A data-driven based active power distribution network topology identification method

Abstract

The application relates to a data-driven active power distribution network topology identification method, and belongs to the field of power distribution network topology identification. The method linearizes a power flow equation, adopts ridge regression to calculate an admittance matrix, combines a corresponding flag coefficient of each element in the admittance matrix, processes the admittance matrix, and gives topology information according to an electric conductance matrix and an electric capacitance matrix output finally. The active power distribution network topology identification method provided by the application is beneficial to providing decision basis for power supply service command and repair and power distribution network theoretical network loss analysis, improving the intelligent level of operation maintenance management and customer satisfaction, and fully utilizing electric energy to reduce energy loss.

Description

TECHNICAL FIELD

[0001] The present application belongs to the field of power distribution network topology identification, and particularly relates to a data-driven active power distribution network topology identification method. BACKGROUND

[0002] With the application of emerging industries such as 5G technology, cloud computing, Internet of Things, artificial intelligence, etc. in engineering practice, the traditional power grid and information technology are accelerating integration, and the power data is booming. As the carrier of the operation of the information system of the smart grid and the power Internet of Things, informatization, scientization, intelligentization and dataization have become an indispensable key infrastructure for the last mile of the construction of the smart grid.

[0003] With the integration of the power Internet of Things and the traditional power grid, the fine management of the low-voltage distribution network has become a trend for power companies to operate and build to continuously improve the power quality protection and service level of power users. High-sensitivity sensors, stable communication methods, and reliable control means add wings to each link under the operation and management of the power grid, provide new power for the continuous and efficient development of various industries, and increase intelligent solutions for power companies to provide thoughtful services to power users.

[0004] With the construction of these systems, the power grid company collects massive data, and at the same time establishes the connection between the power grid company and the user in terms of energy flow, information flow, data flow and business flow, realizing the scientization, reliability and completeness of the power industry. Massive data lays a foundation for low-voltage distribution network topology structure analysis, line parameter identification, and substation area loss analysis, and improves the reliability of power supply.

[0005] Accurate user belonging substation relationship is the premise of "line-house" and "phase-house" relationship identification, and also the basis for a series of work such as substation line optimization and phase load planning. Unclear distribution network topology structure is not conducive to the modernization of the substation, and also cannot determine whether the three-phase load is balanced. Three-phase imbalance will increase the power loss of the substation line, seriously affect the service life of transformers and various electrical equipment, and reduce the quality of substation power supply. In the entire power transmission and distribution system, the voltage quality problem of low-voltage power grid is the most obvious, the failure rate is the most prominent, and the user complaints are the most serious. Realizing the clear attribution of low-voltage distribution network account information and clear topology structure can improve the reliability of low-voltage power supply, improve the user's power experience and reduce the operation cost of power companies. The identification of the topology of the low-voltage power supply network is beneficial to providing decision basis for power supply service command and repair, improving the intelligent level of operation and maintenance management and customer satisfaction, and fully utilizing electric energy to reduce energy loss. In the distribution network, the topology information may change frequently due to various reasons such as daily maintenance and power grid reconfiguration. Therefore, it is of great significance to study the data-driven active power distribution network topology identification method. SUMMARY

[0006] The present application aims to solve the problems in the background art, and provides a data-driven active power distribution network topology identification method to realize active power distribution network topology identification.

[0007] To achieve the above-mentioned purpose, the technical scheme of the present application is as follows: a data-driven active power distribution network topology identification method, comprising:

[0008] Calculate the flag coefficient corresponding to each element in the admittance matrix composed of the connection branches of the bus of the power distribution network, and compare it with the preset minimum flag coefficient, process the corresponding elements in the admittance matrix according to the comparison result, and obtain the final conductance matrix and the final susceptance matrix, wherein the initial value of the preset minimum flag coefficient is equal to the minimum flag coefficient in the flag coefficient corresponding to each element;

[0009] According to the final conductance matrix and the final susceptance matrix, the topology information of the power distribution network is given.

[0010] In an embodiment of the present application, before the topology information of the power distribution network is given according to the final conductance matrix and the final susceptance matrix, it further comprises: calculating the difference between the modulus of the diagonal element value of each row in the admittance matrix and the sum of the modulus of all non-diagonal element values of each row, if the difference is not zero, increasing the minimum flag coefficient in the flag coefficient corresponding to each element by a preset value as the preset minimum flag coefficient until the difference between the modulus of the diagonal element value of each row in the admittance matrix and the sum of the modulus of all non-diagonal element values of each row is zero, and outputting the final conductance matrix and the final susceptance matrix.

[0011] In an embodiment of the present application, before calculating the flag coefficient corresponding to each element in the admittance matrix composed of the connection branches of the bus of the power distribution network, it further comprises: using ridge regression to iteratively calculate the admittance matrix.

[0012] In an embodiment of the present application, before using ridge regression to iteratively calculate the admittance matrix, it further comprises: linearizing the power flow equation constructed based on the relationship between the conductance matrix and the susceptance matrix.

[0013] In an embodiment of the present application, the specific implementation mode of linearizing the power flow equation constructed based on the relationship between the conductance matrix and the susceptance matrix is as follows:

[0014] The power distribution network is represented as an undirected graph G u =(v, ε), where v represents the bus of the power distribution network v={1, 2, 3, …, N}, N is the total number of buses, and ε represents the connection branch connecting two buses in the power distribution network;

[0015] The admittance of each connection branch (i, k) is y ik =gik +jb ik denotes, where i and k denote distribution network buses, g ik denotes the conductance of the connecting branch connecting the distribution network buses, b ik denotes the susceptance of the connecting branch connecting the distribution network buses; the admittance matrix Y=G+jB is an N×N complex matrix, where G is the conductance matrix and B is the susceptance matrix;

[0016] When active and reactive power is injected at bus i, the relationship between the conductance matrix and the susceptance matrix is expressed by the power flow equation as follows:

[0017]

[0018] where p i and q i denote the active power and the reactive power at bus i, respectively; v i and v k denote the voltages at bus i and bus k, respectively; θ ik denotes the phase angle of the voltages at bus i and bus k; G ik and B ik denote the elements in the conductance matrix and the susceptance matrix, respectively;

[0019] Neglecting the phase angle deviation of adjacent buses in the distribution network and linearizing the power flow equation, we obtain:

[0020]

[0021]

[0022] Let

[0023] G * ik =G ik +θ ik B ik (5)

[0024]

[0025] where G * ik and B * ik denote the conjugate of the elements in the conductance matrix and the susceptance matrix, respectively;

[0026] Equations (3)-(6) are rearranged into matrix form as follows:

[0027]

[0028] where [PV] and [QV] are p iv i , q i v i The ratio of the matrix, G * , B * The conjugate matrix of G, B, [V] is a matrix composed of v k .

[0029] In an embodiment of the present application, the ridge regression is used to iteratively calculate the admittance matrix, and the specific implementation is as follows:

[0030] The ridge regression coefficient calculation formula is:

[0031] β=[A T A] -1 [A] T [B] (9)

[0032] Wherein, A and B represent two calculation parameters in the ridge regression coefficient calculation, and β represents the ridge regression coefficient;

[0033] The regularization term λ is introduced to solve the multicollinearity problem existing in formula (9), that is, the ridge regression coefficient is estimated using the following formula:

[0034] β=[A T A+λI] -1 [A] T [B] (10)

[0035] Wherein, I is a unit matrix;

[0036] Using formula (10) in equations (7) and (8) gives:

[0037] [G * ]=[PV][V] T [V T V+λI] -1 (11)

[0038] [B * ]=-[QV][V] T [V T V+λI] -1 (12)。

[0039] In an embodiment of the present application, the specific implementation of calculating the flag coefficient corresponding to each element in the admittance matrix composed of the connecting branch of the bus of the power distribution network is as follows:

[0040] Based on formula (11) and (12), the matrices G * , B * , the flag coefficient α ik corresponding to each element in the admittance matrix is calculated, and is represented as

[0041]

[0042] wherein, W ik and W ii respectively represent the element value corresponding to the corresponding element in the admittance matrix and the diagonal element value of the row where the corresponding element is located;

[0043] If alpha ik is less than or equal to the preset minimum flag coefficient alpha min , the element value corresponding to the corresponding element in the admittance matrix is set to zero.

[0044] In an embodiment of the present application, the value of lambda is set to 0.001 for the IEEE 33 bus system, and the value of lambda is set to 0.1 for the IEEE 6 bus system.

[0045] In an embodiment of the present application, the value of alpha min is set to 0.03 for the IEEE 33 bus system, and the value of alpha min is set to 0.2 for the IEEE 6 bus system.

[0046] The present application also provides a data-driven active power distribution network topology identification system, comprising a memory, a processor and computer program instructions stored in the memory and capable of being executed by the processor, when the processor executes the computer program instructions, the method steps as described above can be realized.

[0047] Compared with the prior art, the present application has the following beneficial effects: the method of the present application processes the power flow equation by linearization and calculates the admittance matrix by using ridge regression, combines the flag coefficient corresponding to each element in the admittance matrix and processes the admittance matrix, and gives the information of the topology according to the final output of the conductance matrix and the susceptance matrix. The topology identification method proposed by the method of the present application can effectively identify the topology of the power distribution network, and has high accuracy. BRIEF DESCRIPTION OF DRAWINGS

[0048] Figure 1 is a flowchart of the data-driven active power distribution network topology identification method of the present application;

[0049] Figure 2 is the topology structure of the IEEE 6 bus system of the present application;

[0050] Figure 3 is the connectivity matrix heat map of the IEEE 33 bus system of the present application;

[0051] Figure 4 is the connectivity matrix heat map of the IEEE 6 bus system of the present application;

[0052] Figure 5The estimated topology of the IEEE 33 bus system of the present application;

[0053] Figure 6 The estimated topology of the IEEE 6 bus system of the present application. DETAILED DESCRIPTION

[0054] The technical solutions of the present application will be described in detail below with reference to the accompanying drawings.

[0055] It should be noted that the following detailed description is exemplary and is intended to provide further explanation of the present application. Unless otherwise indicated, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which the present application belongs.

[0056] It should be noted that the terms used herein are only for the purpose of describing specific embodiments and are not intended to limit the exemplary embodiments according to the present application. As used herein, the singular form is intended to include the plural form unless the context clearly indicates otherwise, and furthermore, it should be understood that when the terms "comprise" and / or "include" are used in the specification, there is a presence of the features, steps, operations, devices, components and / or combinations thereof.

[0057] The present application provides a data-driven active power distribution network topology identification method, comprising:

[0058] The power flow equation constructed based on the relationship between the conductance matrix and the susceptance matrix is linearized, and the ridge regression is used to iteratively calculate the admittance matrix;

[0059] The flag coefficient corresponding to each element in the admittance matrix formed by the connection branch of the bus of the power distribution network is calculated, and compared with the minimum value of the flag coefficient, and the corresponding element in the admittance matrix is processed;

[0060] The difference between each row diagonal and the sum of all non-diagonal values in the admittance matrix is calculated, if the difference is not zero, the minimum value of the flag coefficient is increased by a preset value until the difference is zero, and the final conductance matrix and the susceptance matrix are output;

[0061] The power distribution network topology information is given according to the final output conductance matrix and the susceptance matrix.

[0062] The present application also provides a data-driven active power distribution network topology identification system, comprising a memory, a processor and computer program instructions stored in the memory and capable of being executed by the processor, when the processor executes the computer program instructions, the method steps as described above can be realized.

[0063] The following is a specific embodiment of the present application.

[0064] As Figure 1As shown, the embodiment provides a data-driven active power distribution network topology identification method, comprising the following steps:

[0065] S1, first, data is obtained from the test equipment of the active power distribution network and normalized, including active power P, reactive power Q and voltage amplitude |V|;

[0066] S2, the power flow equation is linearized, and the admittance matrix is iteratively calculated by using ridge regression;

[0067] S3, the flag coefficient a corresponding to each element in the admittance matrix is calculated ij , and the admittance matrix is processed;

[0068] S4, check the difference between each row diagonal and the sum of all non-diagonal values, if the difference is zero, the algorithm converges; otherwise, the value of a min is increased by 0.01 and the iteration is repeated until convergence is reached;

[0069] S5, the information of the topology is given according to the final conductance matrix and the admittance matrix.

[0070] Further, the linearization of the power flow equation is implemented as follows:

[0071] The power distribution network can be represented as an undirected graph F=(v,ε), where v represents the buses v={1,2,3,...,N} in the power distribution network, N is the total number of buses, and ε represents the branches connecting two buses in the power distribution network;

[0072] The admittance of each connecting branch (i,k) is represented by y ik =g ik +jb ik , where i and k represent the buses of the power distribution network, g ik represents the conductance of the connecting branch connecting the buses of the power distribution network, and b ik represents the admittance of the connecting branch connecting the buses of the power distribution network; Y=G+jB represents the admittance matrix, which is an N×N complex matrix, where G is the conductance matrix and B is the admittance matrix;

[0073] When active and reactive power is injected at bus i, the relationship between the conductance and admittance matrices can be represented by the power flow equation,

[0074]

[0075] where p i and q i represent the active and reactive power at bus i, respectively; v i and v k represent the voltages at bus i and bus k, respectively; and θ ikdenotes the voltage phase angle at bus i and bus k; G ik and B ik denote the elements in the conductance and susceptance matrices, respectively. By monitoring and collecting the active power and reactive power (p i and q i ) at each bus, as well as the voltage (v i and v k ) and phase angle (θ ik ) at each bus, the elements G ik and B ik in the conductance and susceptance matrices can be deduced by using the power flow equations, and thus the topology of the power grid can be inferred.

[0076] In a distribution network, the voltage phase angle deviation between adjacent buses is very small, so the approximation excludes the deviation of the voltage phase angle and linearizes the equation,

[0077]

[0078] Let

[0079] G * ik = G ik + θ ik B ik (5)

[0080]

[0081] where G * ik and B * ik denote the conjugate of the elements in the conductance and susceptance matrices, respectively;

[0082] Equations (3) to (6) are arranged in matrix form as follows:

[0083]

[0084]

[0085] where [PV] and [QV] are the matrices of the ratios of p i / v i and q i / v i , respectively, G * and B * are the conjugate matrices of G and B, respectively, and [V] is the matrix of v k .

[0086] Further, the admittance matrix is calculated iteratively using ridge regression, which is implemented as follows:

[0087] The ridge regression coefficients can be computed as,

[0088] β = [A T A+ λI] -1 [A] T [B] (9)

[0089] where A and B represent two computational parameters in the ridge regression coefficient computation, and β represents the ridge regression coefficients;

[0090] The least square model coefficients represented by β are extremely sensitive to the random errors in the response variable represented by [B], which leads to multicollinearity;

[0091] Ridge regression is a technique to solve the problem of multicollinearity by introducing a regularization term λ in equation (9), which estimates the regression coefficients using the following formula:

[0092] β = [A T A+ λI] -1 [A] T [B] (10)

[0093] where λ is a tuning parameter, and I is the identity matrix, which improves the condition of the problem by adding a small positive value λ;

[0094] Therefore, the ridge regression has been used for the results of equations (7) and (8),

[0095] [G * ] = [PV] [V] T [V T V+ λI] -1 (11)

[0096] [B * ] = -[QV] [V] T [V T V+ λI] -1 (12)

[0097] For the IEEE 33-bus system, the value of λ is set to 0.001, and for the IEEE 6-bus system, it is set to 0.1;

[0098] Further, the flag coefficients corresponding to each element in the admittance matrix are calculated as follows:

[0099] Based on the values of matrices G * , B * obtained using equations (11) and (12), the flag coefficients α ik corresponding to each element in the admittance matrix are calculated, which are represented as

[0100]

[0101] where W ik and W ii denote the element value corresponding to the respective element in the admittance matrix and the diagonal element value of the respective element's row, respectively;

[0102] If the calculated flag coefficient a ik is less than or equal to the pre-set minimum flag coefficient a min , the element value corresponding to the respective element in the admittance matrix is set to zero;

[0103] The value of a min is set to 0.03 for the IEEE 33-bus system and 0.2 for the IEEE 6-bus system;

[0104] Further, the difference between the sum of the diagonal and all off-diagonal values of each row is checked, and if the difference is zero, the algorithm has reached convergence; otherwise, the value of a min is increased by 0.01 and the iteration is repeated until convergence is reached;

[0105] Finally, the topological information is given according to the conductance matrix G * and the susceptance matrix B * .

[0106] The performance of the proposed method is tested on the standard IEEE 33-bus radial distribution system and the IEEE 6-bus test system (as in Figure 2 ), and the 24-hour smart meter data sets for P and Q measurements are obtained from MATPOWER 7.0 to obtain the corresponding voltage magnitudes. The experiments are performed on MATLAB R2019a installed on an Intel(R) Core(TM) i7-3770 CPU @ 3.40 GHz with 8 GB RAM.

[0107] Figure 3 The 33x33 estimated connection matrix for the IEEE 33-bus test system is given in the form of a heat map, Figure 4 the 6x6 estimated connection matrix for the IEEE 6-bus test system is given in the form of a heat map, and Figure 5 and Figure 6 the estimated topologies for the IEEE 33 and IEEE 6 bus systems are given in graphical form, respectively. In addition, in the heat maps, yellow indicates connected buses, while green indicates unconnected buses. Furthermore, the x and y axes represent the buses, while the color bar represents the intensity of the color, with a range of 0 to 1, where 1 indicates connected and 0 indicates unconnected. As can be seen from the figures, all branches in the power grid are correctly estimated. Therefore, the accuracy of the calculation using equation (13) is 100%.

[0108]

[0109] wherein E b denotes the estimated number of branches, T b denotes the total number of branches.

[0110] The application also provides a data-driven active power distribution network topology identification system, comprising a memory, a processor and computer program instructions stored in the memory and capable of being executed by the processor, when the processor executes the computer program instructions, the method steps as described above can be implemented.

[0111] Those skilled in the art will appreciate that embodiments of the application can be supplied as methods, systems, or computer program products. Accordingly, the application can be embodied in the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the application can be embodied in the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk memory, CD-ROMs, optical storage media, etc.) having computer usable program code embodied therein.

[0112] The application is described with reference to the flowcharts and / or block diagrams of the methods, apparatus (systems), and computer program products according to embodiments of the application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, as well as combinations of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general purpose computer, a special purpose computer, an embedded processor, or other programmable data processing apparatus to produce a machine, so that the instructions, which are executed via the processor of the computer or other programmable data processing apparatus, generate an apparatus that implements the functions specified in the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocks Figure 1 means for performing the function specified by the flow or flows and / or block or blocks.

[0113] These computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture including instructions means that implement the function specified in the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocks Figure 1 means for performing the function specified by the flow or flows and / or block or blocks.

[0114] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions executed on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocks Figure 1steps of the functions specified in the block or blocks.

[0115] The above descriptions are only the preferred embodiments of the present application, not intended to limit the present application to other forms described. Any person skilled in the art may make changes or modifications to the equivalent embodiments with the disclosed technical contents. However, any simple modification, equivalent change and modification made to the above embodiments without departing from the technical solution of the present application, and according to the technical essence of the present application, still belong to the protection scope of the present application.

Claims

1. A data-driven based active power distribution network topology identification method, characterized in that, The method comprises the following steps: comparing each element of the admittance matrix of the connection branch of the distribution network bus with a preset minimum flag coefficient, and processing the corresponding element in the admittance matrix according to the comparison result to obtain a final conductance matrix and a final susceptance matrix, wherein the initial value of the preset minimum flag coefficient is equal to the minimum flag coefficient corresponding to each element; the distribution network topology information is given according to the final conductance matrix and the final susceptance matrix; before the distribution network topology information is given according to the final conductance matrix and the final susceptance matrix, the method further comprises the following steps: before calculating the flag coefficient corresponding to each element of the admittance matrix of the connection branch of the distribution network bus, the method further comprises the following step: before the admittance matrix is iteratively calculated by using the ridge regression, the method further comprises the following step: linearizing the power flow equation constructed based on the relationship between the conductance matrix and the susceptance matrix; The power distribution network is represented as an undirected graph G u = (v, e), where v represents the power distribution network buses v = {1, 2, 3,..., N}, N is the total number of buses, and e represents the connecting branches connecting two buses in the power distribution network; The admittance of each connection branch (i, k) is denoted by y ik = g ik + jb ik where i and k denote distribution network buses, g ik denotes the conductance of the connection branch connecting the distribution network buses, b ik denotes the susceptance of the connection branch connecting the distribution network buses; the admittance matrix Y = G + jB is a NxN complex matrix, where G is the conductance matrix and B is the susceptance matrix; the specific implementation of linearizing the power flow equation constructed based on the relationship between the conductance matrix and the susceptance matrix is as follows: where p i and q i represent the active and reactive power at bus i, respectively; v i and v k represent the voltage at bus i and bus k, respectively; θ ik represents the phase angle of the voltage at bus i and bus k; G ik and B ik represent the elements in the conductance and susceptance matrices, respectively; when active and reactive power is injected at bus i, the relationship between the conductance matrix and the susceptance matrix is represented by the power flow equation as follows: ignoring the voltage phase angle deviation of adjacent buses in the distribution network and linearizing the power flow equation to obtain: G * ik = G ik + θ ik B ik (5) where G * ik and B * ik denote the conjugate of the elements in the conductance and susceptance matrices, respectively. let Wherein, [PV], [QV] are respectively p i / v i , q i / v i The ratio of the matrix, G * , B * are respectively the conjugate matrix of G, B, [V] is v k The matrix composed of.

2. The data-driven based active power distribution network topology identification method of claim 1, wherein, arranging equations (3) to (6) into a matrix form as follows: the specific implementation of iteratively calculating the admittance matrix by using the ridge regression is as follows: β = [A T A] -1 [A] T [B] (9) the ridge regression coefficient calculation formula is as follows: wherein A and B represent two calculation parameters in the ridge regression coefficient calculation, and β represents the ridge regression coefficient; β = [A T A + λI] -1 [A] T [B] (10) a regularization term λ is introduced to solve the multicollinearity problem existing in equation (9), that is, the ridge regression coefficient is estimated by using the following formula: wherein I is a unit matrix; [G * ] = [PV] [V] T [V T V + λI] -1 (11) [B * ] = -[QV] [V] T [V T V + λI] -1 (12).

3. The data-driven based active power distribution network topology identification method of claim 2, wherein, equations (10) are used in equations (7) and (8) as follows: Based on the equations (11) and (12), the matrix G is obtained * , B * , the flag coefficient a corresponding to each element in the admittance matrix is calculated ik , is expressed as wherein W ik and W ii respectively represent the element value corresponding to the respective element in the admittance matrix and the diagonal element value of the row in which the respective element is located. If α ik is less than or equal to a preset minimum flag coefficient α min , the element value corresponding to the element in the admittance matrix is set to zero.

4. The data-driven based active power distribution network topology identification method of claim 2, wherein, the specific implementation of calculating the flag coefficient corresponding to each element of the admittance matrix of the connection branch of the distribution network bus is as follows:

5. The data-driven based active power distribution network topology identification method of claim 3, wherein, IEEE 33 bus system's a min set to 0.03, IEEE 6 bus system's a min set to 0.

2.

6. A data-driven based active power distribution network topology identification system, characterized in that, for the IEEE 33 bus system, the value of λ is set to 0.001, and for the IEEE 6 bus system, the value of λ is set to 0.

1. The computer program instructions stored in the memory and capable of being executed by the processor can realize the method steps of any one of claims 1-5 when the processor executes the computer program instructions. The computer program instructions stored in the memory and capable of being executed by the processor can realize the method steps of any one of claims 1-5 when the processor executes the computer program instructions.

Images