Non-intrusive topology identification method for transparent low-voltage distribution network

Abstract

The present application relates to the transparent low-voltage distribution network-oriented non-intrusive topology identification method, relates to electric power engineering technical field, the present application converts the spatial correlation of mass unmeasurable node into equivalent impedance attribute through voltage similarity analysis, effectively resolves the topology structure uncertainty caused by physical information loss, determines the membership relation through correlation coefficient preliminary, and combines voltage drop consistency model to remove invalid nodes, gradually compresses the original huge solution space, significantly reduces the calculation amount of single matching, solves the industry pain point of low search efficiency and easy to fall into local optimum, simultaneously introduces tree constraint and power conservation as hard constraint filter in the search process, removes the logical solution not meeting the power operation rule through the improved particle swarm algorithm, ensures the consistency of the final output topology on the mathematical fitting degree and physical feasibility, and adopts the non-intrusive identification technology, completely relies on existing meter data, and significantly reduces the operation and maintenance cost.

Description

Technical Field

[0001] This invention relates to the field of power engineering technology, and in particular to a non-intrusive topology identification method for transparent low-voltage distribution networks. Background Technology

[0002] Low-voltage distribution networks are located at the end of the power system, directly facing end users and closest to load centers, undertaking the critical tasks of power distribution and supply. In the entire power system chain of "generation-transmission-transformation-distribution-consumption", low-voltage distribution networks have the widest coverage and the strongest user connections, and their operational quality directly affects power supply reliability and user satisfaction.

[0003] Low-voltage distribution networks exhibit a typical "complete at both ends, missing in the middle" measurement architecture. In low-voltage distribution networks, smart measurement devices are deployed only on the busbar at the distribution transformer (gate side) and at the user's incoming meter (user side) for transformer monitoring and electricity billing purposes. However, intermediate nodes such as poles and branch meter boxes generally lack effective measurement points, creating a systemic measurement blind spot. Therefore, low-voltage distribution networks are only partially observable, meaning that the acquired measurement data cannot fully cover the status information of all topological nodes.

[0004] Due to the presence of unmeasurable nodes, the solution space of the topology structure is abnormally expanded, leading to a highly complex definition of the structure-data matching degree function and severely limiting the search efficiency and accuracy of the optimal matching topology. Furthermore, the emergence of observation blind spots makes it impossible to construct matching equations for electrical quantities between topology nodes to achieve reconstruction-based topology identification. Clearly, the uneven distribution of systemic measurement points in low-voltage distribution networks has become one of the major challenges in topology identification. To address the aforementioned technical deficiencies, a solution is proposed. Summary of the Invention

[0005] The purpose of this invention is to transform the spatial correlation of massive unmeasurable nodes into equivalent impedance attributes through voltage similarity analysis, effectively resolving the uncertainty of topology caused by the lack of physical information. By initially determining the membership relationship through correlation coefficients and eliminating invalid nodes by combining voltage drop consistency models, the original huge solution space is compressed step by step, significantly reducing the computational load of a single matching. This solves the industry pain points of low search efficiency and easy getting trapped in local optima. At the same time, tree constraints and power conservation are introduced as hard constraints for filtering during the search process. By improving the particle swarm algorithm, logical solutions that do not conform to the power operation law are automatically eliminated, ensuring the consistency of the final output topology in terms of mathematical fit and physical feasibility. Furthermore, non-intrusive identification technology is used, relying entirely on existing meter data, which significantly reduces operation and maintenance costs.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: a non-intrusive topology identification method for transparent low-voltage distribution networks, comprising the following steps: S1. Based on the distributed data of the low-voltage distribution network, set up several measurable nodes and mark the unmeasurable nodes. Collect voltage and power measurement sequences at the measurable nodes and use the correlation coefficient matrix to preliminarily determine the electrical energy affiliation between the measurable nodes. S2. The spatial correlation between measurable and unmeasurable nodes is determined by voltage similarity analysis. By defining virtual aggregation nodes, the black box branch containing multiple unmeasurable nodes is simplified into an equivalent branch with impedance properties. After integration, the node distribution branch diagram is obtained. S3. Establish a structure-data matching degree evaluation model. By introducing the branch power conservation and voltage drop equation, adopting the objective function based on the consistency of branch voltage drop, and simplifying invalid nodes based on the node distribution branch map, the amount of calculation for a single matching is reduced, so as to obtain the effective node distribution map. S4. Based on the tree-like constraints of the physical topology, during the search of the effective node distribution graph, an improved particle swarm optimization algorithm is used to automatically eliminate solutions that do not conform to common sense in electrical engineering, further compressing the solution space until the optimal matching topology is obtained by iterative solution within the physical constraints.

[0007] Furthermore, the specific process for determining the electrical affiliation between measurable nodes is as follows: S101. Mark each measurable node as I, obtain the voltage normalized sequence UI=[U{I,1},U{I,2},...,U{I,n}] of each measurable node I within the sampling period T, and extract the voltage change rate sequence. Defined as: , where t is the sampling time; S102. Select any two measurable nodes I, labeled Ia and Ib respectively, calculate the Pearson correlation coefficient between measurable nodes Ia and Ib, and construct the correlation coefficient matrix as follows: Where Cov is the covariance. Standard deviation; This is the voltage change rate sequence corresponding to the meaable node Ia. This is the voltage change rate sequence corresponding to the measurable node Ib; The value range of is [-1, 1]. The closer it is to 1, the higher the degree of coupling between mea and Ib in the electrical topology; S103. Based on the correlation coefficient matrix, perform the following membership derivation: S1031, Trunk-Branch Affiliation Determination: Preset membership threshold, if If the membership degree is greater than the membership threshold, it is preliminarily determined that mea and Ib are located on the same branch and have a direct parent-child cascade relationship. S1032. Topological Distance Quantization: Utilizing the inverse relationship between electrical distance and correlation, define the virtual electrical distance between nodes: Dab = 1 - ; S1033. Eliminate inter-phase / inter-line interference: For node pairs with Pearson correlation coefficients close to 0 or negative, they are directly determined to be electrically isolated and removed from each other's candidate affiliation list in the initial stage. S104. Output an initial logical topology matrix, which marks the hierarchical depth of each measurable node relative to the transformer side and marks the electrical energy affiliation between the measurable nodes.

[0008] Furthermore, the specific process for determining the spatial correlation patterns between measurable and unmeasurable nodes is as follows: S211. Mark the measurable node as Im, and the unmeasurable nodes adjacent to Im as Iu. Obtain the voltage sequence Um of the measurable node Im and the mean sequence of the reference voltage in the neighborhood. Calculate the voltage sequence Um and the reference voltage mean sequence. The correlation coefficient; S212. Introduce a dynamic time warping algorithm to calculate the alignment cost between the measurable node Im and the voltage trajectory within its associated candidate region. S213. Establish the following spatial correlation discrimination criteria: Strong correlation determination: If the voltage sequence similarity between measurable node Im1 and measurable node Im2 is higher than the preset similarity threshold, and there is a linear voltage drop difference, then it is determined that the unmeasurable node Iu between the two only has the longitudinal impedance attribute and does not have the branch attribute; Weak correlation determination: If the voltage fluctuation waveforms of measurable node Im1 and measurable node Im2 are similar but there is phase lag and nonlinear deviation, it indicates that there is a lateral branch at unmeasurable node Iu, thereby logically identifying the T-connection and cross-connection spatial structure of the unmeasurable node.

[0009] Furthermore, the specific process of obtaining the node distribution branch map is as follows: S221. Based on the power membership relationship between measurable nodes determined in S1, identify all sets of unmeasurable nodes located between two adjacent measurable nodes Im and In, and define the topological region where the set of unmeasurable nodes is located as a black box branch. The boundary of the black box branch is locked by nodes with known measurement data. S222. Obtain the voltage deviation and current vector between measurable node Im and measurable node In, and calculate the total equivalent impedance of the black box branch according to Ohm's law. S223. Based on the spatial correlation discrimination criterion, a continuous unmeasurable node segment with strong correlation is defined as a virtual aggregation node. The virtual aggregation node does not focus on the specific location of a single unmeasurable node inside, but is equivalent to an impedance body with accumulated impedance. The virtual aggregation node serves as an impedance transfer point for power transmission and does not reflect specific physical location coordinates. S224. Assign complex impedance properties to each simplified equivalent branch, and integrate the simplified equivalent branches with the original measurable node connection relationships to form a node distribution branch diagram.

[0010] Furthermore, the specific process for obtaining the effective node distribution map is as follows: S301. Based on Kirchhoff's current law and voltage law, for each equivalent branch in the node distribution branch diagram, the branch power conservation equation and voltage drop equation are introduced: Power conservation constraint: Where P l P is the theoretical voltage value. i P is the measured voltage value. LOSS This represents the theoretical loss of the equivalent branch under the current topological assumptions. Simultaneously define the theoretical value of voltage drop at measurable nodes at both ends of the branch. Compared with the measured voltage drop The residual function between; S302. Constructing an objective function based on voltage drop consistency. The objective function described below is used to evaluate the effectiveness of the current topology branch: , where t is the sampling time and T is the preset sampling period; When the fluctuation deviation of the objective function in the time series is less than the preset fluctuation threshold, it is determined that the equivalent branch and the virtual node connected to it have a high electrical matching degree, and it is represented as a deterministic path. S303. Based on the calculation results of the objective function, perform the following simplification actions on the node distribution branch graph to remove invalid nodes: Homogeneous node merging: If the voltage drop gradient between adjacent virtual aggregation nodes and measurable nodes is close to zero and the power flow is stable, it is determined that there is no significant load or branch between them, and invalid intermediate nodes are merged and simplified. Light load branch truncation: For isolated virtual nodes whose objective function residuals exceed the preset residual threshold for a long period of time and whose power measurement values ​​are lower than the preset noise level, they are marked as invalid noise nodes and removed from the node distribution branch diagram. Impedance equivalent substitution: simplifying multiple continuous, branchless links between measurable nodes and virtual nodes into a single characteristic impedance edge; S304. Output an effective node distribution diagram, which retains key measurable nodes that have a decisive influence on the topology and core virtual branches with clear impedance properties.

[0011] Furthermore, the specific process of obtaining the optimal matching topology is as follows: S401. Obtain the effective node distribution map, and encode the connection status between each node in the effective node distribution map and the impedance parameters of the equivalent branch into the position vector of the particle. S402. Based on the correlation coefficient determination results in S1, the node connection relationships with correlation are set as the dominant genes of the initial population, guiding the particle swarm to concentrate in high-probability regions. S403. Through tree-structured constraint pruning logic, invalid solutions that do not conform to the operating logic of the power system are automatically eliminated, including the following cases: The topological structure represented by the particles is detected in real time. If an isolated node appears, the solution is directly determined to be invalid and its fitness function is set to infinity. Based on the total power on the transformer side, if the identification results show that the total power of the downstream branches is greater than the power of the upstream branches and the current flow direction is reversed, the solution will be forcibly discarded. If the calculated equivalent branch impedance per unit length exceeds the physical threshold of a conventional conductor, it is considered an abnormal solution. S404. Substitute each generation of particles into the structure-data matching evaluation model established in S3 for scoring, and obtain the residual function value. The smaller the residual function value, the better the theoretical voltage value calculated under the topological structure matches the measured voltage value. S405. When the rate of change of the optimal fitness value of the population for several consecutive generations meets the preset physical residual threshold, the iteration stops, and the particle with the highest fitness is decoded as the final physical connection graph, thus obtaining the optimal matching topology for transparent low-voltage distribution network.

[0012] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are: This non-intrusive topology identification method for transparent low-voltage distribution networks transforms the spatial correlation of massive unmeasurable nodes into equivalent impedance attributes through voltage similarity analysis, effectively mitigating the uncertainty in topology structure caused by the lack of physical information. It preliminarily determines membership relationships through correlation coefficients and eliminates invalid nodes by combining voltage drop consistency models, gradually compressing the original massive solution space and significantly reducing the computational load of a single match. This solves the industry pain points of low search efficiency and susceptibility to local optima. At the same time, tree constraints and power conservation are introduced as hard constraint filters during the search process. By improving the particle swarm algorithm, logical solutions that do not conform to the laws of power operation are automatically eliminated, ensuring the consistency of the final output topology in terms of mathematical fit and physical feasibility. Furthermore, the non-intrusive identification technology relies entirely on existing meter data, significantly reducing operation and maintenance costs. Attached Figure Description

[0013] Figure 1 A schematic diagram of the overall method flow of the present invention is shown. Detailed Implementation

[0014] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0015] Example: like Figure 1 As shown, a non-intrusive topology identification method for transparent low-voltage distribution networks includes the following steps: S1. Based on the distributed data of the low-voltage distribution network, set up several measurable nodes and mark the unmeasurable nodes. Collect voltage and power measurement sequences at the measurable nodes and use the correlation coefficient matrix to preliminarily determine the electrical energy affiliation between the measurable nodes. The specific process for determining the electrical energy affiliation between measurable nodes is as follows: S101. Mark each measurable node as I, obtain the voltage normalized sequence UI=[U{I,1},U{I,2},...,U{I,n}] of each measurable node I within the sampling period T, and extract the voltage change rate sequence. Defined as: , where t is the sampling time; S102. Select any two measurable nodes I, labeled Ia and Ib respectively, calculate the Pearson correlation coefficient between measurable nodes Ia and Ib, and construct the correlation coefficient matrix as follows: Where Cov is the covariance. Standard deviation; This is the voltage change rate sequence corresponding to the meaable node Ia. This is the voltage change rate sequence corresponding to the measurable node Ib; The value range of is [-1, 1]. The closer it is to 1, the higher the degree of coupling between mea and Ib in the electrical topology; S103. Based on the correlation coefficient matrix, perform the following membership derivation: S1031, Trunk-Branch Route Attribution Determination: A preset membership threshold (usually 0.85-0.95) is set. If the membership degree is greater than the membership threshold, it is preliminarily determined that mea and Ib are located on the same branch and have a direct parent-child cascade relationship. S1032. Topological Distance Quantization: Utilizing the inverse relationship between electrical distance and correlation, define the virtual electrical distance between nodes: Dab = 1 - ; S1033. Eliminate inter-phase / inter-line interference: For node pairs with Pearson correlation coefficients close to 0 or negative, they are directly determined to be electrically isolated and removed from each other's candidate affiliation list in the initial stage. S104. Output an initial logical topology matrix, which marks the hierarchical depth of each measurable node relative to the transformer side and marks the electrical energy affiliation between the measurable nodes.

[0016] S2. The spatial correlation between measurable and unmeasurable nodes is determined by voltage similarity analysis. By defining virtual aggregation nodes, the black box branch containing multiple unmeasurable nodes is simplified into an equivalent branch with impedance properties. After integration, the node distribution branch diagram is obtained. The specific process for determining the spatial correlation patterns between measurable and unmeasurable nodes is as follows: S211. Label the measurable node as Im, and the unmeasurable node adjacent to Im as Iu. If the measurable node Im is located upstream of the unmeasurable node Iu or within a very close electrical distance, and their voltage time series have high coherence in terms of fluctuation trend, frequency distribution, and phase characteristics, obtain the voltage series Um of the measurable node Im and the mean reference voltage series in the neighborhood. Calculate the voltage sequence Um and the reference voltage mean sequence. The correlation coefficient is used to determine the background fluctuation benchmark for the region; S212. Introduce a dynamic time warping algorithm to calculate the alignment cost of the measurable node Im and the voltage trajectory in its associated candidate region. The lower the alignment cost, the closer the measurable node is to a specific cluster of unmeasurable nodes in the spatial topology. S213. Establish the following spatial correlation discrimination criteria: Strong correlation determination: If the voltage sequence similarity between measurable node Im1 and measurable node Im2 is higher than the preset similarity threshold, and there is a linear voltage drop difference, then it is determined that the unmeasurable node Iu between the two only has the longitudinal impedance attribute and does not have the branch attribute; Weak correlation determination: If the voltage fluctuation waveforms of measurable node Im1 and measurable node Im2 are similar but there is phase lag and nonlinear deviation, it indicates that there is a lateral branch at unmeasurable node Iu, thereby logically identifying the T-connection and cross-connection spatial structure of the unmeasurable node.

[0017] The specific process of obtaining the node distribution branch map is as follows: S221. Based on the power membership relationship between measurable nodes determined in S1, identify all sets of unmeasurable nodes located between two adjacent measurable nodes Im and In, and define the topological region where the set of unmeasurable nodes is located as a black box branch. The boundary of the black box branch is locked by nodes with known measurement data. S222. Obtain the voltage deviation and current vector between measurable node Im and measurable node In, and calculate the total equivalent impedance of the black box branch according to Ohm's law. S223. Based on the spatial correlation discrimination criterion, a continuous unmeasurable node segment with strong correlation is defined as a virtual aggregation node. The virtual aggregation node does not focus on the specific location of a single unmeasurable node inside, but is equivalent to an impedance body with accumulated impedance. The virtual aggregation node serves as an impedance transfer point for power transmission and does not reflect specific physical location coordinates. S224. Assign complex impedance properties to each simplified equivalent branch, and integrate the simplified equivalent branches with the original measurable node connection relationships to form a node distribution branch diagram.

[0018] In the node distribution branch diagram, the original complex "branch-blind zone-terminal" structure is simplified into a skeleton topology composed of "known nodes + equivalent impedance branches". The node distribution branch diagram removes a large number of redundant variables without measurement support, connects and combines the originally infinitely possible unmeasurable nodes, and converges into a branch model with finite parameters, providing a simplified physical carrier for the matching degree calculation in the subsequent S3 step.

[0019] S3. Establish a structure-data matching degree evaluation model. By introducing the branch power conservation and voltage drop equation, adopting the objective function based on the consistency of branch voltage drop, and simplifying invalid nodes based on the node distribution branch map, the amount of calculation for a single matching is reduced, so as to obtain the effective node distribution map. The specific process for obtaining the effective node distribution map is as follows: S301. Based on Kirchhoff's current law and voltage law, for each equivalent branch in the node distribution branch diagram, the branch power conservation equation and voltage drop equation are introduced: Power conservation constraint: Where P l P is the theoretical voltage value. i P is the measured voltage value. LOSS This represents the theoretical loss of the equivalent branch under the current topological assumptions. Simultaneously define the theoretical value of voltage drop at measurable nodes at both ends of the branch. Compared with the measured voltage drop The residual function between; S302. Constructing an objective function based on voltage drop consistency. The objective function described below is used to evaluate the effectiveness of the current topology branch: , where t is the sampling time and T is the preset sampling period; When the fluctuation deviation of the objective function in the time series is less than the preset fluctuation threshold, it is determined that the equivalent branch and the virtual node connected to it have a high electrical matching degree, and it is represented as a deterministic path. S303. Based on the calculation results of the objective function, perform the following simplification actions on the node distribution branch graph to remove invalid nodes: Homogeneous node merging: If the voltage drop gradient between adjacent virtual aggregation nodes and measurable nodes is close to zero and the power flow is stable, it is determined that there is no significant load or branch between them, and invalid intermediate nodes are merged and simplified. Light load branch truncation: For isolated virtual nodes whose objective function residuals exceed the preset residual threshold for a long period of time and whose power measurement values ​​are lower than the preset noise level, they are marked as invalid noise nodes and removed from the node distribution branch diagram. Impedance equivalent substitution: simplifying multiple continuous, branchless links between measurable nodes and virtual nodes into a single characteristic impedance edge; S304. Output an effective node distribution diagram, which retains key measurable nodes that have a decisive influence on the topology and core virtual branches with clear impedance properties.

[0020] S4. Based on the tree-like constraints of the physical topology, during the search of the effective node distribution graph, an improved particle swarm optimization algorithm is used to automatically eliminate solutions that do not conform to common sense in electrical engineering, further compressing the solution space until the optimal matching topology is obtained by iterative solution within the physical constraints.

[0021] The specific process of obtaining the optimal matching topology is as follows: S401. Obtain the effective node distribution map, and encode the connection status between each node in the effective node distribution map and the impedance parameters of the equivalent branch into the position vector of the particle. S402. Based on the correlation coefficient determination results in S1, the node connection relationship with correlation is set as the dominant gene of the initial population, guiding the particle swarm to concentrate in the high probability region and avoiding search divergence caused by random initialization. S403. Through tree-structured constraint pruning logic, invalid solutions that do not conform to the operating logic of the power system are automatically eliminated, including the following cases: The topological structure represented by the particles is detected in real time. If an isolated node appears, the solution is directly determined to be invalid and its fitness function is set to infinity. Based on the total power on the transformer side, if the identification results show that the total power of the downstream branches is greater than the power of the upstream branches and the current flow direction is reversed, the solution will be forcibly discarded. If the calculated equivalent branch impedance per unit length exceeds the physical threshold of a conventional conductor, it is considered an abnormal solution. Adaptive adjustment of the inertia weights of the particle swarm: In the early stages of the search, a larger inertial weight is used to enhance the global search capability and traverse multiple possible topological combinations; In the later stages of the search, as the voltage drop residual function converges, the weights are dynamically reduced to perform a fine search within the neighborhood of the optimal solution.

[0022] S404. Substitute each generation of particles into the structure-data matching evaluation model established in S3 for scoring, and obtain the residual function value. The smaller the residual function value, the better the theoretical voltage value calculated under the topological structure matches the measured voltage value. S405. When the rate of change of the optimal fitness value of the population for several consecutive generations meets the preset physical residual threshold, the iteration stops, and the particle with the highest fitness is decoded as the final physical connection graph, thus obtaining the optimal matching topology for transparent low-voltage distribution network.

[0023] This invention transforms the spatial correlation of massive unmeasurable nodes into equivalent impedance attributes through voltage similarity analysis, effectively mitigating the uncertainty in topology caused by the lack of physical information. It preliminarily determines membership relationships through correlation coefficients and eliminates invalid nodes using a voltage drop consistency model, progressively compressing the original massive solution space and significantly reducing the computational load of a single matching. This addresses the industry pain points of low search efficiency and susceptibility to local optima. Furthermore, it introduces tree-like constraints and power conservation as hard constraints during the search process, automatically eliminating logical solutions that do not conform to power operation laws through an improved particle swarm optimization algorithm, ensuring the consistency of the final output topology in terms of mathematical fit and physical feasibility. Finally, it employs non-intrusive identification technology, relying entirely on existing meter data, significantly reducing operation and maintenance costs.

[0024] The size of the interval and threshold is set to facilitate comparison. The size of the threshold depends on the amount of sample data and the number of bases set by those skilled in the art for each set of sample data; as long as it does not affect the ratio between the parameter and the quantized value.

[0025] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters in the formulas are set by those skilled in the art according to the actual situation. The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A non-intrusive topology identification method for transparent low-voltage distribution networks, characterized in that, Includes the following steps: S1. Based on the distributed data of the low-voltage distribution network, set up several measurable nodes and mark the unmeasurable nodes. Collect voltage and power measurement sequences at the measurable nodes and use the correlation coefficient matrix to preliminarily determine the electrical energy affiliation between the measurable nodes. S2. The spatial correlation between measurable and unmeasurable nodes is determined by voltage similarity analysis. By defining virtual aggregation nodes, the black box branch containing multiple unmeasurable nodes is simplified into an equivalent branch with impedance properties. After integration, the node distribution branch diagram is obtained. S3. Establish a structure-data matching degree evaluation model. By introducing the branch power conservation and voltage drop equation, adopting the objective function based on the consistency of branch voltage drop, and simplifying invalid nodes based on the node distribution branch map, the computational amount of a single matching is reduced to obtain the effective node distribution map. S4. Based on the tree-like constraints of the physical topology, during the search of the effective node distribution graph, an improved particle swarm optimization algorithm is used to automatically eliminate solutions that do not conform to common sense in electrical engineering, further compressing the solution space until the optimal matching topology is obtained by iterative solution within the physical constraints.

2. The non-intrusive topology identification method for transparent low-voltage distribution networks according to claim 1, characterized in that, The specific process for determining the electrical energy affiliation between measurable nodes is as follows: S101. Mark each measurable node as I, obtain the voltage normalized sequence UI=[U{I,1},U{I,2},...,U{I,n}] of each measurable node I within the sampling period T, and extract the voltage change rate sequence. Defined as: , where t is the sampling time; S102. Select any two measurable nodes I, labeled Ia and Ib respectively, calculate the Pearson correlation coefficient between measurable nodes Ia and Ib, and construct the correlation coefficient matrix as follows: Where Cov is the covariance. Standard deviation; This is the voltage change rate sequence corresponding to the meaable node Ia. This is the voltage change rate sequence corresponding to the measurable node Ib; The value range of is [-1, 1]. The closer it is to 1, the higher the degree of coupling between mea and Ib in the electrical topology; S103. Based on the correlation coefficient matrix, perform the following membership derivation: S1031, Trunk-Branch Affiliation Determination: Preset membership threshold, if If the membership degree is greater than the membership threshold, it is preliminarily determined that mea and Ib are located on the same branch and have a direct parent-child cascade relationship. S1032. Topological Distance Quantization: Utilizing the inverse relationship between electrical distance and correlation, define the virtual electrical distance between nodes: Dab = 1 - ; S1033. Eliminate inter-phase / inter-line interference: For node pairs with Pearson correlation coefficients close to 0 or negative, they are directly determined to be electrically isolated and removed from each other's candidate affiliation list in the initial stage. S104. Output an initial logical topology matrix, which marks the hierarchical depth of each measurable node relative to the transformer side and marks the electrical energy affiliation between the measurable nodes.

3. The non-intrusive topology identification method for transparent low-voltage distribution networks according to claim 1, characterized in that, The specific process for determining the spatial correlation patterns between measurable and unmeasurable nodes is as follows: S211. Mark the measurable node as Im, and the unmeasurable nodes adjacent to Im as Iu. Obtain the voltage sequence Um of the measurable node Im and the mean sequence of the reference voltage in the neighborhood. Calculate the voltage sequence Um and the reference voltage mean sequence. The correlation coefficient; S212. Introduce a dynamic time warping algorithm to calculate the alignment cost between the measurable node Im and the voltage trajectory within its associated candidate region. S213. Establish the following spatial correlation discrimination criteria: Strong correlation determination: If the voltage sequence similarity between measurable node Im1 and measurable node Im2 is higher than the preset similarity threshold, and there is a linear voltage drop difference, then it is determined that the unmeasurable node Iu between the two only has the longitudinal impedance attribute and does not have the branch attribute; Weak correlation determination: If the voltage fluctuation waveforms of measurable node Im1 and measurable node Im2 are similar but there is phase lag and nonlinear deviation, it indicates that there is a lateral branch at unmeasurable node Iu, thereby logically identifying the T-connection and cross-connection spatial structure of the unmeasurable node.

4. The non-intrusive topology identification method for transparent low-voltage distribution networks according to claim 3, characterized in that, The specific process of obtaining the node distribution branch map is as follows: S221. Based on the power membership relationship between measurable nodes determined in S1, identify all sets of unmeasurable nodes located between two adjacent measurable nodes Im and In, and define the topological region where the set of unmeasurable nodes is located as a black box branch. The boundary of the black box branch is locked by nodes with known measurement data. S222. Obtain the voltage deviation and current vector between measurable node Im and measurable node In, and calculate the total equivalent impedance of the black box branch according to Ohm's law. S223. Based on the spatial correlation discrimination criterion, a continuous unmeasurable node segment with strong correlation is defined as a virtual aggregation node. The virtual aggregation node does not focus on the specific location of a single unmeasurable node inside, but is equivalent to an impedance body with accumulated impedance. The virtual aggregation node serves as an impedance transfer point for power transmission and does not reflect specific physical location coordinates. S224. Assign complex impedance properties to each simplified equivalent branch, and integrate the simplified equivalent branches with the original measurable node connection relationships to form a node distribution branch diagram.

5. The non-intrusive topology identification method for transparent low-voltage distribution networks according to claim 1, characterized in that, The specific process for obtaining the effective node distribution map is as follows: S301. Based on Kirchhoff's current law and voltage law, for each equivalent branch in the node distribution branch diagram, the branch power conservation equation and voltage drop equation are introduced: Power conservation constraint: Where P l P is the theoretical voltage value. i P is the measured voltage value. LOSS This represents the theoretical loss of the equivalent branch under the current topological assumptions. Simultaneously define the theoretical value of the voltage drop at the measurable nodes at both ends of the branch. Compared with the measured voltage drop The residual function between; S302. Constructing an objective function based on voltage drop consistency. The objective function described below is used to evaluate the effectiveness of the current topology branch: , where t is the sampling time and T is the preset sampling period; When the fluctuation deviation of the objective function in the time series is less than the preset fluctuation threshold, it is determined that the equivalent branch and the virtual node connected to it have a high electrical matching degree, and it is represented as a deterministic path. S303. Based on the calculation results of the objective function, perform the following simplification actions on the node distribution branch graph to remove invalid nodes: Homogeneous node merging: If the voltage drop gradient between adjacent virtual aggregate nodes and measurable nodes is close to zero and the power flow is stable, it is determined that there is no significant load or branch between them, and invalid intermediate nodes are merged and simplified. Light load branch truncation: For isolated virtual nodes whose objective function residuals exceed the preset residual threshold for a long period of time and whose power measurement values ​​are lower than the preset noise level, they are marked as invalid noise nodes and removed from the node distribution branch diagram. Impedance equivalent substitution: simplifying multiple continuous, branchless links between measurable nodes and virtual nodes into a single characteristic impedance edge; S304. Output an effective node distribution diagram, which retains key measurable nodes that have a decisive influence on the topology and core virtual branches with clear impedance properties.

6. The non-intrusive topology identification method for transparent low-voltage distribution networks according to claim 1, characterized in that, The specific process of obtaining the optimal matching topology is as follows: S401. Obtain the effective node distribution map, and encode the connection status between each node in the effective node distribution map and the impedance parameters of the equivalent branch into the position vector of the particle. S402. Based on the correlation coefficient determination results in S1, the node connection relationships with correlation are set as the dominant genes of the initial population, guiding the particle swarm to concentrate in high-probability regions. S403. Through tree-structured constraint pruning logic, invalid solutions that do not conform to the operating logic of the power system are automatically eliminated, including the following cases: The topological structure represented by the particles is detected in real time. If an isolated node appears, the solution is directly determined to be invalid and its fitness function is set to infinity. Based on the total power on the transformer side, if the identification results show that the total power of the downstream branches is greater than the power of the upstream branches and the current flow direction is reversed, the solution will be forcibly discarded. If the calculated equivalent branch impedance per unit length exceeds the physical threshold of a conventional conductor, it is considered an abnormal solution. S404. Substitute each generation of particles into the structure-data matching evaluation model established in S3 for scoring, and obtain the residual function value. The smaller the residual function value, the better the theoretical voltage value calculated under the topological structure matches the measured voltage value. S405. When the rate of change of the optimal fitness value of the population for several consecutive generations meets the preset physical residual threshold, the iteration stops, and the particle with the highest fitness is decoded as the final physical connection graph, thus obtaining the optimal matching topology for transparent low-voltage distribution network.

Images