A power distribution network fault section locating method and system based on feeder terminal data

Abstract

The application relates to the technical field of power distribution network operation and maintenance, and discloses a power distribution network fault section positioning method and system based on feeder terminal data. The method comprises the following steps: purifying and regularizing fault recording data uploaded by a feeder terminal in a power distribution network to obtain a fault feature sequence; synchronizing the time axis of the fault feature sequence to obtain a time sequence fault propagation graph; screening the electrical quantity amplitude in the time sequence fault propagation graph to obtain an abnormal monitoring point, and determining a suspected fault section according to a topological connection relationship and the abnormal monitoring point; calculating the cross-correlation coefficient of a first end transient current waveform and a last end transient current waveform in the suspected fault section, comparing the cross-correlation coefficient with a preset similarity threshold, and obtaining a fault section to be verified; and according to a fault phase propagation logic, topologically adapting and checking the phase of zero sequence voltage on both sides of the fault section to be verified to obtain the fault section. The application can improve the efficiency of power distribution network fault section positioning.

Description

Technical Field

[0001] This invention relates to the field of distribution network operation and maintenance technology, and in particular to a method and system for locating fault sections in a distribution network based on feeder terminal data. Background Technology

[0002] Existing methods for locating fault sections in distribution networks directly use the original fault recording data from feeder terminals for analysis. However, the original sampled data is not processed for missing data or noise removal. Data collected from multiple terminals has a time stamp synchronization problem, making it impossible to construct a fault propagation correlation under a unified time sequence dimension. The fault range is determined solely by a single electrical quantity amplitude change, which is susceptible to sampling distortion and time sequence offset interference, resulting in a large number of misjudged points. The fault section coarse screening process has weak fault tolerance.

[0003] Existing fault location methods lack a layered, multi-dimensional verification mechanism. They fail to combine transient current waveform characteristics and zero-sequence voltage phase topological constraints to verify candidate sections layer by layer. This results in a single fault judgment dimension, insufficient section identification accuracy, and a long overall computation and analysis time for fault location, leading to low fault response efficiency. Therefore, improving the accuracy and computational efficiency of fault location in distribution networks has become an urgent problem to be solved. Summary of the Invention

[0004] This invention provides a method and system for locating fault sections in a distribution network based on feeder terminal data, in order to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides a method for locating fault sections in a distribution network based on feeder terminal data, comprising: D1. Obtain fault waveform data uploaded by feeder terminals in the distribution network, and perform purification and normalization processing on the fault waveform data to obtain the fault characteristic sequence of the distribution network. D2. Based on the topological connection relationship of the distribution network, the fault characteristic sequence is synchronized on the time axis to obtain the time-series fault propagation map of the distribution network; D3. Perform difference screening on the electrical quantity amplitudes in the time-series fault propagation map to obtain abnormal monitoring points in the time-series fault propagation map, and determine the suspected fault sections of the distribution network based on the topological connection relationship and the abnormal monitoring points. D4. Calculate the cross-correlation coefficient between the transient current waveform at the beginning and the transient current waveform at the end of the suspected fault section, and compare the cross-correlation coefficient with a preset similarity threshold to obtain the fault section to be verified in the distribution network. D5. Based on the fault phase propagation logic of the distribution network, perform topology adaptation verification on the zero-sequence voltage phases on both sides of the fault section to be verified to obtain the fault section of the distribution network.

[0006] In a preferred embodiment, the step of acquiring fault waveform data uploaded by feeder terminals in the distribution network and performing purification and normalization processing on the fault waveform data to obtain the fault characteristic sequence of the distribution network includes: The fault waveform data uploaded by the feeder terminal in the distribution network is obtained, the sampling point missing detection is performed on the fault waveform data, and the fault waveform data with missing sampling points is interpolated and supplemented to obtain the normalized fault waveform data of the distribution network. Based on the amplitude distribution of the sampling points in the normalized fault waveform data, outlier sampling points are screened out of the normalized fault waveform data to obtain the screened fault waveform data of the distribution network. The fault recording data after screening is rearranged in time sequence to obtain the fault characteristic sequence of the distribution network.

[0007] In a preferred embodiment, the step of synchronizing the fault characteristic sequence with a time axis according to the topological connection relationship of the distribution network to obtain the time-series fault propagation map of the distribution network includes: The sampling points in the fault feature sequence are timestamped to obtain the time-attribute nodes of the fault feature sequence; Based on the topological connection relationship of the power distribution network, the spatial location of the nodes carrying time attributes is mapped to obtain the spatial topological data of the fault feature sequence; Based on the spatial topology data, the nodes carrying time attributes are time-series aligned to obtain the time-aligned data of the power distribution network. The time-aligned data is serialized and encapsulated to obtain the time-series fault propagation map of the distribution network.

[0008] In a preferred embodiment, the step of performing time-series alignment on the nodes carrying time attributes based on the spatial topology data to obtain the time-aligned data of the distribution network includes: Select a reference node from the nodes carrying time attributes, and use the timestamp of the reference node as a global reference timeline; Based on the global reference time axis, the timestamps of the nodes carrying time attributes are compared by difference to obtain the deviation sequence of the timestamps; Based on the deviation sequence, the nodes carrying time attributes are repositioned and mapped to obtain the time alignment data of the power distribution network.

[0009] In a preferred embodiment, the step of performing difference screening on the electrical quantity amplitudes in the time-series fault propagation map to obtain abnormal monitoring points in the time-series fault propagation map, and determining the suspected fault sections of the distribution network based on the topological connection relationship and the abnormal monitoring points, includes: Traverse the nodes in the time-series fault propagation graph and extract the electrical quantity amplitude of the nodes at the fault time; Based on the electrical quantity amplitude, abrupt change point screening is performed on the time-series fault propagation map to obtain abnormal monitoring points in the time-series fault propagation map; Based on the aforementioned topological connections, path tracing is performed on the abnormal monitoring points to obtain suspected fault sections of the power distribution network.

[0010] In a preferred embodiment, the step of calculating the cross-correlation coefficient between the transient current waveforms at the beginning and end of the suspected fault section, and comparing the cross-correlation coefficient with a preset similarity threshold to obtain the fault section to be verified in the distribution network includes: Extract the transient current waveforms at the beginning and end of the suspected fault section; The similarity quantization of the first-end transient current waveform and the last-end transient current waveform is performed to obtain the cross-correlation coefficient between the first-end transient current waveform and the last-end transient current waveform; Based on the comparison results between the cross-correlation coefficient and the preset similarity threshold, the suspected fault sections are initially screened to obtain the fault sections to be verified in the distribution network.

[0011] In a preferred embodiment, the step of performing similarity quantization on the first-end transient current waveform and the last-end transient current waveform to obtain the cross-correlation coefficient between the first-end transient current waveform and the last-end transient current waveform includes: The first-end transient current waveform and the last-end transient current waveform are subjected to mean-removal processing to obtain the first-end mean-removed waveform and the last-end mean-removed waveform in the suspected fault section. The sequence lengths of the mean-reduced waveforms at the beginning and end are statistically analyzed to obtain the total number of waveform data points in the suspected fault section. Based on the mean-reduced waveform at the beginning and the mean-reduced waveform at the end, an integral operation is performed on the suspected fault section to obtain the cross-correlation coefficient between the transient current waveform at the beginning and the transient current waveform at the end.

[0012] In a preferred embodiment, the cross-correlation coefficient is calculated using the following formula: ; in, Let the cross-correlation coefficient be... The transient current waveform at the beginning is described. The terminal transient current waveform is described above. , The mean value of the terminal transient current waveform. The total number of waveform data points.

[0013] In a preferred embodiment, the step of performing topology adaptation verification on the zero-sequence voltage phases on both sides of the fault section to be verified according to the fault phase propagation logic of the distribution network to obtain the fault section of the distribution network includes: Phase extraction is performed on the zero-sequence voltage waveforms uploaded by the feeder terminals on both sides of the fault section to be verified to obtain the zero-sequence voltage phase at the beginning and end of the fault section to be verified, and the actual phase difference of the fault section to be verified is determined. Based on the fault phase propagation logic of the distribution network, the theoretical phase difference interval of the fault section to be verified under the fault state is determined; Based on the theoretical phase difference range, the actual phase difference is checked by interval matching to obtain the degree of deviation of the phase difference in the fault segment to be verified. Based on the degree of deviation, the fault characteristics of the fault section to be verified are determined, and the fault sections of the distribution network are obtained.

[0014] To address the above problems, the present invention also provides a distribution network fault section location system based on feeder terminal data, the system comprising: The waveform recording and normalization module is used to acquire fault waveform recording data uploaded by feeder terminals in the distribution network, and to perform purification and normalization processing on the fault waveform recording data to obtain the fault characteristic sequence of the distribution network. The time-series graph construction module is used to synchronize the fault feature sequence with the time axis according to the topological connection relationship of the distribution network to obtain the time-series fault propagation graph of the distribution network. An abnormal section screening module is used to screen the differences in electrical quantity amplitudes in the time-series fault propagation map, obtain abnormal monitoring points in the time-series fault propagation map, and determine the suspected fault sections of the distribution network based on the topological connection relationship and the abnormal monitoring points. The waveform correlation verification module is used to calculate the cross-correlation coefficient between the transient current waveform at the beginning and the transient current waveform at the end of the suspected fault section, and compare the cross-correlation coefficient with a preset similarity threshold to obtain the fault section to be verified in the distribution network. The phase topology segmentation module is used to perform topology adaptation verification on the zero-sequence voltage phases on both sides of the fault segment to be verified according to the fault phase propagation logic of the distribution network, thereby obtaining the fault segment of the distribution network. Compared with the prior art, the present invention has the following beneficial effects: 1. This invention performs multi-level data preprocessing using complete fault recording data from feeder terminals. Through sampling point interpolation, outlier removal, and time-series rearrangement, the data is purified and regularized to construct a standardized fault feature sequence. Then, combined with the distribution network topology, a time-series fault propagation map is generated synchronously from the time axis of data from multiple terminals. This can completely restore the entire time-series propagation process of the fault along the feeder, accurately capture the electrical quantity change information of each monitoring node, and significantly improve the completeness of fault anomaly point identification and data utilization efficiency. It also eliminates the positioning error caused by asynchronous terminal sampling and data loss at the underlying data level.

[0015] 2. This invention employs a three-tiered progressive judgment logic: amplitude screening for coarse localization, transient current waveform morphology comparison for fine screening, and zero-sequence voltage phase topology adaptation verification for final confirmation. This logic sequentially relies on electrical quantity mutations to lock in suspected sections, on current polarity and distortion differences to narrow down the scope to be verified, and on fault phase propagation logic to complete the section authenticity verification. This layered and progressive approach compresses the fault investigation scope, taking into account both transient waveform details and zero-sequence phase topology constraints. This effectively improves the accuracy of fault section determination, shortens the time required for fault location in distribution networks, and provides reliable and accurate section determination results for rapid fault handling in distribution networks. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating a method for locating fault sections in a distribution network based on feeder terminal data, provided in an embodiment of the present invention. Figure 2 This is a functional block diagram of a distribution network fault section location system based on feeder terminal data, provided in an embodiment of the present invention. The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0017] It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.

[0018] This application provides a method for locating fault sections in a distribution network based on feeder terminal data. The execution entity of this method includes, but is not limited to, at least one of the following electronic devices that can be configured to execute the method provided in this application: a server, a terminal, etc. In other words, the method for locating fault sections in a distribution network based on feeder terminal data can be executed by software or hardware installed on a terminal device or a server device. The server includes, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster. The server can be an independent server or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDNs), and big data and artificial intelligence platforms.

[0019] Reference Figure 1 The diagram shown is a flowchart illustrating a method for locating fault sections in a distribution network based on feeder terminal data, according to an embodiment of the present invention. In this embodiment, the method for locating fault sections in a distribution network based on feeder terminal data includes: D1. Obtain fault waveform data uploaded by feeder terminals in the distribution network, and perform purification and normalization processing on the fault waveform data to obtain the fault characteristic sequence of the distribution network. In this embodiment of the invention, the step of acquiring fault waveform data uploaded by feeder terminals in the distribution network and performing purification and normalization processing on the fault waveform data to obtain the fault feature sequence of the distribution network includes: The fault waveform data uploaded by the feeder terminal in the distribution network is obtained, the sampling point missing detection is performed on the fault waveform data, and the fault waveform data with missing sampling points is interpolated and supplemented to obtain the normalized fault waveform data of the distribution network. Based on the amplitude distribution of the sampling points in the normalized fault waveform data, outlier sampling points are screened out of the normalized fault waveform data to obtain the screened fault waveform data of the distribution network. The fault recording data after screening is rearranged in time sequence to obtain the fault characteristic sequence of the distribution network.

[0020] The system receives complete fault waveform data uploaded in real time from all feeder terminals within the distribution network. It retrieves all sampling records stored in a single feeder terminal for sampling point missing detection, and completely traverses all continuously arranged sampling positions within a single set of fault waveform data. It checks whether there is corresponding data at each preset sampling position, identifies all sampling points without data filling, and performs interpolation and supplementation operations on fault waveform data with missing sampling points determined during the detection process. It fills the data content of the missing sampling positions according to the electrical quantity change trend of the records of adjacent normal sampling points. After completing the data filling of all missing points, it summarizes all data records that have undergone filling processing to generate normalized fault waveform data.

[0021] Completely retrieve the electrical quantity amplitude records corresponding to all sampling points within the normalized fault waveform data. Organize the amplitudes of all sampling points within the same set of normalized fault waveform data to form a complete amplitude distribution set. Use the 3σ criterion to identify outliers. Delineate the concentrated distribution interval using the arithmetic mean of all sampling point amplitudes as the center and three times the standard deviation as the boundary. Sampling points deviating from the overall concentrated distribution interval within the set are identified as outliers. Directly remove the entire sampling record corresponding to all identified outliers, retaining all sampling data that conforms to the overall amplitude distribution pattern. After all removal operations are completed, integrate the remaining valid sampling records to generate the filtered fault waveform data.

[0022] The acquisition time identifier of each sampling record in the fault waveform data after screening is fully read. The order of all sampling records is readjusted according to the fixed sorting rule of the acquisition time identifier from first to last. The positions of the disordered sampling records are uniformly corrected to ensure that all sampling records are arranged continuously in strict accordance with the time sequence of the fault occurrence. After the order of all sampling records is adjusted, all the orderly arranged sampling data is integrated to generate a fault feature sequence.

[0023] This process sets up a continuous and standardized processing procedure to address three types of original data defects in feeder terminal fault recording data: sampling gaps, abnormal interference points, and timing disorder. By filling in data points, it eliminates information gaps caused by data gaps; by removing abnormal sampling points, it removes internal interference information from the data; and by reordering the timing sequence, it unifies the data layout logic. Finally, it outputs a standardized fault feature sequence, which can provide complete, clean, and time-ordered basic data for subsequent timing synchronization and fault map construction. It avoids fault location deviations caused by defects in the original data from the source, and stably ensures the reliability of the data foundation for subsequent distribution network fault section analysis.

[0024] D2. Based on the topological connection relationship of the distribution network, the fault characteristic sequence is synchronized on the time axis to obtain the time-series fault propagation map of the distribution network; In this embodiment of the invention, the step of synchronizing the fault characteristic sequence with a time axis according to the topological connection relationship of the distribution network to obtain the time-series fault propagation map of the distribution network includes: The sampling points in the fault feature sequence are timestamped to obtain the time-attribute nodes of the fault feature sequence; Based on the topological connection relationship of the power distribution network, the spatial location of the nodes carrying time attributes is mapped to obtain the spatial topological data of the fault feature sequence; Based on the spatial topology data, the nodes carrying time attributes are time-series aligned to obtain the time-aligned data of the power distribution network. The time-aligned data is serialized and encapsulated to obtain the time-series fault propagation map of the distribution network.

[0025] The step of performing time-series alignment on the nodes carrying time attributes based on the spatial topology data to obtain the time-aligned data of the distribution network includes: Select a reference node from the nodes carrying time attributes, and use the timestamp of the reference node as a global reference timeline; Based on the global reference time axis, the timestamps of the nodes carrying time attributes are compared by difference to obtain the deviation sequence of the timestamps; Based on the deviation sequence, the nodes carrying time attributes are repositioned and mapped to obtain the time alignment data of the power distribution network.

[0026] Each sampling record in the fault feature sequence is read completely, and each independent sampling point in the fault feature sequence is separated. The real collection time information recorded when the sampling point is generated is retrieved, and each independent sampling point is configured with a unique timestamp. This ensures that each sampling point is accompanied by a time identification mark accurate to the moment of collection. All sampling points that have completed the timestamp binding process are integrated and stored according to the original arrangement rules of the fault feature sequence. After all sampling points have completed the timestamp binding operation, all sampling units with time information are summarized to generate a time attribute node of the fault feature sequence.

[0027] The system retrieves pre-entered and complete data on the interconnection logic of all feeders, sectionalizing switches, feeder terminals, and branch lines within the distribution network. It then iterates through the physical acquisition device (feeder terminal) corresponding to each time-attribute node, searching for the fixed spatial location information of that feeder terminal within the overall distribution network topology. A unique binding relationship is established between each time-attribute node and its corresponding spatial location in the topology, ensuring that each time-attribute node accurately corresponds to its actual installation location on the physical power grid line. After all time-attribute nodes have completed spatial location matching and binding, the system uniformly collects and stores the corresponding association information between all nodes and their topological spatial locations, generating spatial topology data of fault characteristic sequences.

[0028] The spatial topology data generated in the previous step is fully retrieved, and all time-related node information stored within the spatial topology data is extracted. Following the established execution logic, the entire time alignment process is carried out in sequence, including selecting the reference node, constructing the global reference time axis, comparing the timestamp differences of multiple nodes, and mapping all time-related nodes back to their positions. No steps are omitted in the entire time alignment process. All nodes complete the time scale calibration work in a unified manner. All time-related nodes that have undergone unified time correction and eliminated the clock offset are fully integrated and collected to generate the time alignment data of the distribution network.

[0029] All standardized node information contained in the time-aligned data obtained after time-series calibration is extracted. Based on the physical spatial direction of the fault electrical signal propagating along the feeder line after a distribution network fault occurs, and combined with the time sequence of unified calibration completion, all standardized node information is subjected to layered and orderly packaging and encapsulation operations. The encapsulation process completely retains the unique time identifier bound to each node, the mapped topological spatial location, and the complete acquisition information of various electrical quantities obtained in the original acquisition, without any information loss or deletion. After all layered encapsulation operations are completed, all layered data packets are integrated into a complete structured data set to generate a time-series fault propagation map of the distribution network.

[0030] The process involves traversing all time-attribute nodes recorded in the spatial topology data, verifying the completeness of the sampled data and the start time of fault signal acquisition for each node, and selecting the single node with the earliest fault signal acquisition time, no missing sample records, and complete and unmissing timestamps as the baseline node. If multiple nodes meet the criteria, the node at the substation outlet closest to the power supply side of the distribution network is selected as the baseline node. If all nodes have varying degrees of data loss, the node with the lowest percentage of missing sample points is selected as the baseline node. The unique timestamps bound to this baseline node are extracted separately, and the time scale corresponding to these timestamps is used as a unified reference standard to build a global reference timeline covering all nodes.

[0031] Using the standard time scale of the global reference time axis generated by the benchmark node as a unified reference, the timestamps bound to each node with time attributes in the spatial topology data (excluding the benchmark node) are retrieved sequentially. The time scale corresponding to the timestamp of a single node is compared with the standard scale of the global reference time axis item by item to obtain the overall offset difference of the node relative to the standard time axis. The independent offset difference corresponding to each node with time attributes on all branches of the distribution network is recorded sequentially. Then, according to the arrangement order of the distribution network topology lines from the power source end to the line end, the offset differences corresponding to all nodes are sequentially arranged and integrated to generate a complete and continuous offset sequence of timestamps.

[0032] The deviation sequence of the fully generated timestamps is retrieved, and the unique time offset corresponding to each node carrying time attributes stored within the deviation sequence is read line by line. Using the standard scale of the global reference time axis as the adjustment target, the current timestamp mark of the corresponding node is read, and the time scale contained in the timestamp mark of the node is shifted and corrected in reverse according to the offset of the node. This ensures that the time scale of the adjusted node is completely aligned with the unified standard of the global reference time axis, and there is no situation where any node retains the original offset time scale. After all nodes carrying time attributes in the distribution network have completed the unified correction of their time scales, the complete information records of all calibrated nodes with completely unified time scales are integrated and collected to generate the time alignment data of the distribution network.

[0033] This complete process can achieve unified spatiotemporal processing of heterogeneous data collected by multiple feeder terminals based on the actual physical topology of the distribution network. By configuring dedicated timestamps for all sampling points, it provides precise time identification for each segment of fault data. Through spatial location mapping, it achieves one-to-one binding between abstract data nodes and physical power grid line locations. Based on the selected benchmark nodes, it builds a globally unified reference time axis to completely eliminate timing deviations caused by inconsistent hardware acquisition clocks of different feeder terminals. It uses the deviation sequence to complete the unified repositioning and calibration of all nodes. Finally, through structured serialization and encapsulation, it generates a time-series fault propagation map that can synchronously display the time evolution and spatial propagation path of the fault. The original fault acquisition electrical information is completely preserved without deletion, solving the problem of time sequence misalignment and spatial information disconnection of multi-terminal data from the root. It provides a standardized analysis carrier with completely unified spatiotemporal dimensions and complete information without omissions for subsequent screening of electrical quantity amplitude differences in time-series fault propagation maps, greatly improving the reliability of subsequent anomaly monitoring point identification and reducing the omission and misjudgment of fault sections caused by spatiotemporal data mismatch.

[0034] D3. Perform difference screening on the electrical quantity amplitudes in the time-series fault propagation map to obtain abnormal monitoring points in the time-series fault propagation map, and determine the suspected fault sections of the distribution network based on the topological connection relationship and the abnormal monitoring points. In this embodiment of the invention, the step of performing difference screening on the electrical quantity amplitudes in the time-series fault propagation map to obtain abnormal monitoring points in the time-series fault propagation map, and determining the suspected fault sections of the distribution network based on the topological connection relationship and the abnormal monitoring points, includes: Traverse the nodes in the time-series fault propagation graph and extract the electrical quantity amplitude of the nodes at the fault time; Based on the electrical quantity amplitude, abrupt change point screening is performed on the time-series fault propagation map to obtain abnormal monitoring points in the time-series fault propagation map; Based on the aforementioned topological connections, path tracing is performed on the abnormal monitoring points to obtain suspected fault sections of the power distribution network.

[0035] The entire standardized nodes stored within the time-series fault propagation graph are traversed. Following the arrangement order formed by the layered encapsulation of the graph, the various types of collected information stored within each independent node are retrieved in sequence. The collected record entries corresponding to the moment of the fault occurrence in each node are accurately located. The complete electrical quantity amplitude data of each type of electrical quantity corresponding to the fault occurrence in each entry are extracted separately. All electrical quantity amplitudes corresponding to the fault moment of a single node are collected and stored separately. After completing the traversal operation of all nodes in the time-series fault propagation graph, the electrical quantity amplitude information corresponding to the fault moment of all nodes is summarized and collected.

[0036] The electrical quantity amplitude data at the time of the fault, which has been collected from all nodes, is retrieved. The range of electrical quantity amplitude changes that are stably maintained at each monitoring point under normal operation of the distribution network is used as a reference benchmark. The complete electrical quantity amplitude corresponding to each node in the time-series fault propagation map is compared and verified one by one. The node positions where the electrical quantity amplitude deviates from the normal stable change range at the time of the fault are identified. All the map nodes corresponding to the identified amplitude mutations are uniformly marked and classified. After the entire amplitude comparison and verification process is completed, all the marked mutation corresponding nodes are integrated to generate the abnormal monitoring points in the time-series fault propagation map.

[0037] The system retrieves the topological connection data of all line segments, switching equipment, and feeder terminals of the distribution network that are pre-stored in the system. It then locates the physical line points corresponding to all identified abnormal monitoring points in the topological structure. Using each abnormal monitoring point as the starting point for path tracing, it continuously extends forward along the bidirectional line route of the distribution network feeder to search for adjacent topological nodes. It continuously collects all interconnected line segment units on the tracing path. The entire set of line segments containing abnormal monitoring points within the same continuous connected path is uniformly divided and included. After the path tracing work of all connected lines corresponding to all abnormal monitoring points is completed, all divided continuous line segment sets are integrated to generate the suspected fault section of the distribution network.

[0038] This process relies on a time-series fault propagation map based on a unified spatiotemporal benchmark to complete a layered and progressive coarse fault location process. It fully extracts the electrical quantity amplitude values ​​of all nodes in the map at the time of the fault to ensure that no data retrieval is missed. By relying on the normal electrical amplitude range to complete the abrupt change screening, it can accurately locate all abnormal monitoring points with electrical disturbances. Combined with the complete distribution network topology connection relationship, it carries out bidirectional path tracing to achieve complete collection of fault-related line segments. It can quickly delineate the range of all lines suspected of having faults without complex multi-level calculations, narrowing the analysis scope of subsequent waveform comparison and verification, reducing the total amount of data processing in the overall fault location process, and shortening the overall time required for the initial location of distribution network faults.

[0039] D4. Calculate the cross-correlation coefficient between the transient current waveform at the beginning and the transient current waveform at the end of the suspected fault section, and compare the cross-correlation coefficient with a preset similarity threshold to obtain the fault section to be verified in the distribution network. In this embodiment of the invention, the step of calculating the cross-correlation coefficient between the transient current waveform at the beginning and the transient current waveform at the end of the suspected fault section, and comparing the cross-correlation coefficient with a preset similarity threshold to obtain the fault section to be verified in the distribution network, includes: Extract the transient current waveforms at the beginning and end of the suspected fault section; The similarity quantization of the first-end transient current waveform and the last-end transient current waveform is performed to obtain the cross-correlation coefficient between the first-end transient current waveform and the last-end transient current waveform; Based on the comparison results between the cross-correlation coefficient and the preset similarity threshold, the suspected fault sections are initially screened to obtain the fault sections to be verified in the distribution network.

[0040] The step of performing similarity quantization on the first-end transient current waveform and the last-end transient current waveform to obtain the cross-correlation coefficient between the first-end transient current waveform and the last-end transient current waveform includes: The first-end transient current waveform and the last-end transient current waveform are subjected to mean-removal processing to obtain the first-end mean-removed waveform and the last-end mean-removed waveform in the suspected fault section. The sequence lengths of the mean-reduced waveforms at the beginning and end are statistically analyzed to obtain the total number of waveform data points in the suspected fault section. Based on the mean-reduced waveform at the beginning and the mean-reduced waveform at the end, an integral operation is performed on the suspected fault section to obtain the cross-correlation coefficient between the transient current waveform at the beginning and the transient current waveform at the end.

[0041] The formula for calculating the cross-correlation coefficient is as follows: ; in, Let the cross-correlation coefficient be... The transient current waveform at the beginning is described. The terminal transient current waveform is described above. , The mean value of the terminal transient current waveform. The total number of waveform data points.

[0042] Retrieve the complete set of time-series fault propagation map storage information corresponding to the single suspected fault section after the previous path tracing processing. Based on the complete topology connection relationship of the distribution network, locate the two feeder terminal acquisition devices installed at both ends of the line space of the suspected fault section. Locate the feeder terminal storage unit on the side of the line closer to the power input line and read all the current time-series sampling records of the entire process of transient disturbance at the moment the fault occurs. Integrate the continuous time-series sampling records to form a complete and complete first-end transient current waveform. Then locate the feeder terminal storage unit on the side of the line farther from the power input line and read all the current time-series sampling records of the entire process of fault disturbance under the synchronous time scale of the unit. Integrate the continuous time-series sampling records to form a complete and complete last-end transient current waveform. The extraction process completely retains the original current acquisition values ​​corresponding to each time-series sampling point of the two waveforms, without deleting or truncating the acquisition data in any waveform segment, ensuring that the two waveforms cover the same fault duration range.

[0043] The process involves fully reading the original current values ​​corresponding to all time-series sampling points within the extracted first-end transient current waveform, summing all sampling point values, and then dividing the sum by the number of sampling points in the waveform to obtain the overall average value of the first-end transient current waveform. This process is repeated for each time-series sampling point within the first-end transient current waveform. The original current value at that point is subtracted from the overall average value of the waveform segment, and the new value obtained after the subtraction replaces the original stored value at that point. After replacing all point values, a new sampling sequence with an ordered time sequence is integrated to obtain the first-end de-averaged waveform in the suspected fault section. Simultaneously, the same processing logic is used to read all sampling values ​​of the last-end transient current waveform to calculate the overall average. The replacement operation of subtracting the original value from the average is performed point by point. The processed ordered sampling sequence is then integrated to obtain the last-end de-averaged waveform in the suspected fault section. The de-averaging process completely removes the constant DC offset component inherent in both waveforms, preventing the constant offset component from interfering with the quantitative judgment result of the morphological similarity between the two waveforms.

[0044] The number of all independent sampling points arranged continuously in chronological order within the mean-reduced waveform at the beginning is counted sequentially. Simultaneously, the number of all independent sampling points arranged according to the same time scale within the mean-reduced waveform at the end is counted. The number of sampling points in the two waveforms is verified to be completely consistent. The total number of unified sampling points is recorded and saved as the total number of waveform data points in the suspected fault section. This total number of unified data points can provide a unified calculation scale for subsequent integration operations, ensuring that the calculation intervals of the two waveforms are completely matched and that there will be no distortion of quantization results caused by inconsistent time lengths.

[0045] The sampling points corresponding to the same time scale within the first and last time-series mean-reduced waveforms are aligned sequentially. The mean-reduced waveform values ​​at the same time position are extracted and multiplied. The results of multiplying the two sets of waveform values ​​corresponding to all time-series points are obtained sequentially. Along the fault duration interval completely covered by the two waveforms, the results of multiplying all time-series points are continuously accumulated to complete the overall integration operation. The start and end boundaries of the integration operation are constrained by a uniform total number of waveform data points throughout the process. After the integration operation is completed, a unique standardized value is output. This standardized value is directly defined as the cross-correlation coefficient between the first and last transient current waveforms. This coefficient can objectively quantify the degree of overlap of the overall shape of the two transient current waveforms, and the value directly corresponds to the degree of waveform similarity.

[0046] This is the cross-correlation coefficient between the initial transient current waveform and the final transient current waveform output from this process. This coefficient is calculated based on standardized data from waveforms collected at both ends of the same suspected fault section. The transient current waveform at the beginning of the suspected fault section was obtained from previous extractions. This waveform was formed by integrating fault current sampling records collected from the feeder terminal near the power supply side of the section throughout the entire time period. The transient current waveform at the end of the suspected fault section was obtained from previous extractions. This waveform was formed by integrating fault current sampling records collected during the synchronous duration of the feeder terminal far from the power supply side of the section. The mean value of the first-end waveform is obtained by taking the overall average value of all sampled values ​​of the transient current waveform at the first end. The mean value of the terminal waveform is obtained by taking the overall average value of all sampled values ​​of the terminal transient current waveform. The total number of waveform data points whose values ​​are completely consistent after synchronous statistical analysis of two waveform segments is obtained. This number is recorded and stored after the number of sampling points of the first and last waveforms are counted one by one and the number of sampling points is verified and consistent.

[0047] The entire calculation process first subtracts the corresponding mean from the transient current waveforms at the beginning and end of the line to perform mean-reduction processing, resulting in mean-reduction waveforms at the beginning and end of the line. Then, the sampled values ​​of the two mean-reduction waveforms at the same time point are multiplied pairwise, and the results of all multiplications are summed to obtain the numerator. The sampled values ​​of the mean-reduction waveform at the beginning are squared separately, summed, and then the sum is squared. The same process is repeated for the mean-reduction waveform at the end, and the sum is squared. The results of the two square-root operations are multiplied to obtain the denominator. The numerator is divided by the denominator to obtain the cross-correlation coefficient. This coefficient objectively quantifies the degree of overlap between the overall shape of the transient current waveforms at the beginning and end of the line. The total number of waveform data points is used throughout the calculation to ensure complete dimensional consistency. This coefficient is specifically used to quantify the similarity of the current waveforms at the beginning and end of a suspected fault section, providing a unique quantitative basis for subsequent comparison with a preset similarity threshold to screen for potential fault sections.

[0048] When the overlap between two transient current waveforms increases, the changes in the values ​​of the two sets of waveforms at the same time point remain highly synchronized. The cumulative value of the numerator will continue to increase, while the change in the value of the denominator after multiplying the square roots of the sum of the two squares will be much smaller than the change in the numerator. The calculated cross-correlation coefficient will continue to approach the upper limit of the value. When the overlap between two transient current waveforms decreases, the changes in the values ​​of the two sets of waveforms at the same time point will show a significant deviation. The cumulative value of the numerator will continue to decrease, while the change in the denominator will be relatively stable. The calculated cross-correlation coefficient will continue to approach the lower limit of the value. The closer the cross-correlation coefficient is to the upper limit, the higher the similarity of the transient current waveforms at the beginning and end of the suspected fault section. The closer the cross-correlation coefficient is to the lower limit, the lower the similarity of the transient current waveforms at the beginning and end of the suspected fault section.

[0049] The system retrieves pre-set, fixed similarity threshold criteria. The preset similarity threshold is set to 0.6, determined based on the statistical pattern observed in the field that the similarity of the first and last current waveforms in healthy sections under metallic grounding faults in distribution networks is generally higher than 0.8, while the similarity is generally lower than 0.5 in faulty sections. The cross-correlation coefficient obtained through integration is compared item by item with the preset similarity threshold. When the cross-correlation coefficient is lower than the preset similarity threshold, it indicates a significant difference in the overall shape of the transient current waveforms at the beginning and end of the suspected faulty section, indicating that the fault disturbance point is located in the current faulty section. Within the suspected fault section, line segments that meet the retention criteria are marked and retained for the next stage of phase topology verification. When the cross-correlation coefficient is higher than or equal to the preset similarity threshold, it means that the transient current waveforms at both ends of the suspected fault section are highly overlapping, indicating that the fault disturbance point does not fall within the current line segment interval. Such line segments are directly removed and no longer participate in the subsequent in-depth verification process. After completing the cross-correlation coefficient and threshold comparison and classification of all suspected fault sections, all marked and retained line segments are uniformly collected to generate the distribution network's unverified fault sections.

[0050] This process involves standardized, full-process similarity quantification of transient current waveforms at both ends of suspected faulty sections. By completely extracting the original current waveforms covering the same fault duration, the analysis data is ensured to be complete and without omissions. Unified mean-reduction processing eliminates judgment interference caused by DC offset of the waveform. Unified constraints on the total number of waveform data points ensure the objective and consistent calculation results of cross-correlation coefficients. Automated section screening without human intervention is achieved using fixed thresholds. This can accurately screen out line sections without faults, significantly reducing the number of sections that need to be processed in subsequent zero-sequence voltage phase topology adaptation verification. This effectively reduces the data processing scale and computational consumption of the back-end verification process. At the same time, waveform morphology quantification indicators distinguish between faulty and non-faulty areas, improving the discrimination capability of the pre-screening stage of faulty sections. From the process level, this reduces the overall time consumption of faulty section location and improves the response speed of distribution network fault handling.

[0051] D5. Based on the fault phase propagation logic of the distribution network, perform topology adaptation verification on the zero-sequence voltage phases on both sides of the fault section to be verified to obtain the fault section of the distribution network.

[0052] In this embodiment of the invention, the step of performing topology adaptation verification on the zero-sequence voltage phases on both sides of the fault section to be verified according to the fault phase propagation logic of the distribution network to obtain the fault section of the distribution network includes: Phase extraction is performed on the zero-sequence voltage waveforms uploaded by the feeder terminals on both sides of the fault section to be verified to obtain the zero-sequence voltage phase at the beginning and end of the fault section to be verified, and the actual phase difference of the fault section to be verified is determined. Based on the fault phase propagation logic of the distribution network, the theoretical phase difference interval of the fault section to be verified under the fault state is determined; Based on the theoretical phase difference range, the actual phase difference is checked by interval matching to obtain the degree of deviation of the phase difference in the fault segment to be verified. Based on the degree of deviation, the fault characteristics of the fault section to be verified are determined, and the fault sections of the distribution network are obtained.

[0053] The system locates the feeder terminal equipment at both ends of the single fault section to be verified along the line extension direction. It reads all zero-sequence voltage waveform acquisition records uploaded and stored by the two feeder terminals during the complete fault period, locks the time interval corresponding to the fault disturbance within the waveform, and analyzes the phase change pattern corresponding to the voltage fluctuation within the waveform segment by segment according to the time arrangement. It separates the voltage phase value collected by the terminal on the side of the line closer to the power source as the first zero-sequence voltage phase of the fault section to be verified, and separates the voltage phase value collected by the terminal on the side of the line farther from the power source as the last zero-sequence voltage phase of the fault section to be verified. It performs a difference operation on the two sets of extracted phase values ​​and calculates the fixed difference between the two sets of phase values ​​as the actual phase difference of the fault section to be verified. The entire time sequence sampling data of the zero-sequence voltage waveform is completely preserved without truncating the effective recording interval of the waveform, ensuring that the phase extraction is based on the complete fault disturbance waveform and avoiding the distortion of phase value extraction caused by the absence of local waveforms.

[0054] The system retrieves pre-entered fault phase propagation logic data covering all types of distribution network line structures and grounding fault types. For ungrounded neutral distribution networks, the phase difference between the fundamental zero-sequence voltage upstream and downstream of the fault point under a single-phase grounding fault is close to 180°. For neutral distribution networks grounded through an arc suppression coil, the phase difference between the zero-sequence voltages on both sides of the fault point ranges from 150° to 210°. In healthy sections, the phase difference between the zero-sequence voltages on both sides does not exceed 15°. This data fully records the inherent variation law of phase shift and lag when the fault signal is transmitted in the line medium. Combined with the inherent transmission characteristics corresponding to the line length, line erection type, and wiring method of the equipment at both ends of the current fault section to be verified, the system deduces the stable value range of the zero-sequence voltage phases at both ends when the fault occurs inside this line section according to the fault phase propagation logic. This stable value range is fully defined and recorded to form the theoretical phase difference interval of the fault section to be verified under the fault state. The theoretical phase difference interval is generated in complete accordance with the physical transmission attributes of this line section itself, and there is no judgment deviation caused by a uniform and fixed range.

[0055] The actual phase difference of the faulty section to be verified obtained from the previous step is retrieved in its entirety. The value corresponding to the actual phase difference is placed into the theoretical phase difference interval that has been derived to complete the interval matching verification. The distance relationship between the actual phase difference value and the upper and lower boundaries of the theoretical phase difference interval is compared. The absolute value of the difference between the actual phase difference and the center value of the theoretical interval is used as the quantitative indicator of the degree of deviation. Three different situations are distinguished: the actual phase difference falls within the interval, exceeds the upper limit of the interval, and is lower than the lower limit of the interval. The degree of deviation of the phase difference in the current faulty section to be verified is completely quantified according to the distance between the value and the interval boundary. The quantified deviation degree identifier is completely retained for subsequent fault qualitative judgment. The interval matching verification does not simplify the phase difference value comparison steps throughout the process, and fully distinguishes various deviation scenarios to ensure that the deviation degree judgment is complete and clear.

[0056] The phase difference deviation indicator of the fault section to be verified is read from the output of the interval matching verification step. The fault characterization operation is carried out by comparing it with the pre-set qualitative judgment rules in the fault phase propagation logic of the distribution network. When the deviation is less than 30°, it is determined to meet the phase characteristics of the fault section. When the deviation of the phase difference falls within the no-deviation standard range, it means that the phase change law of the zero-sequence voltage at both ends of the current fault section to be verified is completely consistent with the transmission characteristics of the fault occurring inside this line segment. The line segment is directly marked as the line section with a real fault. When the deviation of the phase difference exceeds the allowable standard range, it means that the source of the fault disturbance is not inside the current fault section to be verified. The line segment is directly discarded and not included in the final fault location result. The deviation comparison and qualitative classification operation of all fault sections to be verified are completed one by one. All line segments that are determined to have real faults are uniformly collected to generate the fault sections of the distribution network.

[0057] This invention establishes a hierarchical zero-sequence voltage phase verification system based on the inherent fault phase propagation law of distribution network lines. By completely extracting the zero-sequence voltage waveforms at both ends of the section and accurately analyzing the phase values, the true phase difference value is obtained. Combined with the physical attributes of the section itself, a unique theoretical phase difference interval is generated to replace the unified fixed judgment standard. The degree of phase difference deviation is quantified by interval matching to complete the fine verification. The degree of deviation is used to complete the automated fault characterization. The multi-level phase verification logic can eliminate the misjudged sections left in the initial screening of transient current waveform similarity. Finally, it outputs a precise and unique true fault line interval, comprehensively improving the reliability of the final judgment result of the entire fault location process, reducing the probability of misjudgment in the location of distribution network fault sections, and providing accurate and reliable line location basis for operation and maintenance personnel to carry out line maintenance.

[0058] like Figure 2 The diagram shown is a functional block diagram of a distribution network fault section location system based on feeder terminal data provided in an embodiment of the present invention.

[0059] The distribution network fault section location system 100 based on feeder terminal data described in this invention can be installed in electronic equipment. Depending on the functions implemented, the distribution network fault section location system 100 based on feeder terminal data may include a waveform recording and normalization module 101, a time series graph construction module 102, an abnormal section initial screening module 103, a waveform correlation verification module 104, and a phase topology segmentation module 105. The modules described in this invention can also be called units, referring to a series of computer program segments that can be executed by the processor of an electronic device and perform a fixed function, stored in the memory of the electronic device.

[0060] In this embodiment, the functions of each module / unit are as follows: The waveform recording and normalization module 101 is used to acquire fault waveform recording data uploaded by the feeder terminal in the distribution network, and to perform purification and normalization processing on the fault waveform recording data to obtain the fault feature sequence of the distribution network. The time-series map construction module 102 is used to synchronize the fault feature sequence with the time axis according to the topological connection relationship of the distribution network to obtain the time-series fault propagation map of the distribution network. The abnormal section screening module 103 is used to screen the differences in the electrical quantity amplitudes in the time-series fault propagation map, obtain the abnormal monitoring points in the time-series fault propagation map, and determine the suspected fault sections of the distribution network based on the topological connection relationship and the abnormal monitoring points. The waveform correlation verification module 104 is used to calculate the cross-correlation coefficient between the transient current waveform at the beginning and the transient current waveform at the end of the suspected fault section, and compare the cross-correlation coefficient with a preset similarity threshold to obtain the fault section to be verified in the distribution network. The phase topology segmentation module 105 is used to perform topology adaptation verification on the zero-sequence voltage phases on both sides of the fault section to be verified according to the fault phase propagation logic of the distribution network, so as to obtain the fault section of the distribution network.

[0061] In the several embodiments provided by this invention, it should be understood that the disclosed methods and systems can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and other division methods may be used in actual implementation.

[0062] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0063] Furthermore, the functional modules in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in the form of hardware plus software functional modules.

[0064] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0065] This application embodiment can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0066] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for locating fault sections in a distribution network based on feeder terminal data, characterized in that, The method includes: D1. Obtain fault waveform data uploaded by feeder terminals in the distribution network, and perform purification and normalization processing on the fault waveform data to obtain the fault characteristic sequence of the distribution network. D2. Based on the topological connection relationship of the distribution network, the fault characteristic sequence is synchronized on the time axis to obtain the time-series fault propagation map of the distribution network; D3. Perform difference screening on the electrical quantity amplitudes in the time-series fault propagation map to obtain abnormal monitoring points in the time-series fault propagation map, and determine the suspected fault sections of the distribution network based on the topological connection relationship and the abnormal monitoring points. D4. Calculate the cross-correlation coefficient between the transient current waveform at the beginning and the transient current waveform at the end of the suspected fault section, and compare the cross-correlation coefficient with a preset similarity threshold to obtain the fault section to be verified in the distribution network. D5. Based on the fault phase propagation logic of the distribution network, perform topology adaptation verification on the zero-sequence voltage phases on both sides of the fault section to be verified to obtain the fault section of the distribution network.

2. The method for locating fault sections in a distribution network based on feeder terminal data as described in claim 1, characterized in that, The process of acquiring fault waveform data uploaded by feeder terminals in the distribution network and performing purification and normalization processing on the fault waveform data to obtain the fault characteristic sequence of the distribution network includes: The fault waveform data uploaded by the feeder terminal in the distribution network is obtained, the sampling point missing detection is performed on the fault waveform data, and the fault waveform data with missing sampling points is interpolated and supplemented to obtain the normalized fault waveform data of the distribution network. Based on the amplitude distribution of the sampling points in the normalized fault waveform data, outlier sampling points are screened out of the normalized fault waveform data to obtain the screened fault waveform data of the distribution network. The fault recording data after screening is rearranged in time sequence to obtain the fault characteristic sequence of the distribution network.

3. The method for locating fault sections in a distribution network based on feeder terminal data as described in claim 1, characterized in that, The step of synchronizing the fault characteristic sequence with a time axis based on the topological connection relationship of the distribution network to obtain the time-series fault propagation map of the distribution network includes: The sampling points in the fault feature sequence are timestamped to obtain the time-attribute nodes of the fault feature sequence; Based on the topological connection relationship of the power distribution network, the spatial location of the nodes carrying time attributes is mapped to obtain the spatial topological data of the fault feature sequence; Based on the spatial topology data, the nodes carrying time attributes are time-series aligned to obtain the time-aligned data of the power distribution network. The time-aligned data is serialized and encapsulated to obtain the time-series fault propagation map of the distribution network.

4. The method for locating fault sections in a distribution network based on feeder terminal data as described in claim 3, characterized in that, The step of performing time-series alignment on the nodes carrying time attributes based on the spatial topology data to obtain the time-aligned data of the distribution network includes: Select a reference node from the nodes carrying time attributes, and use the timestamp of the reference node as a global reference timeline; Based on the global reference time axis, the timestamps of the nodes carrying time attributes are compared by difference to obtain the deviation sequence of the timestamps; Based on the deviation sequence, the nodes carrying time attributes are repositioned and mapped to obtain the time alignment data of the power distribution network.

5. The method for locating fault sections in a distribution network based on feeder terminal data as described in claim 1, characterized in that, The step of screening for differences in the electrical quantity amplitudes in the time-series fault propagation map to obtain abnormal monitoring points in the time-series fault propagation map, and determining the suspected fault sections of the distribution network based on the topological connection relationship and the abnormal monitoring points, includes: Traverse the nodes in the time-series fault propagation graph and extract the electrical quantity amplitude of the nodes at the fault time; Based on the electrical quantity amplitude, abrupt change point screening is performed on the time-series fault propagation map to obtain abnormal monitoring points in the time-series fault propagation map; Based on the aforementioned topological connections, path tracing is performed on the abnormal monitoring points to obtain suspected fault sections of the power distribution network.

6. The method for locating fault sections in a distribution network based on feeder terminal data as described in claim 1, characterized in that, The calculation of the cross-correlation coefficient between the transient current waveforms at the beginning and end of the suspected fault section, and the comparison of the cross-correlation coefficient with a preset similarity threshold, to obtain the fault section to be verified in the distribution network, includes: Extract the transient current waveforms at the beginning and end of the suspected fault section; The similarity quantization of the first-end transient current waveform and the last-end transient current waveform is performed to obtain the cross-correlation coefficient between the first-end transient current waveform and the last-end transient current waveform; Based on the comparison results between the cross-correlation coefficient and the preset similarity threshold, the suspected fault sections are initially screened to obtain the fault sections to be verified in the distribution network.

7. The method for locating fault sections in a distribution network based on feeder terminal data as described in claim 6, characterized in that, The step of performing similarity quantization on the first-end transient current waveform and the last-end transient current waveform to obtain the cross-correlation coefficient between the first-end transient current waveform and the last-end transient current waveform includes: The first-end transient current waveform and the last-end transient current waveform are subjected to mean-removal processing to obtain the first-end mean-removed waveform and the last-end mean-removed waveform in the suspected fault section. The sequence lengths of the mean-reduced waveforms at the beginning and end are statistically analyzed to obtain the total number of waveform data points in the suspected fault section. Based on the mean-reduced waveform at the beginning and the mean-reduced waveform at the end, an integral operation is performed on the suspected fault section to obtain the cross-correlation coefficient between the transient current waveform at the beginning and the transient current waveform at the end.

8. The method for locating fault sections in a distribution network based on feeder terminal data as described in claim 7, characterized in that, The formula for calculating the cross-correlation coefficient is as follows: ； in, Let the cross-correlation coefficient be... The transient current waveform at the beginning is described. The terminal transient current waveform is described above. , The mean value of the terminal transient current waveform. The total number of waveform data points.

9. The method for locating fault sections in a distribution network based on feeder terminal data as described in claim 1, characterized in that, The step of performing topology adaptation verification on the zero-sequence voltage phases on both sides of the fault section to be verified according to the fault phase propagation logic of the distribution network to obtain the fault section of the distribution network includes: Phase extraction is performed on the zero-sequence voltage waveforms uploaded by the feeder terminals on both sides of the fault section to be verified to obtain the zero-sequence voltage phase at the beginning and end of the fault section to be verified, and the actual phase difference of the fault section to be verified is determined. Based on the fault phase propagation logic of the distribution network, the theoretical phase difference interval of the fault section to be verified under the fault state is determined; Based on the theoretical phase difference range, the actual phase difference is checked by interval matching to obtain the degree of deviation of the phase difference in the fault segment to be verified. Based on the degree of deviation, the fault characteristics of the fault section to be verified are determined, and the fault sections of the distribution network are obtained.

10. A distribution network fault section location system based on feeder terminal data, characterized in that, The system for implementing the distribution network fault section location method based on feeder terminal data as described in claim 1 includes: The waveform recording and normalization module is used to acquire fault waveform recording data uploaded by feeder terminals in the distribution network, and to perform purification and normalization processing on the fault waveform recording data to obtain the fault characteristic sequence of the distribution network. The time-series graph construction module is used to synchronize the fault feature sequence with the time axis according to the topological connection relationship of the distribution network to obtain the time-series fault propagation graph of the distribution network. An abnormal section screening module is used to screen the differences in electrical quantity amplitudes in the time-series fault propagation map, obtain abnormal monitoring points in the time-series fault propagation map, and determine the suspected fault sections of the distribution network based on the topological connection relationship and the abnormal monitoring points. The waveform correlation verification module is used to calculate the cross-correlation coefficient between the transient current waveform at the beginning and the transient current waveform at the end of the suspected fault section, and compare the cross-correlation coefficient with a preset similarity threshold to obtain the fault section to be verified in the distribution network. The phase topology segmentation module is used to perform topology adaptation verification on the zero-sequence voltage phases on both sides of the fault segment to be verified according to the fault phase propagation logic of the distribution network, so as to obtain the fault segment of the distribution network.

Images