Transformer area topology identification method and device based on multi-feature fusion

Abstract

The invention discloses a transformer area topology identification method and device based on multi-feature fusion, and the method comprises the steps: S1, distributing an initial branch node for each user, constructing an initial topological structure identification model, and obtaining a preliminary branch node-user association relation adjacency matrix; s2, constructing a topological optimization model based on difference characteristics, and enabling the error between branch node power difference and user power difference combinations to be minimum; s3, constructing a topological optimization model based on peak-valley feature matching, and enabling the matching degree of the peak-valley feature combination of the branch node and the peak-valley feature combination of the user to be maximum; s4, constructing an optimization model based on total fitting error correction, and enabling the total error between the actual power and the fitting power of the branch node to be minimum; s5, judging whether undistributed users exist or not, if not, obtaining a final topology identification result, and otherwise, performing S1 to S4 again; according to the method, the area topology connection relation is effectively identified.

Description

Technical Field

[0001] The present invention belongs to the field of metropolitan area topology identification methods, and in particular relates to a metropolitan area topology identification method and device based on multi-feature fusion. Background Art

[0002] In today's energy transition, the power system is undergoing unprecedented transformation. With the increasing penetration of renewable energy and the large-scale integration of distributed power sources, traditional distribution networks face increasingly complex operating environments and management challenges. Particularly at the distribution substation level, as the critical link connecting the power system to end users, accurate identification of its topology has become a key issue that needs to be addressed in smart grid construction.

[0003] The topology of a distribution network refers to the electrical connections between transformers, branch nodes, users, and other equipment within the network. Accurately understanding these connections is crucial for grid operation and management. However, due to a deficient low-voltage distribution network ledger information system, problems such as incomplete archival data, untimely line modification updates, and discrepancies between actual connections and design plans have led to inaccurate or even missing topology information. This problem is particularly prominent in older residential communities and rural areas.

[0004] Traditional topology identification methods rely primarily on manual on-site surveys and document verification, which are not only labor-intensive and inefficient, but also easily affected by subjective factors and the external environment, making it difficult to guarantee the accuracy of the identification results. With the widespread application of smart meters and distribution automation equipment, topology identification methods based on measurement data have gradually become a research hotspot. Currently, academia and industry have proposed a variety of topology identification methods based on data analysis, including correlation analysis, state estimation, and harmonic feature analysis. However, these methods often only utilize a single type of feature information. When faced with the complex and changing actual power grid environment, the identification performance is easily affected by factors such as noise interference, load fluctuations, and unbalanced operation, making it difficult to meet practical requirements.

[0005] Furthermore, existing methods fail to fully consider the multi-dimensional correlation between user power curves within a substation and branch node power curves, and lack effective iterative optimization mechanisms, resulting in instability and inadequate adaptability in identification results. In practical applications, how to comprehensively utilize the various time series features contained in power data to establish a robust topology identification model has become a technical bottleneck that urgently needs to be overcome.

[0006] Therefore, developing a substation topology identification method that can fully exploit the multi-dimensional characteristics of power data and has strong anti-interference and adaptability has important theoretical value and practical significance for improving the operating efficiency of the distribution network, enhancing fault location capabilities, and supporting precise load management. Summary of the Invention

[0007] In order to solve the problems existing in the background technology, the present invention aims to propose a multi-stage optimized substation topology identification method and device based on multi-feature fusion. By constructing a multi-stage feature optimization model and comprehensively utilizing the differential features, peak-valley features and overall fitting features in the power curve, high-precision identification of the connection relationship between users and branch boxes in the distribution substation area is achieved; this method aims to overcome the limitations of existing single feature identification methods, reduce dependence on physical models and prior knowledge, and improve the robustness and adaptability of identification, especially in complex environments with severe load fluctuations and noisy measurement data; the core goal of the present invention is to accurately reveal the topological connection relationship hidden behind the electrical measurement data in a data-driven manner, to provide technical support for improving the global observability of the distribution substation area, strengthening the ability to quickly locate faults, and optimizing the operation and management of the distribution network, and ultimately achieve a dual improvement in the operation efficiency of the distribution network and the reliability of power supply.

[0008] The present invention adopts the following technical solutions:

[0009] The method for identifying the topology of a substation area based on multi-feature fusion includes the following steps:

[0010] S1. Assign an initial branch node to each user, ensuring that each branch node has at least one user. Build an initial topology structure recognition model to maximize the correlation between the branch node power and the user power combination. Obtain a preliminary branch node-user association relationship adjacency matrix through sorting optimization.

[0011] S2. Based on the preliminary branch node-user association relationship adjacency matrix obtained in step S1, a topology optimization model based on differential features is constructed to minimize the error between the branch node power difference and the user power difference combination. By performing ergodic optimization and solving, a branch node-user association relationship matrix corrected based on differential features is obtained.

[0012] S3. Based on the branch node-user association relationship matrix corrected based on the differential features obtained in step S2, a topology optimization model based on peak-valley feature matching is constructed to maximize the matching degree between the branch node peak-valley features and the user peak-valley features. By performing traversal optimization and solving, a branch node-user association relationship matrix that is further optimized and corrected based on the peak-valley features is obtained.

[0013] S4. Based on the branch node-user association relationship matrix obtained in step S3 and optimized based on peak-valley characteristics, an optimization model based on overall fitting error correction is constructed to minimize the overall error between the actual power and the fitted power of the branch node. By performing traversal optimization and solving, a branch node-user association relationship matrix based on comprehensive fitting error correction is obtained.

[0014] S5. Determine whether there are unassigned users. If not, obtain the final topology recognition result; otherwise, repeat steps S1 to S4.

[0015] Further, the objective function of the initial topology structure recognition model is to maximize the correlation between the branch node power and the user power combination, which can be expressed as:

[0016]

[0017] In the formula, M is the number of branch nodes; N is the number of users; corr i,j is the correlation coefficient between the power curves of branch node i and user node j; C i,j is the connection variable, which takes the value of 1 when user j is connected to branch node i, and 0 otherwise;

[0018] The constraint conditions of the initial topology structure recognition model include: each user is only connected to one branch node; each branch node is connected to at least one user.

[0019] Further, the calculation formula for the correlation coefficient between user nodes and branch nodes is:

[0020]

[0021] In the formula, P i,t is the power consumption of branch node i at time t; P j,t is the power consumption of user node j at time t; and are the average powers of branch node i and user node j respectively; T is the length of the time series.

[0022] Further, the objective function of the topology optimization model based on differential features is to minimize the error between the branch node power difference and the user power difference combination, expressed as:

[0023]

[0024] In the formula, T is the length of the time series, M is the number of branch nodes, ΔP i,t is the power difference value of branch node i at time t, ΔP i,t =P i,t -P i,t-1 P i,t is the power consumption of branch node i at time t; ΔY i,t is the regression power difference value of branch node i, and its expression can be represented as:

[0025]

[0026] In the formula, P j,tThe power consumption of user node j at time t, A i Is the set of user nodes associated with branch node i.

[0027] Furthermore, the objective function of the topology optimization model based on peak-valley feature matching is to maximize the matching degree between the peak-valley features of branch nodes and the peak-valley features of users, expressed as:

[0028]

[0029] In the formula, M is the number of branch nodes, P i peaks Is the set of peak features of the power curve of branch node i, P j peaks Is the set of peak features of the power curve of user node j, P i valley Is the set of valley features of the power curve of branch node i, P j valley Is the set of valley features of the power curve of user node j, A i Is the set of user nodes associated with branch node i, and Mat is the evaluation function for peak and valley feature matching degrees.

[0030] Furthermore, the evaluation functions for peak and valley feature matching degrees are respectively:

[0031]

[0032] Among them, Is the peak matching accuracy of branch node i, Is the valley matching accuracy of branch node i, Is the peak matching recall rate of branch node i, Is the valley matching recall rate of branch node i, where:

[0033]

[0034] In the formula, Represents the number of peaks in the power curve of branch node i that are successfully matched with the peaks of the power curve of user node, Represents the number of valleys in the power curve of branch node i that are successfully matched with the valleys of the power curve of user node, and the calculation formula is:

[0035]

[0036] In the formula, K i P Is the number of peaks in the power curve of branch node i; J j P Is the number of peaks in the power curve of user node j; Indicates the time point of the kth peak of the power curve of branch node i; represents the time point of the hth peak of the power curve of user node j; τ is the time allowed threshold; K i V is the number of valley values of the power curve of branch node i; J j V is the number of valleys in the power curve of user node j; Indicates the time point of the kth valley value of the power curve of branch node i; represents the time point of the hth valley of the power curve of user node j; δ(·) is an explicit function, which takes 1 when the condition is met and 0 otherwise;

[0037] Indicates the number of peaks in the user combined power curve that fail to match the peaks of the branch node power curve. Indicates the number of valley values in the user combination power curve that fail to match the valley values of the branch node power curve. The expression is:

[0038]

[0039] Indicates the number of peaks in the branch node power curve that are not matched by the peak of the user combined power curve. Indicates the number of valleys in the branch node power curve that are not matched by the valleys of the user combined power curve. The expression is:

[0040]

[0041] Furthermore, the objective function of the overall fitting error correction optimization model is to minimize the overall error between the actual power and the fitting power of the branch node, which is expressed as:

[0042]

[0043] Where Y i,t is the fitting power of branch node i at time t, and its expression is:

[0044]

[0045] Among them, P i,t is the power consumption of branch node i at time t; P j,t is the power consumption of user node j at time t; T is the length of the time series; M is the number of branch nodes; A i The set of user nodes associated with branch node i.

[0046] The device for identifying the topology of the substation based on multi-feature fusion includes:

[0047] The initial topology recognition model construction and solution module is used to allocate initial branch nodes to each user, while ensuring that each branch node has at least one user, construct an initial topology recognition model to maximize the correlation between the branch node power and the user power combination, and obtain a preliminary branch node-user association relationship adjacency matrix through sorting optimization and solution;

[0048] The topology optimization model construction and solution module based on differential features is used to construct a topology optimization model based on differential features on the basis of the preliminary branch node-user association relationship adjacency matrix obtained by the initial topology recognition model construction and solution module, to minimize the error between the branch node power difference and the user power difference combination, and obtain a branch node-user association relationship matrix corrected based on differential features through traversal optimization and solution;

[0049] The topology optimization model construction and solution module based on peak-valley feature matching is used to construct a topology optimization model based on peak-valley feature matching on the basis of the branch node-user association relationship matrix corrected by differential features obtained by the topology optimization model construction and solution module based on differential features, to maximize the matching degree between the branch node peak-valley feature and the user peak-valley feature combination, and obtain a branch node-user association relationship matrix further optimized and corrected based on peak-valley features through traversal optimization and solution;

[0050] The overall fitting error correction and optimization model construction and solution module is used to construct an overall fitting error correction and optimization model on the basis of the branch node-user association relationship matrix optimized and corrected by peak-valley features obtained by the topology optimization model construction and solution module based on peak-valley feature matching, to minimize the overall error between the actual power of the branch node and the fitting power, and obtain a branch node-user association relationship matrix corrected based on comprehensive fitting error through traversal optimization and solution;

[0051] The judgment module is used to judge whether there are unallocated users. If not, obtain the final topology recognition result. Otherwise, re-execute the initial topology recognition model construction and solution module, the topology optimization model construction and solution module based on differential features, the topology optimization model construction and solution module based on peak-valley feature matching, and the overall fitting error correction and optimization model construction and solution module in sequence.

[0052] A non-transitory computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it implements the above-mentioned multi-feature fusion-based substation area topology identification method.

[0053] An electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, and when the processor executes the program, it implements the above-mentioned multi-feature fusion-based substation area topology identification method.

[0054] The beneficial technical effects of the present invention are as follows:

[0055] The present invention proposes a method for identifying the topology of a distribution area based on multi-feature fusion, which fully explores the multi-dimensional correlation features between the user power curve and the branch box power curve in the distribution area, reduces the dependence on physical models and prior knowledge, efficiently identifies the association relationship between the branch nodes and user nodes in the distribution area, constructs an initial topology structure recognition model based on correlation analysis, a topology optimization model based on differential features, a topology optimization model based on peak-valley feature matching, and a final optimization model based on the overall fitting error, realizes the effective identification of the topology connection relationship of the distribution area, thereby improving the overall observability performance of the distribution area and providing a good topology transparency basis for the operation analysis of the distribution area. Description of the Drawings

[0056] Figure 1 It is a schematic flow chart of the method for identifying the topology of a distribution area based on multi-feature fusion in an embodiment of the present invention;

[0057] Figure 2 It is the topology structure diagram of the distribution area in Embodiment 1 of the present invention;

[0058] Figure 3 It is the load power diagram of the branch nodes in the distribution area in Embodiment 1 of the present invention;

[0059] Figure 4 It is the load power diagram of the user nodes in the distribution area in Embodiment 1 of the present invention;

[0060] Figure 5 It is the power curve diagram of some branch nodes and user nodes in Embodiment 1 of the present invention. Detailed Embodiment

[0061] The following further clearly and completely describes a method and device for clearing the power capacity market considering flexibility provided by the present invention with reference to the drawings:

[0062] Embodiment 1

[0063] A method for identifying the topology of a distribution area based on multi-feature fusion, the method includes the following steps:

[0064] S1. Assign an initial branch node to each user, and at the same time ensure that each branch node has at least one user, construct an initial topology structure recognition model to maximize the correlation between the branch node power and the user power combination, and obtain the preliminary adjacency matrix of the branch node-user association relationship through sorting optimization;

[0065] S2. On the basis of the preliminary branch node-user association relationship adjacency matrix obtained in step S1, construct a topology optimization model based on differential features to minimize the error between the combination of branch node power differentials and user power differentials. Through traversal optimization and solution, obtain the branch node-user association relationship matrix corrected based on differential features;

[0066] S3. On the basis of the branch node-user association relationship matrix corrected by differential features obtained in step S2, construct a topology optimization model based on peak-valley feature matching to maximize the matching degree between the peak-valley features of branch nodes and the combination of user peak-valley features. Through traversal optimization and solution, obtain the branch node-user association relationship matrix further optimized and corrected based on peak-valley features;

[0067] S4. On the basis of the branch node-user association relationship matrix optimized and corrected by peak-valley features obtained in step S3, construct an overall fitting error correction and optimization model to minimize the overall error between the actual power of branch nodes and the fitting power. Through traversal optimization and solution, obtain the branch node-user association relationship matrix corrected based on comprehensive fitting error;

[0068] S5. Determine whether there are unassigned users. If not, obtain the final topology recognition result; otherwise, repeat steps S1 to S4.

[0069] It should be noted that before performing the above steps of the substation area topology identification method based on multi-feature fusion, it is necessary to perform:

[0070] (1). Data acquisition and preprocessing: Obtain branch node power data and user power data, obtain the number of users, the number of time points, and the number of branch nodes, and perform time series segmentation processing on the user power data and branch node power data;

[0071] (2). Feature extraction: Extract multi-dimensional features of the power curve, including first-order differential features, peak features, and valley features, to enhance the time series characteristics of the data;

[0072] First, global topology exploration based on correlation: Calculate the correlation coefficient between the power of branch nodes and user nodes, and perform initial topology construction. Assign initial branch nodes to each user, and at the same time ensure that each branch node has at least one user; specifically, the objective function of the initial topology structure recognition model is to maximize the correlation between the combination of branch node power and user power, which can be expressed as:

[0073]

[0074] In the formula, M is the number of branch nodes; N is the number of users; corr i,j is the correlation coefficient between the power curves of branch node i and user node j; Ci,j is a connection variable, which takes the value 1 when user j is connected to branch node i, otherwise it takes the value 0;

[0075] The constraints of the initial topology structure recognition model include: each user is connected to only one branch node, and each branch node is connected to at least one user, which can be expressed as:

[0076]

[0077]

[0078] The calculation formula for the correlation coefficient between user nodes and branch nodes is:

[0079]

[0080] Where, P i,t is the power consumption of branch node i at time t; P j,t is the power consumption of user node j at time t; and are the mean power of branch node i and user j respectively; T is the length of the time series.

[0081] When solving the initial topology structure recognition model, an adjacency matrix R is constructed for all branch nodes and user nodes. The rows represent branch nodes and the columns represent user nodes. If branch node i is associated with user node j, the corresponding adjacency matrix element in row i and column j is set to 1, otherwise, it is set to 0.

[0082]

[0083] To solve the element corresponding to branch node i in the adjacency matrix, sort the correlation coefficients between branch node i and each user node calculated above from large to small, and select the users corresponding to the largest correlation coefficients as being associated with branch node i. The cutoff condition is that the sum of the average power of the selected users does not exceed the mean power of the branch node. For example, if users a and b have the largest correlation coefficients with branch node i, and the sum of the mean power of the users is exactly the mean power of the branch node, then the corresponding elements in row i, column a, and row i, column b of the adjacency matrix are set to 1.

[0084]

[0085] The objective function of the topology optimization model based on differential features is to minimize the error between the branch node power difference and the user power difference combination, which can be expressed as:

[0086]

[0087] Where ΔP i,tis the power difference value of branch node i at time t, ΔP i,t = P i,t - P i,t-1 ; ΔY i,t is the regression power difference value of branch node i, and its expression can be expressed as:

[0088]

[0089] In the formula, A i is the set of user nodes associated with branch node i;

[0090] When solving the topology optimization model based on differential features, calculate the power differential feature matching degree under the current configuration, and process the branch nodes in the order of the differential error from large to small, try to exchange the users between different branch nodes, and retain the exchange results that can significantly reduce the differential error; specifically:

[0091] a) Calculate the measured differential power and the fitting differential power error of each branch node under the current configuration respectively;

[0092] b) Based on the users to which the branch nodes are initially determined to belong, traverse and exchange the user nodes in pairs, and calculate the differential feature error after the exchange;

[0093] c) Compare the change in the differential feature error of the branch nodes before and after the user exchange. When the reduction in the differential feature error of both branch nodes after the exchange exceeds 5% before the exchange, and the power of the users to which the branch nodes belong after the exchange does not exceed the sum of the powers of the branch nodes, then perform user exchange correction;

[0094] d) Immediately update the adjacency matrix of the branch node-user association relationship after each successful exchange.

[0095] The objective function of the topology optimization model based on peak-valley feature matching is to maximize the matching degree of the peak-valley features of the branch nodes and the combination of the peak-valley features of the users, which is expressed as:

[0096]

[0097] In the formula, P i peaks is the set of peak features of the power curve of branch node i, P j peaks is the set of peak features of the power curve of user node j, P i valley is the set of valley features of the power curve of branch node i, P j valley is the set of valley features of the power curve of user node j, A iis the set of user nodes associated with branch node i, and Mat is the peak and valley feature matching evaluation function, which is expressed as:

[0098]

[0099] in, is the peak matching accuracy of branch node i, is the valley value matching accuracy of branch node i, is the peak matching recall rate of branch node i, is the recall rate of valley matching of branch node i, where:

[0100]

[0101] Where, Indicates the number of peaks in the power curve of branch node i that successfully match the peaks of the power curve of the user node, It represents the number of valleys in the power curve of branch node i that successfully match the peak of the power curve of the user node. The calculation formula is:

[0102]

[0103] Where K i P is the peak number of the power curve of branch node i; J j P is the number of peaks in the power curve of user node j; Indicates the time point of the kth peak of the power curve of branch node i; represents the time point of the hth peak of the power curve of user node j; τ is the time allowed threshold; K i V is the number of valley values of the power curve of branch node i; J j V is the number of valleys in the power curve of user node j; Indicates the time point of the kth valley value of the power curve of branch node i; represents the time point of the hth valley value of the power curve of user node j; δ(·) is an explicit function, which takes the value of 1 when the condition is met and 0 otherwise.

[0104] Indicates the number of peaks in the user combined power curve that fail to match the peaks of the branch node power curve. Indicates the number of valley values in the user combination power curve that fail to match the valley values of the branch node power curve. The expression is:

[0105]

[0106] Indicates the number of peaks in the branch node power curve that are not matched by the peaks of the user combined power curve, Indicates the number of valleys in the branch node power curve that are not matched by the valleys of the user combined power curve. The expression is:

[0107]

[0108] The solution of the topology optimization model based on peak-valley feature matching is equivalent to local refinement optimization based on peak-valley features: extracting peak and valley features, calculating the feature matching degree score, evaluating the peak-valley feature matching degree before and after user swapping based on the peak-valley feature matching function, and retaining the swapping results that can significantly improve the matching degree; specifically:

[0109] a) Calculate the peak and valley feature matching degrees of each branch node and its affiliated user nodes under the current configuration respectively;

[0110] b) Based on the users affiliated to the branch nodes determined in the previous step of the model, traverse and swap the user nodes pairwise, and calculate the peak and valley feature matching degrees after swapping;

[0111] c) Compare the changes in the peak and valley feature matching degrees of the branch nodes before and after user swapping. When the growth of the peak and valley feature matching degrees of both branch nodes after swapping exceeds 5% before swapping, and the sum of the powers of the users affiliated to the branch nodes after swapping does not exceed the power of the branch node, then perform user swapping correction;

[0112] d) Immediately update the adjacency matrix of the branch node - user association relationship after each successful swap.

[0113] The objective function of the optimization model based on overall fitting error correction is to minimize the overall error between the actual power and the fitting power of the branch nodes, expressed as:

[0114]

[0115] In the formula, Y i,t is the fitting power of branch node i at time t, and its expression is:

[0116]

[0117] When solving the above optimization model based on overall fitting error correction, calculate the fitting error of each branch node under the current topological structure, try to move the users to reduce the overall fitting error, and verify whether the movement brings significant improvement; specifically:

[0118] a) Calculate the error between the measured power and the fitting power of each branch node under the current configuration respectively;

[0119] b), Based on the users to whom the branch nodes are determined in the previous step of the model, traverse and exchange the user nodes to other branch nodes, and calculate the overall error between the measured power and the fitted power of each branch node after the exchange;

[0120] c), Compare the change in the overall error between the measured power and the fitted power of the branch nodes before and after the user exchange. When the overall error between the measured power and the fitted power after the exchange is reduced by 5% compared to before the exchange, and the sum of the powers of the users to whom the branch nodes belong after the exchange does not exceed the power of the branch nodes, then perform user exchange correction;

[0121] d), Immediately update the adjacency matrix of the branch node - user association relationship after each successful exchange.

[0122] As an example, in this embodiment, in order to identify the branch node - user node association relationship of the incoming substation area, a structure model of 12 branch nodes and 36 user nodes is respectively constructed, and the algorithm is verified using the measured power data of a certain substation area. The topological structure of the substation area is as Figure 2 shown.

[0123] Based on the load power change curve of the open - source data set, power loads are configured for each user node, and the power consumption for each user node is set at 15 - minute time intervals. The load change situation of the branch nodes generated is as Figure 3 shown, Figure 4 showing the power distribution of the user nodes in the substation area.

[0124] Based on the designed substation area scenario, the methods described in this embodiment, the power similarity algorithm, and the least - squares method are respectively used to identify the association relationship between the branch nodes and the user nodes. The identification results are shown in Table 1. The verification results show that the algorithm of this embodiment is superior to the other two algorithms. Figure 5 The partial user nodes and branch node load powers after identification are shown. It can be seen from the figure that the power change range of the users is large, with obvious power peak - valley characteristics, and the proposed algorithm is based on power difference characteristics and peak - valley characteristics to identify the branch node - user association relationship, which can better adapt to such a scenario and give full play to its unique advantages.

[0125] Table 1 Identification accuracy result table of the substation area user - branch node association relationship under different strategies

[0126]

[0127]

[0128] Embodiment 2

[0129] This embodiment provides a substation area topology identification device based on multi - feature fusion, including:

[0130] The initial topology recognition model construction and solution module is used to assign initial branch nodes to each user, while ensuring that each branch node has at least one user, construct an initial topology recognition model to maximize the correlation between the branch node power and the user power combination, and obtain a preliminary branch node-user association relationship adjacency matrix through sorting optimization and solution;

[0131] The topology optimization model construction and solution module based on differential features is used to construct a topology optimization model based on differential features on the basis of the preliminary branch node-user association relationship adjacency matrix obtained by the initial topology recognition model construction and solution module, to minimize the error between the branch node power difference and the user power difference combination, and obtain a branch node-user association relationship matrix corrected based on differential features through traversal optimization and solution;

[0132] The topology optimization model construction and solution module based on peak-valley feature matching is used to construct a topology optimization model based on peak-valley feature matching on the basis of the branch node-user association relationship matrix corrected by differential features obtained by the topology optimization model construction and solution module based on differential features, to maximize the matching degree between the branch node peak-valley feature and the user peak-valley feature combination, and obtain a branch node-user association relationship matrix further optimized and corrected based on peak-valley features through traversal optimization and solution;

[0133] The overall fitting error correction and optimization model construction and solution module is used to construct an overall fitting error correction and optimization model on the basis of the branch node-user association relationship matrix optimized and corrected by peak-valley features obtained by the topology optimization model construction and solution module based on peak-valley feature matching, to minimize the overall error between the actual power of the branch node and the fitting power, and obtain a branch node-user association relationship matrix corrected based on comprehensive fitting error through traversal optimization and solution;

[0134] The judgment module is used to judge whether there are unassigned users. If not, the final topology recognition result is obtained. Otherwise, the initial topology recognition model construction and solution module, the topology optimization model construction and solution module based on differential features, the topology optimization model construction and solution module based on peak-valley feature matching, and the overall fitting error correction and optimization model construction and solution module are executed again in sequence.

[0135] A non-transitory computer-readable storage medium stores a computer program thereon. When the computer program is executed by a processor, it implements the above-mentioned multi-feature fusion-based substation area topology identification method.

[0136] Furthermore, the present invention adopts the following technical solutions:

[0137] An electronic device includes a memory, a processor, and a computer program stored on the memory and executable on the processor. When the processor executes the program, it implements the above-mentioned method for identifying the topology of a power distribution area based on multi-feature fusion.

[0138] Through the description of the above embodiments, those skilled in the art can clearly understand that the facilities of the present invention can be implemented by means of software plus a necessary general hardware platform. The embodiments of the present invention can be implemented using existing processors, or by dedicated processors used for this purpose or other purposes in a suitable system, or by a hardwired system. The embodiments of the present invention also include non-transitory computer-readable storage media, which include machine-readable media for carrying or having machine-executable instructions or data structures stored thereon; such machine-readable media can be any available medium accessible by a general or dedicated computer or other machine having a processor. For example, such machine-readable media can include RAM, ROM, EPROM, EEPROM, CD-ROM or other optical disc memories, magnetic disk memories or other magnetic storage devices, or any other medium that can be used to carry or store the required program code in the form of machine-executable instructions or data structures and can be accessed by a general or dedicated computer or other machine with a processor. When information is transmitted or provided to a machine through a network or other communication connection (hardwired, wireless, or a combination of hardwired and wireless), this connection is also regarded as a machine-readable medium.

[0139] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

Claims

1. A method for identifying the topology of a power distribution area based on multi-feature fusion, characterized in that, The method includes the following steps: S1. Assign an initial branch node to each user, and at the same time ensure that each branch node has at least one user, construct an initial topology structure recognition model to maximize the correlation between the branch node power and the user power combination, and obtain a preliminary branch node-user association relationship adjacency matrix through sorting optimization solution; S2. On the basis of the preliminary branch node-user association relationship adjacency matrix obtained in step S1, construct a topology optimization model based on differential features to minimize the error between the branch node power difference and the user power difference combination, and obtain a branch node-user association relationship matrix corrected based on differential features through traversal optimization solution; S3. On the basis of the branch node-user association relationship matrix corrected by differential features obtained in step S2, construct a topology optimization model based on peak-valley feature matching to maximize the matching degree between the branch node peak-valley feature and the user peak-valley feature combination, and obtain a branch node-user association relationship matrix further optimized and corrected based on peak-valley features through traversal optimization solution; S4. On the basis of the branch node-user association relationship matrix optimized and corrected by peak-valley features obtained in step S3, construct an overall fitting error correction optimization model to minimize the overall error between the actual power of the branch node and the fitting power, and obtain a branch node-user association relationship matrix corrected based on comprehensive fitting error through traversal optimization solution; S5. Determine whether there are unassigned users. If not, obtain the final topology recognition result; otherwise, repeat steps S1 to S4.

2. The method for identifying the substation area topology based on multi-feature fusion according to claim 1, wherein, The objective function of the initial topology structure recognition model is to maximize the correlation between the branch node power and the user power combination, which can be expressed as: where M is the number of branch nodes; N is the number of users; corr i,j is the correlation coefficient of the power curves of branch node i and user node j; C i,j is the connection variable, which takes the value of 1 when user j is connected to branch node i, and 0 otherwise; The constraint conditions of the initial topology structure recognition model include: each user is only connected to one branch node; Each branch node is connected to at least one user.

3. The method for identifying the substation area topology based on multi-feature fusion according to claim 2, wherein The calculation formula for the correlation coefficient between the user node and the branch node is: Wherein, P i,t is the power consumption of branch node i at time t; P j,t is the power consumption of user node j at time t; and are the average powers of branch node i and user node j respectively; T is the length of the time series.

4. The method for identifying the topological structure of a transformer substation area based on multi-feature fusion according to claim 1, wherein The objective function of the topology optimization model based on differential features is to minimize the error between the branch node power difference and the user power difference combination, which is expressed as: where T is the length of the time series, M is the number of branch nodes, and ΔP i,t is the power difference value of branch node i at time t, and ΔP i,t = P i,t - P i,t-1 , where P i,t is the power consumption of branch node i at time t; ΔY i,t is the regression power difference value of branch node i, and its expression can be represented as: Where, P j,t is the power consumption of user node j at time t, and A i is the set of user nodes associated with branch node i.

5. The method for identifying the topological structure of a transformer substation area based on multi-feature fusion according to claim 1, characterized in that, The objective function of the topology optimization model based on peak-valley feature matching is to maximize the matching degree between the branch node peak-valley feature and the user peak-valley feature combination, which is expressed as: Where M is the number of branch nodes, and P i peaks is the peak feature set of the power curve of branch node i, and P j peaks is the peak feature set of the power curve of user node j, and P i valley is the trough feature set of the power curve of branch node i, and P j valley is the trough feature set of the power curve of user node j, and A i is the set of user nodes associated with branch node i, and Mat is the peak and trough feature matching degree evaluation function.

6. The method for identifying the substation area topology based on multi-feature fusion according to claim 5, wherein, The peak and valley value feature matching degree evaluation functions are respectively: Among them, is the peak matching accuracy of branch node i, is the valley matching accuracy of branch node i, is the peak matching recall rate of branch node i, is the valley matching recall rate of branch node i, where: Wherein, represents the number of peaks in the power curve of branch node i that are successfully matched with the peak value of the power curve of the user node, represents the number of troughs in the power curve of branch node i that are successfully matched with the peak value of the power curve of the user node, and the calculation formula is: In the formula, is the number of peaks of the power curve of branch node i; is the number of peaks of the power curve of user node j; represents the time point of the k-th peak of the power curve of branch node i; represents the time point of the h-th peak of the power curve of user node j; τ is the time allowance threshold; K i V is the number of valleys of the power curve of branch node i; J j V is the number of valleys of the power curve of user node j; represents the time point of the k-th valley of the power curve of branch node i; represents the time point of the h-th valley of the power curve of user node j; δ(·) is a dominant function, which takes 1 when the condition is satisfied and 0 otherwise. Indicates the number of peaks in the combined user power curve that do not match the peak value of the branch node power curve, Indicates the number of troughs in the combined user power curve that do not match the trough value of the branch node power curve. The expression is: Indicates the number of peaks in the branch node power curve that are not matched by the peaks of the user combined power curve, Indicates the number of valleys in the branch node power curve that are not matched by the valleys of the user combined power curve. The expression is:

7. The method for identifying the topological structure of a transformer substation area based on multi-feature fusion according to claim 1, wherein The objective function of the overall fitting error correction optimization model is to minimize the overall error between the actual power of the branch node and the fitting power, which is expressed as: where Y i,t is the fitted power of branch node i at time t, and its expression is: Among them, P i,t is the power consumption of branch node i at time t; P j,t is the power consumption of user node j at time t; T is the length of the time series; M is the number of branch nodes; A i is the set of user nodes associated with branch node i.

8. The device for identifying the topology of the metropolitan area based on multi-feature fusion is characterized by: including: An initial topology structure recognition model construction and solution module, which is used to assign an initial branch node to each user, and at the same time ensure that each branch node has at least one user, construct an initial topology structure recognition model to maximize the correlation between the branch node power and the user power combination, and obtain a preliminary branch node-user association relationship adjacency matrix through sorting optimization solution; The topology optimization model construction and solution module based on differential features is used to construct a topology optimization model based on differential features on the basis of the preliminary branch node-user association relationship adjacency matrix obtained by the initial topology structure recognition model construction and solution module, so that the error between the combination of branch node power differences and user power differences is minimized. Through traversal optimization and solution, a branch node-user association relationship matrix corrected based on differential features is obtained; The topology optimization model construction and solution module based on peak-valley feature matching is used to construct a topology optimization model based on peak-valley feature matching on the basis of the branch node-user association relationship matrix corrected by differential features obtained by the topology optimization model construction and solution module based on differential features, so that the matching degree between the combination of branch node peak-valley features and user peak-valley features is maximized. Through traversal optimization and solution, a branch node-user association relationship matrix further optimized and corrected based on peak-valley features is obtained; The overall fitting error correction optimization model construction and solution module is used to construct an overall fitting error correction optimization model on the basis of the branch node-user association relationship matrix optimized and corrected by peak-valley features obtained by the topology optimization model construction and solution module based on peak-valley feature matching, so that the overall error between the actual power of the branch node and the fitting power is minimized. Through traversal optimization and solution, a branch node-user association relationship matrix corrected based on comprehensive fitting error is obtained; The judgment module is used to judge whether there are unassigned users. If not, the final topology recognition result is obtained. Otherwise, the initial topology structure recognition model construction and solution module, the topology optimization model construction and solution module based on differential features, the topology optimization model construction and solution module based on peak-valley feature matching, and the overall fitting error correction optimization model construction and solution module are executed in sequence again.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the multi-feature fusion-based substation area topology identification method according to any one of claims 1 to 7.

10. An electronic device, comprising a memory, a processor, and a computer program stored on the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the multi-feature fusion-based substation area topology identification method according to any one of claims 1 to 7.

Images