A medium voltage distribution network topology identification method considering regional similarity

Abstract

The present invention discloses a medium-voltage distribution network topology identification method considering regional similarity. It belongs to the field of electric energy, and includes the following contents: first, based on the measurement information and historical data obtained on the low-voltage side, a measurement aggregation data set with high accuracy is obtained by aggregation on the medium-voltage side; through the measurement data on the medium-voltage side of the distribution network obtained by aggregation, a reverse flow model that does not involve the prior conditions of the distribution network topology is established, and a medium-voltage distribution network topology identification method based on a constant parameter ratio homogeneous is proposed to obtain the preliminary topology identification result of the medium-voltage distribution network; and considering the problem that the preliminary topology identification result is not accurate enough, a distribution network unit system planning method is adopted, and the proposed method is improved through the regional similarity of the medium-voltage distribution network unit system planning. The present invention is closely combined with the actual situation of the distribution network, and solves the problems that the existing distribution network topology identification scheme is not robust, and the identification model has low identification accuracy and is greatly affected by noise during actual application.

Description

Technical Field

[0001] The present invention belongs to the field of electric energy, and in particular, relates to a medium voltage distribution network topology identification method considering the similarity of distribution network areas, which can improve the robustness and accuracy of existing distribution network topology identification schemes. Background Art

[0002] Under the requirements of smart grid development and energy transformation, the power system's dependence on communication technology and electronic technology continues to deepen. Today's distribution system has developed into a highly coupled cyber-physical distribution cyber-physical system (CPDS). In addition, with the continuous development of distributed generation technology, the high-penetration distributed power grid connection has transformed the traditional single-power radial distribution network into a multi-power system with power sources and loads. The high requirements of active distribution networks for coordinated control capabilities have also made the distribution network structure increasingly complex. In this context, accurate real-time topological information of the distribution network will have an important impact on the planning, operation, and analysis of the distribution network.

[0003] At present, the research on distribution network topology identification methods has not considered the topological characteristics of distribution networks of different voltage levels. The proposed topology identification methods are limited to independent application to a single voltage level and do not consider the formation of a bottom-up progressive identification of low-voltage and medium-voltage distribution network topologies.

[0004] Considering the consistency of urban regional development planning requirements and the similarity of power demand, medium-voltage distribution networks are usually planned in a unit system. Therefore, there is a certain degree of regional similarity in the topological structure of medium-voltage distribution networks, which is specifically manifested in: the similarity of the overall scale of medium-voltage distribution stations; the similarity of the mean values ​​of the out-degree and in-degree of all nodes in each distribution network station; similarity indicators such as the singular value, spectral radius, and matrix norm of the adjacency matrix, etc. These regional similarity indicators provide a reference for improving the accuracy and robustness of medium-voltage distribution network identification. Summary of the invention

[0005] Technical problem: In view of the shortcomings of existing medium-voltage distribution network topology identification methods, the technical problem to be solved by the present invention is: to propose a medium-voltage distribution network topology identification method that takes into account the similarity of distribution network areas, so as to solve the problems that existing distribution network topology identification schemes are usually not robust, the recognition accuracy of the identification model is not high in actual application, and it is greatly affected by noise.

[0006] Technical solution: To solve the above technical problems, the technical solution adopted by the present invention is:

[0007] A method for identifying medium voltage distribution network topology considering the similarity of distribution network regions comprises the following steps:

[0008] Based on the user-side power measurement information and historical data obtained on the low-voltage side, a high-precision measurement aggregation data set is obtained on the medium-voltage side, mainly including: active power, reactive power, and voltage data.

[0009] By aggregating the measured data on the medium-voltage side of the distribution network, a reverse power flow model that does not involve the prior conditions of the distribution network topology is established, and linearization and parameter simplification operations are performed on it. A homogeneous medium-voltage distribution network topology identification method based on fixed line parameter ratio is proposed to obtain the preliminary topology identification results of the medium-voltage distribution network.

[0010] Furthermore, taking into account the problem of insufficient accuracy of the above-mentioned preliminary topology identification results, the proposed method is improved based on the distribution network unit planning method and the regional similarity existing in the medium-voltage distribution network unit planning.

[0011] A medium voltage distribution network topology identification method considering the similarity of distribution network regions is based on the following assumptions:

[0012] Different from the high-precision requirements of power flow data in the distribution network situation awareness work, the goal of the distribution network topology identification work is to determine whether there is a physical connection relationship between the nodes. Its requirements for data accuracy and real-time performance are not as strict as those for situation awareness work.

[0013] 1) Ignore the active and reactive power losses that may exist in the network. In fact, the active and reactive power losses in the distribution network are more obvious than those in the transmission network, but they account for a very small proportion compared to the actual power transmitted in the line. Looking at the existing research methods on distribution network topology identification, most of them adopt this assumption;

[0014] 2) Assuming that the coverage rate of IoT devices in the low-voltage distribution network is high (the device coverage rate is different from the proportion of energy-consuming devices), it is possible to basically achieve full regional coverage of the low-voltage distribution network and ensure the data source required for topology identification on the medium-voltage side.

[0015] 3) The possible physical faults, information faults or information-physical coupling faults in the power distribution cyber-physical system are not considered. Under the protection of the three lines of defense of the power system, the distribution network currently has sufficient elastic support to deal with ordinary faults. Even in the event of low probability and high loss extreme events, the power distribution cyber-physical system also has a certain elastic scheduling capability to eliminate or reduce the impact of the fault in time.

[0016] The measured data on the medium voltage side is aggregated, ignoring network losses. The active power of the nodes on the medium voltage side is equal to the sum of the active power consumed by the IoT devices and non-IoT devices on the low voltage side. The active power of IoT devices can be obtained in real time, but there is a lack of active data information on other non-IoT power-consuming devices. The historical data of the distribution network contains the temporal and spatial correlation characteristics between the operating states of the distribution network. When the user's electricity consumption behavior is not affected by major uncertain factors, the proportion of power-consuming device types should not fluctuate significantly within a certain time range. Therefore, the electricity consumption behavior of distribution network users can be explored through their electricity consumption data, and the proportion of different types of power-consuming devices can be obtained, and then the active power measurement data of IoT devices on the low voltage side can be proportionally aggregated to the medium voltage side.

[0017] In the reverse power flow model of medium voltage distribution network, the task of power flow calculation is to find the corresponding state variables according to the given control variables for certain disturbance variables under the condition of known network topology (usually represented by node admittance matrix). Distribution network topology identification is to identify the most likely topology of distribution network based on the existing voltage and power flow data when the network topology and basic parameters are unknown, so as to verify the rationality of the known voltage and power flow data.

[0018] The medium voltage distribution network power flow model can be expressed as:

[0019]

[0020]

[0021]

[0022]

[0023]

[0024]

[0025] The above formulas respectively represent the active and reactive power flow equation conditions; the equation conditions that the node voltage and branch current must satisfy; and the branch current and node voltage constraints. For any radial medium-voltage distribution network, the above general constraints should be satisfied. Based on the above assumptions, other constraints of the distribution network are not considered.

[0026] Since the above formulas all contain branch current or node voltage square terms, the model is a nonlinear model, which makes it difficult to obtain the subsequent topology identification characteristic parameters. and It can be linearized into the following model:

[0027]

[0028]

[0029] Based on the topology identification method of homogeneous medium-voltage distribution network with fixed line parameter ratio, ignoring line loss, the nonlinear power flow model is transformed into an expression containing the physical connection relationship required for topology identification:

[0030]

[0031] in The matrix represents the square of the voltage amplitude at each node, the p and q matrices represent the active and reactive injected powers at the nodes, respectively, and the A matrix is ​​an adjacency matrix related to the topology of the distribution network, whose elements are shown in the following formula. The R and X matrices represent the line resistance and reactance parameters, and each line parameter is represented in the form of the corresponding diagonal elements.

[0032]

[0033] [a0,A T ] T ∈{0,±1} (n+1)×n

[0034] The active and reactive injection powers of each node are known, and the elements of the adjacency matrix A related to the topological structure of the distribution network are unknown. For a radial distribution network, the adjacency matrix A must be a non-singular matrix, and A -T a0 must be the unit matrix, denoted by A -T a0=E n Therefore, the power flow model (*) can be transformed into:

[0035]

[0036] Among them, AX -1 A T That is, the node admittance matrix Y, whose elements directly reflect the physical connection relationship of the distribution network. If there is no physical connection relationship between node i and node j, then y ij =0; if node i is directly connected to node j, then y ij ≠0. Therefore, if the elements in the node admittance matrix Y can be obtained according to (12), the topological connection mode of the medium voltage distribution network can be preliminarily determined.

[0037] p, q, are all known quantities. The key to finding the elements in Y is how to determine AX -1 RA -1 Note that X -1 R is actually the ratio of line resistance to reactance parameters. Let’s assume that the ratio is a certain value. This assumption is basically consistent with the actual power line parameter setting. That is:

[0038]

[0039]

[0040] At this time, the formula (*) becomes:

[0041]

[0042] Therefore, the medium voltage distribution network topology identification method can be summarized as follows: the active power and reactive power obtained by the medium voltage side are measured, and the elements in the distribution network node admittance matrix Y are obtained, so as to obtain the physical connection relationship between the corresponding nodes. The above power flow formula can be transformed into a reverse power flow problem that does not require any prior conditions (such as topological connection relationship and line parameters).

[0043] However, although the accuracy, robustness and operability of medium voltage distribution network topology identification using the reverse power flow model have been significantly improved compared with the traditional distribution network topology identification method, its topology identification results are still difficult to ensure uniqueness. The reason for the possibility of multiple topological structures is that the unconstrained node admittance matrix Y is greatly affected by the accuracy of the active and reactive power measurement data, and the determination of the physical connection relationship in topology identification is uniquely determined by the node admittance matrix Y, which reduces the robustness of the network structure analysis score.

[0044] The regional similarity of the unit planning of medium-voltage distribution networks. For urban medium-voltage distribution networks, due to the consistency of urban regional planning requirements and the similarity of power demand, medium-voltage distribution networks are currently usually planned in a unit system, so medium-voltage distribution networks usually have a certain degree of similarity. This similarity is specifically reflected in the following aspects:

[0045] 1) Similarity in the overall scale of medium-voltage distribution substations. Although different distribution substations are located in different geographical environments, due to the voltage level restrictions of substations, the transmission distance and power of distribution substations are limited, so that the expansion range and overall scale of different distribution substations in the medium-voltage distribution network are similar.

[0046] 2) Similarity of the average out-degree and in-degree of all nodes in each distribution network area. Under the requirements of urban development planning, the regional users supplied by the same medium-voltage distribution network belong to the same functional area, and their physical distribution mode and regional structure are consistent, which greatly affects the installation plan of power lines, resulting in the natural similarity of the out-degree and in-degree of node lines in the substation area.

[0047] 3) Similarity indicators such as adjacency matrix singular value, spectral radius, matrix norm, etc. Some literatures use mathematical tools to directly calculate the similarity index of distribution network substations based on the adjacency matrix singular value as an important indicator to describe the essence of topology, and give the mathematical quantitative relationship between the similarity of power grid topology and the power supply reliability of distribution network.

[0048] Combined with the above analysis, the present invention fully mines the connection information in the known substation topology based on the similarity of the medium-voltage distribution network topology, and provides reasonable hyperparameter constraints for the proposed identification method. This hyperparameter comprehensively considers the node level, out-degree, and in-degree of the known substation, and provides a reference for the distribution network area with unknown topology structure. The distribution network area to be determined can also use this hyperparameter to eliminate unreasonable topology identification results.

[0049] This hyperparameter is expressed as:

[0050]

[0051] Where M is the total number of nodes in the known medium voltage distribution network area; R n Represents the level parameter of the substation node. The node directly connected to the medium voltage transformer is recorded as the first-level node, the non-transformer bus node directly adjacent to the first-level node is the second-level node, the non-first-level node directly adjacent to the second-level node is recorded as the third-level node, and so on; γ n and δ n They represent the out-degree and in-degree values ​​of the nodes in the area respectively.

[0052] After obtaining the possible topological structures by the medium-voltage distribution network topology identification method, the hyperparameter σ value of each topological structure is calculated respectively, and the topological structures with a large difference in σ value from the known substation of the medium-voltage distribution network are excluded, and finally the optimal topological identification result is obtained.

[0053] Beneficial effects: Compared with the prior art, the present invention has the following characteristics:

[0054] Based on the similarity of the topological structure of the medium-voltage distribution network, the present invention fully explores the connection information in the known substation topology and provides reasonable hyperparameter constraints for the proposed identification method. The hyperparameter comprehensively considers the node level, out-degree, and in-degree of the known substation, and provides a reference for the distribution network area with unknown topological structure. The distribution network area to be determined can also use the hyperparameter to eliminate unreasonable topological identification results. It solves the problems that the existing distribution network topology identification scheme is usually not robust, the recognition accuracy of the identification model is not high in actual application, and it is greatly affected by noise. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 A block diagram of a medium voltage distribution network topology identification method considering the similarity of distribution network areas according to the present invention;

[0056] Figure 2 A block diagram of medium voltage distribution network topology identification considering the similarity of distribution network areas according to the present invention;

[0057] Figure 3 This is a steady-state operation mode diagram of the medium voltage distribution network according to the present invention;

[0058] Figure 4 The schematic diagram of the IEEE-37 node power distribution system is used to verify the example of the present invention;

[0059] Figure 5 This is a topology identification result diagram of a 10-node medium-voltage distribution network model in a certain area of ​​the present invention. DETAILED DESCRIPTION

[0060] In order to make the purpose, features and advantages of the present invention more obvious and easy to understand, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the embodiments described below 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.

[0061] A method for identifying medium voltage distribution network topology considering the similarity of distribution network regions, characterized in that it comprises the following steps:

[0062] Based on the user-side power measurement information and historical data obtained on the low-voltage side, a high-precision measurement aggregation data set is obtained on the medium-voltage side, mainly including: active power, reactive power, voltage data, etc.

[0063] By aggregating the measured data on the medium-voltage side of the distribution network, a reverse power flow model that does not involve the prior conditions of the distribution network topology is established, and operations such as linearization and parameter simplification are performed on it. A homogeneous medium-voltage distribution network topology identification method based on fixed line parameter ratio is proposed to obtain the preliminary topology identification results of the medium-voltage distribution network.

[0064] Furthermore, taking into account the problem of insufficient accuracy of the above-mentioned preliminary topology identification results, the proposed method is improved based on the distribution network unit planning method and the regional similarity existing in the medium-voltage distribution network unit planning.

[0065] A medium voltage distribution network topology identification method considering the similarity of distribution network regions is characterized by being based on the following assumptions:

[0066] Different from the high-precision requirements of power flow data in the distribution network situation awareness work, the goal of the distribution network topology identification work is to determine whether there is a physical connection relationship between the nodes. Its requirements for data accuracy and real-time performance are not as strict as those for situation awareness work.

[0067] 1) Ignore the active and reactive power losses that may exist in the network. In fact, the active and reactive power losses in the distribution network are more obvious than those in the transmission network, but they account for a very small proportion compared to the actual power transmitted in the line. Looking at the existing research methods on distribution network topology identification, most of them adopt this assumption;

[0068] 2) Assuming that the coverage rate of IoT devices in the low-voltage distribution network is high (the device coverage rate is different from the proportion of energy-consuming devices), it is possible to basically achieve full regional coverage of the low-voltage distribution network and ensure the data source required for topology identification on the medium-voltage side.

[0069] 3) The possible physical faults, information faults or information-physical coupling faults in the power distribution cyber-physical system are not considered. Under the protection of the three lines of defense of the power system, the distribution network currently has sufficient elastic support to deal with ordinary faults. Even in the event of low probability and high loss extreme events, the power distribution cyber-physical system also has a certain elastic scheduling capability to eliminate or reduce the impact of the fault in time.

[0070] The measurement data aggregation on the medium voltage side is characterized by that, if network losses are not taken into account, the active power of the medium voltage side node is equal to the sum of the active power consumed by the IoT devices and non-IoT devices on the low voltage side. The active power of the IoT device can be obtained in real time, while there is a lack of active data information on other non-IoT power-consuming devices. The historical data of the distribution network contains the temporal and spatial correlation characteristics between the operating states of the distribution network. When the user's electricity consumption behavior is not affected by major uncertain factors, the proportion of power-consuming equipment types should not fluctuate significantly within a certain time range. Therefore, the electricity consumption behavior of distribution network users can be explored through their electricity consumption data, and the proportion of different types of power-consuming equipment can be obtained, and then the active power measurement data of the IoT devices on the low voltage side can be proportionally aggregated to the medium voltage side.

[0071] The reverse power flow model of medium voltage distribution network is characterized by that the task of power flow calculation is to find the corresponding state variables according to the given control variables for certain disturbance variables under the condition of known network topology (usually represented by node admittance matrix). The topology identification of distribution network is to identify the most likely topology structure of distribution network based on the existing voltage and power flow data when the network topology and basic parameters are unknown, so as to verify the rationality of known voltage and power flow data. It can be said that topology identification is the inverse process of power flow calculation, and the process of topology identification is essentially a kind of "reverse power flow calculation".

[0072] The medium voltage distribution network power flow model can be expressed as:

[0073]

[0074]

[0075]

[0076]

[0077]

[0078]

[0079] The above formulas respectively represent the active and reactive power flow equation conditions; the equation conditions that the node voltage and branch current must satisfy; and the branch current and node voltage constraints. For any radial medium-voltage distribution network, the above general constraints should be satisfied. Based on the above assumptions, other constraints of the distribution network are not considered.

[0080] Since the above formulas all contain branch current or node voltage square terms, the model is a nonlinear model, which makes it difficult to obtain the subsequent topology identification characteristic parameters. and It can be linearized into the following model:

[0081]

[0082]

[0083] The method for topology identification of a homogeneous medium-voltage distribution network based on a fixed line parameter ratio is characterized in that, ignoring line losses, the nonlinear power flow model is transformed into an expression containing the physical connection relationship required for topology identification:

[0084]

[0085] in The matrix represents the square of the voltage amplitude at each node, the p and q matrices represent the active and reactive injected powers at the nodes, respectively, and the A matrix is ​​an adjacency matrix related to the topology of the distribution network, whose elements are shown in the following formula. The R and X matrices represent the line resistance and reactance parameters, and each line parameter is represented in the form of the corresponding diagonal elements.

[0086]

[0087] [a0,A T ] T ∈{0,±1} (n+1)×n

[0088] The active and reactive injection powers of each node are known, and the elements of the adjacency matrix A related to the topological structure of the distribution network are unknown. For a radial distribution network, the adjacency matrix A must be a non-singular matrix, and A -T a0 must be the unit matrix, denoted by A -T a0=E n Therefore, the power flow model (*) can be transformed into:

[0089]

[0090] Among them, AX -1 A TThat is, the node admittance matrix Y, whose elements directly reflect the physical connection relationship of the distribution network. If there is no physical connection relationship between node i and node j, then y ij =0; if node i is directly connected to node j, then y ij ≠0. Therefore, if the elements in the node admittance matrix Y can be obtained according to (12), the topological connection mode of the medium voltage distribution network can be preliminarily determined.

[0091] p, q, are all known quantities. The key to finding the elements in Y is how to determine AX -1 RA -1 Note that X -1 R is actually the ratio of line resistance to reactance parameters. Let’s assume that the ratio is a certain value. This assumption is basically consistent with the actual power line parameter setting. That is:

[0092]

[0093]

[0094] At this time, the formula (*) becomes:

[0095]

[0096] Therefore, the medium voltage distribution network topology identification method can be summarized as follows: the active power and reactive power obtained by the medium voltage side are measured, and the elements in the distribution network node admittance matrix Y are obtained, so as to obtain the physical connection relationship between the corresponding nodes. The above power flow formula can be transformed into a reverse power flow problem that does not require any prior conditions (such as topological connection relationship and line parameters).

[0097] In one example of the present invention, the above topology identification method is applied to a 10-node medium-voltage distribution network model in a certain area, and the topology identification result is as follows: Figure 5 As shown, the recognition probability of the corresponding topological structure is shown in Table 1.

[0098] Table 1-10 Identification probability of each topological structure of the medium-voltage distribution network

[0099]

[0100] It is not difficult to see from the above examples that although the use of the reverse flow model for medium-voltage distribution network topology identification has significantly improved accuracy, robustness, and operability compared to traditional distribution network topology identification methods, its topology identification results are still difficult to ensure uniqueness. The reason for the possibility of multiple topological structures is that the unconstrained node admittance matrix Y is greatly affected by the accuracy of the active and reactive measurement data, and the determination of the physical connection relationship in topology identification is uniquely determined by the node admittance matrix Y, which reduces the robustness of the network structure analysis score.

[0101] The regional similarity of the unit planning of medium-voltage distribution networks is characterized by the fact that, for urban medium-voltage distribution networks, due to the consistency of urban regional planning requirements and the similarity of power demand, medium-voltage distribution networks are currently usually planned in a unit system, so medium-voltage distribution networks usually have a certain degree of similarity. This similarity is specifically reflected in the following aspects:

[0102] 1) Similarity in the overall scale of medium-voltage distribution substations. Although different distribution substations are located in different geographical environments, due to the voltage level restrictions of substations, the transmission distance and power of distribution substations are limited, so that the expansion range and overall scale of different distribution substations in the medium-voltage distribution network are similar.

[0103] 2) Similarity of the average out-degree and in-degree of all nodes in each distribution network area. Under the requirements of urban development planning, the regional users supplied by the same medium-voltage distribution network belong to the same functional area, and their physical distribution mode and regional structure are consistent, which greatly affects the installation plan of power lines, resulting in the natural similarity of the out-degree and in-degree of node lines in the substation area.

[0104] 3) Similarity indicators such as adjacency matrix singular value, spectral radius, matrix norm, etc. Some literatures use mathematical tools to directly calculate the similarity index of distribution network substations based on the adjacency matrix singular value as an important indicator to describe the essence of topology, and give the mathematical quantitative relationship between the similarity of power grid topology and the power supply reliability of distribution network.

[0105] Combined with the above analysis, the present invention fully mines the connection information in the known substation topology based on the similarity of the medium-voltage distribution network topology, and provides reasonable hyperparameter constraints for the proposed identification method. This hyperparameter comprehensively considers the node level, out-degree, and in-degree of the known substation, and provides a reference for the distribution network area with unknown topology structure. The distribution network area to be determined can also use this hyperparameter to eliminate unreasonable topology identification results.

[0106] This hyperparameter is expressed as:

[0107]

[0108] Where M is the total number of nodes in the known medium voltage distribution network area; R n Represents the level parameter of the substation node. The node directly connected to the medium voltage transformer is recorded as the first-level node, the non-transformer bus node directly adjacent to the first-level node is the second-level node, the non-first-level node directly adjacent to the second-level node is recorded as the third-level node, and so on; γ n and δ n They represent the out-degree and in-degree values ​​of the nodes in the area respectively.

[0109] After obtaining the possible topological structures by the medium-voltage distribution network topology identification method, the hyperparameter σ value of each topological structure is calculated respectively, and the topological structures with a large difference in σ value from the known substation of the medium-voltage distribution network are excluded, and finally the optimal topological identification result is obtained.

[0110] like Figure 4 As shown, in one example of the present invention, the area to be identified is the red solid line marked part in the figure. It is also assumed that the topological structure of the red dashed area in the figure is known. The known area can provide topological similarity parameters for the area to be identified, thereby improving the accuracy of topological identification. Based on the same assumptions in the previous example, the active and reactive power aggregation data and device proportion data of the Internet of Things devices collected by the Internet of Things enterprise data center at nodes 11-19 are shown in the following table:

[0111] Table 2 Example measurement data collection results and equipment proportion data

[0112]

[0113] To quantitatively illustrate the accuracy of the topology recognition method described in the present invention, the following table compares the topology recognition accuracy of the 9 nodes in the T1 area with or without considering the hyperparameter σ value constraint. It can be seen that after considering the hyperparameter σ value similarity constraint, the topology recognition accuracy is significantly improved.

[0114] Table 3 Comparison of topology recognition accuracy before and after considering the hyperparameter σ value constraint

[0115]

[0116]

[0117] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit its protection scope. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that after reading the present invention, those skilled in the art can still make various changes, modifications or equivalent substitutions to the specific implementation methods of the invention, but these changes, modifications or equivalent substitutions are all within the protection scope of the pending claims of the invention.

Claims

1. A medium voltage distribution network topology identification method considering the similarity of distribution network regions, characterized in that: The steps include: Based on the user-side power measurement information and historical data obtained on the low-voltage side, a high-precision measurement aggregation data set is obtained on the medium-voltage side, including: active power, reactive power, and voltage data; Through the aggregated data set of the distribution network medium-voltage side measurement, a reverse power flow model that does not involve the prior conditions of the distribution network topology is established, and linearization and parameter simplification operations are performed on it. A homogeneous medium-voltage distribution network topology identification method based on a fixed line parameter ratio is proposed. Through the measurement of active power and reactive power obtained by aggregation on the medium-voltage side, each element in the distribution network node admittance matrix is ​​calculated, so as to obtain the physical connection relationship between the corresponding nodes, so as to obtain the preliminary topology identification result of the medium-voltage distribution network. Considering the problem of insufficient accuracy of the preliminary topology identification results, the proposed method is improved based on the distribution network unit planning method and the regional similarity of the medium voltage distribution network unit planning. The medium voltage distribution network power flow model is expressed as: Where i, j are the node numbers of the start and end points, g is the line conductance, b is the line susceptance, r is the line resistance, and x is the line reactance; N is the node set, and E is the branch set; Equations (1) and (2) represent the active and reactive power flow equation conditions, Equations (3) and (4) represent the equation conditions that the node voltage and branch current must satisfy, respectively, and Equations (5)-(6) are the branch current and node voltage constraint conditions; For any radial medium-voltage distribution network, the general constraints (1)-(6) should be satisfied; Since equations (1)-(4) all contain square terms of branch current or node voltage, the model is a nonlinear model, which makes it difficult to obtain the characteristic parameters for subsequent topology identification. and Equations (1)-(4) can be linearized into the following model: Based on the method of topology identification of homogeneous medium-voltage distribution network with fixed line parameter ratio, ignoring line loss, (7) and (8) are transformed into expression (9) containing the physical connection relationship required for topology identification: in The matrix represents the square of the voltage amplitude of each node, the p and q matrices represent the active injection power and reactive injection power of the node respectively, the A matrix is ​​the adjacency matrix related to the topological structure of the distribution network, and its elements are shown in equations (10) and (11); the R and X matrices represent the line resistance and reactance parameters. In equation (10), each line parameter is expressed in the form of corresponding diagonal elements; [a0,A T ] T ∈{0,±1} (n+1)×n (11) In formula (9), the active and reactive injection powers of each node are known, and the elements of the adjacency matrix A related to the topological structure of the distribution network are unknown. For the radial distribution network, the adjacency matrix A must be a non-singular matrix, and A -T a0 must be the unit matrix, denoted by A -T a0=E n ; Therefore, (9) can be transformed into: Among them, AX -1 A T That is, the node admittance matrix Y, whose elements directly reflect the physical connection relationship of the distribution network. If there is no physical connection relationship between node i and node j, then Y ij =0; Node i is directly connected to node j, then Y ij ≠0, according to (12), the elements in the node admittance matrix Y can be obtained, and the topological connection mode of the medium voltage distribution network can be obtained; p, q, are all known quantities. The key to finding the elements in Y is how to determine AX -1 RA -1 The value of X -1 R is actually the ratio of line resistance to reactance parameters. Assuming the ratio is a certain value, this assumption is consistent with the actual power line parameter setting; that is: At this time, formula (12) becomes: The medium voltage distribution network topology identification method can be summarized as follows: the active power and reactive power obtained by aggregation on the medium voltage side are measured, and each element in the admittance matrix Y of the distribution network node is obtained, so as to obtain the physical connection relationship between the corresponding nodes.

2. A medium voltage distribution network topology identification method considering the similarity of distribution network areas according to claim 1, characterized in that: In the medium-voltage side measurement aggregation data set, the active power of the medium-voltage side node is equal to the sum of the active power consumed by the IoT devices and non-IoT devices on the low-voltage side. The active power of the IoT devices can be obtained in real time, while there is a lack of active data information on other non-IoT power-consuming devices. The historical data of the distribution network contains the spatiotemporal correlation characteristics between the operating states of the distribution network. When the user's electricity consumption behavior is not affected by uncertain factors, the proportion of power-consuming equipment types should not fluctuate significantly within a certain time range. Through the distribution network user's electricity consumption data, their electricity consumption behavior is explored, and the proportion of different types of power-consuming equipment is obtained, and then the active power measurement data of the IoT devices on the low-voltage side are proportionally aggregated to the medium-voltage side.

3. A medium voltage distribution network topology identification method considering the similarity of distribution network areas according to claim 1, characterized in that: The regional similarities in the unit planning of medium-voltage distribution network include the similarity of the overall scale of medium-voltage distribution substations, the similarity of the mean out-degree and in-degree of all nodes in each distribution network substation, and the similarity indicators of adjacency matrix singular values, spectral radius, and matrix norm.

4. A medium voltage distribution network topology identification method considering the similarity of distribution network areas according to claim 1, characterized in that: Similarity in the overall scale of medium-voltage distribution substations: Due to the voltage level restrictions of substations, the transmission distance and power of distribution substations are limited, so that the expansion range and overall scale of different distribution substations in the medium-voltage distribution network are similar.

5. A method for identifying medium voltage distribution network topology considering the similarity of distribution network regions according to claim 4, characterized in that: The similarity of the mean out-degree and in-degree of all nodes in each substation of the distribution network. The regional users supplied by the same medium-voltage distribution network belong to the same functional area, and their physical distribution mode and regional structure are consistent, which affects the installation plan of power lines and leads to the natural similarity of the out-degree and in-degree of the node lines in the substation.

6. A method for identifying medium voltage distribution network topology considering the similarity of distribution network regions according to claim 4, characterized in that: Adjacency matrix singular value, spectral radius, and matrix norm similarity indicators: Some literature uses mathematical tools to directly calculate the similarity index of distribution network substations based on the adjacency matrix singular value as an important indicator to describe the essence of topology, and gives the mathematical quantitative relationship between the similarity of power grid topology and the power supply reliability of distribution network.

7. A method for identifying medium voltage distribution network topology considering the similarity of distribution network regions according to claim 4, characterized in that: Based on the similarity of the medium-voltage distribution network topology, the connection information in the known substation topology is fully mined to provide reasonable hyperparameter constraints for the proposed identification method. The hyperparameters comprehensively consider the node level, out-degree and in-degree of the known substation to provide a reference for the distribution network area with unknown topology structure. The distribution network area to be determined can also use this hyperparameter to eliminate unreasonable topology identification results.

Images