A method for identifying line-transformer relationships in distribution networks based on GIS and voltage data fusion

Abstract

The present invention relates to a method for identifying the relationship between distribution network lines and transformers based on GIS and voltage data fusion, which obtains line voltage sequence, distribution transformer outlet voltage and distribution transformer geographical distribution data from a data acquisition and monitoring system and a geographic information system, and calculates the distribution transformer outlet voltage with three-phase load imbalance to obtain a distribution transformer voltage sequence; extracts multiple feature quantities from the line voltage sequence and the distribution transformer voltage sequence, and constructs a multi-feature vector, uses the multi-feature vector to train an autoencoder, and constructs an abnormal feeder-transformer relationship identification model for line-transformer relationship identification; and jointly uses the distribution transformer geographical distribution and its voltage characteristic relationship to verify and correct the line-transformer relationship. By constructing a multi-feature vector, the present invention can better characterize the abnormal characteristics of the feeder-transformer relationship and overcome the limitations of the traditional threshold judgment method, and verifies and corrects the line-transformer relationship through the distribution transformer geographical distribution data, thereby improving the accuracy of line-transformer relationship identification.

Description

Technical Field

[0001] The present invention relates to the technical field of online monitoring of electric power metering, and in particular to a method for identifying line-transformer relationships in a distribution network based on GIS and voltage data fusion. Background Art

[0002] In the distribution network of the power system, the distribution network topology data is the key information describing the physical connection and electrical characteristics of the distribution network, which is crucial for power grid companies to carry out efficient management, optimize operation and improve service quality. Among them, the connection relationship between the power supply line and the distribution transformer (commonly referred to as "distribution transformer") (referred to as "line-to-transformer relationship") is one of the most critical links.

[0003] In the actual operation of the distribution network, due to the transformation, reverse load and sudden failure of the distribution network, the adjusted distribution network line-transformer relationship information is manually verified and changed. This process is prone to untimely updates and maintenance errors. These errors may cause the archive data in the geographic information system (GIS) to be inconsistent with the actual power grid structure, directly affecting the smooth development of multiple businesses such as distribution network fault location, line loss management and power flow analysis. Therefore, timely identification and verification of distribution network line-transformer relationship data is of great significance to ensure the normal operation of related business systems, improve work efficiency and save operating costs. At present, the traditional method of verifying the distribution network line-transformer relationship by manual line patrol is labor-intensive and inefficient, and it is difficult to meet the needs of modern power grids. Intelligent identification of distribution network line-transformer relationship through GIS and voltage data fusion is an effective analysis method. Summary of the invention

[0004] The traditional manual line patrol or power carrier to check abnormal line-transformer relationship has the disadvantages of heavy workload, low efficiency and high cost, which makes it difficult to adapt to the increasingly complex and frequently changing modern distribution network. A distribution network line-transformer relationship identification method based on GIS (geographic information system) and voltage data fusion is proposed, which can verify and correct the line-transformer relationship of the distribution network.

[0005] The present invention is implemented by the following technical solution: a method for identifying line-transformer relationships in a distribution network based on GIS and voltage data fusion, the steps of which are as follows:

[0006] S1: Obtain the line voltage sequence, distribution transformer outlet voltage and distribution transformer geographical distribution data from the data acquisition and monitoring system and the geographic information system, and calculate the distribution transformer outlet voltage with three-phase load imbalance to obtain the distribution transformer voltage sequence;

[0007] S2: Extract multiple feature quantities from the line voltage sequence and the distribution transformer voltage sequence, and construct a multi-feature vector. Use the multi-feature vector to train the autoencoder and build an abnormal feeder-transformer relationship recognition model for line-transformer relationship recognition.

[0008] S3: Verify and correct the line-to-line relationship by using the geographical distribution of distribution transformers and their voltage characteristics;

[0009] According to the identification results of line-transformer relationship, the abnormal distribution transformer and its geographical distribution latitude and longitude are determined. The geographical location of the abnormal distribution transformer is taken as the center and the latitude and longitude are taken as the center. 0 A circular area is formed with a radius of 1, and the lines including the distribution transformer in the circular area are regarded as possible lines of the abnormal distribution transformer. Then, the Pearson correlation coefficient and dynamic time bending distance measurement value of the abnormal distribution transformer and the possible lines and distribution transformer are calculated and sorted. The line with the largest Pearson correlation coefficient and the smallest dynamic time bending distance measurement value is regarded as the line of the abnormal distribution transformer to form a revised line-transformer relationship. At the same time, the revised line-transformer relationship is recalculated and the daily line loss rate of the line is tracked to ensure the correctness of the line-transformer relationship correction.

[0010] Specifically, in step S1, the voltage at the distribution transformer outlet with unbalanced three-phase load is calculated as follows:

[0011] A, B, C are three-phase distribution transformer outlets, N is the neutral point, N' is the neutral point when offset. When the neutral point is offset, the A, B, C three-phase distribution transformer outlet voltages are AN', BN', CN' respectively. After calculating the distribution transformer outlet voltage, the distribution transformer voltage sequence is obtained:

[0012] (1);

[0013] (2);

[0014] (3);

[0015] in, represents the cosine of the angle between side BN' and side AN', represents the cosine of the angle between side BN' and side CN', represents the cosine value of the angle between side CN' and side AN', AB represents the pressure difference between phase A and phase B, BC represents the pressure difference between phase B and phase C, and AC represents the pressure difference between phase A and phase C;

[0016] According to formula (1) - formula (3), we can get:

[0017] (4);

[0018] Use the iterative method to obtain the values ​​of AB, BC, and AC. AN, BN and CN, that is, the three-phase distribution transformer outlet voltage after conversion, can be calculated to form the distribution transformer voltage sequence.

[0019] Specifically, the multiple feature vectors in step S2 include Pearson correlation coefficient, dynamic time warping distance metric, relative coefficient of variation, skewness and kurtosis.

[0020] Specifically, the Pearson correlation coefficient r is expressed by the following formula:

[0021] (5);

[0022] in, and represent the i-th data point of the voltage sequence of distribution transformer X and distribution transformer Y respectively; and represents the mean of the voltage series of distribution transformer X and distribution transformer Y respectively, and n is the number of data points; the Pearson correlation coefficient series of the target distribution transformer and other distribution transformers on the same line is calculated according to formula (5): , then calculate the average of the Pearson correlation coefficient series .

[0023] Specifically, the dynamic time bending distance metric of two distribution transformer voltage sequences is expressed as follows:

[0024] (6);

[0025] in, and denotes the ith data point of the voltage sequence of distribution transformer X and distribution transformer Y, respectively. is a mapping function.

[0026] Specifically, the relative coefficient of variation is expressed by the following formula:

[0027] (7) ;

[0028] (8);

[0029] in, and They represent the coefficient of variation of the target distribution transformer and line voltage series, and They represent the standard deviation and mean value of the target distribution transformer voltage series, and They represent the standard deviation and mean value of the line voltage series, Represents the relative coefficient of variation of the target distribution transformer and the voltage sequence of the line where it is located.

[0030] Specifically, skewness It is expressed by the following formula:

[0031] p sk = E [ ( V i − μ i σ i ) 3 ] (9);

[0032] Among them, E[•] represents the expected value operation, Represents the i-th data point of the target distribution transformer voltage sequence.

[0033] Specifically, kurtosis It is expressed by the following formula:

[0034] f ku = E [ ( V i − μ i σ i ) 4 ] (10).

[0035] Specifically, the reconstruction error of the autoencoder is used to detect the line-transformer relationship, and the root mean square error is used to measure the reconstruction error. By setting a threshold, if the reconstruction error is greater than the threshold, the feeder-transformer relationship is wrong, and if the reconstruction error is not greater than the threshold, the feeder-transformer relationship is correct, thereby realizing the identification of abnormal line-transformer relationship.

[0036] By constructing multiple feature vectors, the present invention can better characterize the abnormal characteristics of the feeder-transformer relationship and overcome the limitations of the traditional threshold judgment method. The line-transformer relationship is verified and corrected through the geographical distribution data of distribution transformers, thereby improving the accuracy of line-transformer relationship identification. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Picture 1 This is a schematic diagram of the neutral point displacement on the high-voltage side of the Yyn0 connection distribution transformer.

[0038] Picture 2 Schematic diagram of the line-variable relationship identification process. DETAILED DESCRIPTION

[0039] The present invention is further described below in conjunction with embodiments. It is necessary to point out that the following embodiments are only used to further illustrate the present invention and cannot be understood as limiting the scope of protection of the present invention. Some non-essential improvements and adjustments made by technical personnel in this field based on the above invention content still fall within the scope of protection of the present invention.

[0040] A method for identifying line-transformer relationships in distribution networks based on GIS and voltage data fusion, the steps are as follows:

[0041] S1: Obtain the line voltage sequence, distribution transformer outlet voltage and distribution transformer geographical distribution data from the data acquisition and monitoring system and the geographic information system, and calculate the distribution transformer outlet voltage with three-phase load imbalance to obtain the distribution transformer voltage sequence;

[0042] S2: Extract multiple feature quantities from the line voltage sequence and distribution transformer voltage sequence, and construct a multi-feature vector. In order to better characterize the abnormal characteristics of the feeder-transformer relationship and overcome the limitations of the traditional threshold judgment method, the multi-feature vector is used to train the autoencoder algorithm and construct an abnormal feeder-transformer relationship recognition model for line-transformer relationship recognition;

[0043] S3: Verify and correct the line-to-line relationship by using the geographical distribution of distribution transformers and their voltage characteristics.

[0044] Specifically, refer to Picture 1 In step S1, the three-phase load unbalance distribution transformer outlet voltage is calculated, specifically:

[0045] Picture 1 In the figure, A, B, and C are three-phase distribution transformer outlets, N is the neutral point, and N' is the neutral point when offset. When the neutral point is offset, the A, B, and C three-phase distribution transformer outlet voltages are AN', BN', and CN' respectively. After calculating the distribution transformer outlet voltage, the distribution transformer voltage sequence is obtained:

[0046] (1);

[0047] (2);

[0048] (3);

[0049] in, represents the cosine of the angle between side BN' and side AN', represents the cosine of the angle between side BN' and side CN', represents the cosine value of the angle between side CN' and side AN', AB represents the pressure difference between phase A and phase B, BC represents the pressure difference between phase B and phase C, and AC represents the pressure difference between phase A and phase C;

[0050] According to formula (1) - formula (3), we can get:

[0051] (4);

[0052] Use the iterative method to obtain the values ​​of AB, BC, and AC. AN, BN and CN, that is, the three-phase distribution transformer outlet voltage after conversion, can be calculated to form the distribution transformer voltage sequence.

[0053] Specifically, the multiple feature vectors in step S2 include Pearson correlation coefficient, dynamic time warping (Dynamic TimeWarping, DTW) distance metric, relative coefficient of variation (Relative Coefficient of Variation, RCV), skewness and kurtosis.

[0054] Specifically, the method for obtaining the Pearson correlation coefficient is:

[0055] The Pearson correlation coefficient, also known as the Pearson product-moment correlation coefficient, is a statistical indicator that measures the degree of linear correlation between two variables. Its value is between -1 and 1, where 1 indicates a perfect positive correlation, -1 indicates a perfect negative correlation, and 0 indicates no linear correlation. The Pearson correlation coefficient r can be expressed as follows:

[0056] (5);

[0057] in, and represent the i-th data point of the voltage sequence of distribution transformer X and distribution transformer Y respectively; and represents the mean of the voltage series of distribution transformer X and distribution transformer Y, respectively, and n is the number of data points. The Pearson correlation coefficient series of the target distribution transformer and other distribution transformers on the same line is calculated according to formula (5): , then calculate the average of the Pearson correlation coefficient series .

[0058] Specifically, the dynamic time warping distance measurement acquisition method is:

[0059] The dynamic time warping distance metric can reflect the fluctuation trend between distribution transformer voltage sequences and has a high sensitivity to the fluctuation trend between different distribution transformer voltage sequences. If the dynamic time warping distance metric distance between two distribution transformer voltage sequences is smaller, the similarity between the two distribution transformer voltage sequences is higher and the electrical distance between the two distribution transformers is closer.

[0060] The dynamic time bending distance metric of two distribution transformer voltage sequences can be expressed as follows:

[0061] (6);

[0062] in, and denotes the ith data point of the voltage sequence of distribution transformer X and distribution transformer Y, respectively. is a mapping function used to find the voltage sequence of the distribution transformer Y that corresponds to The best matching element.

[0063] Specifically, the relative coefficient of variation is obtained as follows:

[0064] The relative coefficient of variation (RCV) is a dimensionless statistical indicator used to compare the discreteness or volatility of two distribution transformer voltage sequences. It can be expressed as follows:

[0065] (7) ;

[0066] (8);

[0067] in, and They represent the coefficient of variation of the target distribution transformer and line voltage series, and They represent the standard deviation and mean value of the target distribution transformer voltage series, and They represent the standard deviation and mean value of the line voltage series, Represents the relative coefficient of variation of the target distribution transformer and the voltage sequence of the line where it is located.

[0068] Specifically, the method for obtaining skewness is:

[0069] Skewness It is a measure of the asymmetry of the statistical data distribution and can be expressed as follows:

[0070] p sk = E [ ( V i − μ i σ i ) 3 ] (9);

[0071] Among them, E[•] represents the expected value operation, Represents the i-th data point of the target distribution transformer voltage sequence.

[0072] Specifically, the kurtosis acquisition method is:

[0073] Kurtosis The characteristic number that characterizes the peak height of the probability density distribution curve at the mean value can be expressed by the following formula:

[0074] f ku = E [ ( V i − μ i σ i ) 4 ] (10);

[0075] Specifically, the autoencoder is trained using multiple feature vectors to construct an abnormal feeder-transformer relationship recognition model for line-transformer relationship recognition, including:

[0076] The autoencoder consists of two parts: the encoder and the decoder. The encoder encodes the original input data into a low-dimensional internal representation, i.e. encoding, and the decoder restores the low-dimensional representation obtained in the encoding stage to an output similar to the original input data. This process is called decoding. The reconstruction error of the autoencoder is used to detect abnormal line-to-transformer relationships, that is, distribution transformer voltage samples that are out of the characteristics of the distribution transformer voltage sequence of the same line. The basic principle is that after training with multiple feature vectors, the autoencoder will have the ability to reconstruct reconstructed feature data that is in the same distribution as the multiple feature vectors. If a reconstructed feature data and a multiple feature vector belong to different distribution laws, there will be a large error between the reconstructed feature data and the multiple feature vectors themselves. After defining the reconstruction error between the input and its reconstruction result, by setting a threshold, if the reconstruction error is greater than the threshold, the feeder-transformer relationship is wrong, and if the reconstruction error is not greater than the threshold, the feeder-transformer relationship is correct, thereby realizing the line-to-transformer relationship recognition.

[0077] Specifically, the trained autoencoder is used to perform line-variant relationship recognition, and the process is as follows:

[0078] Reference Picture 2 , the multi-feature vector constructed by the distribution voltage sequence to be identified is input into the trained autoencoder model for reconstruction. The root mean square error (RMSE) is used to measure the reconstruction error, that is:

[0079] (11);

[0080] in, Multiple feature vectors representing the input And the reconstructed feature data output The root mean square error, is the jth data element of the multi-feature vector, The root mean square error is recorded for the jth data element of the reconstructed feature data, where m is the data amount of the multi-feature vector. When the root mean square error is greater than the threshold, it is determined that the distribution transformer does not belong to the line and the feeder-transformer relationship is wrong. This approach can improve the recognition rate.

[0081] Specifically, in step S3, the geographical distribution of distribution transformers and their voltage characteristic relationship are used together to verify and correct the line transformer relationship, and the process is as follows:

[0082] Collect the geographical distribution data of distribution transformers. Since the geographical distribution of distribution transformers on the same line is similar, the abnormal distribution transformer and its geographical distribution latitude and longitude can be determined according to the line-transformer relationship identification results. The geographical location of the abnormal distribution transformer is taken as the center and the longitude and latitude of the abnormal distribution transformer are taken as the center. 0A circular area is formed with a radius of 0.1, and the lines including the distribution transformer in the circular area are regarded as the possible lines to which the abnormal distribution transformer belongs. Then, the Pearson correlation coefficient and the dynamic time bending distance measurement value of the abnormal distribution transformer and the possible lines and distribution transformers are calculated and sorted. The line with the largest Pearson correlation coefficient and the smallest dynamic time bending distance measurement value is regarded as the line to which the abnormal distribution transformer belongs, forming a revised line-transformer relationship. At the same time, the revised line-transformer relationship is recalculated and the daily line loss rate of the line is tracked to ensure the correctness of the line-transformer relationship correction.

[0083] If the geographical location of the abnormal distribution transformer is not at the edge of all distribution transformers under the line to which it belonged before correction, and is far away from the distribution transformers of other lines, the voltage data of other dates (including the distribution transformer outlet voltage and line voltage) should be re-collected and the line-transformer relationship should be re-identified to avoid errors caused by accidental errors in voltage data.

[0084] In another embodiment, a non-volatile computer storage medium is provided, which stores computer executable instructions, and the computer executable instructions can execute the distribution network line-transformer relationship identification method based on GIS and voltage data fusion in the above embodiment.

[0085] This embodiment provides an electronic device, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute a distribution network line-transformer relationship identification method based on GIS and voltage data fusion.

[0086] The specific implementation scheme described above further illustrates the purpose, technical scheme and technical effect of the present invention in detail. It should be understood that the above description is only a specific implementation scheme of the present invention and is not intended to limit the scope of the present invention. Any equivalent changes and modifications made by any technician in the field without departing from the ideas and principles of the present invention should fall within the scope of protection of the present invention.

Claims

1. A method for identifying line-transformer relationships in a distribution network based on GIS and voltage data fusion, characterized in that: Here are the steps: S1: Obtain the line voltage sequence, distribution transformer outlet voltage and distribution transformer geographical distribution data from the data acquisition and monitoring system and the geographic information system, and calculate the distribution transformer outlet voltage with three-phase load imbalance to obtain the distribution transformer voltage sequence; S2: Extract multiple feature quantities from the line voltage sequence and distribution transformer voltage sequence, and construct a multi-feature vector. The multi-feature vector includes the Pearson correlation coefficient, dynamic time warping distance metric, relative coefficient of variation, skewness, and kurtosis. Use the multi-feature vector to train the autoencoder and build an abnormal feeder-transformer relationship recognition model for line-transformer relationship recognition. The dynamic time bending distance metric of two distribution transformer voltage sequences is expressed as follows: （6）； in, and denotes the ith data point of the voltage sequence of distribution transformer X and distribution transformer Y, respectively. is a mapping function, n is the number of data points; The relative coefficient of variation is expressed by the following formula: （7） ； （8）； in, and They represent the coefficient of variation of the target distribution transformer and line voltage series, and They represent the standard deviation and mean value of the target distribution transformer voltage series, and They represent the standard deviation and mean value of the line voltage series, Represents the relative coefficient of variation of the target distribution transformer and the voltage sequence of the line where it is located; S3: Verify and correct the line-to-line relationship by using the geographical distribution of distribution transformers and their voltage characteristics; According to the line-transformer relationship identification results, the abnormal distribution transformer and its geographical distribution longitude and latitude are determined. A circular area is formed with the geographical location of the abnormal distribution transformer as the center and r0 as the radius. The line to which the distribution transformer is included in the circular area is regarded as the possible line to which the abnormal distribution transformer belongs. Then, the Pearson correlation coefficient and dynamic time bending distance measurement value of the abnormal distribution transformer and the possible lines and distribution transformers are calculated and sorted. The line with the largest Pearson correlation coefficient and the smallest dynamic time bending distance measurement value is regarded as the line to which the abnormal distribution transformer belongs, forming a revised line-transformer relationship. At the same time, the revised line-transformer relationship is recalculated and the daily line loss rate of the line is tracked to ensure the correctness of the line-transformer relationship correction.

2. The method for identifying line-transformer relationships in a distribution network based on GIS and voltage data fusion according to claim 1 is characterized in that: In step S1, the voltage at the distribution transformer outlet with unbalanced three-phase load is calculated as follows: A, B, C are three-phase distribution transformer outlets, N is the neutral point, N' is the neutral point when offset. When the neutral point is offset, the A, B, C three-phase distribution transformer outlet voltages are AN', BN', CN' respectively. After calculating the distribution transformer outlet voltage, the distribution transformer voltage sequence is obtained: （1）； （2）； （3）； in, represents the cosine of the angle between side BN' and side AN', represents the cosine of the angle between side BN' and side CN', represents the cosine value of the angle between side CN' and side AN', AB represents the pressure difference between phase A and phase B, BC represents the pressure difference between phase B and phase C, and AC represents the pressure difference between phase A and phase C; According to formula (1) - formula (3), we can get: （4）； Use the iterative method to find the values ​​of AB, BC, and AC. AN, BN and CN, that is, the three-phase distribution transformer outlet voltage after conversion, can be calculated to form the distribution transformer voltage sequence.

3. The method for identifying line-transformer relationships in a distribution network based on GIS and voltage data fusion according to claim 1 is characterized in that: The Pearson correlation coefficient r is expressed as follows: （5）； in, and represent the i-th data point of the voltage series of distribution transformer X and distribution transformer Y respectively; and represent the mean of the voltage series of distribution transformer X and distribution transformer Y respectively; the Pearson correlation coefficient series of the target distribution transformer and other distribution transformers on the same line is calculated according to formula (5): , then calculate the average of the Pearson correlation coefficient series .

4. The method for identifying line-transformer relationships in a distribution network based on GIS and voltage data fusion according to claim 1 is characterized in that: Skewness It is expressed by the following formula: （9）； Among them, E[•] represents the expected value operation, Represents the i-th data point of the target distribution transformer voltage sequence.

5. The method for identifying line-transformer relationships in a distribution network based on GIS and voltage data fusion according to claim 4 is characterized in that: Kurtosis It is expressed by the following formula: （10）。 6. The method for identifying line-transformer relationships in a distribution network based on GIS and voltage data fusion according to claim 1 is characterized in that: The reconstruction error of the autoencoder is used to detect the linear variation relationship, and the root mean square error is used to measure the reconstruction error; By setting a threshold, if the reconstruction error is greater than the threshold, the feeder-transformer relationship is wrong, and if the reconstruction error is not greater than the threshold, the feeder-transformer relationship is correct, thereby realizing the identification of abnormal line-transformer relationship.

7. The method for identifying line-transformer relationships in a distribution network based on GIS and voltage data fusion according to claim 1 is characterized in that: If the geographical location of the abnormal distribution transformer is not at the edge of all distribution transformers under the line to which it belongs before correction, and is far away from the locations of distribution transformers on other lines, the voltage data of other dates should be re-collected to re-identify the line-transformer relationship.

Images