sample based on the mean and standard deviation, retaining only the data samples whose standardized values ​​are less than the preset threshold.

8. A system for evaluating the error of smart meters and the line loss of distribution transformers, characterized in that, include: The data preprocessing module is used to collect the electricity metering data from the master table and each sub-table to obtain the electricity metering dataset. From the electricity metering dataset, multiple effective incremental blocks that maintain physical continuity in the data are extracted, and outlier cleaning is performed on the data of each effective incremental block. The data of each effective incremental block includes the electricity increment sequence of the master table and each sub-table. The model building module is used to build a dynamic line loss model that varies with load based on the principle of energy conservation and the incremental energy data of the main meter and each sub-meter. The iterative solution module is used to solve the dynamic line loss model by using each effective incremental block as an independent iterative unit and employing a gradual memory recursive algorithm based on a dynamic forgetting factor strategy. This yields the line loss of the transformer area and the relative error of each sub-meter within the time period corresponding to each effective incremental block.

9. An electronic device, characterized in that, The method includes a processor and a memory, wherein the memory stores a computer program, and the processor executes the steps of the method as described in any one of claims 1 to 7 by calling the computer program stored in the memory.

10. A computer-readable storage medium for storing a computer program for evaluating smart meter errors and transformer area line losses, characterized in that, The computer program, when run on a computer, performs the steps of the method as described in any one of claims 1 to 7.

Images