A TLCC-based transformer area topology relationship identification method and device and computer equipment

Abstract

The application discloses a kind of based on TLCC's transformer area topological relation identification method, device and computer equipment, the method includes: determine hysteresis search range and the marker node monitoring module of several monitoring modules;Mobile processing is carried out to second voltage sequence in hysteresis search range, and the voltage sequence after moving is obtained;Correlation quantization calculation is carried out using the single-phase voltage sequence in the voltage sequence after moving and first voltage sequence, and the similarity set is obtained;Clustering processing is carried out based on similarity set, and the topological relationship between marker node monitoring module and electric energy meter is obtained.The algorithm of voltage similarity by time hysteresis cross-correlation calculation can effectively depict the similarity between different time series, so as to improve the accuracy of topological identification.

Description

Technical Field

[0001] This invention relates to the field of low-voltage distribution network topology identification technology, and in particular to a method, device and computer equipment for identifying transformer substation topology relationships based on TLCC. Background Technology

[0002] Low-voltage distribution networks are end-point power supply networks that directly serve users and are an important component of the power grid. Understanding the topology of low-voltage distribution networks is crucial for identifying the topological structures between transformer substations and meter boxes, branches and meter boxes, and meter boxes and users, thus enabling refined management of low-voltage distribution networks.

[0003] In related technologies, the correlation coefficient is generally calculated directly to determine the voltage similarity of the electricity meter data, so as to obtain the connection relationship between the user's electricity meter and the distribution area transformer, thereby identifying the low-voltage distribution network topology relationship corresponding to the electricity meter.

[0004] However, in low-voltage power distribution networks, users' electricity meters with different communication methods may have asynchronous data collection. Directly using correlation coefficients to judge the similarity between different users will lead to large errors in the calculation results, ultimately resulting in poor identification of topological relationships. Summary of the Invention

[0005] This invention aims to at least partially solve one of the technical problems in related technologies. To this end, the first objective of this invention is to propose a method for identifying transformer substation topology relationships based on TLCC (Transformer Transmission Control Center). This method solves the problem of asynchronous data acquisition times among nodes in low-voltage distribution networks by using an algorithm to calculate voltage similarity through time-lag cross-correlation, accurately determining the similarity between different time series, and improving the accuracy of topology identification.

[0006] The second objective of this invention is to propose a TLCC-based device for identifying the topological relationship of transformer substations.

[0007] The third objective of this invention is to provide a computer device.

[0008] The fourth objective of this invention is to provide a computer-readable storage medium.

[0009] To achieve the above objectives, a first aspect of this invention proposes a TLCC-based method for identifying transformer substation topology relationships, applied to a low-voltage distribution network. The low-voltage distribution network includes several monitoring modules and electricity meters connected to the monitoring modules. A time lag exists between a first voltage sequence collected by the monitoring modules and a second voltage sequence collected by the electricity meters. The several monitoring modules include a master node monitoring module and a marker node monitoring module. The TLCC-based method for identifying transformer substation topology relationships includes: determining the lag search range and the marker node monitoring modules of the several monitoring modules; wherein the marker node monitoring modules are located on child nodes of the master node monitoring modules; moving the second voltage sequence within the lag search range to obtain a moved voltage sequence; performing correlation quantization calculations using the moved voltage sequence and single-phase voltage sequences in the first voltage sequence to obtain a similarity set; and performing clustering based on the similarity set to obtain the topological relationship between the marker node monitoring modules and the electricity meters, thereby identifying electricity meters located on the same power grid branch as the marker node monitoring modules.

[0010] According to the TLCC-based transformer substation topology identification method of this invention, the hysteresis search range of the voltage data of the monitoring module and the energy meter is determined based on the sampling interval of the voltage data of the monitoring module and the energy meter. Then, the voltage similarity between the energy meter and the monitoring module is calculated based on the hysteresis search range, resulting in a similarity set of time-lag cross-correlation. Clustering is performed based on the similarity set, and the topological relationship between the monitoring module and the energy meter can be determined according to the results of the clustering. This invention's algorithm for calculating voltage similarity through time-lag cross-correlation can effectively characterize the similarity between different time series, improving the accuracy of topology identification.

[0011] According to one embodiment of the present invention, the method for determining the marker node monitoring module includes: determining a first average correlation matrix; wherein, the elements in the first average correlation matrix correspond to any two monitoring modules other than the master node monitoring module among the plurality of monitoring modules, and are the average correlation coefficients between any two monitoring modules other than the master node monitoring module among the plurality of monitoring modules; performing clustering processing based on the first average correlation matrix to obtain a first clustering result; wherein, the first clustering result includes a plurality of first clusters; and determining the marker node monitoring module among the monitoring modules included in the first cluster according to the correlation coefficients between the master node monitoring module and the monitoring modules included in the first cluster.

[0012] According to one embodiment of the present invention, determining the first average correlation matrix includes: determining a correlation coefficient matrix based on the first voltage sequences collected by any two monitoring modules other than the master node monitoring module; determining target elements in the correlation coefficient matrix; wherein the target elements are used to represent the correlation degree between single-phase voltage sequences in the same phase in the first voltage sequences collected by the arbitrary two monitoring modules; and determining the elements in the first average correlation matrix by averaging the target elements and the number of effective phases; wherein the number of effective phases is used to represent the number of target elements in an effective state.

[0013] According to one embodiment of the present invention, clustering the first average correlation matrix to obtain the first clustering result includes: using an affinity propagation algorithm or a silhouette coefficient method to cluster the first average correlation matrix to obtain the first clustering result.

[0014] According to one embodiment of the present invention, the hysteresis search range is determined based on the hysteresis sample interval between the second voltage sequence and the first voltage sequence; the step of moving the second voltage sequence within the hysteresis search range to obtain the moved voltage sequence includes: moving the second voltage sequence according to a preset step size within the hysteresis search range to obtain a plurality of the moved voltage sequences.

[0015] According to one embodiment of the present invention, the step of performing correlation quantization calculation using the shifted voltage sequence and the single-phase voltage sequence in the first voltage sequence to obtain a similarity set includes: obtaining the sequence correlation coefficient between any of the shifted voltage sequences and the single-phase voltage sequence in the first voltage sequence; determining a target sequence correlation coefficient that meets a preset condition from the sequence correlation coefficients; and constructing the similarity set based on the target sequence correlation coefficient.

[0016] According to one embodiment of the present invention, the step of performing clustering processing based on the similarity set to obtain the topological relationship between the marker node monitoring module and the energy meter includes: determining the equivalent distance set between the second voltage sequence and the single-phase voltage sequence in the first voltage sequence based on the similarity set; and performing hierarchical clustering processing on the equivalent distance set to obtain the topological relationship between the marker node monitoring module and the energy meter.

[0017] According to one embodiment of the present invention, the step of performing hierarchical clustering on the equivalent distance set to obtain the topological relationship between the marker node monitoring module and the electricity meter includes: performing hierarchical clustering on the equivalent distance set according to a preset number of clusters to obtain a second clustering result; wherein the second clustering result includes a plurality of second clusters; determining the clusters to be processed in the second clustering result; wherein the number of electricity meters in the clusters to be processed is less than a threshold for the number of electricity meters, or the clusters to be processed do not include the marker node monitoring module; adding the electricity meters in the clusters to be processed to the second clusters in the second clustering result that include one of the marker node monitoring modules to obtain a new second clustering result; and determining the topological relationship between the marker node monitoring module and the electricity meter based on the new second clustering result.

[0018] According to one embodiment of the present invention, clustering is performed based on the similarity set to obtain several second clusters. The electricity meters included in the second clusters and the marker node monitoring modules included in the second clusters are located on the same power grid branch. The first cluster to which the marker node monitoring modules included in the second clusters belong is denoted as the target cluster. The method further includes: taking the phase corresponding to the marker node monitoring modules included in the second clusters as the target phase corresponding to the second clusters; wherein, the phase category of the second voltage sequence collected by the electricity meters included in the second clusters is the target phase; performing correlation calculation based on the single-phase voltage sequence of the monitoring module in the target cluster on the target phase and the second voltage sequence collected by the electricity meters included in the second clusters to obtain the correlation coefficient between the monitoring module in the target cluster and the electricity meters included in the second cluster on the target phase; and determining, based on the correlation coefficient on the target phase, that the electricity meters included in the second clusters and the monitoring module in the target clusters have a topological relationship on the target phase.

[0019] To achieve the above objectives, a second aspect of the present invention proposes a transformer substation topology identification device based on TLCC, applied to a low-voltage distribution network. The low-voltage distribution network includes several monitoring modules and electricity meters connected to the monitoring modules. A time lag exists between a first voltage sequence collected by the monitoring modules and a second voltage sequence collected by the electricity meters. The several monitoring modules include a master node monitoring module and a marker node monitoring module. The device includes: a determination module for determining the lag search range and the marker node monitoring modules of the several monitoring modules; wherein the marker node monitoring modules are located on child nodes of the master node monitoring modules; a movement processing module for moving the second voltage sequence within the lag search range to obtain a moved voltage sequence; a correlation calculation module for performing correlation quantification calculation using the moved voltage sequence and single-phase voltage sequences in the first voltage sequence to obtain a similarity set; and a clustering processing module for performing clustering based on the similarity set to obtain the topological relationship between the marker node monitoring modules and the electricity meters, thereby identifying electricity meters located on the same power grid branch as the marker node monitoring modules.

[0020] According to an embodiment of the present invention, a transformer substation topology identification device based on TLCC (Time Lag Cross-Correlation) determines the hysteresis search range of the electricity meter voltage data based on the sampling interval of the voltage data of the monitoring module and the electricity meter. Then, it calculates the voltage similarity between the electricity meter and the monitoring module based on the hysteresis search range, obtaining a similarity set of time lag cross-correlation. Clustering is then performed based on the similarity set, and the topological relationship between the monitoring module and the electricity meter can be determined according to the results of the clustering. This embodiment of the invention, through the algorithm of calculating voltage similarity using time lag cross-correlation, can effectively characterize the similarity between different time series, improving the accuracy of topology identification.

[0021] To achieve the above objectives, a third aspect of the present invention provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the TLCC-based method for identifying transformer topology relationships according to any of the above embodiments.

[0022] According to the computer device of the present invention, when the processor executes a computer program, an algorithm for calculating voltage similarity by time lag cross-correlation can effectively characterize the similarity between different time series, thereby improving the accuracy of topology identification.

[0023] To achieve the above objectives, a fourth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, wherein the computer program, when executed by a processor, implements the TLCC-based method for identifying transformer area topology relationships according to any of the above embodiments.

[0024] According to the computer-readable storage medium of the present invention, when a computer program is executed by a processor, an algorithm for calculating voltage similarity through time-lag cross-correlation can effectively characterize the similarity between different time series, thereby improving the accuracy of topology identification.

[0025] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description

[0026] Figure 1 This is a schematic diagram illustrating an application scenario of a TLCC-based transformer topology identification method provided according to one embodiment of this specification.

[0027] Figure 2 This is a flowchart of a TLCC-based method for identifying transformer topology relationships according to one embodiment of this specification.

[0028] Figure 3 This is a flowchart of a method for determining a marker node monitoring module according to one embodiment of this specification.

[0029] Figure 4 This is a flowchart of a method for determining a first average correlation matrix according to one embodiment of this specification.

[0030] Figure 5 This is a flowchart of a correlation quantification calculation method provided according to one embodiment of this specification.

[0031] Figure 6 This is a flowchart of a method for identifying the topological relationship between a marker node monitoring module and an energy meter according to one embodiment of this specification.

[0032] Figure 7a This is a flowchart of a TLCC-based method for identifying transformer topology relationships according to one embodiment of this specification.

[0033] Figure 7b This is a schematic diagram of a low-voltage power distribution network topology provided according to one embodiment of this specification.

[0034] Figure 7c This is a schematic diagram of a low-voltage power distribution network topology provided according to one embodiment of this specification.

[0035] Figure 8 This is a structural block diagram of a TLCC-based transformer topology identification device provided according to one embodiment of this specification.

[0036] Figure 9This is a structural block diagram of a computer device provided according to one embodiment of this specification. Detailed Implementation

[0037] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and intended to explain the present invention, and should not be construed as limiting the present invention.

[0038] The main types of electricity meters used in low-voltage power distribution networks are HPLC electricity meters and RS485 electricity meters. Due to differences in their communication principles, these two types of meters exhibit certain differences in data synchronization and topology calculation. HPLC electricity meters are based on power line carrier communication. These meters use clock synchronization to ensure that the collected current, voltage, and other electricity consumption information are received at the same time. RS485 electricity meters are based on RS485 serial communication. These meters use the serial port to control and transmit data with a secondary data acquisition unit. Because RS485 communication speed is relatively low, and the secondary data acquisition unit needs to control its subordinate meters in a time-sharing manner, the voltage, current, and other electricity consumption data received and reported by the same secondary data acquisition unit may not be from the same time. This often results in a certain acquisition delay, meaning that data from one time point may not be reported until the next. Generally, when performing topology calculations on power distribution networks containing RS485 electricity meters, correlation coefficients cannot be directly calculated; otherwise, the time deviation will lead to significantly different calculation results and very poor identification performance. Therefore, in this case, topology calculations need to consider aligning the collected data to a time series first.

[0039] Related technologies also employ methods to identify voltage curve similarity by calculating dynamic time warping (DTW) distance, thus avoiding distance measurement problems between time-misaligned sequences. However, in low-voltage distribution networks, the DTW distance between voltage curves of the same phase and different phases may not show obvious clustering or distance differences, resulting in inability to characterize the true curve similarity and leading to poor calculation performance.

[0040] Figure 1 This is a schematic diagram illustrating the application scenario of the TLCC-based transformer topology identification method provided in this manual. Taking a residential area as an example, residential areas are typically equipped with low-voltage distribution networks to supply electricity to residents. To ensure the normal operation of the low-voltage distribution network, monitoring modules are installed at each branch node of the low-voltage distribution network for electrical monitoring, fault monitoring, temperature sensing, and other functions. Figure 1This is a schematic diagram using a three-layer topology of a low-voltage distribution network, including transformers, branch lines, and meter boxes. Figure 1 As shown, a transformer monitoring module is installed at the transformer substation, branch line monitoring modules are installed at multiple branch lines, and meter box monitoring modules and multiple single-phase energy meters connected to the meter box monitoring modules are installed in multiple meter boxes. The transformer monitoring module, branch line monitoring module, and meter box monitoring module can be any monitoring module used in the embodiments described in this specification. For example, the transformer monitoring module can be a main node monitoring module, the branch line monitoring module can be a marker node monitoring module, and the meter box monitoring module can be a monitoring module located on a child node of the marker node monitoring module. In this scenario example, the single-phase energy meter can include an RS485 energy meter based on RS485 serial communication.

[0041] In some feasible implementations, the monitoring module can be an LTU (Line Terminal Unit), which can acquire the three-phase voltage sequence on the monitored line. In this example scenario, phases A, B, and C are used to represent the three-phase voltages.

[0042] In this scenario example, the topological relationship between the monitoring modules is first calculated. Then, based on the time lag cross-correlation of the voltage sequences of the electricity meter and the monitoring modules (excluding the master node monitoring module), the topological relationship between the electricity meter and the monitoring modules is calculated, and finally the complete topological relationship of the low-voltage distribution network is obtained.

[0043] This example illustrates how to calculate the topological relationships between monitoring modules. A low-voltage distribution network contains several monitoring modules. Any two monitoring modules, excluding the main node monitoring module (transformer monitoring module), are denoted as monitoring module L1 and monitoring module L2. The voltage sequence collected by monitoring module L1 is denoted as U1, including the A-phase voltage sequence L... A1 The phase B voltage sequence L on phase B B1 and the C-phase voltage sequence L on phase C. C1 The voltage sequence acquired by monitoring module L2 is denoted as U2, including the A-phase voltage sequence L on phase A. A2 The phase B voltage sequence L on phase B B2 and the C-phase voltage sequence L on phase C. C2 .

[0044] The correlation coefficient between any two single-phase voltage sequences is calculated based on the phase voltage sequences acquired by any two monitoring modules, forming a correlation coefficient matrix. For example, based on the voltage sequence U1 acquired by monitoring module L1 and the voltage sequence U2 acquired by monitoring module L2, the correlation coefficient matrix is ​​calculated for the voltage sequence L... A1 With voltage sequence L A2The correlation coefficient between them is denoted as P. A1A2 ; Calculate the voltage sequence L B1 With voltage sequence L B2 The correlation coefficient between them is denoted as P. B1B2 Voltage sequence L C1 With voltage sequence L C2 The correlation coefficient between them is denoted as P. C1C2 For example, the correlation coefficient matrix between the voltage sequence U1 collected by monitoring module L1 and the voltage sequence U2 collected by monitoring module L2 is as follows.

[0045]

[0046] The correlation coefficients of single-phase voltage sequences in the same phase of the correlation coefficient matrix are taken as the target element. For example, the correlation coefficient P of single-phase voltage sequences in the same phase of the correlation coefficient matrix P is taken. A1A2 P B1B2 and P C1C2 As the target element, the number of valid phases n in the target element is counted. It should be noted that when the monitoring module collects voltage sequences, there may be cases where the voltage sequence for a certain phase is not collected, or the voltage sequence value for a certain phase collected is an outlier. In the embodiments of this specification, since null and outlier voltage values ​​are removed when sampling the voltage data collected by the monitoring module, phases where no voltage sequence is collected or phases where outliers are collected can be understood as invalid phases. For example, if the monitoring module L1 does not collect the voltage sequence for phase A, or the collected voltage sequence for phase A is an outlier, then the voltage sequence L... A1 For voltage sequences with invalid phases, the value can be set to 0. It is understandable that the correlation coefficient P between the voltage sequences acquired by monitoring module L1 and monitoring module L2 on phase A is considered within the target element. A1A2 If the value is 0, then phase A in the target element is an invalid phase, while phases B and C are valid phases, and the number of valid phases n is 2.

[0047] The average correlation coefficient between any two monitoring modules (excluding the master node monitoring module) is calculated using the target elements and the number of effective phases, resulting in an average correlation matrix composed of these average correlation coefficients. The average correlation coefficient is calculated as follows.

[0048]

[0049] Using the average correlation matrix as the similarity for clustering, the monitoring modules (excluding the master node monitoring module) are clustered to obtain several clusters of monitoring modules. For example, the affinity propagation algorithm can be used for clustering.

[0050] It should be noted that the voltage at each node on the same branch in a low-voltage distribution network is affected by impedance and load power. When impedance is expressed as electrical distance, under the same active load, the closer the electrical distance, the higher the voltage similarity; conversely, for the same electrical distance, the larger the active load, the higher the voltage similarity. In a single-source line, at the same cross-section, the voltage amplitude at nodes along the line gradually decreases. When the overall load characteristics differ between lines, the similarity between users located on the same outgoing line and with a closer electrical distance will be higher than the similarity between users located on different outgoing lines, and the closer the electrical distance, the higher the similarity between users.

[0051] In this scenario example, through clustering based on the average correlation matrix, monitoring modules located on the same branch line as the main node monitoring module can be grouped into the same cluster. Using the method described above for calculating the average correlation coefficient avoids correlation coefficient errors caused by phase anomalies, thereby improving the accuracy of topology identification.

[0052] In this scenario example, the correlation between the master node monitoring module and the monitoring modules contained in several clusters (obtained through clustering based on the average correlation matrix) is calculated to determine the labeled node monitoring modules in the clusters. For example, the correlation between the master node monitoring module and the monitoring modules contained in the clusters can be calculated using the average correlation coefficient. It is understood that the closer a monitoring module is electrically to the master node monitoring module, the greater its correlation with the master node monitoring module.

[0053] In this scenario example, for the branch lines of the main node monitoring module, the branch line monitoring module directly connected to the main node monitoring module has the highest correlation with the main node monitoring module, and can be identified as the marker node monitoring module. The other monitoring modules in the cluster are meter box monitoring modules. Through the above correlation calculation method, the topological relationship of the transformer monitoring module, branch line monitoring module, and meter box monitoring module can be determined.

[0054] This example illustrates how to calculate the topological relationship between the energy meter and the monitoring module (excluding the master node monitoring module) based on the time lag cross-correlation of the voltage sequences.

[0055] Voltage data from electricity meters and monitoring modules at marked nodes, determined through the aforementioned topology calculations, are sampled in the low-voltage distribution network. The lag sampling time between the electricity meters and the marked node monitoring modules is statistically analyzed, and the lag search range for the voltage sequence is set based on this lag sampling time. For example, the lag sampling time between the electricity meters and the marked node monitoring modules is denoted as the lag sampling interval N. i The maximum lag interval N = MAX(N) is obtained statistically. iLet ), where i is the sample number. The lag search range for the voltage sequence is set to SCOPE = [-w*N, w*N] based on the maximum lag interval. Typically, the energy meter sequence lags by one sampling time, so the maximum lag interval N can be 1. w is the search space expansion factor, which is usually 2. Therefore, the lag search range SCOPE = [-2, 2].

[0056] Multiple shifted voltage sequences are obtained by shifting the elements in the voltage sequence of the electricity meter or the voltage sequence of the marker node monitoring module according to the step size within the SCOPE range. For example, the voltage sequence of the electricity meter is denoted as V, and the voltage sequence of the marker node monitoring module is denoted as U. For the hysteresis search range SCOPE = [-2, 2], the shift step size can be s = {-2, -1, 0, 1, 2}. The data at the first sampling time t1 in the voltage sequence V is denoted as V0. t1 The data at the second sampling time t2 is denoted as V. t2 The data at the third sampling time t3 is denoted as V. t3 And so on. Each element in the voltage sequence V is shifted by a step size s. For example, when the step size s is -1, it means that all elements in the voltage sequence V are shifted one position to the previous time step, resulting in the shifted voltage sequence V'. In the shifted voltage sequence V', the data at the first sampling time t1 is V. (t-1)1 The data at the second sampling time t2 is V (t-1)2 The data at the third sampling time t3 is V (t-1)3 And so on. In some cases, the sampled data corresponding to certain sampling times in the shifted voltage sequence V' may be missing. After shifting, a value of 0 can be inserted at the corresponding missing positions to ensure that the number of elements in the two sequences is consistent when calculating the correlation coefficient between the shifted voltage sequence V' and the first voltage sequence U, thus avoiding calculation anomalies. For energy meters that have built-in clock synchronization functions to ensure that the current, voltage, and other electricity consumption information collected by the energy meter and the distribution branch lines are at the same time, the lag interval can be set to 0.

[0057] Calculate the correlation coefficients between multiple shifted voltage sequences and another voltage sequence. For example, calculate the correlation coefficients between the five sequences of the electricity meter voltage sequence V shifted by a step size s = {-2, -1, 0, 1, 2} and the A-phase voltage sequence of the voltage sequence U in the marker node monitoring module. Among the five calculated correlation coefficients, the maximum correlation coefficient is taken as the similarity between the electricity meter voltage sequence V and the A-phase voltage sequence of the voltage sequence U in the marker node monitoring module. The similarity between the electricity meter voltage sequence V and the B-phase and C-phase voltage sequences of the voltage sequence U in the marker node monitoring module can be calculated using the same method.

[0058] Based on the correlation coefficients between the voltage sequences of electricity meters and the voltage sequences of the marked node monitoring modules in a low-voltage distribution network on a single phase, a three-phase correlation coefficient set is generated. This set is used to cluster and identify the electricity meters and marked node monitoring modules, thus revealing the topological relationship between them. It should be noted that for electricity meters clustered based on the three-phase correlation coefficients, the meter can be assigned to any one of the three phases (A, B, and C) of the marked node monitoring module. In this scenario example, by statistically analyzing the lag sampling interval between the electricity meter sampling and the monitoring module sampling, and using a search space expansion factor to determine the lag search range, the lag correlation between the data collected by the electricity meter and the data collected by the monitoring module is calculated. This avoids the problems of other similarity measures such as DTW and effectively characterizes the similarity between different sequences with sampling time lags.

[0059] At this point, the topological relationship between the electricity meters and the marker node monitoring modules has been determined, i.e., the target electricity meters located in the same power grid branch as the marker node monitoring modules. Further determination of the topological relationship between these target electricity meters and the meter box monitoring modules is needed. Specifically, voltage data from these target electricity meters and voltage data from the meter box monitoring modules, determined through the aforementioned topological relationship calculations, are sampled. Within the clusters of the marker node monitoring modules, each cluster includes both the marker node monitoring modules and several meter box monitoring modules, which can be located on child nodes of the marker node monitoring modules. The correlation coefficients between the voltage sequences of these target electricity meters and the voltage sequences of these meter box monitoring modules on a single phase are calculated to determine the topological relationship between the electricity meters and the meter box monitoring modules.

[0060] In this scenario example, from one perspective, based on the previously established topological relationship between the transformer monitoring module, branch line monitoring module, and meter box monitoring module, the topological relationship between the electricity meter and the meter box monitoring module can be added to obtain the complete topological relationship of the low-voltage distribution network. From another perspective, based on the previously established topological relationship between the transformer monitoring module, branch line monitoring module, and meter box monitoring module, and based on the previously established topological relationship between the electricity meter and the marker node monitoring module, the target electricity meter on the same grid branch as the marker node monitoring module can be identified. Then, by combining the topological relationship between the electricity meter and the meter box monitoring module, the target electricity meter can be connected to its corresponding meter box monitoring module to obtain the complete topological relationship of the low-voltage distribution network. In this scenario example, the method of first clustering the monitoring modules into large clusters and then clustering the electricity meters into smaller clusters avoids the problem of uneven voltage similarity distance density within the transformer substation.

[0061] Figure 2This is a flowchart illustrating a TLCC-based transformer topology identification method provided in this specification. This specification provides a TLCC-based transformer topology identification method applied to low-voltage distribution networks. The low-voltage distribution network includes several monitoring modules and energy meters connected to the monitoring modules. There is a time lag between the first voltage sequence collected by the monitoring modules and the second voltage sequence collected by the energy meters. The monitoring modules include a master node monitoring module and a marker node monitoring module; (reference...) Figure 2 As shown, the method may include the following steps.

[0062] S210, a marker node monitoring module that determines the lag search range and several monitoring modules; wherein, the marker node monitoring module is located on a child node of the main node monitoring module.

[0063] The hysteresis search range is used to represent the range of values ​​for the moving step size of the voltage sequence. In some cases, low-voltage distribution networks are equipped with electricity meters based on serial communication. Due to the inherently low communication speed of this method, the voltage, current, and other electricity consumption data reported by such electricity meters received by the monitoring module at a certain moment may not be current, typically resulting in a certain acquisition delay. Generally, when performing topology calculations on the monitoring module and the electricity meter, the correlation coefficient between the two cannot be directly calculated; otherwise, the time deviation in data acquisition will lead to significantly different calculation results, ultimately resulting in poor identification performance. Therefore, it is necessary to determine the hysteresis search range to help more accurately characterize the time lag correlation between the voltage sequence acquired by the monitoring module and the voltage sequence acquired by the electricity meter. In some embodiments, the hysteresis search range can be determined based on the hysteresis sampling interval between the second voltage sequence and the first voltage sequence.

[0064] The marker node monitoring module can be used to represent the monitoring module with the highest similarity to the main node monitoring module in the low-voltage distribution network. The main node monitoring module can be a transformer monitoring module used to represent the transformer in the distribution area. To determine the connection relationship between the marker node monitoring module and the energy meter, the marker node monitoring module needs to be selected from several monitoring modules. Specifically, the several monitoring modules include the main node monitoring module and the marker node monitoring module. The similarity between each monitoring module and the main node monitoring module is calculated, and the marker node monitoring module is determined based on the calculated similarity.

[0065] S220. The second voltage sequence is shifted within the hysteresis search range to obtain the shifted voltage sequence.

[0066] The second voltage sequence can be single-phase voltage time-series data collected by the electricity meter within a preset time period on a single day. For example, the second voltage sequence can be voltage data of the electricity meter within a certain 5-minute or 15-minute period on a single day, with no fewer than 70 voltage collection points. For example, if the number of sampling points is 80, then voltage data collected by all electricity meters in the distribution area within the same time period on the same day can be obtained, and each voltage sequence includes 80 voltage values.

[0067] Specifically, by shifting the voltage values ​​in the second voltage sequence forward or backward by a corresponding number of time steps according to the step size within the hysteresis search range, multiple shifted voltage sequences with different step size values ​​can be obtained. It should be noted that multiple shifted voltage sequences can have the same length.

[0068] S230. Correlation quantification calculation is performed using the shifted voltage sequence and the single-phase voltage sequence in the first voltage sequence to obtain a similarity set.

[0069] The single-phase voltage sequence in the first voltage sequence can be any one of the following: the phase A voltage sequence on phase A, the phase B voltage sequence on phase B, and the phase C voltage sequence on phase C, all collected by the monitoring module. The first voltage sequence can also be three-phase voltage data collected by the monitoring module within a preset time period on a single day. Specifically, it can be three-phase voltage data collected by the monitoring module within a certain 5-minute or 15-minute period on a single day, with at least 70 voltage collection points. For example, if the number of sampling points is 80, then three-phase voltage data collected by all monitoring modules in the transformer area within the same time period on the same day can be obtained, and each single-phase voltage sequence includes 80 voltage values.

[0070] The correlation quantification calculation means calculating the correlation coefficient between each shifted voltage sequence and the A, B, and C phase voltage sequences in the first voltage sequence, respectively. This yields a similarity set consisting of the correlation coefficients of all shifted voltage sequences and the first voltage sequence in the three phases.

[0071] In some cases, the voltage amplitude at any node on a line is related to the voltage of that phase bus, the electrical distance of the node from the head end, and the line load distribution. At adjacent moments, the voltage change (direction and amplitude) at any node on the line is mainly related to the change in active power flowing through upstream segments (overall line load time characteristics), the length of upstream segments, and the voltage amplitude of upstream nodes. In low-voltage distribution networks, the voltage of nodes on the same branch is affected by impedance and load power. When impedance is reflected by electrical distance, under the same active load, the closer the electrical distance, the higher the voltage similarity; for the same electrical distance, the larger the active load, the higher the voltage similarity. For single-source lines, at the same moment, the voltage amplitude at nodes along the line shows a gradually decreasing trend. When there are differences in the overall load characteristics between lines, the similarity between users located on the same outgoing line and with a closer electrical distance will be higher than the similarity between users located on different outgoing lines, and the closer the electrical distance, the higher the similarity between users. In some embodiments, to avoid the influence of absolute voltage differences on the calculation results, the Pearson correlation coefficient can be introduced to measure the similarity of voltage changes.

[0072] The Pearson correlation coefficient R(X,Y) is a statistical reflection of the degree of linear correlation between two sequences X and Y.

[0073]

[0074] Where Cov(X, Y) represents the covariance of sequence X and sequence Y; σ(X) and σ(Y) are the standard deviations of sequence X and sequence Y, respectively. The correlation of voltages is quantified, considering the time-series variation of voltage. The Pearson correlation coefficients of the voltage time-series curves at nodes u and v are:

[0075]

[0076] Among them, U u,t U v,t These represent the voltages at time t, respectively, of the voltage time-series curves of nodes u and v. The monitoring module and energy meter in the embodiments of this specification can be nodes used to calculate the Pearson correlation coefficient.

[0077] S240. Clustering is performed based on the similarity set to obtain the topological relationship between the marker node monitoring module and the electricity meter, so as to determine the electricity meter located on the same power grid branch as the marker node monitoring module.

[0078] Here, the similarity set represents the set of similarities between the voltage sequence of the energy meter calculated in step S230 and the voltage sequence of the marker node monitoring module across three phases. Clustering processing refers to clustering the energy meters so that they can be grouped according to the similarity between their voltage sequences and those of the marker node monitoring modules. Specifically, the number of clusters for energy meter clustering can be set according to the number of marker node monitoring modules and the number of phase items. Based on the aforementioned similarity set, the energy meters are clustered to obtain the energy meter cluster with the highest similarity to the marker node monitoring module in a single phase. The number of phase items can be 3. A single phase can be any one of phase A, phase B, or phase C.

[0079] The topological relationship between the marker node monitoring module and the electricity meter is represented by clustering the electricity meters to obtain clusters, and then connecting the electricity meters in each cluster to the branch line of the marker node monitoring module with the highest similarity, thus obtaining the meter affiliation and the topological relationship between the marker node monitoring module and the electricity meter. Specifically, the electricity meters in the clusters obtained after clustering can be connected to the branch lines of the marker node monitoring module corresponding to the phases with the highest similarity to the marker node monitoring module.

[0080] Identifying energy meters located on the same power grid branch as the marker node monitoring module means treating the branch line where the marker node monitoring module is located as a power grid branch of the low-voltage distribution network. Based on the obtained topological relationship between the marker node monitoring module and the energy meter, the power grid branch to which the energy meter belongs can be determined. Specifically, the energy meter can be included in the power grid branch where the marker node monitoring module is located, based on the branch line of the marker node monitoring module. It should be noted that the branch line where the marker node monitoring module is located can be a branch line in the low-voltage distribution network where the marker node monitoring module is the master node, and the branch line where the marker node monitoring module's child nodes are located can also be a branch line where the marker node monitoring module's child nodes are located.

[0081] In some implementations, reference Figure 3 As shown, the method for determining the marker node monitoring module may include the following steps.

[0082] S310. Determine the first average correlation matrix.

[0083] Among them, the elements in the first average correlation matrix correspond to any two monitoring modules other than the master node monitoring module, and are the average correlation coefficients between any two monitoring modules other than the master node monitoring module.

[0084] The average correlation coefficient represents the average correlation between the voltage sequences of any two monitoring modules (excluding the main node monitoring module) in a low-voltage distribution network across their effective in-phase phases. In some cases, when a monitoring module collects data, it may miss collecting voltage sequences on certain phases, or the collected voltage sequence values ​​on certain phases may be outliers. This can lead to errors in calculating the correlation coefficient between monitoring modules due to phase anomalies. Therefore, the average correlation coefficient is used to eliminate the impact of phase anomalies on the correlation calculation between monitoring modules. Specifically, the correlation coefficient matrix between any two monitoring modules is calculated based on the voltage sequences collected across all phases. Then, the correlation coefficients across the in-phase phases in this matrix are summed to obtain the summed correlation coefficient. Finally, the average correlation coefficient between the two monitoring modules is calculated by averaging the summed correlation coefficients based on the number of effective phases of each module.

[0085] The first average correlation matrix can be formed by calculating the average correlation coefficient between any two monitoring modules in a low-voltage power distribution network, excluding the main node monitoring module.

[0086] S320. Clustering is performed based on the first average correlation matrix to obtain the first clustering result; wherein, the first clustering result includes several first clusters.

[0087] The clustering process based on the first average correlation matrix means that the monitoring modules are clustered based on the average correlation coefficient between any two monitoring modules, so that monitoring modules with high similarity are grouped together. The first clustering result represents the clustering of the monitoring modules and the number of clusters obtained after the clustering process. The first clustering indicates the clustering of monitoring modules after clustering, where the similarity between any two monitoring modules within a cluster is greater than the similarity between any two monitoring modules between different clusters.

[0088] In some cases, users in a low-voltage distribution network located on the same outgoing line and with close electrical distance will have higher similarity than users located on different outgoing lines. Therefore, affinity clustering of monitoring modules can be performed using branch line correlation. Specifically, affinity propagation algorithms or silhouette coefficient methods can be used for evaluation. Based on the first average correlation matrix, monitoring modules are divided according to their average similarity to each other, ultimately obtaining the optimal clustering of monitoring modules and the number of clusters. The number of clusters can be used to set the number of clusters for clustering energy meters.

[0089] S330. Based on the correlation coefficient between the master node monitoring module and the monitoring modules included in the first cluster, determine the marker node monitoring module among the monitoring modules included in the first cluster.

[0090] In this context, the marker node monitoring module represents the monitoring module in the first cluster that has the greatest similarity to the master node monitoring module. In some cases, the voltage of nodes on the same branch in a low-voltage distribution network is affected by impedance and load power. When impedance is expressed as electrical distance, under the same active load, the closer the electrical distance, the higher the voltage similarity; conversely, under the same electrical distance, the larger the active load, the higher the voltage similarity. For a single-source line, at the same cross-section, the voltage amplitude of nodes along the line gradually decreases. Therefore, on the same branch in a low-voltage distribution network, when the active load is the same, the monitoring module in the first cluster with the closest electrical distance to the master node monitoring module is the marker node monitoring module; when the electrical distance is the same, the monitoring module with the largest active load in the first cluster is the marker node monitoring module.

[0091] In some embodiments, the correlation coefficient between the master node monitoring module and the monitoring modules included in the first cluster can be calculated using the Pearson correlation coefficient method. The monitoring module corresponding to the maximum correlation coefficient between the master node monitoring module and the monitoring modules included in the first cluster is determined as the marker node monitoring module.

[0092] The method described above for identifying the topological relationship between monitoring modules uses the average correlation coefficient to measure the similarity between any two monitoring modules, thus avoiding correlation coefficient errors caused by phase anomalies.

[0093] In some implementations, reference Figure 4 As shown, determining the first average correlation matrix may include:

[0094] S311. Based on the first voltage sequence collected by any two monitoring modules other than the master node monitoring module, determine the correlation coefficient matrix.

[0095] The correlation coefficient is used to represent the curve similarity of the first voltage sequence collected by any two monitoring modules.

[0096] Specifically, any two monitoring modules include a first monitoring module and a second monitoring module. Pearson correlation coefficients are calculated for any single-phase voltage sequence acquired by the first monitoring module and any single-phase voltage sequence acquired by the second monitoring module. In one example, if the single-phase voltage sequences of phases A, B, and C in the three-phase voltage sequences acquired by the first and second monitoring modules are all valid data, then after calculating the Pearson correlation coefficients, nine Pearson correlation coefficients can be obtained between the first and second monitoring modules. These are the Pearson correlation coefficients between the phase A voltage sequence of the first monitoring module and the phase A, B, and C voltage sequences of the second monitoring module; the Pearson correlation coefficients between the phase B voltage sequence of the first monitoring module and the phase A, B, and C voltage sequences of the second monitoring module; and the Pearson correlation coefficients between the phase C voltage sequence of the first monitoring module and the phase A, B, and C voltage sequences of the second monitoring module. Based on the above calculation theory, multiple Pearson correlation coefficients between any two monitoring modules in the transformer area can be obtained, and all Pearson correlation coefficients can form a correlation coefficient matrix.

[0097] S313. Determine the target element in the correlation coefficient matrix; wherein the target element is used to represent the degree of correlation between the single-phase voltage sequences in the same phase of the first voltage sequence collected by any two monitoring modules.

[0098] Specifically, taking the first monitoring module and the second monitoring module mentioned above as examples, the target element may include the phase A correlation coefficient of the voltage sequence collected by the first monitoring module and the second monitoring module on phase A, the phase B correlation coefficient on phase B, and the phase C correlation coefficient on phase C.

[0099] S315. Calculate the elements in the first average correlation matrix by averaging the target elements and the number of effective phases; where the number of effective phases is used to represent the number of target elements in an effective state.

[0100] In the embodiments described in this specification, when the monitoring module acquires voltage sequences, it may acquire null or outlier values ​​for a certain phase. In such cases, that phase is considered invalid, and the corresponding correlation coefficient calculated for that phase is also invalid. The number of valid phases can be obtained based on the number of valid phases in the three-phase voltage time-series data corresponding to any two monitoring modules. In one possible implementation, all correlation coefficients calculated for invalid phases of the monitoring module can be replaced with the value 0.

[0101] Specifically, taking the first and second monitoring modules mentioned above as examples, calculating the average correlation coefficient between the first and second monitoring modules includes: obtaining the A-phase correlation coefficient, B-phase correlation coefficient, and C-phase correlation coefficient between the first and second monitoring modules as target elements from the correlation coefficient matrix. If any of the three correlation coefficients is 0, then the effective phase number of the first and second monitoring modules is 2. In one example, the A-phase voltage sequence in the voltage sequence collected by the first monitoring module is empty, while all three-phase voltage sequences collected by the second monitoring module are valid data. When calculating the correlation coefficient between the first and second monitoring modules, the correlation coefficient values ​​between the A-phase voltage sequence of the first monitoring module and the A-phase, B-phase, and C-phase voltage sequences of the second monitoring module are all 0, and the remaining correlation coefficients are valid data. Therefore, the A-phase correlation coefficient between the first and second monitoring modules is 0, which is an invalid target element. The B-phase and C-phase correlation coefficients are both valid target elements, so the effective phase number is 2.

[0102] Then, the sum of the A-phase correlation coefficient, B-phase correlation coefficient, and C-phase correlation coefficient of the first monitoring module and the second monitoring module is calculated and denoted as the in-phase correlation coefficient sum. The quotient of the in-phase correlation coefficient sum and the effective phase number is taken as the average correlation coefficient between the first monitoring module and the second monitoring module.

[0103] Based on the above calculation process, the average correlation coefficient between any two monitoring modules in the transformer area can be calculated, and all the average correlation coefficients constitute the first average correlation matrix.

[0104] The above method uses the three-phase correlation coefficient and the number of effective phases of the voltage sequence to calculate the average correlation coefficient between the monitoring modules, thus avoiding errors in the correlation coefficient caused by phase anomalies.

[0105] In some implementations, clustering based on the first average correlation matrix to obtain the first clustering result may include: using an affinity propagation algorithm or a silhouette coefficient method to cluster the first average correlation matrix to obtain the first clustering result.

[0106] Specifically, based on the first average correlation matrix, the affinity propagation algorithm can be used to cluster several monitoring modules to obtain the optimal clustering of monitoring modules and the number of clusters, so as to ensure that the similarity between any two monitoring modules within a cluster is greater than the similarity between any two monitoring modules between clusters.

[0107] In some implementations, the hysteresis search range is determined based on the hysteresis sample interval between the second voltage sequence and the first voltage sequence; shifting the second voltage sequence within the hysteresis search range to obtain a shifted voltage sequence may include: shifting the second voltage sequence according to a preset step size within the hysteresis search range to obtain multiple shifted voltage sequences.

[0108] The lag search range is determined by the maximum lag sampling interval of the data between the statistical energy meter and the marker node monitoring module, and by expanding this maximum lag sampling interval according to the search space expansion factor to obtain the lag search range. Specifically, the lag sampling time of the data between the statistical energy meter and the marker node monitoring module is denoted as the lag sampling interval N. i The maximum lag interval N = MAX(N) is obtained statistically. i ), where i is the sample number. The hysteresis search range of the voltage sequence is set to SCOPE = [-w*N, w*N] based on the maximum hysteresis interval.

[0109] The preset step size within the hysteresis search range represents the number of moments that an element in the voltage sequence moves forward or backward.

[0110] For example, the first voltage sequence may include a voltage sequence U, and the second voltage sequence may include a voltage sequence V. For instance, since the electricity meter sequence typically lags by one sampling time, the maximum lag interval N can be 1. w is the search space expansion factor, which can typically be 2, so the lag search range SCOPE = [-2, 2]. For this lag search range, the preset step size can be s = {-2, -1, 0, 1, 2}. The data at the first sampling time t1 in the voltage sequence V is denoted as V0. t1 The data at the second sampling time t2 is denoted as V. t2 The data at the third sampling time t3 is denoted as V. t3 And so on. Each element in the voltage sequence V is shifted by a step size s. For example, when the step size s is -1, it means that all elements in the voltage sequence V are shifted one position to the previous time step, resulting in the shifted voltage sequence V'. In the shifted voltage sequence V', the data at the first sampling time t1 is V. (t-1)1 The data at the second sampling time t2 is V (t-1)2 The data at the third sampling time t3 is V (t-1)3And so on. In some cases, the sampled data corresponding to certain sampling times in the shifted voltage sequence V' may be missing. After shifting, a value of 0 can be inserted at the corresponding missing positions to ensure that the number of elements in the two sequences is consistent when calculating the correlation coefficient between the shifted voltage sequence V' and the first voltage sequence U, thus avoiding calculation anomalies. In other embodiments, for energy meters that have built-in clock synchronization functions to ensure that the collected current, voltage, and other electricity consumption information are at the same time, such as HPLC energy meters, the lag interval can be set to 0.

[0111] Based on the value of each preset step size within the hysteresis search range, the second voltage sequence is shifted to obtain multiple shifted voltage sequences.

[0112] In some implementations, reference Figure 5 As shown, correlation quantization is performed using the shifted voltage sequence and the single-phase voltage sequence in the first voltage sequence to obtain a similarity set, which may include:

[0113] S231. Obtain the sequence correlation coefficient between any shifted voltage sequence and the single-phase voltage sequence in the first voltage sequence.

[0114] Specifically, taking the first voltage sequence U and the second voltage sequence V as examples, obtaining the sequence correlation coefficient between the shifted voltage sequence V' and the single-phase voltage sequences in the voltage sequence U includes: calculating the A-phase correlation coefficient of the single-phase voltage sequence V' and the three-phase voltage sequence U in phase A, the B-phase correlation coefficient in phase B, and the C-phase correlation coefficient in phase C, respectively, thus obtaining three single-phase correlation coefficients. Based on the above calculation steps, the single-phase correlation coefficient between any shifted voltage sequence and the first voltage sequence can be obtained.

[0115] S233. Determine the target sequence correlation coefficient that meets the preset conditions from the sequence correlation coefficients.

[0116] Here, the preset conditions refer to the conditions that the data obtained from the aforementioned correlation coefficients must meet, as predetermined in the algorithm. The target sequence correlation coefficient represents the data obtained from the aforementioned correlation coefficients according to the preset conditions. In some embodiments, the preset conditions may be to select the correlation coefficient with the largest value among the sequence correlation coefficients.

[0117] S235. Construct a similarity set based on the correlation coefficient of the target sequence.

[0118] The similarity set can be a correlation coefficient matrix composed of the correlation coefficients of the target sequences.

[0119] In some embodiments, a shifted voltage sequence obtained by any single-phase voltage sequence collected by an energy meter within a hysteresis search space by a certain step size can be obtained as a shifted voltage sequence. The Pearson correlation coefficient between this shifted voltage sequence and any single-phase voltage sequence collected by any monitoring module can be calculated to obtain the correlation coefficients of the shifted voltage sequence and the voltage sequence collected by the monitoring module in three phases. Repeating this step yields a set of Pearson correlation coefficients R_offset = {R_offset} between multiple shifted voltage sequences obtained by shifting the single-phase voltage sequence of any energy meter within the hysteresis search space and the voltage sequence of any monitoring module. j}, where j is the search range number. Take the {R} corresponding to all electricity meters. j The set with the largest correlation coefficient constitutes the similarity set between the electricity meter and the monitoring module.

[0120] Specifically, taking the second voltage sequence V and the first voltage sequence U as examples, the second voltage sequence V is moved by a preset step size s = {-2, -1, 0, 1, 2} within the lag search range to obtain 5 moved voltage sequences. The Pearson correlation coefficients between the moved voltage sequences and the voltage sequence U in phases A, B, and C are calculated respectively, resulting in 15 Pearson correlation coefficients. These constitute the correlation coefficient set between the energy meter and the monitoring module corresponding to the voltage sequence U. A preset condition is set to take the largest correlation coefficient in this set as the similarity between the energy meter and the monitoring module corresponding to the voltage sequence U. For example, in this set, the Pearson correlation coefficient between the moved voltage sequence obtained by moving the second voltage sequence V by a step size s = -1 and the voltage sequence U in phase A is the largest. Therefore, this correlation coefficient is taken as the target sequence correlation coefficient and placed into the similarity set P between the energy meter and the monitoring module.

[0121] The above similarity calculation method takes into account the situation where the actual electricity meter data collection is mainly due to the data reporting delay and the collection time deviation is small. It utilizes the fact that the correlation of voltage sequences is the greatest at the time of time alignment, and that the correlation of in-phase sequences is greater than that of different phase sequences. Through the calculation of lag correlation coefficient and dynamic search, the amount of computation when calculating asynchronous time sequences by methods such as DTW can be greatly reduced, which can facilitate deployment in edge devices.

[0122] In some implementations, reference Figure 6 As shown, clustering based on similarity sets yields the topological relationship between the marker node monitoring module and the electricity meter, which may include:

[0123] S241. Determine the equivalent distance set between the second voltage sequence and the single-phase voltage sequence in the first voltage sequence based on the similarity set.

[0124] The equivalent distance can be used to measure the spatial distance relationship of electricity meter distribution and the impact of load distribution on electricity meter clustering. In some embodiments, the equivalent distance can be obtained based on the similarity between the electricity meter and the monitoring module. Specifically, taking the similarity set P between the electricity meter and the monitoring module obtained in step S235 as an example, the equivalent distance set between the second voltage sequence collected by any electricity meter and the first voltage sequence collected by any monitoring module can be calculated using the equivalent distance D = 1 - P.

[0125] S243. Perform hierarchical clustering on the equivalent distance set to obtain the topological relationship between the marked node monitoring module and the energy meter.

[0126] Based on the equivalent distance set obtained above, hierarchical clustering of electricity meters is performed. Specifically, this may include: (1) placing all electricity meters into an initial cluster. The split cluster and the original cluster are left empty as two object clusters; (2) selecting the cluster with the largest diameter from all clusters, finding the element with the largest average distance to other elements in that cluster, and placing it into the split cluster, with the remainder placed into the original cluster; (3) continuously finding an element in the original cluster such that the minimum distance between that element and elements in the split cluster is less than or equal to the minimum distance between that element and other elements in the original cluster. This element is placed into the split cluster until no element satisfying the above condition is found in the original cluster. At this point, the split cluster and the original cluster, together with other clusters, form a new cluster set; (4) repeating steps 2 and 3 until the number of clusters reaches the termination condition. The maximum average distance and minimum distance can be calculated based on the equivalent distance set.

[0127] In some implementations, hierarchical clustering of the equivalent distance set is performed to obtain the topological relationship between the marker node monitoring module and the energy meter. This may include: performing hierarchical clustering of the equivalent distance set according to a preset number of clusters to obtain a second clustering result; wherein the second clustering result includes several second clusters. A cluster to be processed is determined from the second clustering result; wherein the number of energy meters in the cluster to be processed is less than a threshold for the number of energy meters, or the cluster to be processed does not include a marker node monitoring module. The energy meters in the cluster to be processed are added to a second cluster in the second clustering result that includes one marker node monitoring module, resulting in a new second clustering result. Based on the new second clustering result, the topological relationship between the marker node monitoring module and the energy meter is determined.

[0128] The second cluster represents the energy meter cluster that has the greatest voltage similarity to the marker node monitoring module in the same phase.

[0129] In one embodiment, the preset cluster number for hierarchical clustering can be determined by the number of clusters of the monitoring module obtained in step S320 of the TLCC-based transformer topology relationship identification method provided in this specification, and the number of phase terms. Specifically, the cluster number k of the monitoring module is obtained through step S320. The voltage sequence collected by the monitoring module may include voltage sequences with three phases, so the number of phase terms can be 3. Based on the cluster number k and the number of phase terms, the cluster number for hierarchical clustering of the energy meter can be set to 3k. The above-described hierarchical clustering process of the energy meter based on the equivalent distance set terminates the clustering when the number of clusters reaches the preset cluster number condition of 3k. The second clustering result, which divides the energy meter into clusters according to the similarity of the single-phase phase voltage sequence with the monitoring module of the marked node, can be obtained. The single-phase phase may include any one of phase A, phase B, and phase C.

[0130] The above method uses the effective cluster number of the monitoring module for the initial clustering of the electricity meters, instead of directly using the number of meters as the cluster number to cluster the electricity meters. This avoids the identification error problems introduced by the fact that the Pearson correlation coefficient between different electricity meters that are physically close may be larger than that between different electricity meters that are physically far apart, and that the size of the Pearson correlation coefficient is related to the load distribution.

[0131] In some cases, due to calculation errors or other reasons, the clustering results of electricity meters obtained through clustering may not be ideal, requiring re-clustering and identification of the electricity meters in the unsatisfactory clusters. In the embodiments of this specification, a pre-set cluster for electricity meters to be processed is used to store the electricity meters in the clusters with unsatisfactory results after the aforementioned clustering process. In one example, there may be isolated values ​​among the electricity meters after clustering. Isolated values ​​indicate clusters where the number of electricity meters after clustering is much smaller than the preset expected value. The electricity meters in these clusters need to be re-clustered and identified. Specifically, an isolated number threshold m can be set; when the number of electricity meters in a cluster is less than m, the cluster is removed, and the electricity meters in that cluster are added to the cluster to be processed. In another example, the electricity meters in the clusters after clustering may not be assigned to any phase branch line of any monitoring module; in this case, the cluster is removed, and the electricity meters in that cluster are added to the cluster to be processed.

[0132] In some embodiments, adding energy meters from the cluster to be processed to the second clustering result may include: calculating the correlation between the energy meters in the cluster to be processed and other clusters, and determining which cluster to place the energy meters in the cluster to be processed into based on the correlation. Specifically, the average correlation between the energy meters in the cluster to be processed and other energy meters in other clusters may be calculated separately, and the energy meters in the cluster to be processed may be placed into the corresponding cluster based on the highest average correlation obtained. Alternatively, the correlation between the energy meters in the cluster to be processed and any energy meter in other clusters may be calculated, and the energy meters in the cluster to be processed may be placed into the corresponding cluster based on the highest correlation obtained. Finally, a new second clustering result for all energy meters is obtained.

[0133] In some embodiments, determining the topological relationship between the marker node monitoring module and the energy meter based on the new second clustering results may include: placing the energy meter cluster in the new second clustering results into the cluster containing the marker node monitoring module with the highest correlation, and connecting the energy meters in the energy meter cluster to the branch lines of the marker node monitoring module, thereby obtaining the topological relationship between the energy meter and the marker node monitoring module. Specifically, the average correlation coefficients between the single-phase voltage sequence collected by the energy meter in the new second cluster and the three-phase voltage sequence of any marker node monitoring module in phases A, B, and C can be calculated respectively. Based on the calculated maximum average correlation coefficient in a certain phase, the energy meter in the new second cluster is placed into the corresponding phase cluster under the cluster containing the marker node monitoring module, and the energy meter in the second cluster is connected to the corresponding phase branch line of the marker node monitoring module, thereby obtaining the topological relationship between the energy meter and the marker node monitoring module.

[0134] In some cases, the electricity meters in the clustered data may be assigned to multiple phase clusters, including being assigned to more than one phase cluster of the same marker node monitoring module, or being assigned to multiple phase clusters of multiple different marker node monitoring modules. In the embodiments of this specification, the number of phase clusters to which the electricity meters in the cluster are assigned can be used as the number of clusters for re-clustering the electricity meters within the cluster, and the electricity meters in the cluster can be re-clustered. For example, assuming that an electricity meter cluster is simultaneously assigned to the phase clusters of phase A and phase B of the same monitoring module, the number of phase clusters to which the electricity meter cluster is assigned is 2. Setting the number of clusters to 2, the electricity meter cluster is re-clustered, resulting in 2 new electricity meter clusters, each corresponding to only one phase cluster. Finally, the single-phase topology relationship between the electricity meters and the marker node monitoring modules can be obtained.

[0135] The above method identifies the topological relationship between the monitoring modules and electricity meters. It employs a method of first clustering the monitoring modules into large clusters and then clustering the electricity meters into smaller clusters, avoiding the problem of uneven voltage similarity density within the transformer area. Instead of directly using the number of meters as the cluster number for electricity meter clustering, it avoids the possibility that the Pearson correlation coefficient between physically close electricity meters might be larger than that between physically distant electricity meters, and also avoids the identification error introduced by the relationship between the magnitude of the Pearson correlation coefficient and load distribution.

[0136] In some implementations, reference Figure 7a As shown, clustering based on similarity sets yields several second clusters. The electricity meters included in these second clusters and the marker node monitoring modules included in these second clusters are located on the same power grid branch. The first cluster to which the marker node monitoring modules in these second clusters belong is denoted as the target cluster. This TLCC-based transformer substation topology identification method may further include the following steps:

[0137] S410. The phase corresponding to the marker node monitoring module included in the second cluster is taken as the target phase corresponding to the second cluster.

[0138] Among them, the phase category of the second voltage sequence collected by the energy meter included in the second cluster is the target phase.

[0139] The system comprises several second clusters obtained by clustering with a cluster size of 3k. Each second cluster includes a single-phase voltage sequence from a marker node monitoring module. For example, a second cluster may include phase A voltage data from a marker node monitoring module, phase B voltage data from a marker node monitoring module, and phase C voltage data from a marker node monitoring module. Specifically, the system traverses the second clusters and checks the phase corresponding to the marker node monitoring module within each cluster, using this phase as the phase of the energy meter in the second cluster connected to the branch line of that marker node monitoring module.

[0140] For example, if the phase corresponding to the marker node monitoring module included in the second cluster is found to be phase A, then phase A is determined as the target phase corresponding to the second cluster, and the phase category of the second voltage sequence collected by the energy meter included in the second cluster is phase A.

[0141] For example, if the phase corresponding to the marker node monitoring module included in the second cluster is detected to be phase B, then phase B is determined as the target phase corresponding to the second cluster, and the phase category of the second voltage sequence collected by the energy meter included in the second cluster is phase B.

[0142] For example, if the phase corresponding to the marker node monitoring module included in the second cluster is detected to be phase C, then phase C is determined as the target phase corresponding to the second cluster, and the phase category of the second voltage sequence collected by the energy meter included in the second cluster is phase C.

[0143] S420. Based on the single-phase voltage sequence of the monitoring module in the target cluster at the target phase and the second voltage sequence collected by the energy meter included in the second cluster, the correlation coefficient between the monitoring module in the target cluster and the energy meter included in the second cluster at the target phase is calculated.

[0144] In some embodiments, the monitoring module may be a bin monitoring module in the target cluster, excluding the marker node monitoring module, and the bin monitoring module is a child node of the marker node monitoring module. Specifically, refer to... Figure 7b As shown, for a four-level low-voltage distribution network topology consisting of "master node monitoring module - marker node monitoring module - meter box monitoring module - electricity meter", the marker node monitoring module can be an LTU corresponding to the secondary data acquisition unit. After clustering the monitoring modules other than the master node monitoring module to obtain the first cluster, and after identifying the marker node monitoring module in the first cluster in step S330, the monitoring modules other than the marker node monitoring module in the first cluster can be further identified as meter box monitoring modules. The meter box monitoring module is located on a child node of the marker node monitoring module, thus obtaining the topological relationship between the meter box monitoring module and the marker node monitoring module, and thus the topological relationship between "master node monitoring module - marker node monitoring module - meter box monitoring module". Further determination of the topological relationship between the electricity meter and the meter box monitoring module is needed to finally determine the topology of the low-voltage distribution network.

[0145] In other embodiments, the monitoring module may be a bin detection module that includes a marker node monitoring module within the target cluster. Specifically, refer to... Figure 7c As shown, for a three-tiered low-voltage distribution network topology consisting of "master node monitoring module - meter box monitoring module - electricity meter", the marked node monitoring module identified in the first cluster represents the meter box monitoring module that has the greatest similarity to the master node monitoring module among those located on the same outgoing branch as the master node monitoring module. After determining the marked node monitoring module, the topological relationship between "master node monitoring module - meter box monitoring module" can be determined. Further determination of the topological relationship between the electricity meter and all meter box monitoring modules is needed to ultimately determine the topology of the low-voltage distribution network.

[0146] Having already identified the marker node monitoring modules and the electricity meters and monitoring modules located on the same grid branch, it is necessary to further determine the topological relationship (or direct connection relationship) between the electricity meters in the second cluster and the monitoring modules in the target cluster. Specifically, the similarity between the voltage sequences of the electricity meters in the second cluster and the voltage sequences of the monitoring modules in the target cluster is calculated. It should be noted that the phase of the single-phase voltage sequence on the target phase is already determined; when calculating similarity, the similarity between the voltage sequences of the electricity meters and the monitoring modules at that phase is calculated. Furthermore, the correlation coefficient between the voltage sequences of the electricity meters in the second cluster and the single-phase voltage sequences of the monitoring modules in the target cluster at that phase is calculated using the hysteresis correlation calculation method.

[0147] For example, any second cluster is denoted as cluster F. Cluster F includes the single-phase voltage sequence of any phase in the marker node monitoring module X and the second voltage sequence collected by the energy meter. The first cluster to which the marker node monitoring module X belongs in cluster F is denoted as the target cluster. The target phase corresponding to cluster F is phase A. The correlation between the single-phase voltage sequence of the meter box monitoring module in phase A in the target cluster and the second voltage sequence collected by the energy meter in cluster F is calculated to obtain several correlation coefficients. The specific situations of phases B and C will not be elaborated here.

[0148] S430. Based on the correlation coefficient on the target phase, determine that the electricity meters included in the second cluster and the monitoring modules in the target cluster have a topological relationship on the target phase.

[0149] Specifically, based on the correlation coefficient between the voltage sequence of the energy meters in the second cluster calculated at the target phase and the voltage sequence of the monitoring module in the target cluster, the set of energy meters with the highest similarity to the monitoring module at different target phases can be obtained. By connecting the energy meters in this set to the phase branch lines of the corresponding monitoring modules, the topological relationship between the energy meters in the second cluster and the monitoring modules in the target cluster at the target phase can be obtained.

[0150] For example, cluster F includes a marker node monitoring module X, and the target cluster to which marker node monitoring module X belongs includes meter boxes LTU1 and LTU2. If cluster F includes the A-phase voltage sequence collected by marker node monitoring module X, the second voltage sequence collected by energy meter 1, the second voltage sequence collected by energy meter 2, the second voltage sequence collected by energy meter 3, and the second voltage sequence collected by energy meter 4, then the A-phase voltage sequence collected by meter box LTU1 and the A-phase voltage sequence collected by meter box LTU2 are obtained.

[0151] Calculate the correlation coefficient C11 between the second voltage sequence collected by energy meter 1 and the phase A voltage sequence collected by meter box LTU1, and calculate the correlation coefficient C12 between the second voltage sequence collected by energy meter 1 and the phase A voltage sequence collected by meter box LTU2. If the correlation coefficient C11 is greater than the correlation coefficient C21, then connect energy meter 1 to the phase A sequence collected by meter box LTU1.

[0152] Calculate the correlation coefficient C21 between the second voltage sequence collected by energy meter 2 and the phase A voltage sequence collected by meter box LTU1, and calculate the correlation coefficient C22 between the second voltage sequence collected by energy meter 2 and the phase A voltage sequence collected by meter box LTU2. If the correlation coefficient C21 is less than the correlation coefficient C22, then connect energy meter 2 to the phase A sequence collected by meter box LTU2.

[0153] Calculate the correlation coefficient C31 between the second voltage sequence collected by energy meter 3 and the phase A voltage sequence collected by meter box LTU1, and calculate the correlation coefficient C32 between the second voltage sequence collected by energy meter 3 and the phase A voltage sequence collected by meter box LTU2. If the correlation coefficient C31 is greater than the correlation coefficient C32, then connect energy meter 3 to the phase A sequence collected by meter box LTU1.

[0154] Calculate the correlation coefficient C41 between the second voltage sequence collected by energy meter 4 and the phase A voltage sequence collected by meter box LTU1, and calculate the correlation coefficient C42 between the second voltage sequence collected by energy meter 4 and the phase A voltage sequence collected by meter box LTU2. If the correlation coefficient C41 is less than the correlation coefficient C42, then connect energy meter 4 to the phase A sequence collected by meter box LTU2.

[0155] It should be noted that if the correlation coefficient C11 is less than the correlation coefficient C21, then electricity meter 1 will be connected to phase A, which is collected by meter box LTU2. This will not be elaborated further. It is understood that electricity meters 1, 2, 3, and 4 respectively collect the second voltage sequence of phase A.

[0156] For example, cluster F includes a marker node monitoring module X, and the target cluster to which marker node monitoring module X belongs includes meter boxes LTU1 and LTU2. If cluster F includes the B-phase voltage sequence collected by marker node monitoring module X, the second voltage sequence collected by energy meter 1, the second voltage sequence collected by energy meter 2, the second voltage sequence collected by energy meter 3, and the second voltage sequence collected by energy meter 4, then the B-phase voltage sequence collected by meter box LTU1 and the B-phase voltage sequence collected by meter box LTU2 are obtained. Further, the correlation coefficient between the second voltage sequence collected by energy meter 1 and the B-phase voltage sequence collected by meter box LTU1, and the correlation coefficient between the second voltage sequence collected by energy meter 1 and the B-phase voltage sequence collected by meter box LTU2 are calculated, as described above, and will not be repeated here.

[0157] For example, cluster F includes a marker node monitoring module X, and the target cluster to which marker node monitoring module X belongs includes meter boxes LTU1 and LTU2. If cluster F includes the C-phase voltage sequence collected by marker node monitoring module X, the second voltage sequence collected by energy meter 1, the second voltage sequence collected by energy meter 2, the second voltage sequence collected by energy meter 3, and the second voltage sequence collected by energy meter 4, then the C-phase voltage sequence collected by meter box LTU1 and the C-phase voltage sequence collected by meter box LTU2 are obtained. Further, the correlation coefficient between the second voltage sequence collected by energy meter 1 and the C-phase voltage sequence collected by meter box LTU1, and the correlation coefficient between the second voltage sequence collected by energy meter 1 and the C-phase voltage sequence collected by meter box LTU2 are calculated, as described above, and will not be repeated here.

[0158] The above implementation employs the maximum correlation in-phase allocation method within the monitoring module cluster, supporting the identification of the electricity meter topology down to the smallest branch unit, effectively improving the accuracy of topology identification. This method only requires voltage data analysis, significantly reducing the data acquisition pressure on the power communication network and the computing power requirements of the processing unit.

[0159] This specification also provides a TLCC-based transformer topology identification device 800, applied to a low-voltage power distribution network. The low-voltage power distribution network includes several monitoring modules and energy meters connected to the monitoring modules; wherein there is a time lag between the first voltage sequence collected by the monitoring modules and the second voltage sequence collected by the energy meters; the several monitoring modules include a master node monitoring module and a marker node monitoring module. Figure 8 As shown, the TLCC-based transformer area topology relationship identification device 800 includes: a determination module 810, a movement processing module 820, a correlation calculation module 830, and a clustering processing module 840.

[0160] The determination module is used to determine the lag search range and the marker node monitoring unit of several monitoring units; wherein the marker node monitoring unit is located on the child node of the main node monitoring unit.

[0161] The shift processing module is used to shift the second voltage sequence within the hysteresis search range to obtain the shifted voltage sequence.

[0162] The correlation calculation module is used to perform correlation quantification calculations on the shifted voltage sequence and the single-phase voltage sequence in the first voltage sequence to obtain a similarity set.

[0163] The clustering module is used to perform clustering based on similarity sets to obtain the topological relationship between the marked node monitoring unit and the energy meter.

[0164] In some embodiments, the determining module is specifically used for: determining a first average correlation matrix; wherein, the elements in the first average correlation matrix correspond to any two monitoring modules other than the master node monitoring module in a plurality of monitoring modules, and are the average correlation coefficients between any two monitoring modules other than the master node monitoring module in a plurality of monitoring modules; performing clustering processing based on the first average correlation matrix to obtain a first clustering result; wherein, the first clustering result includes a plurality of first clusters; and determining a marker node monitoring module in the monitoring modules included in the first cluster according to the correlation coefficients between the master node monitoring module and the monitoring modules included in the first cluster.

[0165] In some embodiments, the determining module is further specifically used to: determine a first average correlation matrix. This includes: determining a correlation coefficient matrix based on first voltage sequences collected by any two monitoring modules (excluding the master node monitoring module); determining target elements in the correlation coefficient matrix; wherein the target elements represent the degree of correlation between single-phase voltage sequences in the same phase of the first voltage sequences collected by any two monitoring modules; and averaging the target elements and the number of effective phases to determine the elements in the first average correlation matrix; wherein the number of effective phases represents the number of target elements in an effective state.

[0166] In some embodiments, the determining module is further configured to: determine the hysteresis search range based on the hysteresis sample interval between the second voltage sequence and the first voltage sequence.

[0167] In some embodiments, the moving processing module is specifically used to: move the second voltage sequence according to a preset step size within the hysteresis search range to obtain multiple moved voltage sequences.

[0168] In some embodiments, the correlation calculation module is specifically used to: obtain the sequence correlation coefficient between any moved voltage sequence and the single-phase voltage sequence in the first voltage sequence; determine the target sequence correlation coefficient that meets the preset conditions from the sequence correlation coefficients; and construct a similarity set based on the target sequence correlation coefficient.

[0169] In some embodiments, the clustering processing module is specifically used to: determine the equivalent distance set between the second voltage sequence and the single-phase voltage sequence in the first voltage sequence based on the similarity set; perform hierarchical clustering processing on the equivalent distance set to obtain the topological relationship between the marker node monitoring module and the energy meter.

[0170] In some embodiments, the clustering processing module is further configured to perform hierarchical clustering on the equivalent distance set to obtain the topological relationship between the marker node monitoring module and the electricity meter. Specifically, this includes: performing hierarchical clustering on the equivalent distance set according to a preset number of clusters to obtain a second clustering result; wherein the second clustering result includes several second clusters; determining the cluster to be processed in the second clustering result; wherein the number of electricity meters in the cluster to be processed is less than a threshold for the number of electricity meters, or the cluster to be processed does not include a marker node monitoring module; adding the electricity meters in the cluster to be processed to a second cluster in the second clustering result that includes one marker node monitoring module to obtain a new second clustering result; and determining the topological relationship between the marker node monitoring module and the electricity meter based on the new second clustering result.

[0171] In some embodiments, the clustering processing module is further configured to perform clustering processing on the first average correlation matrix to obtain a first clustering result. Specifically, this includes: using an affinity propagation algorithm or a silhouette coefficient method to perform clustering processing on the first average correlation matrix to obtain the first clustering result.

[0172] In some embodiments, clustering is performed based on a similarity set to obtain several second clusters. The electricity meters included in the second clusters and the marker node monitoring modules included in the second clusters are located on the same power grid branch. The first cluster to which the marker node monitoring modules included in the second clusters belong is denoted as the target cluster. The device may also include a target phase determination module, a phase coefficient determination module, and a phase identification module.

[0173] The target phase determination module is used to take the phase corresponding to the marker node monitoring module included in the second cluster as the target phase of the second cluster; wherein, the phase category of the second voltage sequence collected by the energy meter included in the second cluster is the target phase.

[0174] The phase coefficient determination module is used to perform correlation calculations based on the single-phase voltage sequence of the monitoring module in the target cluster at the target phase and the second voltage sequence collected by the energy meters included in the second cluster, to obtain the correlation coefficient between the monitoring module in the target cluster and the energy meters included in the second cluster at the target phase.

[0175] The phase identification module is used to determine, based on the correlation coefficient on the target phase, the topological relationship between the energy meters included in the second cluster and the monitoring modules in the target cluster on the target phase.

[0176] The TLCC-based transformer substation topology identification device described in this specification solves the problem of asynchronous data acquisition times among nodes in low-voltage distribution networks by using a time-lag cross-correlation algorithm to calculate voltage similarity. This avoids the problems associated with other similarity measurement methods such as DTW and effectively characterizes the similarity between different sequences. Furthermore, it employs a method of first clustering monitoring modules into large clusters and then clustering energy meters into smaller clusters, avoiding the problem of uneven voltage similarity distance density within the transformer substation and effectively improving the accuracy of topology identification. This method only requires voltage data analysis, significantly reducing the data acquisition burden on the power communication network and the computational power requirements of the processing unit.

[0177] The embodiments described in this specification also provide a computer device. Figure 9 This is a structural block diagram of a computer device according to one embodiment of this specification, such as... Figure 9 As shown, the computer device 900 includes a memory 910, a processor 920, and a computer program 930 stored in the memory 910 and executable on the processor 920. When the processor 920 executes the computer program 930, it implements the TLCC-based method for identifying the topology relationship of a transformer area according to any of the aforementioned embodiments.

[0178] The computer equipment described in this specification solves the problem of asynchronous data acquisition times among nodes in low-voltage distribution networks by using an algorithm for calculating voltage similarity through time-lag cross-correlation. This avoids the problems associated with other similarity measurement methods such as DTW and effectively characterizes the similarity between different sequences. Furthermore, it employs a method of first clustering monitoring modules into large clusters and then clustering energy meters into smaller clusters, avoiding the problem of uneven voltage similarity distance density within distribution areas and effectively improving the accuracy of topology identification. This method only requires voltage data analysis, significantly reducing the data acquisition burden on the power communication network and the computing power requirements of the processing unit.

[0179] The embodiments of this specification also provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the TLCC-based method for identifying transformer area topology relationships according to any of the foregoing embodiments.

[0180] The computer-readable storage medium described in this specification solves the problem of asynchronous data acquisition times among nodes in low-voltage distribution networks by using an algorithm for calculating voltage similarity through time-lag cross-correlation. This avoids the problems associated with other similarity measurement methods such as DTW and effectively characterizes the similarity between different sequences. Furthermore, it employs a method of first clustering monitoring modules into large clusters and then clustering energy meters into smaller clusters, avoiding the problem of uneven voltage similarity distance density within distribution areas and effectively improving the accuracy of topology identification. This method only requires voltage data analysis, significantly reducing the data acquisition burden on power communication networks and the computational power requirements of processing units.

[0181] It should be noted that the logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of computer-readable media include: an electrical connection having one or more wires (electronic device), a portable computer disk drive (magnetic device), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Alternatively, the computer-readable medium may be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in a computer memory.

[0182] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0183] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0184] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this invention, "a plurality of" means at least two, such as two, three, etc., unless otherwise explicitly specified.

[0185] In this invention, unless otherwise explicitly specified and limited, the terms "installation," "connection," "linking," and "fixing," etc., should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral part; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; they can refer to the internal communication of two components or the interaction between two components, unless otherwise explicitly limited. Those skilled in the art can understand the specific meaning of the above terms in this invention according to the specific circumstances.

[0186] Although embodiments of the present invention have been shown and described above, it is understood that the above embodiments are exemplary and should not be construed as limiting the present invention. Those skilled in the art can make changes, modifications, substitutions and variations to the above embodiments within the scope of the present invention.

Claims

1. A method for identifying transformer substation topology relationships based on TLCC, characterized in that, The method is applied to a low-voltage power distribution network, which includes several monitoring modules and energy meters connected to the monitoring modules; wherein there is a time lag between a first voltage sequence collected by the monitoring modules and a second voltage sequence collected by the energy meters; the several monitoring modules include a master node monitoring module and a marker node monitoring module; the method includes: Determine the lag search range and the marker node monitoring module of the plurality of monitoring modules; wherein, the marker node monitoring module is located on a child node of the main node monitoring module; Within the hysteresis search range, the second voltage sequence is shifted according to a preset step size within the hysteresis search range to obtain multiple shifted voltage sequences; Correlation quantification calculation is performed using the shifted voltage sequence and the single-phase voltage sequence in the first voltage sequence to obtain a similarity set; Clustering is performed based on the similarity set to obtain the topological relationship between the marker node monitoring module and the electricity meter, so as to identify the electricity meter located on the same power grid branch as the marker node monitoring module. The step of performing clustering based on the similarity set to obtain the topological relationship between the marker node monitoring module and the energy meter includes: Based on the similarity set, determine the equivalent distance set between the second voltage sequence and the single-phase voltage sequence in the first voltage sequence; Hierarchical clustering is performed on the equivalent distance set to obtain the topological relationship between the marker node monitoring module and the energy meter; Hierarchical clustering is performed on the equivalent distance set according to a preset number of clusters to obtain a second clustering result; wherein, the second clustering result includes several second clusters; In the second clustering result, a cluster to be processed is determined; wherein, the number of electricity meters in the cluster to be processed is less than the electricity meter number threshold, or, the cluster to be processed does not include the marker node monitoring module; The energy meters in the cluster to be processed are added to the second cluster result, which includes one of the marked node monitoring modules, to obtain a new second cluster result; Based on the new second clustering results, the topological relationship between the marker node monitoring module and the electricity meter is determined.

2. The method according to claim 1, characterized in that, The method for determining the marker node monitoring module includes: Determine the first average correlation matrix; wherein, the elements in the first average correlation matrix correspond to any two monitoring modules other than the master node monitoring module among the plurality of monitoring modules, and are the average correlation coefficients between any two monitoring modules other than the master node monitoring module among the plurality of monitoring modules; Clustering is performed based on the first average correlation matrix to obtain a first clustering result; wherein, the first clustering result includes several first clusters; Based on the correlation coefficient between the master node monitoring module and the monitoring modules included in the first cluster, the marker node monitoring module is determined among the monitoring modules included in the first cluster.

3. The method according to claim 2, characterized in that, Determining the first average correlation matrix includes: Based on the first voltage sequence collected by any two monitoring modules other than the master node monitoring module, the correlation coefficient matrix is ​​determined. A target element is determined in the correlation coefficient matrix; wherein the target element is used to represent the degree of correlation between the single-phase voltage sequences in the same phase of the first voltage sequence collected by any two monitoring modules; The elements in the first average correlation matrix are determined by averaging the target elements and the number of effective phases; wherein the number of effective phases is used to represent the number of target elements in an effective state.

4. The method according to claim 2, characterized in that, The step of clustering the first average correlation matrix to obtain the first clustering result includes: The first average correlation matrix is ​​clustered using the affinity propagation algorithm or the silhouette coefficient method to obtain the first clustering result.

5. The method according to any one of claims 1 to 4, characterized in that, The hysteresis search range is determined based on the hysteresis sample interval between the second voltage sequence and the first voltage sequence.

6. The method according to any one of claims 1 to 4, characterized in that, The correlation quantization calculation is performed using the shifted voltage sequence and the single-phase voltage sequence in the first voltage sequence to obtain a similarity set, including: Obtain the sequence correlation coefficient between any of the shifted voltage sequences and the single-phase voltage sequences in the first voltage sequence; Determine the target sequence correlation coefficient that meets the preset conditions from the sequence correlation coefficients; The similarity set is constructed based on the correlation coefficient of the target sequence.

7. The method according to claim 1, characterized in that, Clustering is performed based on the similarity set to obtain several second clusters. The electricity meters included in the second clusters and the marker node monitoring modules included in the second clusters are located on the same power grid branch. The first cluster to which the marker node monitoring modules included in the second clusters belong is denoted as the target cluster. The method further includes: The phase corresponding to the marker node monitoring module included in the second cluster is taken as the target phase corresponding to the second cluster; wherein, the phase category of the second voltage sequence collected by the energy meter included in the second cluster is the target phase; The correlation coefficient between the monitoring module in the target cluster and the energy meter in the second cluster on the target phase is obtained by performing correlation calculation based on the single-phase voltage sequence of the monitoring module in the target cluster and the second voltage sequence collected by the energy meter in the second cluster on the target phase. Based on the correlation coefficient on the target phase, it is determined that the electricity meters included in the second cluster and the monitoring modules in the target cluster have a topological relationship on the target phase.

8. A transformer substation topology identification device based on TLCC, characterized in that, The transformer area topology identification device is used to perform the steps of the method according to any one of claims 1 to 7 and is applied to a low-voltage distribution network, the low-voltage distribution network including a plurality of monitoring modules and an energy meter connected to the monitoring modules; wherein, there is a time lag between the first voltage sequence collected by the monitoring modules and the second voltage sequence collected by the energy meter; the plurality of monitoring modules include a master node monitoring module and a marker node monitoring module; the device includes: A determination module is used to determine the hysteresis search range and the marker node monitoring module of the plurality of monitoring modules; wherein, the marker node monitoring module is located on a child node of the main node monitoring module; The shift processing module is used to shift the second voltage sequence within the hysteresis search range to obtain the shifted voltage sequence. The correlation calculation module is used to perform correlation quantification calculation using the shifted voltage sequence and the single-phase voltage sequence in the first voltage sequence to obtain a similarity set. A clustering module is used to perform clustering based on the similarity set to obtain the topological relationship between the marker node monitoring module and the electricity meter, so as to determine the electricity meter located on the same power grid branch as the marker node monitoring module.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 7.

10. A 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 steps of the method according to any one of claims 1 to 7.

Images