Low-voltage area line impedance accurate calculation method and system

Abstract

This invention discloses a method and system for accurate calculation of line impedance in low-voltage distribution areas. The method includes the following steps: acquiring user signal data from the low-voltage distribution area and generating a standardized dataset; extracting user characteristics and performing cluster analysis on users within the area; calculating the Pearson correlation coefficient between the voltage curve of abnormal users and the bus voltage curve of adjacent distribution areas, and performing triple verification by combining the topological distance between the user and the distribution area and the matching degree of the meter identification signal; determining the true user affiliation based on the verification results; constructing a lumped parameter equivalent model for single-branch lines and a distributed parameter equivalent model for multi-branch lines; establishing a hierarchical model of line impedance in the distribution network of the area; and inverting to obtain the resistance and reactance parameters of each branch line. Using this technical solution, abnormal users can be automatically identified, the distribution area topology can be accurately constructed, and high-precision calculation of complex line impedance can be achieved, providing a reliable basis for distribution network loss analysis, fault location, and operation optimization.

Description

Technical Field

[0001] This invention belongs to the field of power system technology and relates to a method and system for accurate calculation of line impedance in low-voltage distribution areas. Background Technology

[0002] As the terminal power supply unit of the power system, the accurate calculation of line impedance parameters of low-voltage distribution substations is of great significance for line loss analysis, fault location, operational status assessment, and lean management. Traditional line impedance calculation methods mostly rely on manual measurement or simplified models, which suffer from problems such as computational complexity, poor adaptability, and low accuracy. Especially when the relationship between customers and transformers is unclear and the topology is complex, traditional methods are difficult to achieve accurate impedance calculation.

[0003] With the widespread adoption of smart meters and electricity consumption data collection systems, massive amounts of electricity data have made data-driven impedance calculations possible. However, existing methods still suffer from slow convergence speed and susceptibility to local optima when dealing with nonlinear, multi-branch circuit models. Summary of the Invention

[0004] The purpose of this invention is to address the aforementioned problems in existing technologies by proposing a method and system for accurate calculation of low-voltage distribution area line impedance.

[0005] To achieve the above objectives, the basic solution of this invention is: a method for accurately calculating the line impedance of low-voltage distribution areas, comprising the following steps:

[0006] S1: Obtain user profile information, electricity consumption data, and meter identification signal data of the low-voltage distribution transformer area, and perform cleaning and time-series alignment to generate a standardized dataset;

[0007] S2, extract user voltage curves, load curves, and electricity consumption period characteristics, and use weighted K-means clustering algorithm to perform cluster analysis on users within the transformer area;

[0008] The baseline cluster is dynamically updated based on a sliding window mechanism to identify abnormal users who deviate from the baseline cluster for three consecutive periods.

[0009] S3, calculate the Pearson correlation coefficient between the abnormal user voltage curve and the bus voltage curve of the adjacent transformer area, and perform triple verification by combining the topological distance between the user and the transformer area and the matching degree of the meter identification signal;

[0010] Determine the user's true affiliation based on the verification results, and construct and visualize the topological relationships of the transformer area;

[0011] S4. Based on the topology of the distribution area, construct a lumped parameter equivalent model for single-branch lines and a distributed parameter equivalent model for multi-branch lines, and establish a bottom-up hierarchical model of the line impedance of the distribution network in the distribution area.

[0012] S5 uses an improved Newton-Raphson algorithm to adaptively solve the impedance model and invert the resistance and reactance parameters of each branch line.

[0013] The working principle and beneficial effects of this basic solution are as follows: This technical solution accurately obtains the line impedance of low-voltage distribution areas through data acquisition, abnormal user identification, topology verification, impedance modeling and inversion, realizing the automation and intelligence of impedance calculation.

[0014] By combining cluster analysis with a sliding window mechanism to identify anomalous users, the accuracy of topology construction can be improved when the user-transformer relationship is unclear. A triple verification mechanism (voltage correlation, topological distance, and signal matching) is employed to comprehensively determine the true user affiliation, enhancing the reliability of anomalous user identification. An improved Newton-Raphson algorithm is used for impedance inversion, improving convergence speed and stability and avoiding local optima.

[0015] Furthermore, the user profile information, electricity consumption data, and meter identification signal data of the low-voltage distribution substation are cleaned and time-series aligned to generate a standardized dataset.

[0016] The static attribute set of the user is obtained from the marketing system, including user ID, meter asset number, installation location, transformer ownership information, phase attribute (A / B / C phase), and rated capacity. ;

[0017] The electricity consumption information collection system collects time-series data from each user's electricity meter at fixed intervals, including voltage. Current Active power reactive power This constitutes an electricity consumption time series dataset. ;

[0018] The communication characteristic signals between the meter and the concentrator / transformer, including the signal-to-noise ratio, are acquired in real time using a high-speed power line carrier (HPLC) or low-power wireless communication module. Signal strength Communication success rate This constitutes the identification signal dataset. ;

[0019] If electricity consumption data is missing in time period t, cubic spline interpolation is used to complete it:

[0020] ,

[0021] in, This is the set of measured voltage values ​​for user i at several valid times before and after the missing time t. It is a cubic spline interpolation function. The interpolated voltage estimate for user i at the missing time t;

[0022] If consecutive missing data exceeds a preset threshold, it is marked as an abnormal data segment;

[0023] Outlier identification is based on box plot method, with upper and lower boundaries as follows:

[0024] Lower boundary: ,

[0025] Upper boundary: ,in, The interquartile range is used; values ​​exceeding the boundary are considered outliers and replaced with the mean of adjacent time periods. It is the third quartile, which is the value at the 75th percentile of the data. This is the first quartile, which is the value at the 25th percentile of the data.

[0026] Wavelet thresholding is used to denoise time-series data such as voltage and current. Since the frequencies of electricity consumption data (15-minute granularity) and signal data (1-minute granularity) differ, timestamp alignment is required. After alignment, the feature vector of each user i at time t is... for:

[0027] ,

[0028] To eliminate the influence of dimensions, numerical features are standardized using Z-score:

[0029] ,

[0030] in, These are the mean and standard deviation of feature f across all users and all time periods, respectively. This represents the value of user i at time t on feature f (such as voltage, current, etc.); This represents the standardized value of user i on feature f at time t;

[0031] The final result is a standardized dataset that can be used for subsequent clustering and impedance calculations.

[0032] ,

[0033] in, To align the unified time points, For user static attributes, This is the aligned and standardized temporal feature vector.

[0034] This paper proposes interpolation methods and outlier identification mechanisms for missing data to improve data integrity and reliability. It clarifies the acquisition and cleaning methods for multi-source data (archives, electricity consumption data, and signal data), providing high-quality standardized datasets for subsequent analysis. Through time-series alignment and standardization, it eliminates temporal and dimensional differences between different data sources, facilitating subsequent feature extraction and modeling.

[0035] Furthermore, user voltage curves, load curves, and electricity consumption period characteristics are extracted, specifically:

[0036] Voltage curve characteristics include:

[0037] The average voltage of user i;

[0038] : Standard deviation of voltage for user i;

[0039] : The voltage peak-to-valley difference for user i;

[0040] in, , These are the lowest and highest voltage values ​​collected by user i within this period, respectively.

[0041] The characteristics of the load curve include:

[0042] The daily load factor of user i reflects the stability of the load.

[0043] The peak-to-valley ratio for user i reflects the magnitude of load fluctuations;

[0044] in, , , These represent the maximum active power, minimum active power, and average active power of user i within the statistical period, respectively.

[0045] Characteristics of electricity consumption periods include:

[0046] The percentage of electricity consumption by user i during peak hours;

[0047] The percentage of electricity consumption by user i during off-peak hours;

[0048] in, These are the sets of peak and off-peak periods defined by the system. : Active power of user i at time t.

[0049] By extracting three types of features—voltage, load, and electricity consumption period—a comprehensive characterization of user electricity consumption behavior is achieved, providing multidimensional evidence for cluster analysis and facilitating the interpretation of cluster results and the causes of anomalies.

[0050] Furthermore, the steps for clustering users within the transformer area using the weighted K-means clustering algorithm are as follows:

[0051] K initial cluster centers are randomly selected. ;

[0052] Define the weighted distance between user i and the center of cluster k. :

[0053] ,

[0054] in, Let j be the eigenvalue of the center of cluster k. Let j be the j-th feature value of user i. This represents the weight of the j-th feature. .

[0055] Weighted K-means clustering can be used to assign different weights based on the importance of features, thereby improving the sensitivity of clustering results to key features (such as voltage).

[0056] Furthermore, based on the sliding window mechanism, the steps for dynamically updating the baseline cluster and identifying abnormal users who deviate from the baseline cluster for three consecutive periods are as follows:

[0057] Assign users to the nearest cluster:

[0058] ,

[0059] in, Let be the center vector of cluster k′ in the t-th iteration. Let be the set of users belonging to cluster k in the t-th iteration;

[0060] Update cluster centers:

[0061] ,

[0062] in, Let i be the feature vector of user i. The number of users in cluster k;

[0063] Repeatedly assign users to the nearest cluster and update cluster centers until the change in cluster centers is less than the threshold ε;

[0064] Suppose the sliding window length is L periods, and clustering is performed once in each period (e.g., one day) to obtain the baseline cluster set for that period. ;

[0065] If a cluster has a stable user composition over M consecutive periods, it is marked as a stable baseline cluster.

[0066] If user i does not belong to any baseline cluster for 33 consecutive periods, that is:

[0067] ,

[0068] Then mark it as an abnormal user;

[0069] in, For user i in the first Eigenvectors of each period The abnormal distance threshold can be set to three standard deviations of the historical normal user distance to the cluster center. For periodic indexes, For the first The center of cluster k in each period.

[0070] A sliding window mechanism is introduced to dynamically update the baseline cluster, adapting to the time-varying nature of distribution area load and user behavior. Abnormal users are identified by continuous periodic deviations, avoiding false positives caused by single fluctuations and improving the stability of anomaly detection.

[0071] Furthermore, the Pearson correlation coefficient between the abnormal user voltage curve and the bus voltage curve of the adjacent transformer area is calculated. Triple verification is then performed, combining the topological distance between the user and the transformer area, and the meter identification signal matching degree. The specific steps are as follows:

[0072] Calculate the user voltage curve Voltage curves of adjacent transformer substations and busbars Pearson correlation coefficient:

[0073] ,

[0074] like If so, it is considered that the voltage correlation is strong;

[0075] in, This represents the average voltage of busbar m in the transformer substation area. The average voltage of user i, Let be the voltage of busbar m in transformer substation at time t. Let i be the voltage of user i at time t. The correlation coefficient between user i and the bus voltage of transformer area m; T is the time series length;

[0076] Based on the GIS system, the line topology distance from user i to each transformer m is obtained. ,like:

[0077] If so, the topological distance is considered reasonable;

[0078] in, This is the topological distance threshold;

[0079] Extract the signal characteristics of user electricity meters, match them with the signal characteristics of concentrators in each distribution area, and calculate the signal matching degree. :

[0080] ,

[0081] in, , These are the weighting coefficients; These are the maximum signal strength and signal-to-noise ratio reference values ​​set by the system, respectively. , These are the signal-to-noise ratio and signal strength indication of the signal received by the user's electricity meter, respectively; if If so, the signal is considered matched. This is the preset signal matching threshold.

[0082] The triple verification mechanism integrates multiple dimensions of information, including electrical, topological, and communication data, to improve the comprehensiveness and accuracy of user attribution determination.

[0083] Furthermore, based on the verification results, the steps to determine the user's true affiliation and construct and visualize the transformer area topology are as follows:

[0084] Based on the combined results of the triple validation, the attribution score is defined as follows:

[0085] ,

[0086] in, The overall score for user i belonging to the district m. As weight, satisfying ; This represents the maximum topological distance allowed by the system.

[0087] Judgment rules:

[0088] ,

[0089] If the maximum score is higher than the threshold If yes, the attribution is determined; otherwise, it is marked as "pending on-site verification".

[0090] By using a comprehensive attribution score formula, the correlation strength between users and transformer stations can be quantified, making it easier to set thresholds for automated judgment.

[0091] Furthermore, based on the transformer substation topology, a lumped parameter equivalent model is constructed for single-branch lines, and a distributed parameter equivalent model is constructed for multi-branch lines. A bottom-up hierarchical model of the substation distribution network's line impedance is established, specifically as follows:

[0092] For a single-branch line (such as transformer → concentrator → user group), establish a π-type equivalent circuit:

[0093] ,

[0094] Where R is the line resistance, X is the line reactance, G and B represent the parallel admittance, and j is the characteristic value number; For parallel admittance of the line, This refers to the series impedance of the line.

[0095] For complex multi-branch lines, a multi-segment distributed parameter model is adopted:

[0096] ,

[0097] in, The unit length admittance matrix, The impedance matrix is ​​a unit length matrix. Let n be the reactance per unit length of the nth line segment. Let n be the resistance per unit length of the nth segment of the line, where n represents the total number of segments in the line. Let n be the conductance per unit length of the nth line segment. Let be the susceptance per unit length of the nth line segment; Let be the complex admittance per unit length of the nth line segment. Let be the complex impedance per unit length of the nth line segment; Let x be the voltage vector and current vector along the line at position x, where x represents the distance variable along the line.

[0098] Aggregation occurs hierarchically from the user's perspective upwards:

[0099] User layer: Each user is considered a load node;

[0100] Branch layer: Users under the same branch are merged into an equivalent load;

[0101] Main trunk layer: all branches converge to the main trunk line;

[0102] Transformer layer: The main line connects to the transformer;

[0103] Each layer establishes a corresponding impedance-admittance parameter matrix, forming a layered parameter set:

[0104] ,

[0105] Where L is the total number of layers. Let l be the admittance matrix of the l-th layer. Let be the impedance matrix of the l-th layer.

[0106] Differentiated modeling strategies are adopted for different line types (single-branch / multi-branch) to balance computational complexity and model accuracy. Bottom-up hierarchical modeling conforms to the actual structure of the distribution network and facilitates step-by-step parameter aggregation and verification.

[0107] Furthermore, an improved Newton-Raphson algorithm is used to adaptively solve the impedance model, and the resistance and reactance parameters of each branch line are obtained through inversion. The specific steps are as follows:

[0108] Establish an impedance inversion optimization problem:

[0109] ,

[0110] in, Calculate the voltage vector for the model. represents the measured voltage vector (at all nodes); J represents the objective function, and T is the total length of the time series.

[0111] Improved Newton-Raphson iteration:

[0112] ,

[0113] in, For the residual vector, ; Given the Jacobian matrix, calculate the partial derivatives of voltage with respect to impedance / admittance; For Jacobian matrices, These are the impedance and admittance increments for the k-th iteration, respectively.

[0114] like Then reduce the step size:

[0115] ,

[0116] in, η is the step size factor for the k-th iteration; η is the step size decay coefficient;

[0117] Impedance smoothing constraints are added to prevent overfitting.

[0118] ,

[0119] Where λ is the regularization coefficient. The prior impedance value for the line type. Let the regularization objective function be used.

[0120] Iteration termination condition:

[0121] ,

[0122] in, The residual threshold, The threshold for relative impedance change; This refers to the impedance vector or impedance matrix obtained in the k-th iteration.

[0123] Output branch resistors Reactance Line impedance angle , impedance model confidence level.

[0124] An impedance inversion optimization problem is established, transforming impedance calculation into a numerical solution problem, enabling data-driven high-precision parameter identification. An improved Newton-Raphson algorithm introduces adaptive step size adjustment and regularization constraints to enhance convergence performance and prevent overfitting.

[0125] The present invention also provides a system for accurate calculation of line impedance in low-voltage distribution areas, including a data acquisition unit and a processing unit. The data acquisition unit is used to acquire user profile information and power consumption data of low-voltage distribution areas and transmit them to the processing unit.

[0126] The processing unit executes the method described in this invention to accurately calculate the line impedance of the low-voltage distribution area.

[0127] This system has a clear structure, accurately calculates the line impedance of low-voltage distribution areas, is easy to integrate into existing distribution network management platforms, and supports large-scale distribution area impedance calculation tasks. Attached Figure Description

[0128] Figure 1 This is a flowchart illustrating the method for accurate calculation of low-voltage distribution area line impedance according to the present invention. Detailed Implementation

[0129] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0130] In the description of this invention, it should be understood that the terms "longitudinal", "lateral", "up", "down", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on this invention.

[0131] In the description of this invention, unless otherwise specified and limited, it should be noted that the terms "installation", "connection" and "linking" should be interpreted broadly. For example, they can refer to mechanical or electrical connections, or internal connections between two components. They can be direct connections or indirect connections through an intermediate medium. Those skilled in the art can understand the specific meaning of the above terms according to the specific circumstances.

[0132] This invention discloses a method for accurate calculation of line impedance in low-voltage distribution areas. The entire process of data acquisition, processing, and analysis is automated, requiring no manual intervention. It obtains accurate impedance parameters that can be used for line loss calculation, providing a reliable basis for loss reduction optimization. Figure 1 As shown, the method for accurately calculating the line impedance of a low-voltage distribution area includes the following steps:

[0133] S1: Obtain user profile information, electricity consumption data, and meter identification signal data of the low-voltage distribution transformer area, and perform cleaning and time-series alignment to generate a standardized dataset;

[0134] S2, extract user voltage curves, load curves, and electricity consumption period characteristics, and use weighted K-means clustering algorithm to perform cluster analysis on users within the transformer area;

[0135] The baseline cluster is dynamically updated based on a sliding window mechanism to identify abnormal users who deviate from the baseline cluster for three consecutive periods.

[0136] S3, calculate the Pearson correlation coefficient between the abnormal user voltage curve and the bus voltage curve of the adjacent transformer area, and perform triple verification by combining the topological distance between the user and the transformer area and the matching degree of the meter identification signal;

[0137] Determine the user's true affiliation based on the verification results, and construct and visualize the topological relationships of the transformer area;

[0138] S4. Based on the topology of the distribution area, construct a lumped parameter equivalent model for single-branch lines and a distributed parameter equivalent model for multi-branch lines, and establish a bottom-up hierarchical model of the line impedance of the distribution network in the distribution area.

[0139] S5 uses an improved Newton-Raphson algorithm to adaptively solve the impedance model, and inversely obtains the resistance and reactance parameters of each branch line. It can be used for practical applications such as line loss analysis, fault location, and network reconfiguration to improve the economic operation level of the distribution network.

[0140] In a preferred embodiment of the present invention, the method for generating a standardized dataset by cleaning and time-series alignment of user profile information, electricity consumption data, and meter identification signal data of a low-voltage distribution substation is as follows:

[0141] The static attribute set of the user is obtained from the marketing system, including user ID, meter asset number, installation location, transformer ownership information, phase attribute (A / B / C phase), and rated capacity. ;

[0142] The electricity consumption information collection system collects time-series data from each user's electricity meter at fixed intervals, including voltage. Current Active power reactive power This constitutes an electricity consumption time series dataset. ;

[0143] The communication characteristic signals between the meter and the concentrator / transformer, including the signal-to-noise ratio, are acquired in real time using a high-speed power line carrier (HPLC) or low-power wireless communication module. Signal strength Communication success rate This constitutes the identification signal dataset. ;

[0144] If electricity consumption data is missing in time period t, cubic spline interpolation is used to complete it:

[0145] ,

[0146] in, This is the set of measured voltage values ​​for user i at several valid times before and after the missing time t. It is a cubic spline interpolation function. The interpolated voltage estimate for user i at the missing time t;

[0147] If consecutive missing data exceeds a preset threshold, it is marked as an abnormal data segment;

[0148] Outlier identification is based on box plot method, with upper and lower boundaries as follows:

[0149] Lower boundary: ,

[0150] Upper boundary: ,

[0151] in, The interquartile range is used; values ​​exceeding the boundary are considered outliers and replaced with the mean of adjacent time periods. It is the third quartile, which is the value at the 75th percentile of the data. This is the first quartile, which is the value at the 25th percentile of the data.

[0152] Wavelet thresholding is used to denoise time-series data such as voltage and current. Since the frequencies of electricity consumption data (15-minute granularity) and signal data (1-minute granularity) differ, timestamp alignment is required. After alignment, the feature vector of each user i at time t is... for:

[0153] ,

[0154] To eliminate the influence of dimensions, numerical features are standardized using Z-score:

[0155] ,

[0156] in, These are the mean and standard deviation of feature f across all users and all time periods, respectively. This represents the value of user i at time t on feature f (such as voltage, current, etc.); This represents the standardized value of user i on feature f at time t;

[0157] The final result is a standardized dataset that can be used for subsequent clustering and impedance calculations.

[0158] ,

[0159] in, To align the unified time points, For user static attributes, This is the aligned and standardized temporal feature vector.

[0160] In a preferred embodiment of the present invention, the extraction of user voltage curves, load curves, and electricity consumption time characteristics specifically includes:

[0161] Voltage curve characteristics include:

[0162] The average voltage of user i;

[0163] : Standard deviation of voltage for user i;

[0164] : The voltage peak-to-valley difference for user i;

[0165] in, , These are the lowest and highest voltage values ​​collected by user i within this period, respectively.

[0166] The characteristics of the load curve include:

[0167] The daily load factor of user i reflects the stability of the load.

[0168] The peak-to-valley ratio for user i reflects the magnitude of load fluctuations;

[0169] in, , , These represent the maximum active power, minimum active power, and average active power of user i within the statistical period, respectively.

[0170] Characteristics of electricity consumption periods include:

[0171] The percentage of electricity consumption by user i during peak hours;

[0172] The percentage of electricity consumption by user i during off-peak hours;

[0173] in, These are the sets of peak and off-peak periods defined by the system. : Active power of user i at time t.

[0174] In a preferred embodiment of the present invention, the step of using the weighted K-means clustering algorithm to perform cluster analysis on users within the transformer area is as follows:

[0175] K initial cluster centers are randomly selected. ;

[0176] Define the weighted distance between user i and the center of cluster k. :

[0177] ,

[0178] in, Let j be the eigenvalue of the center of cluster k. Let j be the j-th feature value of user i. This represents the weight of the j-th feature. .

[0179] In a preferred embodiment of the present invention, the steps for dynamically updating the reference cluster based on the sliding window mechanism and identifying abnormal users who deviate from the reference cluster for three consecutive periods are as follows:

[0180] Assign users to the nearest cluster:

[0181] ,

[0182] in, Let be the center vector of cluster k′ in the t-th iteration. Let be the set of users belonging to cluster k in the t-th iteration;

[0183] Update cluster centers:

[0184] ,

[0185] in, Let i be the feature vector of user i. The number of users in cluster k;

[0186] Repeatedly assign users to the nearest cluster and update cluster centers until the change in cluster centers is less than the threshold ε;

[0187] Suppose the sliding window length is L periods, and clustering is performed once in each period (e.g., one day) to obtain the baseline cluster set for that period. ;

[0188] If a cluster has a stable user composition over M consecutive periods, it is marked as a stable baseline cluster.

[0189] If user i does not belong to any baseline cluster for 33 consecutive periods, that is:

[0190] ,

[0191] Then mark it as an abnormal user;

[0192] in, For user i in the first Eigenvectors of each period The abnormal distance threshold can be set to three standard deviations of the historical normal user distance to the cluster center. For periodic indexes, For the first The center of cluster k in each period.

[0193] In a preferred embodiment of the present invention, the Pearson correlation coefficient between the voltage curve of the abnormal user and the bus voltage curve of the adjacent transformer area is calculated, and triple verification is performed by combining the topological distance between the user and the transformer area and the matching degree of the meter identification signal. The specific steps are as follows:

[0194] Calculate the user voltage curve Voltage curves of adjacent transformer substations and busbars Pearson correlation coefficient:

[0195] ,

[0196] like If so, it is considered that the voltage correlation is strong;

[0197] in, This represents the average voltage of busbar m in the transformer substation area. The average voltage of user i, Let be the voltage of busbar m in transformer substation at time t. Let i be the voltage of user i at time t. The correlation coefficient between user i and the bus voltage of transformer area m; T is the time series length;

[0198] Based on the GIS system, the line topology distance from user i to each transformer m is obtained. ,like:

[0199] If so, the topological distance is considered reasonable;

[0200] in, This is the topological distance threshold;

[0201] Extract the signal characteristics of user electricity meters, match them with the signal characteristics of concentrators in each distribution area, and calculate the signal matching degree. :

[0202] ,

[0203] in, , These are the weighting coefficients; These are the maximum signal strength and signal-to-noise ratio reference values ​​set by the system, respectively. , These are the signal-to-noise ratio and signal strength indication of the signal received by the user's electricity meter, respectively; if If so, the signal is considered matched. This is the preset signal matching threshold.

[0204] In a preferred embodiment of the present invention, the steps of determining the user's true affiliation based on the verification results and constructing and visualizing the transformer area topology are as follows:

[0205] Based on the combined results of the triple validation, the attribution score is defined as follows:

[0206] ,

[0207] in, The overall score for user i belonging to the district m. As weight, satisfying ; This represents the maximum topological distance allowed by the system.

[0208] Judgment rules:

[0209] ,

[0210] If the maximum score is higher than the threshold If yes, the attribution is determined; otherwise, it is marked as "pending on-site verification".

[0211] In a preferred embodiment of the present invention, based on the transformer substation topology, a lumped parameter equivalent model is constructed for single-branch lines, and a distributed parameter equivalent model is constructed for multi-branch lines. A bottom-up hierarchical model of the substation distribution network's line impedance is established, specifically as follows:

[0212] For a single-branch line (such as transformer → concentrator → user group), establish a π-type equivalent circuit:

[0213] ,

[0214] Where R is the line resistance, X is the line reactance, G and B represent the parallel admittance, and j is the characteristic value number; For parallel admittance of the line, This refers to the series impedance of the line.

[0215] For complex multi-branch lines, a multi-segment distributed parameter model is adopted:

[0216] ,

[0217] in, The unit length admittance matrix, The impedance matrix is ​​a unit length matrix. Let n be the reactance per unit length of the nth line segment. Let n be the resistance per unit length of the nth segment of the line, where n represents the total number of segments in the line. Let n be the conductance per unit length of the nth line segment. Let be the susceptance per unit length of the nth line segment; Let be the complex admittance per unit length of the nth line segment. Let be the complex impedance per unit length of the nth line segment; Let x be the voltage vector and current vector along the line at position x, where x represents the distance variable along the line.

[0218] Aggregation occurs hierarchically from the user's perspective upwards:

[0219] User layer: Each user is considered a load node;

[0220] Branch layer: Users under the same branch are merged into an equivalent load;

[0221] Main trunk layer: all branches converge to the main trunk line;

[0222] Transformer layer: The main line connects to the transformer;

[0223] Each layer establishes a corresponding impedance-admittance parameter matrix, forming a layered parameter set:

[0224] ,

[0225] Where L is the total number of layers. Let l be the admittance matrix of the l-th layer. Let be the impedance matrix of the l-th layer.

[0226] In a preferred embodiment of the present invention, an improved Newton-Raphson algorithm is used to adaptively solve the impedance model, and the resistance and reactance parameters of each branch line are obtained by inversion. The specific steps are as follows:

[0227] Establish an impedance inversion optimization problem:

[0228] ,

[0229] in, Calculate the voltage vector for the model. represents the measured voltage vector (at all nodes); J represents the objective function, and T is the total length of the time series.

[0230] Improved Newton-Raphson iteration:

[0231] ,

[0232] in, For the residual vector, ; Given the Jacobian matrix, calculate the partial derivatives of voltage with respect to impedance / admittance; For Jacobian matrices, These are the impedance and admittance increments for the k-th iteration, respectively.

[0233] like Then reduce the step size:

[0234] ,

[0235] in, η is the step size factor for the k-th iteration; η is the step size decay coefficient;

[0236] Impedance smoothing constraints are added to prevent overfitting.

[0237] ,

[0238] Where λ is the regularization coefficient. The prior impedance value for the line type. Let the regularization objective function be used.

[0239] Iteration termination condition:

[0240] ,

[0241] in, The residual threshold, The threshold for relative impedance change; This refers to the impedance vector or impedance matrix obtained in the k-th iteration.

[0242] Output branch resistors Reactance Line impedance angle , impedance model confidence level.

[0243] This invention also provides a system for accurately calculating the line impedance of a low-voltage distribution area, comprising a data acquisition unit and a processing unit. The data acquisition unit acquires user profile information and electricity consumption data of the low-voltage distribution area and transmits them to the processing unit. The processing unit executes the method described in this invention to accurately calculate the line impedance of the low-voltage distribution area.

[0244] Preferably, the data acquisition unit collects power consumption data and signal data in real time through a power line carrier and wireless communication module, and the processing unit is equipped with the algorithm described in this invention to automatically perform data cleaning, cluster analysis, topology construction, impedance modeling and inversion, and displays the topology diagram and impedance parameter table through a human-machine interface.

[0245] This system has a clear structure, accurately calculates the line impedance of low-voltage distribution areas, and is easy to integrate into existing distribution network management platforms, supporting large-scale distribution area impedance calculation tasks. The system supports seamless integration with existing marketing systems, GIS systems, and electricity consumption information collection platforms, facilitating large-scale deployment.

[0246] The specific embodiments described herein are merely illustrative examples of the present invention. Those skilled in the art can make various modifications or additions to the described embodiments or use similar methods to substitute them, without departing from the technology of the present invention or exceeding the scope defined by the appended claims.

Claims

1. A method for accurately calculating the line impedance of a low-voltage distribution area, characterized in that, Includes the following steps: S1: Obtain user profile information, electricity consumption data, and meter identification signal data of the low-voltage distribution transformer area, and perform cleaning and time-series alignment to generate a standardized dataset; S2, extract user voltage curves, load curves, and electricity consumption period characteristics, and use weighted K-means clustering algorithm to perform cluster analysis on users within the transformer area; The baseline cluster is dynamically updated based on a sliding window mechanism to identify abnormal users who deviate from the baseline cluster for three consecutive periods. S3, calculate the Pearson correlation coefficient between the abnormal user voltage curve and the bus voltage curve of the adjacent transformer area, and perform triple verification by combining the topological distance between the user and the transformer area and the matching degree of the meter identification signal; Determine the user's true affiliation based on the verification results, and construct and visualize the topological relationships of the transformer area; S4. Based on the topology of the distribution area, construct a lumped parameter equivalent model for single-branch lines and a distributed parameter equivalent model for multi-branch lines, and establish a bottom-up hierarchical model of the line impedance of the distribution network in the distribution area. S5 uses an improved Newton-Raphson algorithm to adaptively solve the impedance model and invert the resistance and reactance parameters of each branch line.

2. The method for accurate calculation of low-voltage distribution area line impedance according to claim 1, characterized in that, The method for generating a standardized dataset by cleaning and time-series alignment of user profile information, electricity consumption data, and meter identification signal data for low-voltage distribution transformer areas is as follows: The static attribute set of the user is obtained from the marketing system, including user ID, meter asset number, installation location, transformer ownership information, phase attribute (A / B / C phase), and rated capacity. ; The electricity consumption information collection system collects time-series data from each user's electricity meter at fixed intervals, including voltage. Current Active power reactive power This constitutes an electricity consumption time series dataset. ; The communication characteristic signals between the meter and the concentrator / transformer, including the signal-to-noise ratio, are acquired in real time using a high-speed power line carrier (HPLC) or low-power wireless communication module. Signal strength Communication success rate This constitutes the identification signal dataset. ; If electricity consumption data is missing in time period t, cubic spline interpolation is used to complete it: ， in, This is the set of measured voltage values ​​for user i at several valid times before and after the missing time t. It is a cubic spline interpolation function. The interpolated voltage estimate for user i at the missing time t; If consecutive missing data exceeds a preset threshold, it is marked as an abnormal data segment; Outlier identification is based on box plot method, with upper and lower boundaries as follows: Lower boundary: , Upper boundary: , in, The interquartile range is used; values ​​exceeding the boundary are considered outliers and replaced with the mean of adjacent time periods. It is the third quartile, which is the value at the 75th percentile of the data. This is the first quartile, which is the value at the 25th percentile of the data. Wavelet thresholding is used to denoise time-series data such as voltage and current. Since the frequency of power consumption data is different from that of signal data, timestamp alignment is required. After alignment, the feature vector of each user i at time t is obtained. for: ， To eliminate the influence of dimensions, numerical features are standardized using Z-score: ， in, These are the mean and standard deviation of feature f across all users and all time periods, respectively. This represents the value of user i at time t on feature f; This represents the standardized value of user i on feature f at time t; The final result is a standardized dataset that can be used for subsequent clustering and impedance calculations. ， in, To align the unified time points, For user static attributes, This is the aligned and standardized temporal feature vector.

3. The method for accurate calculation of low-voltage distribution area line impedance according to claim 1, characterized in that, Extract user voltage curves, load curves, and electricity consumption time characteristics, specifically: Voltage curve characteristics include: The average voltage of user i; : Standard deviation of voltage for user i; : The voltage peak-to-valley difference for user i; in, , These are the lowest and highest voltage values ​​collected by user i within this period, respectively. The characteristics of the load curve include: The daily load factor of user i reflects the stability of the load. The peak-to-valley ratio for user i reflects the magnitude of load fluctuations; in, , , These represent the maximum active power, minimum active power, and average active power of user i within the statistical period, respectively. Characteristics of electricity consumption periods include: The percentage of electricity consumption by user i during peak hours; The percentage of electricity consumption by user i during off-peak hours; in, These are the sets of peak and off-peak periods defined by the system. : Active power of user i at time t.

4. The method for accurate calculation of low-voltage distribution area line impedance according to claim 3, characterized in that, The steps for performing cluster analysis on users within a transformer area using the weighted K-means clustering algorithm are as follows: K initial cluster centers are randomly selected. ; Define the weighted distance between user i and the center of cluster k. : ， in, Let j be the eigenvalue of the center of cluster k. Let j be the j-th feature value of user i. This represents the weight of the j-th feature. .

5. The method for accurate calculation of low-voltage distribution area line impedance according to claim 4, characterized in that, The steps for identifying abnormal users who deviate from the benchmark cluster for three consecutive periods based on a sliding window mechanism for dynamically updating the benchmark cluster are as follows: Assign users to the nearest cluster: ， in, Let be the center vector of cluster k′ in the t-th iteration. Let be the set of users belonging to cluster k in the t-th iteration; Update cluster centers: ， in, Let i be the feature vector of user i. The number of users in cluster k; Repeatedly assign users to the nearest cluster and update cluster centers until the change in cluster centers is less than the threshold ε; Suppose the sliding window length is L periods, and clustering is performed once in each period to obtain the baseline cluster set for that period. ; If a cluster has a stable user composition over M consecutive periods, it is marked as a stable baseline cluster. If user i does not belong to any baseline cluster for 33 consecutive periods, that is: ， Then mark it as an abnormal user; in, For user i in the first Eigenvectors of each period The abnormal distance threshold can be set to three standard deviations of the historical normal user distance to the cluster center. For periodic indexes, For the first The center of cluster k in each period.

6. The method for accurate calculation of low-voltage distribution area line impedance according to claim 1, characterized in that, The Pearson correlation coefficient between the voltage curve of the abnormal user and the bus voltage curve of the adjacent transformer area is calculated. Triple verification is then performed, combining the topological distance between the user and the transformer area, and the matching degree of the meter identification signal. The specific steps are as follows: Calculate the user voltage curve Voltage curves of adjacent transformer substations and busbars Pearson correlation coefficient: ， like If so, it is considered that the voltage correlation is strong; in, This represents the average voltage of busbar m in the transformer substation area. The average voltage of user i, Let be the voltage of busbar m in transformer substation at time t. Let i be the voltage of user i at time t. The correlation coefficient between user i and the bus voltage of transformer area m; T is the time series length; Based on the GIS system, the line topology distance from user i to each transformer m is obtained. ,like: If so, the topological distance is considered reasonable; in, This is the topological distance threshold; Extract the signal characteristics of user electricity meters, match them with the signal characteristics of concentrators in each distribution area, and calculate the signal matching degree. : ， in, , These are the weighting coefficients; These are the maximum signal strength and signal-to-noise ratio reference values ​​set by the system, respectively. , These are the signal-to-noise ratio and signal strength indication of the signal received by the user's electricity meter, respectively; if If so, the signal is considered matched. This is the preset signal matching threshold.

7. The method for accurate calculation of low-voltage distribution area line impedance according to claim 6, characterized in that, The steps for determining the true user affiliation based on the verification results and constructing and visualizing the transformer area topology are as follows: Based on the combined results of the triple validation, the attribution score is defined as follows: ， in, The overall score for user i belonging to the district m. As weight, satisfying ; This represents the maximum topological distance allowed by the system. Judgment rules: ， If the maximum score is higher than the threshold If yes, the attribution is determined; otherwise, it is marked as "pending on-site verification".

8. The method for accurate calculation of low-voltage distribution area line impedance according to claim 1, characterized in that, Based on the transformer substation topology, a lumped parameter equivalent model is constructed for single-branch lines, and a distributed parameter equivalent model is constructed for multi-branch lines. A bottom-up hierarchical model of the substation distribution network's line impedance is then established, specifically as follows: For a single-branch circuit, construct a π-type equivalent circuit: ， Where R is the line resistance, X is the line reactance, G and B represent the parallel admittance, and j is the characteristic value number; For parallel admittance of the line, This refers to the series impedance of the line. For complex multi-branch lines, a multi-segment distributed parameter model is adopted: ， in, The unit length admittance matrix, The impedance matrix is ​​a unit length matrix. Let n be the reactance per unit length of the nth line segment. Let n be the resistance per unit length of the nth segment of the line, where n represents the total number of segments in the line. Let n be the conductance per unit length of the nth line segment. Let be the susceptance per unit length of the nth line segment; Let be the complex admittance per unit length of the nth line segment. Let be the complex impedance per unit length of the nth line segment; Let x be the voltage vector and current vector along the line at position x, where x represents the distance variable along the line. Aggregation occurs hierarchically from the user's perspective upwards: User layer: Each user is considered a load node; Branch layer: Users under the same branch are merged into an equivalent load; Main trunk layer: all branches converge to the main trunk line; Transformer layer: The main line connects to the transformer; Each layer establishes a corresponding impedance-admittance parameter matrix, forming a layered parameter set: ， Where L is the total number of layers. Let l be the admittance matrix of the l-th layer. Let be the impedance matrix of the l-th layer.

9. The method for accurate calculation of low-voltage distribution area line impedance according to claim 8, characterized in that, An improved Newton-Raphson algorithm is used to adaptively solve the impedance model, and the resistance and reactance parameters of each branch line are obtained by inversion. The specific steps are as follows: Establish an impedance inversion optimization problem: ， in, Calculate the voltage vector for the model. Here, J represents the measured voltage vector; J represents the objective function; and T is the total length of the time series. Improved Newton-Raphson iteration: ， in, For the residual vector, ; Given the Jacobian matrix, calculate the partial derivatives of voltage with respect to impedance / admittance; For Jacobian matrices, These are the impedance and admittance increments for the k-th iteration, respectively. like Then reduce the step size: ， in, η is the step size factor for the k-th iteration; η is the step size decay coefficient; Impedance smoothing constraints are added to prevent overfitting. ， Where λ is the regularization coefficient. The prior impedance value for the line type. Let the regularization objective function be used. Iteration termination condition: ， in, The residual threshold, The threshold for relative impedance change; This refers to the impedance vector or impedance matrix obtained in the k-th iteration. Output branch resistors Reactance Line impedance angle , impedance model confidence level.

10. A system for accurately calculating the line impedance of a low-voltage distribution area, characterized in that, It includes a data acquisition unit and a processing unit. The data acquisition unit is used to acquire user profile information and electricity consumption data of the low-voltage distribution transformer area and transmit them to the processing unit. The processing unit executes the method described in any one of claims 1-9 to accurately calculate the line impedance of the low-voltage distribution area.

Images