Low-voltage distribution network customer phase identification method and system based on voltage fluctuation significance analysis

Abstract

The present invention discloses a method and system for identifying the phase of a low-voltage power distribution network customer based on voltage fluctuation significance analysis. The method comprises the following steps: obtaining a complete voltage time series and a complete power time series; using a dynamic sliding window model to extract a significant fluctuation difference segment of the complete voltage time series of a three-phase feeder; performing customer power supply phase matching based on the significant fluctuation difference segment and the complete voltage time series of all customers, and if the matching is successful, obtaining a customer power supply phase label; otherwise, treating the customer who fails in phase matching as a phase fuzzy customer; based on the complete power time series of the three-phase feeder and the complete power time series of all customers, using the power conservation principle, performing power supply phase matching for the phase fuzzy customer, and obtaining a customer power supply phase label. The present invention can realize accurate identification of the customer power supply phase.

Description

Technical Field

[0001] The present invention belongs to the technical field of low-voltage distribution network phase identification, and specifically relates to a low-voltage distribution network customer phase identification method and system based on voltage fluctuation significance analysis. Background Art

[0002] Understanding the power supply phase information of customers in the low-voltage distribution network is conducive to rapid fault location, power outage response and asset management. However, due to differences in customer load types, the difficulty of phase identification in various low-voltage distribution substations varies, and traditional general methods may not be effective in identifying specific substations. In large commercial or industrial buildings, the high power supply density poses greater challenges to data measurement. The loads are diverse and have large power differences. In order to ensure the stability of power supply, most loads are in three-phase power supply mode. The degree of three-phase imbalance in the substation is very low, which increases the difficulty of identification of data-driven methods. Therefore, the study of phase identification technology based on distribution substations with low three-phase imbalance is of great significance for promoting the construction of smart grids, enhancing the stability of power systems, improving the service level of power companies, and enhancing customer experience.

[0003] At present, there are two main research directions for phase identification of low-voltage distribution networks, namely communication signal analysis and electrical quantity data analysis. The use of communication signal analysis for phase identification has high accuracy, but the frequency of the communication signal is very particular and is easily interfered by noise. In areas with large load changes, the identification performance is not very ideal. In addition, this type of method also requires the installation of additional terminal equipment, which will lead to problems such as high investment costs and high operation and maintenance pressure.

[0004] With the continuous improvement of the ability to collect electricity consumption data of low-voltage customers, topological analysis methods based on electrical quantity data have gradually become popular. The current mainstream research methods include: mathematical programming, cluster analysis, deep learning, etc. The principle of mathematical programming is relatively simple, but the calculation complexity is high, and the requirements on the number and quality of data samples are relatively strict. The clustering-based topological identification algorithm is more convenient when the scale of the substation is large. It can quickly classify customers and pair them with corresponding distribution transformers and phases. The disadvantage is that the accuracy of the algorithm is greatly affected by parameters such as features and clustering thresholds. The real-time performance of deep learning algorithms is poor, so their application in practice is hindered to a certain extent. Summary of the invention

[0005] In view of the deficiencies in the prior art, the present invention provides a method and system for identifying the customer phase of a low-voltage distribution network based on voltage fluctuation significance analysis, so as to achieve accurate identification of the customer power supply phase.

[0006] To achieve the above object, the present invention provides the following solutions:

[0007] A method for identifying customer phases in a low-voltage distribution network based on voltage fluctuation significance analysis includes the following steps:

[0008] S1: Obtain the complete voltage time series and complete power time series of the three-phase feeder and all customers;

[0009] S2: Using the dynamic sliding window model, the significant fluctuation difference segments of the complete voltage time series of the three-phase feeder are extracted;

[0010] S3: Based on the significant fluctuation difference segment and the complete voltage time series of all customers, customer power supply phase matching is performed. If the matching is successful, the customer power supply phase label is obtained, and customer phase identification based on voltage fluctuation significance analysis is completed; otherwise, the customer who fails the phase matching is regarded as a phase fuzzy customer, and step S4 is executed;

[0011] S4: Based on the complete power time series of the three-phase feeder and the complete power time series of all customers, the power conservation principle is adopted to perform power supply phase matching for the phase-ambiguous customer, obtain the customer power supply phase label, and complete customer phase identification based on voltage fluctuation significance analysis.

[0012] Preferably, in step S2, the method of extracting the significant fluctuation difference segment of the complete voltage time series of the three-phase feeder using the dynamic sliding window model is:

[0013] Based on the daily load curve of the transformer main port, obtain the high load period;

[0014] During the high load period, a window length candidate set is preset and the step length is fixed;

[0015] Starting from the first voltage data point of the complete voltage time series of the three-phase feeder, respectively calculating the Pearson correlation coefficients between the voltage windows at the head end of the three-phase feeder when the window length takes different values ​​in the window length candidate set;

[0016] The mean value of the Pearson correlation coefficient between the voltage windows at the head end of the three-phase feeder is compared with a preset similarity threshold to obtain a maximum window length;

[0017] The maximum window length is used as the actual window length of the current operation, and the window period of the maximum window length is recorded as the significant fluctuation difference period.

[0018] Preferably, in step S3, the method for performing customer power supply phase matching based on the significant fluctuation difference segment and the complete voltage time series of all customers is:

[0019] Reconnect the significant fluctuation difference segments in time sequence to obtain a three-phase feeder phase detection sequence;

[0020] Calculate the comprehensive similarity coefficient between the complete voltage time series of all customers and the three-phase feeder phase detection sequence, and use the comprehensive similarity coefficient as a determination index to perform customer power supply phase matching;

[0021] When the comprehensive similarity coefficient meets the preset conditions, the customer power supply phase is determined and the customer power supply phase label is obtained; otherwise, the customer who fails to match the phase is regarded as a phase ambiguous customer.

[0022] Preferably, the method for calculating the comprehensive similarity coefficient is:

[0023] Based on the mean values ​​of the customer's complete voltage time series and the three-phase feeder phase detection sequences and the time series lengths, a Pearson correlation coefficient between the sequences is obtained;

[0024] Based on the Pearson correlation coefficient between the sequences, the correlation coefficient matrix is ​​obtained;

[0025] Calculate the average Euclidean distance between sequences to obtain the distance matrix;

[0026] Normalizing the distance matrix to obtain a comprehensive similarity matrix;

[0027] Based on the comprehensive similarity matrix, the comprehensive similarity coefficients of the complete voltage time series of all customers and the three-phase feeder phase detection sequence are obtained.

[0028] Preferably, in step S4, the formula for power supply phase matching for the phase-ambiguous customer is:

[0029]

[0030] Among them, p i,t represents the power of the i-phase feeder at time t; u represents the number of customers whose power supply phase is determined to be i-phase, p j,t represents the power of customer j at time t whose power supply phase is determined to be phase i; p 0,t represents the power of the phase-ambiguous customer at time t; A, B, and C represent the three-phase feeder, Δp i Power supply phase for phase-ambiguous customers.

[0031] The present invention provides a low-voltage distribution network customer phase identification system based on voltage fluctuation significance analysis, which is used to implement the customer phase identification method, including:

[0032] A sequence reading module is used to obtain the complete voltage time series and the complete power time series of the three-phase feeder and all customers;

[0033] A significant fluctuation difference segment extraction module is used to extract the significant fluctuation difference segment of the complete voltage time series of the three-phase feeder using a dynamic sliding window model;

[0034] The first phase matching module is used to perform customer power supply phase matching based on the significant fluctuation difference segment and the complete voltage time series of all customers. If the matching is successful, the customer power supply phase label is obtained to perform customer phase identification based on voltage fluctuation significance analysis; otherwise, the customer who fails the phase matching is regarded as a phase fuzzy customer, and the second phase matching module is executed;

[0035] The second phase matching module is used to perform power supply phase matching for the phase-ambiguous customer based on the complete power time series of the three-phase feeder and the complete power time series of all customers, adopt the power conservation principle, obtain the customer power supply phase label, and complete the customer phase identification based on voltage fluctuation significance analysis.

[0036] Preferably, the significant fluctuation difference segment extraction module includes:

[0037] A high load period selection unit is used to obtain a high load period based on a daily load curve of the transformer main port;

[0038] A window length candidate set establishing unit, used for presetting a window length candidate set and fixing the step length during the high load period;

[0039] A correlation coefficient calculation unit is used to calculate the Pearson correlation coefficient between the voltage windows at the head end of the three-phase feeder when the window length takes different values ​​in the window length candidate set, starting from the first voltage data point of the complete voltage time series of the three-phase feeder;

[0040] A maximum window length acquisition unit is used to compare the mean value of the Pearson correlation coefficient between the voltage windows at the head end of the three-phase feeder with a preset similarity threshold to obtain a maximum window length;

[0041] The significant fluctuation difference period acquisition unit is used to take the maximum window length as the actual window length of the current operation, and record the window period of the maximum window length as the significant fluctuation difference period.

[0042] Preferably, the first phase matching module includes:

[0043] A sequence splicing unit, used for re-splicing the significant fluctuation difference segments in time sequence to obtain a three-phase feeder phase detection sequence;

[0044] A phase matching unit, used to calculate the comprehensive similarity coefficient between the complete voltage time series of all customers and the three-phase feeder phase detection sequence, and use the comprehensive similarity coefficient as a judgment index to perform customer power supply phase matching;

[0045] The phase ambiguity customer acquisition unit is used to determine the customer power supply phase and obtain the customer power supply phase label when the comprehensive similarity coefficient meets the preset conditions; otherwise, the customer who fails the phase matching is regarded as a phase ambiguity customer.

[0046] Compared with the prior art, the beneficial effects of the present invention are as follows: a method for extracting voltage significant fluctuation difference segments based on a dynamic sliding window model and a comprehensive similarity coefficient is proposed. Compared with a fixed sliding window, the dynamic sliding window model balances accuracy and efficiency and can effectively extract phase differences. The proposed comprehensive similarity coefficient adds a measurement of amplitude on the basis of the Pearson correlation coefficient, has good phase distinction, and is conducive to judging the neighbors and upstream and downstream relationships of the same-phase customers.

[0047] The power balance principle is used to perform phase matching on a small number of "fuzzy" customers. Such customers usually have voltage deterioration or poor measurement quality, and the similarity index cannot correctly determine the phase. Power verification is used to further improve recognition accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] In order to more clearly illustrate the technical solution of the present invention, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0049] Figure 1 This is a flow chart of a method for identifying a phase of a low-voltage power distribution network customer based on voltage fluctuation significance analysis according to an embodiment of the present invention;

[0050] Figure 2 Schematic diagram of a voltage time series sliding window model according to an embodiment of the present invention. DETAILED DESCRIPTION

[0051] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0052] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0053] Embodiment 1

[0054] The present invention proposes a low-voltage distribution network customer phase identification method based on voltage fluctuation significance analysis. The design idea is: first, through the power grid flow analysis, the influence of customer power fluctuations on the voltage amplitude of each node is found, and the characteristic of relatively large change trend between phase voltages during high-load periods in the substation area is found. Then, a dynamic sliding window model is used to assist in finding voltage segments with more significant differences between three-phase feeder voltages, and the Pearson correlation coefficient is improved to form a similarity coefficient of comprehensive curve change trend and amplitude size as the main judgment indicator. Finally, for a small number of customer nodes with fuzzy identification of similarity indicators, phase matching is performed based on power balance theory to achieve accurate identification of customer power supply phases.

[0055] like Figure 1 As shown, the specific steps include:

[0056] S1: Read and preprocess the voltage time series and power time series of the three-phase feeder and all customers to obtain the complete voltage time series and complete power time series of the three-phase feeder and all customers; in this embodiment, the voltage and power data collected from the customers and three-phase feeders over a period of time are used as input. There is noise data in the original data, and it is necessary to detect abnormal noise values ​​and replace them. The abnormal values ​​in the voltage time series can be screened by the standard deviation detection method and replaced with the average value over a period of time or the value at the last moment. For missing data, adjacent values ​​are used to fill in, and the missing segments of continuous data are deleted. Missing detection is performed on the time series of all nodes, and the common time period with measurement is selected to finally generate the complete voltage time series and complete power time series including the three-phase feeder and all customers.

[0057] Specifically, in this embodiment, the voltage time series similarity coefficient between customer i and the three-phase feeders A, B, and C and the Pearson correlation coefficient PCC (Pearson correlation coefficient) between the voltage windows at the head end of the three-phase feeders are calculated. The specific calculation process takes the calculation of the PCC and the comprehensive similarity coefficient between the customer voltage time series as an example:

[0058] 1) Selection of comprehensive similarity calculation indicators: The PCC expression between two time series is as follows:

[0059]

[0060] Among them, X and Y represent two time series curves respectively. They represent the means of X and Y respectively, and n represents the length of the time series.

[0061] The present invention proposes an improved solution for PCC, in which PCC is calculated between two of n customer voltage curves to form a correlation coefficient matrix W:

[0062]

[0063] Where m represents the number of customers, ρ ij Represents the PCC between the voltage timing curves of customer i and customer j.

[0064] Then, the average Euclidean distance between the two curves is calculated by formula (3), and the distance matrix D shown in formula (4) is formed:

[0065]

[0066] Where n represents the length of the time series; v i,k 、v j,k Respectively represent the voltage values ​​of customer i and customer j at time section k; l ij Represents the average Euclidean distance between the voltage timing curves of customer i and customer j. The distance matrix D is normalized to the range of [0,1] using the min-max normalization method, and the two matrices are merged to form a comprehensive similarity matrix S:

[0067]

[0068] in, represents the normalized average Euclidean distance matrix; A1 represents a matrix of order m with all elements being 1; the element s in S ij represents the comprehensive similarity coefficient between customer i and customer j; α is the weight coefficient, and its value range is [0,1].

[0069] S2: Using a dynamic sliding window model, extract the significant fluctuation difference segments of the complete voltage time series of the three-phase feeder; the present invention proposes a sliding window model for dynamically selecting the window length, which is used to select the voltage time series segments with large differences in voltage fluctuations of each phase.

[0070] A further implementation method is to use a dynamic sliding window model to extract a significant fluctuation difference segment of the complete voltage time series of the three-phase feeder line:

[0071] Based on the daily load curve of the transformer main port, obtain the high load period;

[0072] Specifically, the daily load curve of the transformer main port is periodic and representative, so the present invention defines the high load period as:

[0073] p hl ≥p max ·σ (6)

[0074] T min ≥k / f (7)

[0075] Among them, p hl Indicates the total transformer power at the time of high load condition; p max Indicates the maximum value of the daily load curve of the total port; σ is the percentage coefficient; T min It indicates the minimum duration of high-load segments, i.e., continuous high-load moments; f indicates the frequency of data collection; and k is a parameter used to determine the length of time that the shortest high-load segment needs to satisfy.

[0076] During the high-load period, a window length candidate set w is preset set ={w1,w2,...,w n} and fix the step size; Figure 2 The following is a schematic diagram of the voltage sequence sliding window, v i It represents the voltage value of the customer at time section i. The figure includes two important parameters, namely the window length (shown as w1 and w2 in the figure) and the step size (step).

[0077] Starting from the first voltage data point of the complete voltage time series of the three-phase feeder, the window lengths are calculated and the candidate window length sets w are taken. set When the values ​​of are different, the Pearson correlation coefficient PCC between the voltage windows at the head end of the three-phase feeder is denoted as ρ AB , BC , AC .

[0078] The mean value of the Pearson correlation coefficient between the voltage windows at the head end of the three-phase feeder is ρ AVE Similarity threshold ρ thres Compare and obtain the maximum window length; specifically, select AVE ≤ρ thres The maximum window length is used as the actual window length of the current operation.

[0079] The maximum window length is taken as the actual window length of the current operation, and the window period of the maximum window length is recorded as the significant fluctuation difference period. Repeat the above steps to determine the window length before each window update.

[0080] S3: Based on the significant fluctuation difference segment and the complete voltage time series of all customers, customer power supply phase matching is performed. If the matching is successful, the customer power supply phase label is obtained; otherwise, the customer who fails the phase matching is regarded as a phase fuzzy customer; in this embodiment, when the cumulative difference segment samples meet the preset requirements, customer phase matching is performed, otherwise the time series is re-read. When the proportion of phase model customers is lower than the preset threshold, step S4 is performed to perform fuzzy customer phase matching based on power balance, otherwise the time series is re-read.

[0081] A further implementation method is that, based on the significant fluctuation difference segment and the complete voltage time series of all customers, the method for performing customer power supply phase matching is:

[0082] The significant fluctuation difference segments are reconnected in time sequence to obtain the three-phase feeder phase detection sequence;

[0083] The comprehensive similarity coefficient of the complete voltage time series of all customers and the phase detection sequence of the three-phase feeder is calculated, and the comprehensive similarity coefficient is used as a judgment index to match the customer power supply phase; in this embodiment, for all customers and three-phase feeders, all fluctuation difference periods within the total data length are reconnected in chronological order, and then the voltage comprehensive similarity matrix (including all customers and phase feeders) is calculated. Based on the voltage comprehensive similarity matrix, the comprehensive similarity coefficient is obtained.

[0084] When the comprehensive similarity coefficient meets the preset conditions, the customer power supply phase is determined and the customer power supply phase label is obtained; otherwise, the customer who fails to match the phase is regarded as a phase ambiguity customer. Specifically, according to the similarity coefficient s between customer i and the three-phase feeder A, B, and C iA 、s iB 、s iC The power supply phase of the customer is determined by the relationship between the three. When the relationship between the three satisfies the formula (8), it is considered that the discrimination is good, and the phase corresponding to the maximum coefficient is the power supply phase of the customer. If a customer does not meet this condition, the customer is temporarily identified as a phase "fuzzy" customer.

[0085] max(s iA ,s iB ,s iC )-median(s iA ,s iB ,s iC )≥0.1 (8)

[0086] Among them, max() is the maximum value function, and median() is the median function.

[0087] A further implementation method is that the method for calculating the comprehensive similarity coefficient is:

[0088] Based on the mean and time series length of the customer's complete voltage time series and the three-phase feeder phase detection sequence, the Pearson correlation coefficient between the sequences is obtained;

[0089] Based on the Pearson correlation coefficient between the sequences, the correlation coefficient matrix is ​​obtained;

[0090] Calculate the average Euclidean distance between sequences to obtain the distance matrix;

[0091] Normalize the distance matrix to obtain a comprehensive similarity matrix;

[0092] Based on the comprehensive similarity matrix, the comprehensive similarity coefficients of the complete voltage time series of all customers and the phase detection sequence of the three-phase feeder are obtained.

[0093] S4: Based on the complete power time series of the three-phase feeder and the complete power time series of all customers, the power conservation principle is used to match the power supply phase for phase-ambiguous customers, obtain the customer power supply phase label, and complete customer phase identification based on voltage fluctuation significance analysis.

[0094] For customers with "fuzzy" phases that may appear during the identification process, the power balance principle is used to match their phases. That is, in the distribution network system, the total power of a phase load in the substation is equal to the sum of the power consumption of all customers in that phase and the network loss power. The low-voltage commercial substation has a low voltage level and short power supply lines, and the network loss is relatively small.

[0095] A further implementation method is that when the ratio of customers with "fuzzy" phase to all customers in the area is less than 5%, the phase of a certain customer can be matched in the form of power difference. That is:

[0096]

[0097] Among them, p i,t represents the power of the i-phase feeder at time t; u represents the number of customers whose power supply phase is determined to be i-phase, p j,t represents the power of customer j at time t whose power supply phase is determined to be phase i; p 0,t represents the power of the phase-ambiguous customer at time t; A, B, and C represent the three-phase feeder, Δp i Power supply phase for phase-ambiguous customers.

[0098] The phase of the customer is the calculated Δp A , Δp B , Δp C It should be emphasized that when calculating Δp i It is necessary to ensure that the above power difference at any time t (t = 1, 2, ..., n) is non-negative, otherwise the customer will obviously not belong to this phase.

[0099] Embodiment 2

[0100] The present invention also provides a low-voltage distribution network customer phase identification system based on voltage fluctuation significance analysis, which is used to implement a customer phase identification method, including:

[0101] A sequence reading module is used to obtain the complete voltage time series and the complete power time series of the three-phase feeder and all customers;

[0102] A significant fluctuation difference segment extraction module is used to extract the significant fluctuation difference segment of the complete voltage time series of the three-phase feeder using a dynamic sliding window model;

[0103] The first phase matching module is used to match the customer power supply phase based on the significant fluctuation difference segment and the complete voltage time series of all customers. If the match is successful, the customer power supply phase label is obtained to complete the customer phase identification based on the voltage fluctuation significance analysis; otherwise, the customer who fails the phase matching is regarded as a phase fuzzy customer, and the second phase matching module is executed;

[0104] The second phase matching module is used to perform power supply phase matching for phase-ambiguous customers based on the complete power time series of the three-phase feeder and the complete power time series of all customers, adopt the power conservation principle, obtain the customer power supply phase label, and complete customer phase identification based on voltage fluctuation significance analysis.

[0105] A further implementation method is that the significant fluctuation difference segment extraction module includes:

[0106] A high load period selection unit is used to obtain a high load period based on a daily load curve of the transformer main port;

[0107] A window length candidate set establishing unit, used to preset a window length candidate set and fix the step length during a high load period;

[0108] A correlation coefficient calculation unit is used to calculate the Pearson correlation coefficient between the voltage windows at the head end of the three-phase feeder when the window length takes different values ​​in the window length candidate set, starting from the first voltage data point of the complete voltage time series of the three-phase feeder;

[0109] A maximum window length acquisition unit is used to compare the mean value of the Pearson correlation coefficient between the voltage windows at the head end of the three-phase feeder with a preset similarity threshold to obtain a maximum window length;

[0110] The significant fluctuation difference period acquisition unit is used to set the maximum window length as the actual window length of the current operation, and record the window period of the maximum window length as the significant fluctuation difference period.

[0111] In a further embodiment, the first phase matching module comprises:

[0112] A sequence splicing unit is used to re-splice the significant fluctuation difference segments in time sequence to obtain a three-phase feeder phase detection sequence;

[0113] The phase matching unit is used to calculate the comprehensive similarity coefficient between the complete voltage time series of all customers and the phase detection sequence of the three-phase feeder, and use the comprehensive similarity coefficient as a judgment index to perform customer power supply phase matching;

[0114] The phase ambiguity customer acquisition unit is used to determine the customer power supply phase and obtain the customer power supply phase label when the comprehensive similarity coefficient meets the preset conditions; otherwise, the customer who fails to match the phase is regarded as a phase ambiguity customer.

[0115] Embodiment 3

[0116] Two typical low-voltage distribution area datasets of different area types are selected to test and evaluate the proposed method. The first dataset comes from a main commercial substation power supply network in a city. It includes data of transformers and their related power supply loads. All loads in this dataset are three-phase loads, and the original data are collected at a high frequency of 1 sample per second. The second dataset comes from the "IEEE European Low Voltage Test Feeder" (hereinafter referred to as TestFeeder), which was provided by the Test Feeders Working Group of the Distribution System Analysis Subcommittee of the Power Systems Analysis, Computing, and Economics in 2015. This dataset contains a total of 906 nodes in the entire substation area, of which 55 nodes represent load customers, all of which are single-phase customers. In addition, the dataset also provides customer power consumption data at a frequency of 1 sample per minute, as well as detailed information on the model and impedance parameters of all lines. The voltage data of the transformer and customers are obtained by performing time-series power flow calculations using OpenDSS software. Taking into account the versatility of the algorithm, the most common data collection frequency, i.e. 1 sample every 15 minutes, was used during the test.

[0117] The key parameters and values ​​involved in the construction of the comprehensive similarity index, the determination of the high-load period, and the dynamic sliding window stage of the present invention are shown in Table 1 (key parameters of the algorithm).

[0118] Table 1

[0119]

[0120] In terms of evaluation indicators, in addition to the commonly used precision in classification accuracy, the inter-phase average correlation coefficient IT-PCC (Inter-phase Average Pearson Correlation Coefficient), inter-phase composite variation ITV (Inter-phase Composite Variance), and inner-phase similarity INS (Inner-phase Similarity) are added. The specific expressions are as follows:

[0121]

[0122] According to the above experimental configuration, training and testing were carried out, and the test results were as follows: Table 2 (comprehensive similarity coefficient on the segment with significant fluctuation difference in the commercial area); Table 3 (comprehensive similarity coefficient on the segment with significant fluctuation difference in the Test Feeder); Table 4 (phase identification results of the two areas).

[0123] Table 2

[0124]

[0125] Table 3

[0126]

[0127] Table 4

[0128]

[0129] The comparative test results in commercial substations dominated by three-phase customers and residential substations dominated by single-phase customers show that there are obvious differences in the comprehensive similarity coefficients between different feeders in the sections with significant fluctuation differences among customers in the two substations, and the comprehensive differences between phases are both above 86%. INS is significantly lower than ITV, and the accuracy rates are both above 97%, indicating that the method proposed in the present invention is applicable to both substations.

[0130] The embodiments described above are only descriptions of the preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the design spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by ordinary technicians in this field should all fall within the protection scope determined by the claims of the present invention.

Claims

1. A method for identifying customer phases in a low-voltage distribution network based on voltage fluctuation significance analysis, characterized in that: The following steps are involved: S1: Obtain the complete voltage time series and complete power time series of the three-phase feeder and all customers; S2: Using a dynamic sliding window model, extract the significant fluctuation difference segment of the complete voltage time series of the three-phase feeder; the method of step S2 is: Based on the daily load curve of the transformer main port, obtain the high load period; During the high load period, a window length candidate set is preset and the step length is fixed; Starting from the first voltage data point of the complete voltage time series of the three-phase feeder, respectively calculating the Pearson correlation coefficients between the voltage windows at the head end of the three-phase feeder when the window length takes different values ​​in the window length candidate set; The mean value of the Pearson correlation coefficient between the voltage windows at the head end of the three-phase feeder is compared with a preset similarity threshold to obtain a maximum window length; Taking the maximum window length as the actual window length of the current operation, and recording the window period of the maximum window length as the significant fluctuation difference period; S3: Based on the significant fluctuation difference segment and the complete voltage time series of all customers, customer power supply phase matching is performed. If the matching is successful, the customer power supply phase label is obtained, and customer phase identification based on voltage fluctuation significance analysis is completed; otherwise, the customer who fails the phase matching is regarded as a phase fuzzy customer, and step S4 is executed; the method for performing customer power supply phase matching in step S3 is: Reconnect the significant fluctuation difference segments in time sequence to obtain a three-phase feeder phase detection sequence; Calculate the comprehensive similarity coefficient between the complete voltage time series of all customers and the three-phase feeder phase detection sequence, and use the comprehensive similarity coefficient as a determination index to perform customer power supply phase matching; When the comprehensive similarity coefficient meets the preset conditions, the customer power supply phase is determined and the customer power supply phase label is obtained; otherwise, the customer who fails to match the phase is regarded as a phase ambiguous customer; The method for calculating the comprehensive similarity coefficient is: Based on the mean and time series length of the customer's complete voltage time series and the three-phase feeder phase detection sequence, respectively, obtaining the Pearson correlation coefficient between the sequences; Based on the Pearson correlation coefficient between the sequences, the correlation coefficient matrix is ​​obtained; Calculate the average Euclidean distance between sequences to obtain the distance matrix; Normalizing the distance matrix to obtain a comprehensive similarity matrix; Based on the comprehensive similarity matrix, obtaining the comprehensive similarity coefficients of the complete voltage time series of all customers and the three-phase feeder phase detection sequence; S4: Based on the complete power time series of the three-phase feeder and the complete power time series of all customers, the power conservation principle is adopted to perform power supply phase matching for the phase-ambiguous customer, obtain the customer power supply phase label, and complete customer phase identification based on voltage fluctuation significance analysis.

2. The method for identifying the phase of a low-voltage power distribution network customer based on voltage fluctuation significance analysis according to claim 1, characterized in that: In step S4, the formula for power supply phase matching for the phase-ambiguous customer is: Among them, p i,t represents the power of the i-phase feeder at time t; u represents the number of customers whose power supply phase is determined to be i-phase, p j,t represents the power of customer j at time t whose power supply phase is determined to be phase i; p 0,t represents the power of the phase-ambiguous customer at time t; A, B, and C represent the three-phase feeder, Δp i Power supply phase for phase-ambiguous customers.

3. A low-voltage distribution network customer phase identification system based on voltage fluctuation significance analysis, used to implement the customer phase identification method according to any one of claims 1-2, characterized in that: include: A sequence reading module is used to obtain the complete voltage time series and the complete power time series of the three-phase feeder and all customers; A significant fluctuation difference segment extraction module is used to extract the significant fluctuation difference segment of the complete voltage time series of the three-phase feeder using a dynamic sliding window model; The first phase matching module is used to perform customer power supply phase matching based on the significant fluctuation difference segment and the complete voltage time series of all customers. If the matching is successful, the customer power supply phase label is obtained to perform customer phase identification based on voltage fluctuation significance analysis; otherwise, the customer who fails the phase matching is regarded as a phase fuzzy customer, and the second phase matching module is executed; The second phase matching module is used to perform power supply phase matching for the phase-ambiguous customer based on the complete power time series of the three-phase feeder and the complete power time series of all customers, adopt the power conservation principle, obtain the customer power supply phase label, and complete the customer phase identification based on voltage fluctuation significance analysis.

4. The low voltage distribution network customer phase identification system based on voltage fluctuation significance analysis according to claim 3 is characterized in that: The significant fluctuation difference segment extraction module comprises: A high load period selection unit is used to obtain a high load period based on a daily load curve of the transformer main port; A window length candidate set establishing unit, used for presetting a window length candidate set and fixing the step length during the high load period; A correlation coefficient calculation unit is used to calculate the Pearson correlation coefficient between the voltage windows at the head end of the three-phase feeder when the window length takes different values ​​in the window length candidate set, starting from the first voltage data point of the complete voltage time series of the three-phase feeder; A maximum window length acquisition unit is used to compare the mean value of the Pearson correlation coefficient between the voltage windows at the head end of the three-phase feeder with a preset similarity threshold to obtain a maximum window length; The significant fluctuation difference period acquisition unit is used to use the maximum window length as the actual window length of the current operation, and record the window period of the maximum window length as the significant fluctuation difference period.

5. The low voltage distribution network customer phase identification system based on voltage fluctuation significance analysis according to claim 3 is characterized in that: The first phase matching module comprises: A sequence splicing unit, used for re-splicing the significant fluctuation difference segments in time sequence to obtain a three-phase feeder phase detection sequence; A phase matching unit, used to calculate the comprehensive similarity coefficient between the complete voltage time series of all customers and the three-phase feeder phase detection sequence, and use the comprehensive similarity coefficient as a judgment index to perform customer power supply phase matching; The phase ambiguity customer acquisition unit is used to determine the customer power supply phase and obtain the customer power supply phase label when the comprehensive similarity coefficient meets the preset conditions; otherwise, the customer who fails the phase matching is regarded as a phase ambiguity customer.

Images