A method, device, equipment and storage medium for evaluating line loss in a distribution network

Abstract

The present invention discloses a distribution network line loss assessment method, apparatus, device, and storage medium. The method includes standardizing target distribution network data to obtain multidimensional standardized data; analyzing the multidimensional standardized data to identify abnormal data points, extracting features from the abnormal data points, and obtaining spatiotemporal characteristic factors; performing high-dimensional mapping on the multidimensional standardized data to obtain NTK similarity; weightedly fusing the spatiotemporal characteristic factors with the NTK similarity to obtain a final similarity; performing nearest neighbor screening on the abnormal data points to obtain nearest neighbor data; repairing the nearest neighbor data based on the final similarity to obtain a first data set; performing integrity assessment, consistency assessment, and accuracy assessment on the first data set to obtain a comprehensive assessment result; and calculating line losses on the first data set based on the comprehensive assessment result. The method of the present invention improves the accuracy of distribution network line loss calculation.

Description

Technical Field

[0001] The present invention relates to the technical field of line loss calculation, and in particular to a method, device, equipment and storage medium for evaluating line loss in a distribution network. Background Art

[0002] From power plants to terminal loads, electric energy goes through multiple links including power generation, transformation, transmission, distribution and consumption. Each link will cause power loss to varying degrees. Accurate line loss calculation and analysis is of great significance.

[0003] The distribution network is the primary source of power system losses. Existing methods for calculating line losses primarily rely on empirical methods such as statistical analysis, equivalent resistance, and root mean square current (RMS) current. These methods rely heavily on empirical parameters and typical daily load curves. Changes in the distribution network structure or operating mode require readjustment of these parameters, resulting in a lack of adaptability and difficulty ensuring accuracy.

[0004] Therefore, how to optimize the existing distribution network line loss calculation method to improve the accuracy of distribution network line loss calculation has become a technical problem that needs to be solved urgently by those skilled in the art. Summary of the Invention

[0005] The present invention provides a distribution network line loss assessment method, device, equipment and storage medium to solve the technical problem of how to optimize the existing distribution network line loss calculation method, thereby achieving the effect of improving the accuracy of distribution network line loss calculation.

[0006] In order to solve the above technical problems, an embodiment of the present invention provides a method for evaluating line loss in a distribution network, comprising:

[0007] Standardize the daily frozen power data, high-frequency current and voltage data, and equipment parameter data of the target distribution network to obtain multi-dimensional standardized data;

[0008] Analyzing the multidimensional standardized data based on a dynamic sliding window, identifying abnormal data points that meet a first preset condition, and extracting spatiotemporal features of the abnormal data points to obtain spatiotemporal feature factors;

[0009] Performing high-dimensional mapping on the multidimensional normalized data based on the neural tangent kernel function to calculate the NTK similarity; performing weighted fusion of the spatiotemporal feature factor and the NTK similarity to obtain the final similarity;

[0010] Performing neighbor screening on the abnormal data points based on an improved KNN algorithm to obtain neighbor data, wherein the neighbor screening is designed to adaptively adjust the number of neighbor screenings based on the standard deviation of the data distribution of the final similarity;

[0011] Repairing the neighboring data according to the final similarity to obtain a first data set, wherein the repairing is designed to determine an optimal filling path based on a Bayesian network;

[0012] Performing integrity assessment, consistency assessment, and accuracy assessment on the first data set in sequence to obtain a comprehensive assessment result;

[0013] A calculation path decision is made based on the comprehensive evaluation result, and a line loss calculation is performed on the first data set based on the decision result to obtain a line loss evaluation result of the target distribution network.

[0014] As one preferred solution, analyzing the multidimensional normalized data based on a dynamic sliding window to identify abnormal data points that meet a first preset condition includes:

[0015] Analyzing the multidimensional normalized data based on a sliding window of a preset window size, and calculating the coefficient of variation of the multidimensional normalized data within the preset window;

[0016] Dynamically adjusting the preset window size based on the coefficient of variation, and calculating the mean and standard deviation of the multidimensional normalized data within the adjusted window;

[0017] Dynamically calculate an abnormal threshold coefficient based on the load deviation of the multi-dimensional normalized data, and identify abnormal data points that meet a first preset condition based on the mean, the standard deviation, and the abnormal threshold coefficient; wherein the first preset condition is expressed as:

[0018]

[0019] in, For data to be identified, is the mean, is the standard deviation, is the abnormal threshold coefficient.

[0020] As one preferred solution, the spatiotemporal characteristic factors include a temporal correlation factor and a spatial topological correlation degree;

[0021] The extracting spatiotemporal features of the abnormal data points to obtain spatiotemporal feature factors includes:

[0022] The time correlation factor is calculated as:

[0023]

[0024] in, is the time correlation factor, Indicates the number of historical windows, represents the sampling interval, represents the Pearson correlation coefficient;

[0025] The calculation of the spatial topological association degree is expressed as:

[0026]

[0027] in, is the spatial topological correlation, is the distance between the electrical appliances, calculated based on the admittance matrix, is the minimum value.

[0028] As one of the preferred solutions, the improved KNN algorithm is used to perform neighbor screening on the abnormal data points to obtain neighbor data, including:

[0029] Dynamically calculating the number of nearest neighbors K based on the standard deviation of the distribution of the final similarity;

[0030] The abnormal data points are arranged in descending order according to the final similarity, and the first K data points are selected as the neighboring data.

[0031] As one preferred solution, repairing the neighboring data according to the final similarity to obtain a first data set includes:

[0032] Constructing a Bayesian network with the final similarity and data importance weight as parent nodes and the repair priority as child nodes, and quantifying the repair order using a conditional probability formula, wherein the data importance weight is determined according to the node type;

[0033] The neighboring data is repaired based on the repair order to obtain the first data set.

[0034] As one preferred solution, the completeness assessment, consistency assessment, and accuracy assessment are performed on the first data set in sequence to obtain a comprehensive assessment result, including:

[0035] Calculating the data integrity rate of the first data set to obtain an integrity assessment result;

[0036] Performing a power balance check, a current-power consistency check, and a device parameter compliance check on the first data set in sequence to obtain a consistency assessment result;

[0037] Calculating the relative error of the first data set to obtain an accuracy evaluation result;

[0038] The integrity assessment result, the consistency assessment result and the accuracy assessment result are weightedly fused to obtain the comprehensive assessment result.

[0039] As one preferred solution, making a calculation path decision based on the comprehensive evaluation result, and performing line loss calculation on the first data set based on the decision result, include:

[0040] If the comprehensive evaluation result of the first data set is the first level, the line loss is calculated using the root mean square current method, where the root mean square current method is expressed as:

[0041]

[0042]

[0043] in, is the RMS current during the operating time, is the resistance of the power grid components, is the line running time, is the correction factor, is the basic calculated value of the RMS current, is the active power, is the reactive power, is the voltage, is the RMS line loss, is the line resistance;

[0044] If the comprehensive evaluation result of the first data set is the second level, the line loss is calculated using the equivalent resistance method, where the equivalent resistance method is expressed as:

[0045]

[0046] in, is the equivalent line loss, For the The loss of the distribution transformer, is the number of empty losses, is the first equivalent resistance, is the second equivalent resistance.

[0047] Another embodiment of the present invention provides a distribution network line loss assessment device, comprising:

[0048] The preprocessing module is used to standardize the daily frozen power data, high-frequency current and voltage data, and equipment parameter data of the target distribution network to obtain multi-dimensional standardized data;

[0049] an extraction module, configured to analyze the multidimensional standardized data based on a dynamic sliding window, identify abnormal data points that meet a first preset condition, and extract spatiotemporal features of the abnormal data points to obtain spatiotemporal feature factors;

[0050] A mapping module is used to perform high-dimensional mapping on the multidimensional normalized data based on a neural tangent kernel function to calculate the NTK similarity; and perform weighted fusion of the spatiotemporal feature factor and the NTK similarity to obtain a final similarity;

[0051] A screening module is configured to perform neighbor screening on the abnormal data points based on an improved KNN algorithm to obtain neighbor data, wherein the neighbor screening is designed to adaptively adjust the number of neighbor screenings based on the standard deviation of the data distribution of the final similarity;

[0052] a repair module, configured to repair the neighboring data according to the final similarity to obtain a first data set, wherein the repair is designed to determine an optimal filling path based on a Bayesian network;

[0053] An evaluation module, configured to sequentially perform integrity evaluation, consistency evaluation, and accuracy evaluation on the first data set to obtain a comprehensive evaluation result;

[0054] A calculation module is used to make a calculation path decision based on the comprehensive evaluation result, and to perform line loss calculation on the first data set based on the decision result to obtain a line loss evaluation result of the target distribution network.

[0055] Another embodiment of the present invention provides a distribution network line loss assessment device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, and when the processor executes the computer program, the distribution network line loss assessment method as described above is implemented.

[0056] Yet another embodiment of the present invention provides a computer-readable storage medium storing a computer program, wherein when the device where the computer-readable storage medium is located executes the computer program, the distribution network line loss assessment method as described above is implemented.

[0057] Compared with the prior art, the embodiments of the present invention have the following advantages:

[0058] 1) This invention uses dynamic thresholds instead of fixed thresholds to identify abnormal data. Combined with secondary verification of topological consistency, this method increases the abnormal data identification rate from 80% to 95%, and improves the efficiency of erroneous data filtering by 50%. The combination of a dynamic sliding window and an adaptive threshold coefficient automatically adjusts detection accuracy based on load fluctuations, reducing line loss calculation errors by 5%-10%, effectively addressing the inaccuracy of traditional methods due to fixed thresholds.

[0059] 2) This invention utilizes sliding window parallel processing technology, reducing terabyte-level data cleaning time from 8 hours to 6 hours. Combined with standardized data interfaces, this technology reduces data integration time across different systems by 20%. Multi-dimensional standardized data processing and a dynamic neighbor screening mechanism enable streamlined data cleaning, repair, and evaluation, improving efficiency by over 25% compared to traditional single-threaded processing.

[0060] 3) By retrofitting existing metrology systems, the system eliminates the need for new advanced measurement equipment and automatically generates standardized reports, reducing manual compilation time from 4 man-days per month to 0.5 man-days per month. The intelligent calculation path decision mechanism eliminates the manual calculation and verification required for traditional advanced algorithms, reducing implementation costs by 80% while maintaining the simplicity of traditional methods. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 1 is a flow chart of a method for evaluating line loss in a distribution network in one embodiment of the present invention;

[0062] Figure 2 1 is a schematic structural diagram of a distribution network line loss assessment device in one embodiment of the present invention;

[0063] Figure 3 is a schematic diagram of a distribution network line loss assessment device in one embodiment of the present invention;

[0064] Reference numerals:

[0065] Among them, 11, preprocessing module; 12, extraction module; 13, mapping module; 14, screening module; 15, repair module; 16, evaluation module; 17, calculation module. DETAILED DESCRIPTION

[0066] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of them. The purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts shall fall within the scope of protection of the present invention.

[0067] In the description of the present invention, the terms "first," "second," "third," etc. are used for descriptive purposes only and should not be understood to indicate or imply relative importance or implicitly specify the number of the technical features indicated. Therefore, a feature specified as "first," "second," "third," etc. may explicitly or implicitly include one or more of the features. In the description of the present invention, unless otherwise specified, "plurality" means two or more.

[0068] In describing the present invention, it should be noted that, unless otherwise defined, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art. The terms used in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. Those skilled in the art will understand the specific meanings of the above terms in the present invention in specific circumstances.

[0069] An embodiment of the present invention provides a method for evaluating line loss in a distribution network. For details, see Figure 1 , Figure 1 The figure shows a flow chart of a method for evaluating line loss in a distribution network in one embodiment of the present invention, which includes steps S1-S7:

[0070] S1: Standardize the daily frozen power data, high-frequency current and voltage data, and equipment parameter data of the target distribution network to obtain multi-dimensional standardized data;

[0071] In this embodiment, real-time synchronous acquisition of key data is achieved through three major systems. Specifically, the electric energy collection system regularly collects daily frozen electricity data from gateway meters and user meters. This data is used to calculate the total amount of electricity supply and consumption, providing basic data support for macro-analysis of line losses. The SCADA system collects analog data such as line current and voltage at high frequency at 15-minute intervals to dynamically reflect the real-time operating status of the power grid and capture details of load fluctuations. The basic ledger database integrates static equipment information such as line length, conductor model, and transformer parameters to provide a physical parameter benchmark for subsequent line loss calculations.

[0072] Among them, the daily frozen electricity data, high-frequency current and voltage data, and equipment parameter data of the target distribution network are standardized, including the unification of time dimensions, physical unit specifications, format conversion, and model unification.

[0073] Specifically, in this embodiment, through timestamp calibration technology, the time tags of each system data are accurate to the minute level, ensuring the consistency of timestamps of data from different sources, solving the problem of time sequence misalignment of multi-source data, and laying the foundation for subsequent spatiotemporal correlation analysis.

[0074] In this embodiment, the unit standardization conversion of power, current, voltage and other data is performed according to international standards and industry specifications:

[0075] The power data is unified into kWh (kilowatt-hour);

[0076] Current data is unified as A (ampere) or kA (kiloampere);

[0077] Voltage data is uniformly expressed as V (volts) or kV (kilovolts);

[0078] This eliminates the heterogeneity of units from different equipment manufacturers and different acquisition systems, ensuring data comparability.

[0079] In this embodiment, a standardized data model (such as the IEC 61970 / 61968 standards) is used to convert structured, semi-structured, and unstructured data into a unified format. For example, real-time SCADA data and static parameters of a ledger database are mapped to the same data model, facilitating subsequent cleaning, verification, and analysis.

[0080] Through this processing, multi-source heterogeneous data is converted into multidimensional standardized data with time alignment, unified units, and standardized formats. This provides a high-quality data foundation for subsequent dynamic data cleaning, outlier identification, and line loss calculation. This step resolves the computational errors caused by data heterogeneity in traditional methods, ensuring the accuracy and consistency of subsequent analysis and being a prerequisite for the high-precision line loss assessment achieved by this invention.

[0081] S2: Analyzing the multidimensional standardized data based on a dynamic sliding window, identifying abnormal data points that meet a first preset condition, and extracting spatiotemporal features of the abnormal data points to obtain spatiotemporal feature factors;

[0082] Preferably, in one embodiment of the present invention, analyzing the multidimensional normalized data based on a dynamic sliding window to identify abnormal data points that meet a first preset condition includes:

[0083] Analyzing the multidimensional normalized data based on a sliding window of a preset window size, and calculating the coefficient of variation of the multidimensional normalized data within the preset window;

[0084] Dynamically adjusting the preset window size based on the coefficient of variation, and calculating the mean and standard deviation of the multidimensional normalized data within the adjusted window;

[0085] Dynamically calculate an abnormal threshold coefficient based on the load deviation of the multi-dimensional normalized data, and identify abnormal data points that meet a first preset condition based on the mean, the standard deviation, and the abnormal threshold coefficient; wherein the first preset condition is expressed as:

[0086]

[0087] in, For data to be identified, is the mean, is the standard deviation, is the abnormal threshold coefficient.

[0088] In this example, a fixed initial window (default 24 points, corresponding to 1 hour of data) is used to perform sliding analysis on the multi-dimensional normalized data. The window length matches the 15-minute sampling frequency of the SCADA system to ensure that high-frequency load fluctuation characteristics are captured. The coefficient of variation of the data within the window is calculated. :

[0089]

[0090] in, The value reflects the degree of data dispersion and is the core basis for dynamically adjusting the window size.

[0091] when When it is >0.15, the load fluctuation is judged to be severe, and the window is narrowed to 12 points (0.5 hours) to capture the abnormal characteristics of rapid changes;

[0092] when When <0.05, the load fluctuation is judged to be gentle, and the window is expanded to 48 points (2 hours) to smooth out noise interference;

[0093] After each window adjustment, the mean of the data in the window is recalculated and standard deviation , providing a dynamic benchmark for subsequent anomaly determination.

[0094] In this embodiment, the abnormal threshold coefficient It is positively correlated with the current load deviation and is calculated as follows:

[0095]

[0096] Among them, the basic coefficient =2.5, Current load and rated load deviation.

[0097] This embodiment provides an adaptive strategy for operating condition perception:

[0098] Light load state (P<0.3 ): k=1.8, relax the threshold to avoid misjudgment;

[0099] Overload state (P>0.9 ): k=3.5, tighten the threshold to capture potential anomalies;

[0100] Sudden response: When the load amplitude suddenly changes When , k=4.0 is instantly increased to enhance abnormal sensitivity.

[0101] based on Identify anomalous data points.

[0102] Preferably, in one embodiment of the present invention, the spatiotemporal characteristic factors include a temporal correlation factor and a spatial topological correlation degree;

[0103] The extracting spatiotemporal features of the abnormal data points to obtain spatiotemporal feature factors includes:

[0104] The time correlation factor is calculated as:

[0105]

[0106] in, is the time correlation factor, Indicates the number of historical windows, represents the sampling interval, represents the Pearson correlation coefficient;

[0107] The calculation of the spatial topological association degree is expressed as:

[0108]

[0109] in, is the spatial topological correlation, is the distance between the electrical appliances, calculated based on the admittance matrix, is the minimum value.

[0110] In this embodiment, when When >0.8, the current data is considered to be highly correlated with the historical pattern, the time dimension weight is increased, and the anomalies are repaired with reference to the historical trend first. When <0.3, reduce the time weight to avoid mismatching of historical data with abnormal fluctuations.

[0111] S3: Performing high-dimensional mapping on the multidimensional normalized data based on the neural tangent kernel function to calculate the NTK similarity; performing weighted fusion of the spatiotemporal feature factor and the NTK similarity to obtain the final similarity;

[0112] Multidimensional normalized data (including electrical parameters such as current, voltage, and power) is mapped through a two-layer fully connected neural network. The activation function uses ReLU to map the original data from a low-dimensional space (such as 3D electrical parameters) to a high-dimensional space (such as a 128-dimensional feature space). This mapping is achieved through the neural tangent kernel function, the core of which is to calculate the inner product of the data points in the high-dimensional space. The formula is:

[0113]

[0114] in, The mapping function defined for the neural network, To standardize the data points, It is the NTK similarity, which can capture nonlinear relationships such as current fluctuation and renewable energy output.

[0115] The final similarity is composed of three weighted components:

[0116] Time correlation factor: quantifies historical periodicity, such as the similarity of daily load curves;

[0117] Spatial topological correlation: electrical distance correlation calculated based on the grid admittance matrix;

[0118] NTK similarity: nonlinear feature similarity in high-dimensional space;

[0119] The spatiotemporal feature factor is weightedly fused with the NTK similarity to obtain a final similarity.

[0120] S4: performing neighbor screening on the abnormal data point based on an improved KNN algorithm to obtain neighbor data, wherein the neighbor screening is designed to adaptively adjust the number of neighbor screenings based on the standard deviation of the data distribution of the final similarity;

[0121] Preferably, in one embodiment of the present invention, performing neighbor screening on the abnormal data points based on the improved KNN algorithm to obtain neighbor data includes:

[0122] Dynamically calculating the number of nearest neighbors K based on the standard deviation of the distribution of the final similarity;

[0123] The abnormal data points are arranged in descending order according to the final similarity, and the first K data points are selected as the neighboring data.

[0124] In this embodiment, the distribution standard deviation of the number of neighbors K and the final similarity It is negatively correlated, and the calculation formula is:

[0125]

[0126] Among them, when When it is small (data is densely distributed), Approaching 1, the K value is reduced to 3 to avoid neighbor redundancy;

[0127] when When it is large (data distribution is sparse), Approaching 0, the K value is increased to 7 to ensure that there are enough neighbors.

[0128] Arrange the final similarity between abnormal data points and all data points in descending order, and give priority to selecting data points with high similarity as neighbors.

[0129] S5: repairing the neighboring data according to the final similarity to obtain a first data set, wherein the repairing is designed to determine an optimal filling path based on a Bayesian network;

[0130] Preferably, in one embodiment of the present invention, repairing the neighboring data according to the final similarity to obtain the first data set includes:

[0131] Constructing a Bayesian network with the final similarity and data importance weight as parent nodes and the repair priority as child nodes, and quantifying the repair order using a conditional probability formula, wherein the data importance weight is determined according to the node type;

[0132] The neighboring data is repaired based on the repair order to obtain the first data set.

[0133] In this embodiment, the final similarity The time correlation factor, spatial topology correlation and NTK similarity are integrated to quantify the similarity between the point to be repaired and its neighboring data. Determined by node type, specifically, the total meter data of the substation area directly affects the calculation of line loss rate. =1; branch line data has a minor impact, =0.6.

[0134] The repair priority of each neighboring data is quantified by the conditional probability model, and the formula is:

[0135]

[0136] Among them, the formula normalizes the product of similarity and weight to ensure that the priority ranking reflects the combined impact of data reliability and importance.

[0137] By using the posterior probability calculation of the Bayesian network, the filling order is transformed into an optimization problem to maximize the repair accuracy. =0.9 and it is the total meter data of the substation area, its priority is calculated as follows:

[0138]

[0139] The larger the value, the higher the repair order.

[0140] According to the repair priority, the neighboring data is dynamically weighted to obtain the first data set, which is expressed as:

[0141]

[0142] in, is the target data value after repair, For the The final similarity of neighbor data, For the The original value of the neighbor data.

[0143] S6: performing integrity assessment, consistency assessment, and accuracy assessment on the first data set in sequence to obtain a comprehensive assessment result;

[0144] Preferably, in one embodiment of the present invention, the step of sequentially performing integrity assessment, consistency assessment, and accuracy assessment on the first data set to obtain a comprehensive assessment result includes:

[0145] Calculating the data integrity rate of the first data set to obtain an integrity assessment result;

[0146] Performing a power balance check, a current-power consistency check, and a device parameter compliance check on the first data set in sequence to obtain a consistency assessment result;

[0147] Calculating the relative error of the first data set to obtain an accuracy evaluation result;

[0148] The integrity assessment result, the consistency assessment result and the accuracy assessment result are weightedly fused to obtain the comprehensive assessment result.

[0149] Specifically, the data integrity rate of the first data set is calculated as:

[0150]

[0151] in, Indicates the number of missing data points. If the continuous missing time is greater than 4 hours, it is considered as severe missing. is the total number of data that should be collected theoretically. When I<0.9, the data integrity is determined to be insufficient and the supplementary collection or repair process needs to be triggered.

[0152] In this embodiment, the power balance check means verifying the balance between the total power supply, total power sales and line loss based on the principle of conservation of energy to ensure accurate power statistics; the current-power consistency check means comparing the current-time integral value with the actual power data through integral calculation to verify the logical consistency of the two; the equipment parameter compliance check means comparing equipment parameters such as conductor impedance and transformer capacity with the standard parameter range to ensure the rationality of the equipment parameters.

[0153] In this embodiment, the relative error calculation is expressed as:

[0154]

[0155] in, It is the reference value of PMU or high-precision measuring device. The upper limit of the measuring range.

[0156] S7: Perform calculation path decision based on the comprehensive evaluation result, and perform line loss calculation on the first data set based on the decision result to obtain a line loss evaluation result of the target distribution network.

[0157] Preferably, in one embodiment of the present invention, performing calculation path decision based on the comprehensive evaluation result, and performing line loss calculation on the first data set based on the decision result, includes:

[0158] If the comprehensive evaluation result of the first data set is the first level, the line loss is calculated using the root mean square current method, where the root mean square current method is expressed as:

[0159]

[0160]

[0161] in, is the RMS current during the operating time, is the resistance of the power grid components, is the line running time, is the correction factor, is the basic calculated value of the RMS current, is the active power, is the reactive power, is the voltage, is the RMS line loss, is the line resistance;

[0162] If the comprehensive evaluation result of the first data set is the second level, the line loss is calculated using the equivalent resistance method, where the equivalent resistance method is expressed as:

[0163]

[0164] in, is the equivalent line loss, For the The loss of the distribution transformer, is the number of empty losses, is the first equivalent resistance, is the second equivalent resistance.

[0165] Specifically, when the comprehensive score Q > 0.8, indicating that the data integrity, consistency, and accuracy all meet the requirements for high-precision calculations, the RMS current method is triggered. In this embodiment, 24-hour continuous current data is collected at the line headend to obtain a complete sequence of current changes over time. As shown in Table 1, Table 1 shows the operating data required to calculate the power loss of power grid components at 35kV and above using the RMS current method.

[0166] Table 1

[0167]

[0168] Specifically, when the comprehensive score Q is less than 0.8, indicating missing data or quality defects, the equivalent resistance method is automatically switched to. In this embodiment, the equivalent resistance of the line is calculated based on parameters such as line length and wire type, and the equivalent current is calculated based on the power supply and operating time. Table 2 shows the operating data required for the equivalent resistance method.

[0169] Table 2

[0170]

[0171] Another embodiment of the present invention provides a distribution network line loss assessment device. For details, see Figure 1 , Figure 1 FIG. 1 is a schematic structural diagram of a distribution network line loss assessment device according to one embodiment of the present invention, which includes:

[0172] The pre-processing module 11 is used to perform standardization processing on the daily frozen power data, high-frequency current and voltage data, and equipment parameter data of the target distribution network to obtain multi-dimensional standardized data;

[0173] An extraction module 12 is configured to analyze the multidimensional standardized data based on a dynamic sliding window, identify abnormal data points that meet a first preset condition, and extract spatiotemporal features of the abnormal data points to obtain spatiotemporal feature factors;

[0174] A mapping module 13 is configured to perform high-dimensional mapping on the multidimensional normalized data based on a neural tangent kernel function to calculate the NTK similarity; and perform weighted fusion of the spatiotemporal feature factor and the NTK similarity to obtain a final similarity;

[0175] A screening module 14 is configured to perform neighbor screening on the abnormal data points based on an improved KNN algorithm to obtain neighbor data, wherein the neighbor screening is designed to adaptively adjust the number of neighbor screenings based on the standard deviation of the data distribution of the final similarity;

[0176] a repairing module 15, configured to repair the neighboring data according to the final similarity to obtain a first data set, wherein the repairing is designed to determine an optimal filling path based on a Bayesian network;

[0177] An evaluation module 16 is configured to sequentially perform integrity evaluation, consistency evaluation, and accuracy evaluation on the first data set to obtain a comprehensive evaluation result;

[0178] The calculation module 17 is configured to make a calculation path decision based on the comprehensive evaluation result, and perform line loss calculation on the first data set based on the decision result to obtain a line loss evaluation result of the target distribution network.

[0179] Another embodiment of the present invention provides a method for evaluating line loss in a distribution network. Figure 3 , which is a structural block diagram of a distribution network line loss assessment device provided in an embodiment of the present invention. The distribution network line loss assessment device provided in an embodiment of the present invention includes a processor 21, a memory 22, and a computer program stored in the memory 22 and configured to be executed by the processor 21. When the processor 21 executes the computer program, the steps in the above-mentioned distribution network line loss assessment method embodiment are implemented, for example Figure 1 or, when the processor 21 executes the computer program, the functions of the modules in the above-mentioned device embodiments, such as the preprocessing module 11, are implemented.

[0180] Exemplarily, the computer program may be divided into one or more modules, which are stored in the memory 22 and executed by the processor 21 to implement the present invention. The one or more modules may be a series of computer program instruction segments capable of performing specific functions, and the instruction segments are used to describe the execution process of the computer program in the distribution network line loss assessment device.

[0181] The distribution network line loss assessment device may include, but is not limited to, a processor 21 and a memory 22. Those skilled in the art will appreciate that the schematic diagram is merely an example of a distribution network line loss assessment device and does not limit the distribution network line loss assessment device. The distribution network line loss assessment device may include more or fewer components than shown in the diagram, or may combine certain components or different components. For example, the distribution network line loss assessment device may also include input and output devices, network access devices, buses, and the like.

[0182] The processor 21 can be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor. The processor 21 is the control center of the distribution network line loss assessment device, and uses various interfaces and lines to connect various parts of the entire distribution network line loss assessment device.

[0183] The memory 22 can be used to store the computer programs and / or modules. The processor 21 implements the various functions of the distribution network line loss assessment device by running or executing the computer programs and / or modules stored in the memory 22 and accessing the data stored in the memory 22. The memory 22 may primarily include a program storage area and a data storage area. The program storage area may store an operating system and at least one application required for a function (such as sound playback or image playback); the data storage area may store data generated based on the use of the mobile phone (such as audio data and a phone book). Furthermore, the memory 22 may include high-speed random access memory (RAM) and non-volatile memory, such as a hard disk, internal memory, a plug-in hard disk, a Smart Media Card (SMC), a Secure Digital (SD) card, a flash memory card, at least one disk storage device, a flash memory device, or other volatile solid-state storage device.

[0184] If the module integrated into the distribution network line loss assessment device is implemented as a software functional unit and sold or used as a standalone product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention can also implement all or part of the process steps in the above-mentioned method embodiments by using a computer program to instruct the relevant hardware. The computer program can be stored in a computer-readable storage medium. When executed by a processor, the computer program can implement the steps of each of the above-mentioned method embodiments. The computer program includes computer program code, which can be in source code form, object code form, executable file, or some intermediate form. The computer-readable medium can include any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a removable hard drive, a magnetic disk, an optical disk, computer memory, read-only memory (ROM), random access memory (RAM), an electrical carrier signal, a telecommunications signal, and software distribution media.

[0185] Those skilled in the art will appreciate that all or part of the processes in the above-described method embodiments can be implemented by instructing related hardware through a computer program. The program can be stored in a computer-readable storage medium, and when executed, the program can include the processes in the above-described method embodiments. The storage medium can be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0186] Accordingly, an embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the steps of the distribution network line loss assessment method as described in the above embodiment, for example Figure 1 Steps S1 to S7 described in .

[0187] Compared with the prior art, the embodiments of the present invention have the following advantages:

[0188] 1) This invention uses dynamic thresholds instead of fixed thresholds to identify abnormal data. Combined with secondary verification of topological consistency, this method increases the abnormal data identification rate from 80% to 95%, and improves the efficiency of erroneous data filtering by 50%. The combination of a dynamic sliding window and an adaptive threshold coefficient automatically adjusts detection accuracy based on load fluctuations, reducing line loss calculation errors by 5%-10%, effectively addressing the inaccuracy of traditional methods due to fixed thresholds.

[0189] 2) This invention utilizes sliding window parallel processing technology, reducing terabyte-level data cleaning time from 8 hours to 6 hours. Combined with standardized data interfaces, this technology reduces data integration time across different systems by 20%. Multi-dimensional standardized data processing and a dynamic neighbor screening mechanism enable streamlined data cleaning, repair, and evaluation, improving efficiency by over 25% compared to traditional single-threaded processing.

[0190] 3) By retrofitting existing metrology systems, the system eliminates the need for new advanced measurement equipment and automatically generates standardized reports, reducing manual compilation time from 4 man-days per month to 0.5 man-days per month. The intelligent calculation path decision mechanism eliminates the manual calculation and verification required for traditional advanced algorithms, reducing implementation costs by 80% while maintaining the simplicity of traditional methods.

[0191] The above-described embodiments merely illustrate several implementations of the present invention, and while their descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art would be able to make numerous variations and improvements without departing from the spirit of the present invention, all of which fall within the scope of protection of the present invention. Therefore, the scope of protection of the present invention shall be determined by the appended claims.

Claims

1. A method for evaluating line loss in a distribution network, characterized in that: include: Standardize the daily frozen power data, high-frequency current and voltage data, and equipment parameter data of the target distribution network to obtain multi-dimensional standardized data; Analyzing the multidimensional standardized data based on a dynamic sliding window, identifying abnormal data points that meet a first preset condition, and extracting spatiotemporal features of the abnormal data points to obtain spatiotemporal feature factors; Performing high-dimensional mapping on the multidimensional normalized data based on the neural tangent kernel function to calculate the NTK similarity; performing weighted fusion of the spatiotemporal feature factor and the NTK similarity to obtain the final similarity; Performing neighbor screening on the abnormal data points based on an improved KNN algorithm to obtain neighbor data, wherein the neighbor screening is designed to adaptively adjust the number of neighbor screenings based on the standard deviation of the data distribution of the final similarity; Repairing the neighboring data according to the final similarity to obtain a first data set, wherein the repairing is designed to determine an optimal filling path based on a Bayesian network; Performing integrity assessment, consistency assessment, and accuracy assessment on the first data set in sequence to obtain a comprehensive assessment result; Performing a calculation path decision based on the comprehensive evaluation result, and performing line loss calculation on the first data set based on the decision result to obtain a line loss evaluation result of the target distribution network; The spatiotemporal characteristic factors include a temporal correlation factor and a spatial topological correlation degree; The extracting spatiotemporal features of the abnormal data points to obtain spatiotemporal feature factors includes: The time correlation factor is calculated as: in, is the time correlation factor, Indicates the number of historical windows, represents the sampling interval, represents the Pearson correlation coefficient; The calculation of the spatial topological association degree is expressed as: in, is the spatial topological correlation, is the distance between the electrical appliances, calculated based on the admittance matrix, is the minimum value.

2. The distribution network line loss assessment method according to claim 1, wherein: The analyzing the multidimensional standardized data based on the dynamic sliding window to identify abnormal data points that meet the first preset condition includes: Analyzing the multidimensional normalized data based on a sliding window of a preset window size, and calculating the coefficient of variation of the multidimensional normalized data within the preset window; Dynamically adjusting the preset window size based on the coefficient of variation, and calculating the mean and standard deviation of the multidimensional normalized data within the adjusted window; Dynamically calculate an abnormal threshold coefficient based on the load deviation of the multi-dimensional normalized data, and identify abnormal data points that meet a first preset condition based on the mean, the standard deviation, and the abnormal threshold coefficient; wherein the first preset condition is expressed as: in, For data to be identified, is the mean, is the standard deviation, is the abnormal threshold coefficient.

3. The distribution network line loss assessment method according to claim 1, wherein: The performing of neighbor screening on the abnormal data points based on the improved KNN algorithm to obtain neighbor data includes: Dynamically calculating the number of nearest neighbors K based on the standard deviation of the distribution of the final similarity; The abnormal data points are arranged in descending order according to the final similarity, and the first K data points are selected as the neighboring data.

4. The method for evaluating line loss in a distribution network according to claim 1, wherein: The repairing of the neighboring data according to the final similarity to obtain a first data set includes: Constructing a Bayesian network with the final similarity and data importance weight as parent nodes and the repair priority as child nodes, and quantifying the repair order using a conditional probability formula, wherein the data importance weight is determined according to the node type; The neighboring data is repaired based on the repair order to obtain the first data set.

5. The method for evaluating line loss in a distribution network according to claim 1, wherein: The step of sequentially performing integrity assessment, consistency assessment, and accuracy assessment on the first data set to obtain a comprehensive assessment result includes: Calculating the data integrity rate of the first data set to obtain an integrity assessment result; Performing a power balance check, a current-power consistency check, and a device parameter compliance check on the first data set in sequence to obtain a consistency assessment result; Calculating the relative error of the first data set to obtain an accuracy evaluation result; The integrity assessment result, the consistency assessment result and the accuracy assessment result are weightedly fused to obtain the comprehensive assessment result.

6. The method for evaluating line loss in a distribution network according to claim 1, wherein: The performing a calculation path decision based on the comprehensive evaluation result, and performing line loss calculation on the first data set based on the decision result, includes: If the comprehensive evaluation result of the first data set is the first level, the line loss is calculated using the root mean square current method, where the root mean square current method is expressed as: in, is the RMS current during the operating time, is the resistance of the power grid components, is the line running time, is the correction factor, is the basic calculated value of the RMS current, is the active power, is the reactive power, is the voltage, is the RMS line loss, is the line resistance; If the comprehensive evaluation result of the first data set is the second level, the line loss is calculated using the equivalent resistance method, where the equivalent resistance method is expressed as: in, is the equivalent line loss, For the The loss of the distribution transformer, is the number of empty losses, is the first equivalent resistance, is the second equivalent resistance.

7. A distribution network line loss assessment device, characterized in that: include: The preprocessing module is used to standardize the daily frozen power data, high-frequency current and voltage data, and equipment parameter data of the target distribution network to obtain multi-dimensional standardized data; an extraction module, configured to analyze the multidimensional standardized data based on a dynamic sliding window, identify abnormal data points that meet a first preset condition, and extract spatiotemporal features of the abnormal data points to obtain spatiotemporal feature factors; A mapping module is used to perform high-dimensional mapping on the multidimensional normalized data based on a neural tangent kernel function to calculate the NTK similarity; and perform weighted fusion of the spatiotemporal feature factor and the NTK similarity to obtain a final similarity; A screening module is configured to perform neighbor screening on the abnormal data points based on an improved KNN algorithm to obtain neighbor data, wherein the neighbor screening is designed to adaptively adjust the number of neighbor screenings based on the standard deviation of the data distribution of the final similarity; a repair module, configured to repair the neighboring data according to the final similarity to obtain a first data set, wherein the repair is designed to determine an optimal filling path based on a Bayesian network; An evaluation module, configured to sequentially perform integrity evaluation, consistency evaluation, and accuracy evaluation on the first data set to obtain a comprehensive evaluation result; a calculation module, configured to make a calculation path decision based on the comprehensive evaluation result, and perform line loss calculation on the first data set based on the decision result to obtain a line loss evaluation result of the target distribution network; The spatiotemporal characteristic factors include a temporal correlation factor and a spatial topological correlation degree; The extracting spatiotemporal features of the abnormal data points to obtain spatiotemporal feature factors includes: The time correlation factor is calculated as: in, is the time correlation factor, Indicates the number of historical windows, represents the sampling interval, represents the Pearson correlation coefficient; The calculation of the spatial topological association degree is expressed as: in, is the spatial topological correlation, is the distance between the electrical appliances, calculated based on the admittance matrix, is the minimum value.

8. A distribution network line loss assessment device, characterized in that: The method comprises a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, the method for evaluating line loss in a distribution network according to any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, wherein when the device where the computer-readable storage medium is located executes the computer program, the distribution network line loss assessment method according to any one of claims 1 to 6 is implemented.

Images