Power distribution network topology change type identification method and system based on second-order gradient constraint heterogeneous correlation theory

Abstract

The invention provides a power distribution network topology change type identification method and system based on a second-order gradient constraint heterogeneous correlation theory. The method comprises the following steps: (1) collecting node voltage data of a power distribution network; (2) constructing a power distribution network voltage matrix and a first-order gradient matrix; (3) respectively calculating a second-order gradient heterogeneous correlation matrix of the voltage matrix and the first-order gradient matrix; (4) respectively calculating an autocorrelation matrix of a second-order gradient heterogeneous correlation matrix of the voltage and the gradient matrix; and (5) respectively calculating topology correlation identification factors according to the autocorrelation coefficients, and identifying the topology change type of the power distribution network according to calculation results. In addition, the invention also discloses a power distribution network topology change type identification system based on the second-order gradient constraint heterogeneous correlation theory, which comprises a voltage data acquisition device, a data processing device and a power distribution network topology change type identification device for executing the power distribution network topology change type identification method.

Description

Technical Field

[0001] This invention relates to a method for identifying the state of a power distribution network, and more particularly to a method and system for identifying the topology change type of a power distribution network based on the second-order gradient constraint heterogeneous correlation theory. Background Technology

[0002] With socio-economic development and rising living standards, the demands for the quality and reliability of power supply are increasing. Smart grids, as an upgraded version of traditional power grids, aim to improve the efficiency, security, and adaptability of the power grid through information technology, automation, and interactivity. Currently, new business models such as demand-side management, microgrids, and the energy internet are emerging, placing higher demands on the flexibility and transparency of distribution systems. State awareness technologies, especially those addressing topology event changes, can provide data support for these new businesses, helping to optimize resource allocation and improve market operational efficiency.

[0003] Therefore, to improve the state awareness capability of distribution networks, it is desirable to obtain a method and system for identifying distribution network topology change types based on second-order gradient-constrained heterogeneous correlation theory. This method can detect and identify the types of topology changes in the network by measuring the voltage amplitude of multiple sensor nodes, and simultaneously provide a quantitative method for identification and differentiation, thus creating conditions for achieving refined management and intelligent decision-making of the power grid. Summary of the Invention

[0004] One of the objectives of this invention is to provide a method and system for identifying topology change types in distribution networks based on second-order gradient constrained heterogeneous correlation theory. By collecting voltage amplitude values ​​from multiple sensors and combining them with second-order gradient constrained heterogeneous correlation theory, the method detects and identifies the types of topology changes in the distribution network. The method achieves good quantitative differentiation results, enables real-time monitoring of the operating status of the distribution network, timely detection of potential faults, and assists the decision-making system in rapid response, effectively improving change efficiency.

[0005] In accordance with the above-mentioned objectives, this invention provides a method and system for identifying the topology change type of a power distribution network based on the second-order gradient constraint heterogeneous correlation theory.

[0006] The steps are as follows:

[0007] (1) Collect node voltage data of the power distribution network;

[0008] (2) Construct the voltage matrix and first-order gradient matrix of the power distribution network;

[0009] (3) Calculate the second-order gradient heterogeneous correlation matrix of the voltage matrix and the first-order gradient matrix respectively;

[0010] (4) Calculate the autocorrelation matrix of the second-order gradient heterogeneous correlation matrix of the voltage and gradient matrices respectively;

[0011] (5) Calculate the topological correlation recognition factors according to the autocorrelation coefficients respectively, and identify the types of topological changes in the distribution network according to the calculation results.

[0012] In the method for identifying the type of topological change in a distribution network based on the second-order gradient constraint heterogeneous correlation theory described in this invention, by perceiving the continuity in multi-node data and using the gradient change to constrain the range of topological change perception, the second-order gradient and characteristic distribution of the voltage data obtained by measurement in the distribution network are studied, and it is found that compared with before the topological change, after the topological change occurs in the network, the characteristic distribution of heterogeneous data under the second-order gradient constraint will change, and different types of topological changes correspond to different characteristic factors. Therefore, by analyzing the transformation situation of the above characteristic factors, that is, the correlation recognition factors, the topological change situation can be identified and classified.

[0013] Further, in step 1, the monitoring sensors of the distribution network are used to collect the voltage data of the topological nodes. In order to save the subsequent calculation cost, the processed voltage data is obtained by using periodic local truncated mean processing.

[0014] Let X = [x1, x2,..., x n be the single-node voltage data, and let the local window size be k (k < n), then the truncated data is X i = [x i , x i+1 ,..., x i+k-1 , and the local mean of the truncated data is obtained Let the processed voltage data X c = [A1, A2,..., A i .

[0015] Further, in step 2, it specifically includes:

[0016] Let the voltage matrix be A (N*R) , and let the first-order gradient matrix be G (R*N) , where N represents the number of sensors or the number of nodes for collecting data, and R represents the length of the processed voltage data X c .

[0017]

[0018]

[0019] Further, the second-order gradient heterogeneous correlation matrix in step 3 is specifically as follows.

[0020] Let the second-order gradient heterogeneous correlation matrix R1 of the voltage matrix be:

[0021] Let R2 be the second-order gradient heterogeneous correlation matrix of the first-order gradient matrix:

[0022] Furthermore, in step 4, the autocorrelation matrix V of the second-order gradient heterogeneous correlation matrix is ​​calculated based on the following formula:

[0023] V=D -1 / 2 CD -1 / 2

[0024] Where D is a diagonal matrix, and its diagonal elements are the square roots of the diagonal elements of the covariance matrix C. -1 / 2 Let be the inverse square root matrix of D.

[0025] Furthermore, in step 5, the topology-related identification factor F is calculated based on the following formula:

[0026]

[0027] Where, ρ r and ρ n The diagonal elements of the correlation matrices R1 and R2 are the correlation coefficients, α and β are the maximum and minimum values ​​of the correlation coefficients, respectively, Γ() is the gamma function, N represents the number of sensors or the number of data acquisition nodes, and R represents the processed voltage data X. c The length.

[0028] Furthermore, the types of changes in the power distribution network topology include: normal, line connection change, node status change, and load transfer change.

[0029] Furthermore, in step 5, the topology-related identification factors are set as follows: F∈[0.75,1.05] represents the normal state, F∈[2.95e4,3.05e4] represents line connection changes, F∈[1.25e6,2.05e7] represents node status changes, and F∈[3.25e8,4.05e8] represents load transfer changes.

[0030] Furthermore, a distribution network topology change type identification system based on second-order gradient constrained heterogeneous correlation theory is characterized in that it includes a voltage data acquisition device, a data processing device, and a distribution network topology change type identification device, wherein the distribution network topology change type identification system based on second-order gradient constrained heterogeneous correlation theory executes the distribution network topology change type identification method.

[0031] This invention, by introducing specific voltage and first-order gradient matrices, can directly derive the optimal structure of matrix variables, thereby significantly improving the accuracy of optimization problems involving change type identification. The use of second-order gradient constraints not only makes the optimization process more efficient but also reveals the characteristics of the matrix equations, making the solution path for the identification problem more direct. Attached Figure Description

[0032] Figure 1 The flowchart of this invention

[0033] Figure 2 Trend chart of recognition performance of the method of the present invention under different local windows

[0034] Figure 3 This is a trend chart of the recognition performance of the method of the present invention under different acquisition lengths. Detailed Implementation

[0035] The following will further explain and illustrate the method for identifying the topology change type of distribution network based on the second-order gradient constraint heterogeneous correlation theory described in this invention, with reference to the accompanying drawings and specific embodiments. However, this explanation and illustration do not constitute an undue limitation on the technical solution of this invention.

[0036] In the distribution network topology change type identification method based on second-order gradient constraint heterogeneous correlation theory described in this invention, the continuity in multi-node data is sensed, and the range of topology change sensing is constrained by gradient change. The second-order gradient and characteristic distribution of voltage data obtained from measurements in the distribution network are studied. It is found that, compared to before the topology change, the characteristic distribution of heterogeneous data changes under the second-order gradient constraint after the topology change occurs in the network, and different types of topology changes correspond to different characteristic factors. Therefore, by analyzing the transformation of the aforementioned characteristic factors, i.e., the relevant identification factors, topology changes can be identified and classified.

[0037] Figure 1 This is a flowchart illustrating the steps of one implementation of the distribution network topology change type identification method based on second-order gradient constraint heterogeneous correlation theory described in this invention.

[0038] like Figure 1 As shown in this embodiment, the distribution network topology change type identification method based on second-order gradient constraint heterogeneous correlation theory of the present invention may include the following steps:

[0039] (1) Collect node voltage data of the power distribution network;

[0040] Voltage data of the topology nodes are collected using monitoring sensors of the power distribution network. To save subsequent calculation costs, periodic local truncation of the mean is used to obtain the processed voltage data.

[0041] Let \(X = [x_1, x_2, \ldots, x n \) be the single - node voltage data. Denote the local window size as \(k(k < n)\), then the truncated data is \(X i = [x i , x i+1 , \ldots, x i+k-1 \). Calculate the local mean of the truncated data Denote the processed voltage data \(X c = [A_1, A_2, \ldots, A i \).

[0042] (2) Construct the distribution network voltage matrix and the first - order gradient matrix; specifically including:

[0043] Denote the voltage matrix as \(A (N*R) \), and denote the first - order gradient matrix as \(G (R*N) \), where \(N\) represents the number of sensors or the number of nodes collecting data, and \(R\) represents the length of the processed voltage data \(X c .

[0044]

[0045]

[0046] (3) Calculate the second - order gradient heterogeneous correlation matrices of the voltage matrix and the first - order gradient matrix respectively; specifically:

[0047] Denote the second - order gradient heterogeneous correlation matrix \(R_1\) of the voltage matrix as:

[0048] Denote the second - order gradient heterogeneous correlation matrix \(R_2\) of the first - order gradient matrix as:

[0049] (4) Calculate the autocorrelation matrices of the second - order gradient heterogeneous correlation matrices of the voltage and gradient matrices respectively; calculate the autocorrelation matrix \(V\) of the second - order gradient heterogeneous correlation matrix:

[0050] V = D -1 / 2 CD -1 / 2

[0051] where \(D\) is a diagonal matrix, the diagonal elements of which are the square roots of the diagonal elements of the covariance matrix \(C\), and \(D -1 / 2 is the inverse square - root matrix of \(D\).

[0052] (5) Calculate the topological - related recognition factors according to the autocorrelation coefficients respectively and identify the types of topological changes in the distribution network according to the calculation results.

[0053] Let F∈[0.75,1.05] represent the normal state, F∈[2.95e4,3.05e4] represent line connection changes, F∈[1.25e6,2.05e7] represent node status changes, and F∈[3.25e8,4.05e8] represent load transfer changes.

[0054] Calculate the topology-related identification factor F:

[0055]

[0056] Where, ρ r and ρ n The diagonal elements of the correlation matrices R1 and R2 are the correlation coefficients, α and β are the maximum and minimum values ​​of the correlation coefficients, respectively, Γ() is the gamma function, N represents the number of sensors or the number of data acquisition nodes, and R represents the processed voltage data X. c The length.

[0057] A distribution network topology change type identification system based on second-order gradient constrained heterogeneous correlation theory includes a voltage data acquisition device, a data processing device, and a distribution network topology change type identification device. The distribution network topology change type identification system based on second-order gradient constrained heterogeneous correlation theory executes the distribution network topology change type identification method.

[0058] To better illustrate the application of the distribution network topology change type identification method based on second-order gradient constraint heterogeneous correlation theory described in this invention, simulation tests were conducted using an IEEE 123-node distribution network system for further explanation.

[0059] A distribution network topology change type identification system was used in the simulation test. This system may include a voltage data acquisition device and a processing module, as well as the distribution network topology change type identification system itself. In this invention, the distribution network topology change type identification system described herein can be used to execute the distribution network topology change type identification method described above. Topology change problems in the distribution network are detected, wherein an IEEE 123-node distribution network model is constructed according to the IEEE standard using simulation software, and experimental verification is performed in the network. Table 1 lists the changes in topology-related identification factors and the mean voltage matrix A under different topology changes in different distribution networks, where the topology changes may include: normal, line connection change, node status change, and load transfer change.

[0060] Topological changes F A normal 0.87 0.94 Changes in line connections 29741 3.14 Node state changes 20178 4.78 Load transfer changes 37614 8.94

[0061] As can be seen from Table 1, before the change in the distribution network topology, i.e. under normal conditions, the value of the topology-related identification factor F and the overall average value of the voltage amplitude are all around 1. However, after the topology changes, the topology-related identification factor will exceed the above range and reach about 3e4. Further quantitative differentiation can effectively identify the type of transformation.

[0062] Meanwhile, to test the impact of local window size on the final recognition result, a fixed acquisition length was used to obtain corresponding accuracy results for different windows, such as... Figure 2 As shown, the length of the acquisition voltage of a single-node sensor is set to 150. Different window sizes show different trends in recognition accuracy under different signal-to-noise ratios.

[0063] It should be noted that the signal-to-noise ratio is calculated as follows:

[0064] SNR = 20 * log(A) S / A N )

[0065] Among them, A S Indicates signal amplitude; A N This indicates the noise amplitude.

[0066] To test the impact of sensor node acquisition length on the final recognition result, the size of the local window was fixed, and the corresponding accuracy was obtained according to different acquisition lengths, such as... Figure 3 As shown, with the local window size set to 8, different window sizes exhibit different trends in recognition accuracy under different signal-to-noise ratios. When the acquisition length is greater than 210, the accuracy remains around 90%, which is difficult to improve, indicating that an acquisition length of 210 is an overfitting length.

[0067] It can be seen that by collecting voltage amplitudes from multiple sensors and combining them with the second-order gradient constraint heterogeneous correlation theory, the topological change types in the distribution network can be detected and identified. The identification and quantitative differentiation results are good, enabling real-time monitoring of the distribution network operation status, timely detection of potential faults, and assistance to the decision-making system for rapid response, effectively improving change efficiency.

[0068] By introducing specific voltage and first-order gradient matrices, the optimal structure of the matrix variables can be directly derived, thereby significantly improving the accuracy of optimization problems involving change type identification. The use of second-order gradient constraints not only makes the optimization process more efficient but also reveals the characteristics of the matrix equations, making the solution path for the identification problem more direct.

[0069] It should be noted that the scope of protection of the prior art in this invention is not limited to the embodiments given in this application. All prior art that does not contradict the solution of this invention, including but not limited to prior patent documents, prior publications, prior public uses, etc., can be included in the scope of protection of this invention.

[0070] Furthermore, the combination of the technical features in this case is not limited to the combination methods described in the claims of this case or the combination methods described in the specific embodiments. All technical features described in this case can be freely combined or combined in any way, unless they contradict each other.

[0071] It should also be noted that the embodiments listed above are merely specific embodiments of the present invention. Obviously, the present invention is not limited to the above embodiments, and similar changes or modifications made thereto are those that can be directly derived or easily conceived by those skilled in the art from the content disclosed in the present invention, and should all fall within the protection scope of the present invention.

Claims

1. A method for identifying the topology change type of a distribution network based on the second-order gradient-constrained heterogeneous correlation theory, characterized in that, Specifically, it includes: S1. Collect voltage data of each topology node through monitoring sensors of the power distribution network, and perform periodic local truncation averaging. S2. Using the processed voltage data, construct two key matrices for the power distribution network: the voltage matrix and the first-order gradient matrix. S3. Calculate the second-order gradient heterogeneous correlation matrix of the voltage matrix and the first-order gradient matrix respectively, to obtain the deep information and features in the voltage data; S4 calculates the autocorrelation matrix of the second-order gradient heterogeneous correlation matrix of the voltage matrix, and the autocorrelation matrix of the second-order gradient heterogeneous correlation matrix of the first-order gradient matrix. S5. Calculate the topology correlation identification factor based on the autocorrelation coefficient and identify the type of topology change in the distribution network based on the calculation results.

2. The method for identifying distribution network topology change types based on second-order gradient-constrained heterogeneous correlation theory according to claim 1, characterized in that, The periodic local truncation mean processing in step S1 is specifically as follows: Let the voltage data of a single topological node be \(X = [x_1, x_2, \cdots, x\) n , \(n\) be the length of the voltage data collected by a single sensor, and the local window size be \(k\), where \(k < n\); then the truncated data \(X\) i = [x i , x i+1 , \cdots, x i+k-1 , and \(i\) is the starting time position of the truncated data; The local mean of the truncated data is calculated using the following formula: The voltage data X after periodic local truncation and mean processing c =[A1,A2,...,A i ].

3. The method for identifying distribution network topology change types based on second-order gradient-constrained heterogeneous correlation theory according to claim 1, characterized in that, Step 2 specifically includes: Let the voltage matrix be A (N*R) Let the first-order gradient matrix be G. (R*N) Where N represents the number of sensors or the number of data acquisition nodes, and R represents the processed voltage data X. c The length.

4. The method for identifying distribution network topology change types based on second-order gradient-constrained heterogeneous correlation theory according to claim 1, characterized in that, Step 3 specifically includes: The second-order gradient heterogeneous correlation matrix R1 of the voltage matrix is: The second-order gradient heterogeneous correlation matrix R2 of the first-order gradient matrix is:

5. The method for identifying distribution network topology change types based on second-order gradient-constrained heterogeneous correlation theory according to claim 1, characterized in that, In step 4, the autocorrelation matrix V of the second-order gradient heterogeneous correlation matrix is ​​calculated: V6D -1 / 2 CD -1 / 2 Where D is a diagonal matrix, and its diagonal elements are the square roots of the diagonal elements of the covariance matrix C. -1 / 2 Let be the inverse square root matrix of D.

6. The method for identifying distribution network topology change types based on second-order gradient-constrained heterogeneous correlation theory according to claim 1, characterized in that, In step 5, the topology-related identification factor F is calculated using the following formula: Where, ρ r and ρ n The diagonal elements of the correlation matrices R1 and R2 are the correlation coefficients, α and β are the maximum and minimum values ​​of the correlation coefficients, respectively, Γ() is the gamma function, N represents the number of sensors or the number of data acquisition nodes, and R represents the processed voltage data X. c The length.

7. The method for identifying the topology change type of a distribution network based on second-order gradient-constrained heterogeneous correlation theory according to claim 1, characterized in that, The types of power distribution network topology changes include: normal, line connection changes, node status changes, and load transfer changes.

8. The method for identifying the topology change type of a distribution network based on second-order gradient-constrained heterogeneous correlation theory according to claim 1, characterized in that, In step 5, the topology-related identification factors are set as follows: F∈[0.75,1.05] represents the normal state, F∈[2.95e4,3.05e4] represents line connection changes, F∈[1.25e6,2.05e7] represents node status changes, and F∈[3.25e8,4.05e8] represents load transfer changes.

9. A distribution network topology change type identification system based on second-order gradient-constrained heterogeneous correlation theory, characterized in that, It includes a voltage data acquisition device, a data processing device, and a distribution network topology change type identification device. The distribution network topology change type identification system based on the second-order gradient constraint heterogeneous correlation theory executes the distribution network topology change type identification method as described in any one of claims 1-8.

Images