An asynchronous multi-measurement-point harmonic responsibility section identification method and system

Abstract

The application discloses the technical field of harmonic responsibility identification and relates to an asynchronous multi-measuring-point harmonic responsibility section identification method and system. The method comprises the following steps: reconstructing the time stamp of each local voltage waveform file according to the matching relationship between the local voltage waveform file and the main station update index record, and constructing multiple groups of synchronous multi-point samples through the nearest neighbor time matching criterion; performing discrete Fourier transform on the synchronous multi-point samples, extracting the fundamental amplitude and the multiple harmonic amplitudes, respectively, calculating the total harmonic distortion rate, calculating the change amount of the multiple harmonic amplitudes, and constructing a unified feature vector; according to the unified feature vector and the branch or section to which each measuring point belongs, solving the optimal fusion weight based on the constraint optimization adaptive objective function, outputting the instantaneous responsibility label and the abnormal mark of the synchronous multi-point sample; and merging the index continuous and abnormal synchronous multi-point samples into an abnormal event, and outputting the abnormal event responsibility label and the responsibility confidence.

Description

Technical Field

[0001] This invention relates to the field of harmonic responsibility identification technology, and in particular to an asynchronous multi-measurement point harmonic responsibility segment identification method and system. Background Technology

[0002] With the large-scale integration of new energy sources, power electronic equipment, and flexible loads, harmonics in modern power systems exhibit characteristics of multi-source, strong coupling, time-varying, and cross-level propagation. Harmonic anomalies in localized scenarios at converter stations may not only be caused by a single nonlinear device, but may also propagate between different equipment branches, valve hall sections, and start-up circuit sections, thus forming multi-point responses with similar amplitudes but different physical locations of responsibility.

[0003] Existing methods for locating harmonic sources and assigning responsibility often rely on synchronous voltage and current phasors, complete network parameters, harmonic impedance identification, or large-scale training samples. In engineering sites, there are usually few key measurement points, and multiple acquisition devices are triggered independently and upload raw waveforms asynchronously. The files themselves often only contain file identifiers and discrete waveforms, lacking a unified sampling time that can be directly used for cross-point comparisons.

[0004] Under the above conditions, if the waveforms of each measurement point are compared directly according to the file generation order or local storage time, time mismatch is likely to occur; if the location of responsibility is determined only based on the peak value of a single point or the amplitude of a certain low harmonic, background fluctuations, instantaneous spikes or common propagation responses are easily misjudged as local responsibility anomalies; if only time series similarity is used without introducing the equipment wiring relationship, it is difficult to further interpret the results as responsibility branches and responsibility sections.

[0005] Therefore, there is a need for a method that can recover the actual sampling time of asynchronous waveforms using master station index information, unify the observations of a limited number of measurement points to the same time reference, and combine local electrical topology to adaptively fuse and hierarchically determine multi-frequency harmonic anomalies, so as to improve the robustness, interpretability and engineering applicability of harmonic responsibility section identification. Summary of the Invention

[0006] The purpose of this invention is to solve the problems of lack of unified time reference for the original waveform of asynchronous multi-measurement points, difficulty in direct cross-point comparison of limited measurement points, easy misjudgment of responsibility due to single-point peak values, inability of pure time similarity to reflect the hierarchical relationship between branches and sections, and sensitivity of fixed empirical thresholds to abnormal peak values. It provides an asynchronous multi-measurement point harmonic responsibility section identification method and system, which can achieve harmonic responsibility section identification with hierarchical physical interpretation under the conditions of limited measurement points, asynchronous waveforms and missing complete impedance parameters, and maintain high stability against weight and threshold perturbations.

[0007] To solve the above-mentioned technical problems, the present invention is implemented using the following technical solution:

[0008] In a first aspect, the present invention provides an asynchronous multi-measurement point harmonic responsibility section identification method, comprising:

[0009] Based on the matching relationship between the local voltage waveform files of each measuring point and the updated index records of the main station, the timestamp of each local voltage waveform file is reconstructed, and multiple sets of synchronous multi-point samples are constructed through the nearest neighbor time matching criterion.

[0010] Perform Discrete Fourier Transform on the sampled value sequence of each measuring point in each group of synchronous multi-point samples, extract the fundamental amplitude and multiple harmonic amplitudes respectively, and calculate the total harmonic distortion rate. Based on the multiple harmonic amplitudes and the total harmonic distortion rate, construct the multi-frequency harmonic anomaly characterization vector of each measuring point in each group of synchronous multi-point samples.

[0011] Normalize the multi-frequency harmonic anomaly characterization vector of each measuring point in each group of synchronous multi-point samples, calculate the first-order and second-order changes of the normalized multiple harmonic amplitudes, and construct a unified feature vector of each measuring point in each group of synchronous multi-point samples.

[0012] Based on the unified feature vector of each measuring point in each group of synchronous multi-point samples and the branch or section to which each measuring point belongs, the optimal fusion weight is solved based on the constraint optimization adaptive objective function, and the abnormal intensity of each measuring point in each group of synchronous multi-point samples is calculated based on the optimal fusion weight.

[0013] Based on the abnormal intensity of each measuring point in each group of synchronous multi-point samples and the branch or section to which each measuring point belongs, responsibility and abnormality are determined, and the instantaneous responsibility label and abnormality mark of each group of synchronous multi-point samples are output.

[0014] Synchronous multi-point samples with continuous and abnormal indices are merged into abnormal events according to the index from smallest to largest. Within an abnormal event, a majority vote is performed based on the instantaneous responsibility labels of all synchronous multi-point samples to obtain the responsibility label of the abnormal event. The mean of the responsibility confidence of all synchronous multi-point samples is calculated to obtain the responsibility confidence of the abnormal event.

[0015] Optionally, the timestamp for reconstructing each local voltage waveform file is achieved using the following formula:

[0016] ,

[0017] ,

[0018] in, Indicates the measurement point index. This indicates the local voltage waveform file index. This indicates that the main site is updating the index record. Indicates the first The first measuring point A local voltage waveform file The reconstructed timestamp , Indicates the first The first measuring point A local voltage waveform file The waveform identifier field, Indicates the first The first measuring point A local voltage waveform file The sequence of sampled values, Indicates the first The total number of updated index records for each test point on the main station; Indicates the first The first measuring point A local voltage waveform file With the Main site update index records The matching relationship, Indicates a match. Indicates a mismatch; Indicates the first The first measuring point Main site update index records timestamp, , Indicates the first The first measuring point Main site update index records The waveform identifier field.

[0019] Optionally, constructing a synchronized multi-point sample set using the nearest neighbor time matching criterion includes:

[0020] Select one measurement point from all measurement points as the reference measurement point, and use the local voltage waveform file of that measurement point as the reference local voltage waveform file;

[0021] For each reference local voltage waveform file, perform the following operation:

[0022] Iterate through each of the remaining measurement points except the reference measurement point, calculate the time deviation between the reference local voltage waveform file and each local voltage waveform file of the remaining measurement points, select the local voltage waveform file with the smallest time deviation as the candidate sample of the remaining measurement point, and record the smallest time deviation.

[0023] If the minimum time deviation is not greater than the allowable synchronization error threshold, then add the remaining measurement points to the candidate measurement point set of the reference local voltage waveform file; otherwise, end the operation.

[0024] After the traversal is completed, it is determined whether the time deviation between any two candidate samples of the candidate measurement points in the candidate measurement point set of the reference local voltage waveform file is not greater than the allowable synchronization error threshold.

[0025] If none of them are greater than, the reference local voltage waveform file and the reconstruction timestamps of all candidate samples are weighted and summed to obtain a unified reference timestamp, and the reference local voltage waveform file and all candidate samples are combined into a set of synchronous multi-point samples according to the unified reference timestamp.

[0026] The time deviation is calculated using the following formula:

[0027] ,

[0028] in, Indicates the first The first measuring point A local voltage waveform file The reconstructed timestamp Indicates the first The first measuring point A local voltage waveform file The reconstructed timestamp express and Time deviation between;

[0029] The expression for the synchronized multi-point samples is as follows:

[0030] ,

[0031] in, Indicates the first Group synchronous multi-point samples, Indicates the first Group synchronous multi-point samples A unified reference timestamp Represents the set of measurement points. Indicates the total number of measuring points. Indicates the first The measuring point at the ... The sequence of sampled values ​​in a group of synchronized multi-point samples.

[0032] Optionally, the discrete Fourier transform of the sampled value sequence of each measuring point in the synchronous multi-point sample is performed by the following formula:

[0033] ,

[0034] in, Indicates frequency index, Indicates the first The measuring point at the ... Frequency in group synchronized multi-point samples The range, This represents the number of sampling points within an observation window. Indicates the first The measuring point at the ... The first group of synchronous multi-point samples Each sample value, represents an imaginary number;

[0035] The extraction of the fundamental frequency amplitude and multiple harmonic amplitudes is achieved through the following formula:

[0036] ,

[0037] in, This indicates the harmonic order index; higher harmonics include the 3rd, 5th, and 7th harmonics. Indicates the first The measuring point at the ... The first group of synchronous multi-point samples Second harmonic amplitude Indicates the first Frequency index of subharmonics Indicates the first The measuring point at the ... Frequency in group synchronized multi-point samples The range;

[0038] The total harmonic distortion rate is calculated using the following formula:

[0039] ,

[0040] in, Indicates the first The measuring point at the ... Total harmonic distortion rate in a group of synchronized multi-point samples Represents the set of harmonic orders. , Indicates the first The measuring point at the ... The fundamental amplitude value in a group of synchronized multi-point samples;

[0041] The expression for the multi-frequency harmonic anomaly characterization vector is as follows:

[0042] ,

[0043] in, Show the first The measuring point at the ... Multi-frequency harmonic anomaly representation vector in a group of synchronized multi-point samples. Indicates the first The measuring point at the ... The amplitude of the third harmonic in the group of synchronized multi-point samples. Indicates the first The measuring point at the ... The amplitude of the 5th harmonic in the group of synchronized multi-point samples, Indicates the first The measuring point at the ... The amplitude of the 7th harmonic in the group of synchronous multi-point samples, where T represents transpose.

[0044] Optionally, the first-order change of the normalized multiple harmonic amplitude is obtained by the following formula:

[0045] ,

[0046] in, Indicates the first The measuring point at the ... The first group of synchronous multi-point samples Second harmonic amplitude Indicates the first The measuring point at the ... The first group of synchronous multi-point samples Second harmonic amplitude Indicates the first The measuring point at the ... The first group of synchronous multi-point samples The first-order change in the amplitude of the second harmonic;

[0047] The second-order change of the normalized multiple harmonic amplitude is obtained by the following formula:

[0048] ,

[0049] in, Indicates the first The measuring point at the ... The first group of synchronous multi-point samples The second-order change in the amplitude of the second harmonic. Indicates the first The measuring point at the ... The first group of synchronous multi-point samples Second harmonic amplitude;

[0050] The expression for the unified feature vector is as follows:

[0051] ,

[0052] in, Indicates the first The measuring point at the ... A unified feature vector in a group of synchronized multi-point samples. Indicates the first The measuring point at the ... Normalized third harmonic amplitude in a group of synchronized multi-point samples. Indicates the first The measuring point at the ... Normalized 5th harmonic amplitude in a group of synchronized multi-point samples Indicates the first The measuring point at the ... Normalized 7th harmonic amplitude in a group of synchronized multi-point samples. Indicates the first The measuring point at the ... The first-order change in the normalized third harmonic amplitude in a group of synchronized multi-point samples. Indicates the first The measuring point at the ... The first-order change in the normalized fifth harmonic amplitude in a group of synchronized multi-point samples. Indicates the first The measuring point at the ... The first-order change in the normalized 7th harmonic amplitude in a group of synchronized multi-point samples. Indicates the first The measuring point at the ... The second-order change of the normalized third harmonic amplitude in a group of synchronized multi-point samples. Indicates the first The measuring point at the ... The second-order change of the normalized fifth harmonic amplitude in a group of synchronized multi-point samples. Indicates the first The measuring point at the ... The second-order change of the normalized 7th harmonic amplitude in a group of synchronized multi-point samples. Indicates the first The measuring point at the ... The normalized total harmonic distortion rate in a group of synchronous multi-point samples, where T represents the transpose.

[0053] Optionally, the expression for the constrained optimization adaptive objective function is as follows:

[0054] ,

[0055] in, To maximize, Indicates the fusion weight. , , For preset weighting coefficients, Indicates based on fusion weights The calculated average branch divisibility, Indicates based on fusion weights The calculated average segment discrimination, Indicates based on fusion weights The calculated average common disturbance intensity, where T represents the transpose, This indicates the second normal form.

[0056] Average branch indexing It can be obtained through the following formula:

[0057] ,

[0058] in, This represents the total number of samples from multiple points simultaneously. Indicates based on fusion weights The calculated first The average branch separability of a group of synchronized multi-point samples , Represents the set of unordered branch pairs. This represents the total number of unordered branch pairs. Indicates based on fusion weights Calculated branches and branch roads In the Branch separability of grouped synchronous multi-point samples , Indicates based on fusion weights Calculated branches In the The abnormal intensity of multiple synchronized samples in a group. , Indicates a branch The set of measurement points on the surface For preset weighting coefficients, Indicates based on fusion weights The calculated first The measuring point at the ... Abnormal intensity in multiple synchronized samples , Indicates the first The measuring point at the ... A unified feature vector in a group of synchronized multi-point samples. Indicates based on fusion weights Calculated branches In the The abnormal intensity of multiple synchronized samples in a group. Preset zero-prevention constant;

[0059] Average segment discrimination It can be obtained through the following formula:

[0060] ,

[0061] in, Indicates based on fusion weights The calculated first The average segment discrimination of a group of synchronized multi-point samples, , Represents an unordered set of pairs of segments. Indicates the total number of unordered segments. Indicates based on fusion weights Calculated segment and section In the Segment discriminability of synchronized multi-point samples, , Indicates based on fusion weights Calculated segment In the The abnormal intensity of multiple synchronized samples in a group. , Indicates a section The set of measurement points on the surface Indicates based on fusion weights Calculated segment In the The abnormal intensity of multiple points of samples in a group;

[0062] Average common disturbance intensity It can be obtained through the following formula:

[0063] ,

[0064] in, Indicates based on fusion weights The calculated first The common perturbation index of multiple synchronous samples in a group. , Represents the set of measurement points. Indicates the total number of measuring points. Indicates based on fusion weights The calculated first The measuring point at the ... Abnormal intensity in group synchronous multi-point samples.

[0065] Optionally, the step of determining responsibility and anomaly based on the abnormal intensity of each measuring point in each group of synchronous multi-point samples and the branch or section to which each measuring point belongs, and outputting the instantaneous responsibility label and whether it is abnormal for each group of synchronous multi-point samples, includes:

[0066] Based on the anomaly intensity of each measuring point in each synchronous multi-point sample and the branch or segment to which each measuring point belongs, calculate the anomaly intensity of each branch or segment in each synchronous multi-point sample, the branch separability of each disordered branch in each synchronous multi-point sample, the segment discriminability of each disordered segment in each synchronous multi-point sample, and the common disturbance index, responsibility confidence and overall anomaly intensity of each synchronous multi-point sample.

[0067] Based on the anomaly intensity of each branch or segment in each synchronous multi-point sample, the branch separability of each disordered branch pair in each synchronous multi-point sample, the segment discrimination of each disordered segment pair in each synchronous multi-point sample, and the common disturbance index and responsibility confidence of each synchronous multi-point sample, the branch separability threshold, segment discrimination threshold, common disturbance threshold, event triggering threshold, and responsibility confidence threshold are constructed using the median and median absolute deviation.

[0068] Based on the anomaly intensity of each branch or segment in each synchronous multi-point sample, the branch separability of each disordered branch pair in each synchronous multi-point sample, the segment discrimination of each disordered segment pair in each synchronous multi-point sample, and the common disturbance index and responsibility confidence of each synchronous multi-point sample, the instantaneous responsibility label of each synchronous multi-point sample is output as common side disturbance, branch side disturbance, segment side disturbance, or insufficient criteria, based on the branch separability threshold, segment discrimination threshold, common disturbance threshold, event triggering threshold, and responsibility confidence threshold.

[0069] Synchronous multi-point samples whose overall anomaly intensity is greater than the event trigger threshold are marked as anomalies.

[0070] Optionally, the branch separability threshold is constructed using the following formula:

[0071] ,

[0072] in, Indicates the branch separability threshold. Denotes the set of branches that are divisible. , Indicates a branch and branch roads In the Branch separability of grouped synchronous multi-point samples Represents the set of unordered branch pairs. This represents the total number of samples from multiple points simultaneously. This indicates the weight of the median absolute deviation of the predefined branch indexing. This represents the function for finding the median. This represents the function for calculating the absolute deviation of the median;

[0073] The segment discrimination threshold is constructed using the following formula:

[0074] ,

[0075] in, Indicates the threshold for segment discrimination. Represents the set of segment discriminant values. , Indicates a section and section In the Segment discriminability of synchronized multi-point samples, Represents an unordered set of pairs of segments. This indicates the median absolute deviation weight of the discrimination of the preset segment;

[0076] The common perturbation threshold is constructed using the following formula:

[0077] ,

[0078] in, Indicates the common disturbance threshold. This represents the set of common disturbance indicators. , Indicates the first The common perturbation index of multiple synchronous samples in a group. This indicates the weight of the median absolute deviation of the preset common disturbance index;

[0079] The event trigger threshold is constructed using the following formula:

[0080] ,

[0081] in, Indicates the event trigger threshold. Represents the set of overall anomaly intensities. , Indicates the first The overall anomaly intensity of the group of synchronized multi-point samples, , Represents the set of measurement points. Indicates the total number of measuring points. Indicates the first The measuring point at the ... Abnormal intensity in multiple synchronized samples This represents the weight of the median absolute deviation of the preset total anomaly intensity;

[0082] The responsibility confidence threshold is constructed using the following formula:

[0083]

[0084] in, Indicates the responsibility confidence threshold. This represents the set of responsibility confidence scores for all synchronized multi-point samples. , Indicates the first Responsibility confidence of group synchronous multi-point samples , Indicates the first The maximum value of the anomaly intensity among all measuring points in the synchronized multi-point sample group. Indicates the first The second largest anomaly intensity among all measuring points in the synchronized multi-point sample group. This represents the median absolute deviation weight of the preset responsibility confidence level.

[0085] Optionally, the instantaneous responsibility label for each group of synchronized multi-point samples output may be a common-side disturbance, a branch-side disturbance, a segment-side disturbance, or insufficient criterion, including:

[0086] If criterion 1 is met, the instantaneous responsibility label of the current synchronized multi-point sample is output as common side disturbance; where the criterion is that the common disturbance index of the current synchronized multi-point sample is not greater than the common disturbance threshold and the responsibility confidence is not greater than the confidence threshold.

[0087] If criterion 1 is not satisfied but criterion 2 is satisfied, the responsible branch of the current synchronized multi-point sample is obtained; where criterion 2 is that there are at least two branches in the current synchronized multi-point sample with a branch separability greater than the branch separability threshold, and the responsible branch is the branch with the largest abnormal intensity among all branches in the current synchronized multi-point sample that satisfy criterion 2.

[0088] If criterion 4 is satisfied, the responsible segment of the current synchronized multi-point sample is obtained, and the instantaneous responsibility label of the current synchronized multi-point sample is output as segment-side disturbance; where, criterion 4 is that there are multiple segments within the responsible branch and criterion 3 is satisfied, criterion 3 is that there are at least two segments whose segment discrimination is greater than the segment discrimination threshold, and the responsible segment is the segment with the largest abnormal intensity among all segments that satisfy criterion 3;

[0089] If criterion 4 is not met, output the instantaneous responsibility label of the current synchronized multi-point sample as branch-side disturbance;

[0090] If neither criterion 1 nor criterion 2 is satisfied, then the instantaneous responsibility label of the current synchronized multi-point sample is output as insufficient.

[0091] In a second aspect, the present invention provides a system for implementing the asynchronous multi-point harmonic responsibility section identification method according to any one of the first aspects, comprising:

[0092] The synchronous multi-point sample construction module is used to reconstruct the timestamp of each local voltage waveform file based on the matching relationship between the local voltage waveform file of each measurement point and the update index record of the master station, and construct multiple sets of synchronous multi-point samples through the nearest neighbor time matching criterion.

[0093] The multi-frequency harmonic anomaly characterization vector construction module is used to perform discrete Fourier transform on the sampled value sequence of each measuring point in each group of synchronous multi-point samples, extract the fundamental amplitude and multiple harmonic amplitudes respectively, calculate the total harmonic distortion rate, and construct the multi-frequency harmonic anomaly characterization vector of each measuring point in each group of synchronous multi-point samples based on the multiple harmonic amplitudes and the total harmonic distortion rate.

[0094] The unified feature vector construction module is used to normalize the multi-frequency harmonic anomaly characterization vector of each measuring point in each group of synchronous multi-point samples, calculate the first-order and second-order changes of the normalized multiple harmonic amplitudes, and construct the unified feature vector of each measuring point in each group of synchronous multi-point samples.

[0095] The module for optimizing the fusion weight and calculating the anomaly intensity is used to solve for the optimal fusion weight based on the unified feature vector of each measuring point in each group of synchronous multi-point samples and the branch or section to which each measuring point belongs, and to calculate the anomaly intensity of each measuring point in each group of synchronous multi-point samples based on the optimal fusion weight.

[0096] The responsibility determination and anomaly determination module is used to determine responsibility and anomaly based on the anomaly intensity of each measuring point in each group of synchronous multi-point samples and the branch or section to which each measuring point belongs, and output the instantaneous responsibility label and whether it is abnormal for each group of synchronous multi-point samples.

[0097] The abnormal event characterization module is used to merge synchronous multi-point samples with continuous and abnormal indices into abnormal events according to the index from smallest to largest. Within an abnormal event, a majority vote is performed based on the instantaneous responsibility labels of all synchronous multi-point samples to obtain the responsibility label of the abnormal event. The mean of the responsibility confidence of all synchronous multi-point samples is calculated to obtain the responsibility confidence of the abnormal event.

[0098] Compared with the prior art, the beneficial effects achieved by the present invention are as follows:

[0099] 1. The asynchronous multi-point harmonic responsibility section identification method proposed in this invention utilizes master station index constraints and nearest neighbor time matching to reconstruct asynchronously uploaded, discrete, and weakly correlated original waveforms into unified time samples that can be compared across points, avoiding mismatches caused by direct comparison according to file order; it adaptively fuses the amplitude, dynamic changes, local mutations, and total harmonic distortion rate of the third, fifth, and seventh harmonics to avoid responsibility judgment relying entirely on a single frequency, a single moment amplitude, or manually fixed weights; it reduces the impact of synchronous multi-point simultaneous increases, isolated spikes, and extreme values ​​on responsibility conclusions through joint judgment of common disturbance suppression, robust statistical thresholds, and responsibility confidence; and it transforms instantaneous fluctuation identification into event-level output containing start and end times, responsibility labels, and confidence levels through event triggering, continuous sample merging, and event-level aggregation, which is more in line with the analysis habits of engineers for harmonic anomaly events.

[0100] 2. The system proposed in this invention for identifying asynchronous multi-measurement point harmonic responsibility sections, through the establishment of a synchronous multi-point sample construction module, a multi-frequency harmonic anomaly characterization vector construction module, a unified feature vector construction module, a fusion weight optimization solution and anomaly intensity calculation module, a responsibility determination and anomaly determination module, and an anomaly event characterization module, under the conditions of limited number of measurement points, asynchronous uploading of original waveforms, and unavailability of complete network impedance parameters, identifies harmonic responsibility branches and responsibility sections based on timestamp reconstruction constrained by the master station index, adaptive multi-feature fusion, time-topology collaborative criteria, and event-driven output. Engineering data shows that this invention can form statistical results of responsibility sections with clear physical meaning in 327 sets of synchronous samples, and maintain label stability coefficients of 0.96, 0.92, and 0.85 when the weights and criteria thresholds are disturbed by ±5%, ±10%, and ±20%, respectively. Attached Figure Description

[0101] Figure 1 A flowchart illustrating the asynchronous multi-measurement point harmonic responsibility section identification method provided in this embodiment of the invention;

[0102] Figure 2 A simplified schematic diagram of the local responsibility topology provided in an embodiment of the present invention;

[0103] Figure 3 This is a schematic diagram of the event-driven responsibility identification process provided in an embodiment of the present invention;

[0104] Figure 4 This is a schematic diagram of the time series of the fifth harmonic amplitude at three measuring points provided in an embodiment of the present invention;

[0105] Figure 5 This is a schematic diagram of the low-order harmonic characteristics of three measuring points provided in an embodiment of the present invention;

[0106] Figure 6 This is a schematic diagram illustrating the evolution of the responsibility share of three measurement points over time, as provided in an embodiment of the present invention.

[0107] Figure 7 This is a statistical diagram illustrating the responsibility labels of all synchronized samples provided in this embodiment of the invention.

[0108] Figure 8 This is a partially enlarged schematic diagram of the fifth harmonic amplitude of a peak anomaly event provided in an embodiment of the present invention. Detailed Implementation

[0109] The technical solution of the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the embodiments and specific features in the embodiments are detailed descriptions of the technical solution of the present application, rather than limitations thereof. In the absence of conflict, the embodiments and technical features in the embodiments can be combined with each other.

[0110] It should be noted that the term "and / or" in this article is merely a description of the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A existing alone, A and B existing simultaneously, or B existing alone. Additionally, the character " / " in this article generally indicates that the preceding and following related objects have an "or" relationship.

[0111] Example 1

[0112] This invention discloses an asynchronous multi-measurement point harmonic responsibility section identification method, with reference to... Figure 1 As shown, it includes:

[0113] S1. Based on the matching relationship between the local voltage waveform files of each measuring point and the update index records of the main station, reconstruct the timestamp of each local voltage waveform file, and construct multiple sets of synchronous multi-point samples through the nearest neighbor time matching criterion.

[0114] S2, Perform Discrete Fourier Transform on the sampling value sequence of each measuring point in each group of synchronous multi-point samples, extract the fundamental amplitude and multiple harmonic amplitudes respectively and calculate the total harmonic distortion rate, and construct the multi-frequency harmonic anomaly characterization vector of each measuring point in each group of synchronous multi-point samples based on the multiple harmonic amplitudes and the total harmonic distortion rate.

[0115] S3, normalize the multi-frequency harmonic anomaly characterization vector of each measuring point in each group of synchronous multi-point samples, calculate the first-order and second-order changes of the normalized multiple harmonic amplitudes, and construct the unified feature vector of each measuring point in each group of synchronous multi-point samples.

[0116] S4. Based on the unified feature vector of each measuring point in each group of synchronous multi-point samples and the branch or section to which each measuring point belongs, solve the optimal fusion weight based on the constraint optimization adaptive objective function, and calculate the abnormal intensity of each measuring point in each group of synchronous multi-point samples based on the optimal fusion weight.

[0117] S5, based on the abnormal intensity of each measuring point in each group of synchronous multi-point samples and the branch or section to which each measuring point belongs, performs responsibility determination and abnormality determination, and outputs the instantaneous responsibility label and whether it is abnormal for each group of synchronous multi-point samples.

[0118] S6. Merge the synchronous multi-point samples with continuous and abnormal indices into abnormal events according to the index from smallest to largest. Within the abnormal event, perform majority voting based on the instantaneous responsibility labels of all synchronous multi-point samples to obtain the responsibility label of the abnormal event. Calculate the mean of the responsibility confidence of all synchronous multi-point samples to obtain the responsibility confidence of the abnormal event.

[0119] In step S1, this embodiment acquires local voltage waveform files and master station update index records for multiple measuring points in the local study area of ​​the converter station, and establishes a matching relationship between the local voltage waveform files and the master station update index records based on the waveform identifier field, reconstructing the actual sampling timestamp corresponding to each local voltage waveform file:

[0120] ,

[0121] ,

[0122] in, Indicates the measurement point index. This indicates the local voltage waveform file index. This indicates that the main site is updating the index record. Indicates the first The first measuring point A local voltage waveform file The reconstructed timestamp , Indicates the first The first measuring point A local voltage waveform file The waveform identifier field, Indicates the first The first measuring point A local voltage waveform file The sequence of sampled values, Indicates the first The total number of updated index records for each test point on the main station; Indicates the first The first measuring point A local voltage waveform file With the Main site update index records The matching relationship, Indicates a match. Indicates a mismatch; Indicates the first The first measuring point Main site update index records timestamp, , Indicates the first The first measuring point Main site update index records The waveform identifier field.

[0123] This embodiment uses the reconstructed timestamp of one measurement point as a reference and employs the nearest neighbor time matching criterion to search for the minimum time deviation sample of other measurement points. When the minimum time deviation is not greater than the allowable synchronization error threshold, the samples from each measurement point are combined into a synchronized multi-point sample under the same unified reference time. Specifically, this includes:

[0124] Select one measurement point from all measurement points as the reference measurement point, and use the local voltage waveform file of that measurement point as the reference local voltage waveform file;

[0125] For each reference local voltage waveform file, perform the following operation:

[0126] Iterate through each of the remaining measurement points except the reference measurement point, calculate the time deviation between the reference local voltage waveform file and each local voltage waveform file of the remaining measurement points, select the local voltage waveform file with the smallest time deviation as the candidate sample of the remaining measurement point, and record the minimum time deviation; the time deviation is calculated using the following formula:

[0127] ,

[0128] in, Indicates the first The first measuring point A local voltage waveform file The reconstructed timestamp Indicates the first The first measuring point A local voltage waveform file The reconstructed timestamp express and Time deviation between;

[0129] If the minimum time deviation is not greater than the allowable synchronization error threshold, then add the remaining measurement points to the candidate measurement point set of the reference local voltage waveform file; otherwise, end the operation.

[0130] After the traversal is completed, it is determined whether the time deviation between any two candidate samples of the candidate measurement points in the candidate measurement point set of the reference local voltage waveform file is not greater than the allowable synchronization error threshold.

[0131] If none of them are greater than, the reference local voltage waveform file and the reconstruction timestamps of all candidate samples are weighted and summed to obtain a unified reference timestamp. Based on this unified reference timestamp, the reference local voltage waveform file and all candidate samples are combined into a set of synchronized multi-point samples. The expression for the synchronized multi-point samples is as follows:

[0132] ,

[0133] in, Indicates the first Group synchronous multi-point samples, Indicates the first Group synchronous multi-point samples A unified reference timestamp Represents the set of measurement points. Indicates the total number of measuring points. Indicates the first The measuring point at the ... The sequence of sampled values ​​in a group of synchronized multi-point samples.

[0134] This embodiment does not infer the sampling order based on the file arrangement order. Instead, it first uses the master station index to establish a deterministic association between the file and the timestamp, and then constrains the nearest neighbor matching with the allowable synchronization error, thereby providing a unified time basis for multi-point frequency domain comparison.

[0135] In step S2, the discrete Fourier transform of the sampled value sequence of each measuring point in the synchronous multi-point sample is performed by the following formula:

[0136] ,

[0137] in, Indicates frequency index, Indicates the first The measuring point at the ... Frequency in group synchronized multi-point samples The range, This represents the number of sampling points within an observation window. Indicates the first The measuring point at the ... The first group of synchronous multi-point samples Each sample value, represents an imaginary number;

[0138] The extraction of the fundamental frequency amplitude and multiple harmonic amplitudes is achieved through the following formula:

[0139] ,

[0140] in, This indicates the harmonic order index; higher harmonics include the 3rd, 5th, and 7th harmonics. Indicates the first The measuring point at the ... The first group of synchronous multi-point samples Second harmonic amplitude Indicates the first Frequency index of subharmonics Indicates the first The measuring point at the ... Frequency in group synchronized multi-point samples The range;

[0141] The total harmonic distortion rate is calculated using the following formula:

[0142] ,

[0143] in, Indicates the first The measuring point at the ... Total harmonic distortion rate in a group of synchronized multi-point samples Represents the set of harmonic orders. , Indicates the first The measuring point at the ... The fundamental amplitude value in a group of synchronized multi-point samples;

[0144] The expression for the multi-frequency harmonic anomaly characterization vector is as follows:

[0145] ,

[0146] in, Show the first The measuring point at the ... Multi-frequency harmonic anomaly representation vector in a group of synchronized multi-point samples. Indicates the first The measuring point at the ... The amplitude of the third harmonic in the group of synchronized multi-point samples. Indicates the first The measuring point at the ... The amplitude of the 5th harmonic in the group of synchronized multi-point samples, Indicates the first The measuring point at the ... The amplitude of the 7th harmonic in the group of synchronous multi-point samples, where T represents transpose.

[0147] In step S3, to eliminate the influence of different characteristic dimensions and value ranges, the characteristics of the multi-frequency anomaly characterization vector at any measurement point are first normalized; to reflect the dynamic fluctuations between adjacent time points, the first-order change of the normalized multiple harmonic amplitude is calculated; the first-order change of the normalized multiple harmonic amplitude is obtained by the following formula:

[0148] ,

[0149] in, Indicates the first The measuring point at the ... The first group of synchronous multi-point samples Second harmonic amplitude Indicates the first The measuring point at the ... The first group of synchronous multi-point samples Second harmonic amplitude Indicates the first The measuring point at the ... The first group of synchronous multi-point samples The first-order change in the amplitude of the second harmonic;

[0150] To reflect the abruptness of local anomaly changes, the second-order change of the normalized multiple harmonic amplitude is calculated; the second-order change of the normalized multiple harmonic amplitude is obtained by the following formula:

[0151] ,

[0152] in, Indicates the first The measuring point at the ... The first group of synchronous multi-point samples The second-order change in the amplitude of the second harmonic. Indicates the first The measuring point at the ... The first group of synchronous multi-point samples Second harmonic amplitude;

[0153] The normalized multiple harmonic amplitudes, first-order variations, second-order variations, and total harmonic distortion rate are organized into a unified eigenvector, the expression of which is as follows:

[0154] ,

[0155] in, Indicates the first The measuring point at the ... A unified feature vector in a group of synchronized multi-point samples. Indicates the first The measuring point at the ... Normalized third harmonic amplitude in a group of synchronized multi-point samples. Indicates the first The measuring point at the ... Normalized 5th harmonic amplitude in a group of synchronized multi-point samples Indicates the first The measuring point at the ... Normalized 7th harmonic amplitude in a group of synchronized multi-point samples. Indicates the first The measuring point at the ... The first-order change in the normalized third harmonic amplitude in a group of synchronized multi-point samples. Indicates the first The measuring point at the ... The first-order change in the normalized fifth harmonic amplitude in a group of synchronized multi-point samples. Indicates the first The measuring point at the ... The first-order change in the normalized 7th harmonic amplitude in a group of synchronized multi-point samples. Indicates the first The measuring point at the ... The second-order change of the normalized third harmonic amplitude in a group of synchronized multi-point samples. Indicates the first The measuring point at the ... The second-order change of the normalized fifth harmonic amplitude in a group of synchronized multi-point samples. Indicates the first The measuring point at the ... The second-order change of the normalized 7th harmonic amplitude in a group of synchronized multi-point samples. Indicates the first The measuring point at the ... The normalized total harmonic distortion rate in a group of synchronous multi-point samples, where T represents the transpose.

[0156] In step S4, to avoid a hierarchical inverse dependency between fusion weights, anomaly intensity of measurement points, and branch and segment discrimination indices, this embodiment unifies feature fusion, anomaly intensity characterization, and temporal topology discrimination into a joint constrained optimization problem, and constructs a constrained optimization adaptive objective function; the expression of the constrained optimization adaptive objective function is as follows:

[0157] ,

[0158] in, To maximize, Indicates the fusion weight. , , For preset weighting coefficients, Indicates based on fusion weights The calculated average branch divisibility, Indicates based on fusion weights The calculated average segment discrimination, Indicates based on fusion weights The calculated average common disturbance intensity, where T represents the transpose, This indicates the second normal form.

[0159] To comprehensively evaluate the branch separability under different fusion weights, the average branch separability is... It can be obtained through the following formula:

[0160] ,

[0161] in, This represents the total number of samples from multiple points simultaneously. Indicates based on fusion weights The calculated first The average branch separability of a group of synchronized multi-point samples , Represents the set of unordered branch pairs. This represents the total number of unordered branch pairs. Indicates based on fusion weights Calculated branches and branch roads In the Branch separability of grouped synchronous multi-point samples , Indicates based on fusion weights Calculated branches In the The abnormal intensity of multiple synchronized samples in a group. , Indicates a branch The set of measurement points on the surface For preset weighting coefficients, Indicates based on fusion weights The calculated first The measuring point at the ... Abnormal intensity in multiple synchronized samples , Indicates the first The measuring point at the ... A unified feature vector in a group of synchronized multi-point samples. Indicates based on fusion weights Calculated branches In the The abnormal intensity of multiple synchronized samples in a group. Preset zero-prevention constant;

[0162] To comprehensively evaluate the segment discrimination capability under different fusion weights, the average segment discrimination score is... It can be obtained through the following formula:

[0163] ,

[0164] in, Indicates based on fusion weights The calculated first The average segment discrimination of a group of synchronized multi-point samples, , Represents an unordered set of pairs of segments. Indicates the total number of unordered segments. Indicates based on fusion weights Calculated segment and section In the Segment discriminability of synchronized multi-point samples, , Indicates based on fusion weights Calculated segment In the The abnormal intensity of multiple synchronized samples in a group. , Indicates a section The set of measurement points on the surface Indicates based on fusion weights Calculated segment In the The abnormal intensity of multiple points of samples in a group;

[0165] To comprehensively evaluate the degree of common perturbation under different fusion weights, the average common perturbation intensity is... It can be obtained through the following formula:

[0166] ,

[0167] in, Indicates based on fusion weights The calculated first The common perturbation index of multiple synchronous samples in a group. , Represents the set of measurement points. Indicates the total number of measuring points. Indicates based on fusion weights The calculated first The measuring point at the ... Abnormal intensity in group synchronous multi-point samples.

[0168] This embodiment is based on a pre-constructed constrained optimization adaptive objective function. The feature fusion weights are used as independent decision variables, and the anomaly intensity of measurement points, branch separability, segment discriminability, and common disturbance index are used as derived constraint quantities determined by the fusion weights. The goal is to enhance branch separability and segment discriminability, suppress common disturbances, and limit the weights from deviating from the prior values. The optimal fusion weights and the corresponding optimal anomaly intensity, branch separability, segment discriminability, and common disturbance index are obtained by solving the problem.

[0169] In step S5, the determination of responsibility and anomaly is performed based on the abnormal intensity of each measuring point in each group of synchronous multi-point samples and the branch or section to which each measuring point belongs, and the instantaneous responsibility label and anomaly mark are output for each group of synchronous multi-point samples, including:

[0170] Based on the anomaly intensity of each measuring point in each synchronous multi-point sample and the branch or segment to which each measuring point belongs, calculate the anomaly intensity of each branch or segment in each synchronous multi-point sample, the branch separability of each disordered branch in each synchronous multi-point sample, the segment discriminability of each disordered segment in each synchronous multi-point sample, and the common disturbance index, responsibility confidence and overall anomaly intensity of each synchronous multi-point sample.

[0171] Based on the anomaly intensity of each branch or segment in each synchronous multi-point sample, the branch separability of each disordered branch pair in each synchronous multi-point sample, the segment discrimination of each disordered segment pair in each synchronous multi-point sample, and the common disturbance index and responsibility confidence of each synchronous multi-point sample, the branch separability threshold, segment discrimination threshold, common disturbance threshold, event triggering threshold, and responsibility confidence threshold are constructed using the median and median absolute deviation.

[0172] Based on the anomaly intensity of each branch or segment in each synchronous multi-point sample, the branch separability of each disordered branch pair in each synchronous multi-point sample, the segment discrimination of each disordered segment pair in each synchronous multi-point sample, and the common disturbance index and responsibility confidence of each synchronous multi-point sample, the instantaneous responsibility label of each synchronous multi-point sample is output as common side disturbance, branch side disturbance, segment side disturbance, or insufficient criteria, based on the branch separability threshold, segment discrimination threshold, common disturbance threshold, event triggering threshold, and responsibility confidence threshold.

[0173] Synchronous multi-point samples whose overall anomaly intensity is greater than the event trigger threshold are marked as anomalies.

[0174] The branch separability threshold is constructed using the following formula:

[0175] ,

[0176] in, Indicates the branch separability threshold. Denotes the set of branches that are divisible. , Indicates a branch and branch roads In the Branch separability of grouped synchronous multi-point samples Represents the set of unordered branch pairs. This represents the total number of samples from multiple points simultaneously. This indicates the weight of the median absolute deviation of the predefined branch indexing. This represents the function for finding the median. This represents the function for calculating the absolute deviation of the median;

[0177] The segment discrimination threshold is constructed using the following formula:

[0178] ,

[0179] in, Indicates the threshold for segment discrimination. Represents the set of segment discriminant values. , Indicates a section and section In the Segment discriminability of synchronized multi-point samples, Represents an unordered set of pairs of segments. This indicates the median absolute deviation weight of the discrimination of the preset segment;

[0180] The common perturbation threshold is constructed using the following formula:

[0181] ,

[0182] in, Indicates the common disturbance threshold. This represents the set of common disturbance indicators. , Indicates the first The common perturbation index of multiple synchronous samples in a group. This indicates the weight of the median absolute deviation of the preset common disturbance index;

[0183] The event trigger threshold is constructed using the following formula:

[0184] ,

[0185] in, Indicates the event trigger threshold. Represents the set of overall anomaly intensities. , Indicates the first The overall anomaly intensity of the group of synchronized multi-point samples, , Represents the set of measurement points. Indicates the total number of measuring points. Indicates the first The measuring point at the ... Abnormal intensity in multiple synchronized samples This represents the weight of the median absolute deviation of the preset total anomaly intensity;

[0186] The responsibility confidence threshold is constructed using the following formula:

[0187]

[0188] in, Indicates the responsibility confidence threshold. This represents the set of responsibility confidence scores for all synchronized multi-point samples. , Indicates the first Responsibility confidence of group synchronous multi-point samples , Indicates the first The maximum value of the anomaly intensity among all measuring points in the synchronized multi-point sample group. Indicates the first The second largest anomaly intensity among all measuring points in the synchronized multi-point sample group. This represents the median absolute deviation weight of the preset responsibility confidence level.

[0189] The instantaneous responsibility label for each group of synchronous multi-point samples output is either a common-side disturbance, a branch-side disturbance, a segment-side disturbance, or there is insufficient criterion, including:

[0190] If criterion 1 is met, the instantaneous responsibility label of the current synchronized multi-point sample is output as common side disturbance; where the criterion is that the common disturbance index of the current synchronized multi-point sample is not greater than the common disturbance threshold and the responsibility confidence is not greater than the confidence threshold.

[0191] If criterion 1 is not satisfied but criterion 2 is satisfied, the responsible branch of the current synchronized multi-point sample is obtained; where criterion 2 is that there are at least two branches in the current synchronized multi-point sample with a branch separability greater than the branch separability threshold, and the responsible branch is the branch with the largest abnormal intensity among all branches in the current synchronized multi-point sample that satisfy criterion 2.

[0192] If criterion 4 is satisfied, the responsible segment of the current synchronized multi-point sample is obtained, and the instantaneous responsibility label of the current synchronized multi-point sample is output as segment-side disturbance; where, criterion 4 is that there are multiple segments within the responsible branch and criterion 3 is satisfied, criterion 3 is that there are at least two segments whose segment discrimination is greater than the segment discrimination threshold, and the responsible segment is the segment with the largest abnormal intensity among all segments that satisfy criterion 3;

[0193] If criterion 4 is not met, output the instantaneous responsibility label of the current synchronized multi-point sample as branch-side disturbance;

[0194] If neither criterion 1 nor criterion 2 is satisfied, then the instantaneous responsibility label of the current synchronized multi-point sample is output as insufficient.

[0195] In step S6, refer to Figure 3 As shown, the final exception event is composed of a quadruple. It means that among them and These are abnormal events. Event start and end index, An abnormal event Abnormal event responsibility label, An abnormal event The confidence level of responsibility for abnormal events.

[0196] The structure of the local study area of ​​the converter station in this embodiment is as follows: Figure 2 As shown, the local study area of ​​the converter station is represented by an electrical topology diagram. ;in, , Represents a set of device nodes. Represents the set of electrical connections between nodes; three measurement points. Defined as:

[0197] ,

[0198] in, This indicates the first measuring point. This indicates the second measuring point. This indicates the 3rd measuring point. The measuring point is located on the wall sleeve side of the VSC1 valve hall. The measuring point is located on the wall sleeve side of the VSC3 valve hall. Located on the VSC3 start-up circuit side.

[0199] In this embodiment, the original C-phase voltage waveforms of three measurement points are selected for verification; each text file corresponds to an independently uploaded voltage waveform, with a sampling frequency of 50 kHz and a sampling duration of 100 ms; the three measurement points and their topological locations are shown in Table 1.

[0200] Table 1. Description of the three measuring points

[0201] m1 565612340224 VSC1 valve hall through-wall sleeve side, phase C VSC1 Branch Road Comparison Point m2 565612340222 VSC3 valve hall through-wall sleeve side, phase C VSC3 Valve Hall Side Section m3 565612340223 On the VSC3 start-up circuit side, between switch 4231 and disconnector switch 42312, phase C VSC3 Start-up circuit side section

[0202] After reconstructing the timestamps based on the main station index records, the three measurement point samples were aligned using the nearest neighbor criterion, resulting in 327 sets of synchronized samples obtained between November 23, 2025 and January 3, 2026. Based on the overall anomaly intensity threshold, 10 high-anomaly events were extracted, including short-term shock anomalies and phased anomalies involving multiple synchronized samples.

[0203] Figure 4 The fifth harmonic amplitude time series is shown, where the horizontal axis represents the sampling time and the vertical axis represents the normalized fifth harmonic amplitude. It shows that from November 23 to November 30, 2025, the harmonic amplitudes at all three measuring points increased to varying degrees, with the VSC1 side and the VSC3 valve hall side being relatively higher. After December 3, 2025, the overall harmonic level decreased, and from December 28, 2025 to January 1, 2026, a new local increase mainly occurred on the VSC3 valve hall side.

[0204] Figure 5The heatmap of low-order harmonic features shown is as follows: the horizontal axis represents time / sample update time, and each vertical row represents a "measuring point + harmonic order" feature. The color is Z-score. h3_amp, h5_amp, and h7_amp represent the amplitudes of the 3rd, 5th, and 7th harmonics, respectively. This indicates that the anomalies of the 3rd, 5th, and 7th harmonics do not persist throughout the entire time period, but rather exhibit obvious stages and spatial selectivity. In the early high-anomaly stage, all three measuring points showed varying degrees of response. In the later stage, the highlighted areas were mainly concentrated in the multi-frequency feature row corresponding to the wall bushing C of the VSC3 valve hall.

[0205] Figure 6 The evolution of responsibility share is shown, where the horizontal axis represents the update time and the vertical axis represents the proportion of each branch's responsibility score to the total responsibility score, ranging from 0 to 1. This indicates that the responsibility share of each measuring point fluctuated frequently in the early high anomaly stage, suggesting that the anomaly intensity was similar and the responsibility was relatively dispersed. In the later stage, the responsibility gradually concentrated and was dominated by the VSC3 valve hall side, which can reflect the process of responsibility migrating over time.

[0206] The responsibility labels of all 327 synchronous samples were statistically analyzed, and the results are as follows: Figure 7 As shown: The horizontal axis represents different responsibility categories, and the vertical axis represents the number of times each category occurs. It indicates that the responsibility of the VSC3 valve hall side is relatively high 165 times, accounting for 50.46%; the responsibility of the VSC1 branch is relatively high 114 times, accounting for 34.86%; the responsibility is not obvious or more information is needed 36 times, accounting for 11.01%; the responsibility of the VSC3 start-up circuit side is relatively high 10 times, accounting for 3.06%; and the common side or common disturbance occurs 2 times, accounting for 0.61%.

[0207] From an event perspective, among the 10 highly anomalous events, 5 were marked as having unclear responsibility or requiring more information, 3 were dominated by the VSC1 branch, and 2 were dominated by the VSC3 valve hall side. Event-level conclusions describe the dominant attributes of individual anomalous events, while sample-level statistics describe the frequency distribution of responsibility throughout the entire time period, and the two complement each other.

[0208] Taking the peak anomaly event around 13:00 to 23:00 on November 24, 2025 as an example, Figure 8 This indicates that the VSC3 valve hall side maintained a higher amplitude in the first half of the event, where the horizontal axis represents the specific update time and the vertical axis represents the amplitude of the 5th harmonic. The VSC1 measuring point exhibited two sharper local peaks in the second half. The maximum amplitude of the 5th harmonic at the VSC1 measuring point reached 204.07, while the amplitudes at the VSC3 valve hall side and the VSC3 start-up loop side were 132.38 and 90.91, respectively. If only instantaneous peak values ​​are considered, the entire event might be attributed to the VSC1 branch. This embodiment combines stage evolution and topology hierarchy to interpret the event as a composite pattern of sustained dominance of the VSC3 valve hall side in the early stage and enhanced local peaks in the VSC1 branch in the later stage.

[0209] Compared with the single-point maximum fifth harmonic amplitude criterion, the method in this embodiment is not sensitive to isolated spikes and can identify the responsible segments that have been dominant for a long time. Compared with the pure time similarity criterion that does not include topological priors, the method in this embodiment improves the time statistical comparison to the hierarchical identification of responsible branches and responsible segments through branch separability, segment discriminability, common disturbance suppression and robustness threshold.

[0210] To evaluate the label consistency when the weight vector and criterion threshold are perturbed, a parameter set Θ and its perturbed parameter set Θ′ are defined. The responsibility label stability coefficient is:

[0211] ,

[0212] in, This represents the responsibility label stability coefficient obtained based on the parameter set Θ and its perturbed parameter set Θ′, where N is the total number of synchronization samples. The result is shown in Table 2 for different parameter perturbation amplitudes, which is the indicator function. Represents the first value obtained based on the parameter set Θ. Instantaneous responsibility labels for synchronized multi-point samples in a group. This represents the first parameter obtained based on the perturbed parameter set Θ′. Instantaneous responsibility labels for group-synchronized multi-point samples.

[0213] Table 2. Stability of Responsibility Labels under Parameter Perturbation

[0214] ±5% Weight vector and criterion threshold 0.96 Highly stable ±10% Weight vector and criterion threshold 0.92 Stablize ±20% Weight vector and criterion threshold 0.85 It has acceptable robustness

[0215] As the magnitude of parameter perturbation increases, label consistency decreases, but remains at a high level overall, indicating that the method in this embodiment has good robustness to reasonable changes in adaptive weights and robust thresholds.

[0216] Example 2:

[0217] This invention discloses a system for implementing the asynchronous multi-measurement point harmonic responsibility section identification method described in Embodiment 1, comprising:

[0218] The synchronous multi-point sample construction module is used to reconstruct the timestamp of each local voltage waveform file based on the matching relationship between the local voltage waveform file of each measurement point and the update index record of the master station, and construct multiple sets of synchronous multi-point samples through the nearest neighbor time matching criterion.

[0219] The multi-frequency harmonic anomaly characterization vector construction module is used to perform discrete Fourier transform on the sampled value sequence of each measuring point in each group of synchronous multi-point samples, extract the fundamental amplitude and multiple harmonic amplitudes respectively, calculate the total harmonic distortion rate, and construct the multi-frequency harmonic anomaly characterization vector of each measuring point in each group of synchronous multi-point samples based on the multiple harmonic amplitudes and the total harmonic distortion rate.

[0220] The unified feature vector construction module is used to normalize the multi-frequency harmonic anomaly characterization vector of each measuring point in each group of synchronous multi-point samples, calculate the first-order and second-order changes of the normalized multiple harmonic amplitudes, and construct the unified feature vector of each measuring point in each group of synchronous multi-point samples.

[0221] The module for optimizing the fusion weight and calculating the anomaly intensity is used to solve for the optimal fusion weight based on the unified feature vector of each measuring point in each group of synchronous multi-point samples and the branch or section to which each measuring point belongs, and to calculate the anomaly intensity of each measuring point in each group of synchronous multi-point samples based on the optimal fusion weight.

[0222] The responsibility determination and anomaly determination module is used to determine responsibility and anomaly based on the anomaly intensity of each measuring point in each group of synchronous multi-point samples and the branch or section to which each measuring point belongs, and output the instantaneous responsibility label and whether it is abnormal for each group of synchronous multi-point samples.

[0223] The abnormal event characterization module is used to merge synchronous multi-point samples with continuous and abnormal indices into abnormal events according to the index from smallest to largest. Within an abnormal event, a majority vote is performed based on the instantaneous responsibility labels of all synchronous multi-point samples to obtain the responsibility label of the abnormal event. The mean of the responsibility confidence of all synchronous multi-point samples is calculated to obtain the responsibility confidence of the abnormal event.

[0224] The specific functions of each module described above are explained in the relevant content of the method in Embodiment 1, and will not be repeated here.

[0225] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0226] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0227] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0228] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0229] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. A method for identifying asynchronous multi-measurement point harmonic responsibility sections, characterized in that, include: Based on the matching relationship between the local voltage waveform files of each measuring point and the updated index records of the main station, the timestamp of each local voltage waveform file is reconstructed, and multiple sets of synchronous multi-point samples are constructed through the nearest neighbor time matching criterion. Perform Discrete Fourier Transform on the sampled value sequence of each measuring point in each group of synchronous multi-point samples, extract the fundamental amplitude and multiple harmonic amplitudes respectively, and calculate the total harmonic distortion rate. Based on the multiple harmonic amplitudes and the total harmonic distortion rate, construct the multi-frequency harmonic anomaly characterization vector of each measuring point in each group of synchronous multi-point samples. Normalize the multi-frequency harmonic anomaly characterization vector of each measuring point in each group of synchronous multi-point samples, calculate the first-order and second-order changes of the normalized multiple harmonic amplitudes, and construct a unified feature vector of each measuring point in each group of synchronous multi-point samples. Based on the unified feature vector of each measuring point in each group of synchronous multi-point samples and the branch or section to which each measuring point belongs, the optimal fusion weight is solved based on the constraint optimization adaptive objective function, and the abnormal intensity of each measuring point in each group of synchronous multi-point samples is calculated based on the optimal fusion weight. Based on the abnormal intensity of each measuring point in each group of synchronous multi-point samples and the branch or section to which each measuring point belongs, responsibility and abnormality are determined, and the instantaneous responsibility label and abnormality mark of each group of synchronous multi-point samples are output. Synchronous multi-point samples with continuous and abnormal indices are merged into abnormal events according to the index from smallest to largest. Within an abnormal event, a majority vote is performed based on the instantaneous responsibility labels of all synchronous multi-point samples to obtain the responsibility label of the abnormal event. The mean of the responsibility confidence of all synchronous multi-point samples is calculated to obtain the responsibility confidence of the abnormal event.

2. The asynchronous multi-measurement point harmonic responsibility section identification method according to claim 1, characterized in that, The timestamp for reconstructing each local voltage waveform file is achieved using the following formula: ， ， in, Indicates the measurement point index. This indicates the local voltage waveform file index. This indicates that the main site is updating the index record. Indicates the first The first measuring point A local voltage waveform file The reconstructed timestamp , Indicates the first The first measuring point A local voltage waveform file The waveform identifier field, Indicates the first The first measuring point A local voltage waveform file The sequence of sampled values, Indicates the first The total number of updated index records for each test point on the main station; Indicates the first The first measuring point A local voltage waveform file With the Main site update index records The matching relationship Indicates a match. Indicates a mismatch; Indicates the first The first measuring point Main site update index records timestamp, , Indicates the first The first measuring point Main site update index records The waveform identifier field.

3. The asynchronous multi-measurement point harmonic responsibility section identification method according to claim 1, characterized in that, The construction of a synchronized multi-point sample set using the nearest neighbor time matching criterion includes: Select one measurement point from all measurement points as the reference measurement point, and use the local voltage waveform file of that measurement point as the reference local voltage waveform file; For each reference local voltage waveform file, perform the following operation: Iterate through each of the remaining measurement points except the reference measurement point, calculate the time deviation between the reference local voltage waveform file and each local voltage waveform file of the remaining measurement points, select the local voltage waveform file with the smallest time deviation as the candidate sample of the remaining measurement point, and record the smallest time deviation. If the minimum time deviation is not greater than the allowable synchronization error threshold, then add the remaining measurement points to the candidate measurement point set of the reference local voltage waveform file; otherwise, end the operation. After the traversal is completed, it is determined whether the time deviation between any two candidate samples of the candidate measurement points in the candidate measurement point set of the reference local voltage waveform file is not greater than the allowable synchronization error threshold. If none of them are greater than, the reference local voltage waveform file and the reconstruction timestamps of all candidate samples are weighted and summed to obtain a unified reference timestamp, and the reference local voltage waveform file and all candidate samples are combined into a set of synchronous multi-point samples according to the unified reference timestamp. The time deviation is calculated using the following formula: ， in, Indicates the first The first measuring point A local voltage waveform file The reconstructed timestamp Indicates the first The first measuring point A local voltage waveform file The reconstructed timestamp express and Time deviation between them; The expression for the synchronized multi-point samples is as follows: ， in, Indicates the first Group synchronous multi-point samples, Indicates the first Group synchronous multi-point samples A unified reference timestamp Represents the set of measurement points. Indicates the total number of measuring points. Indicates the first The measuring point at the ... The sequence of sampled values ​​in a group of synchronized multi-point samples.

4. The asynchronous multi-measurement point harmonic responsibility section identification method according to claim 1, characterized in that, The discrete Fourier transform of the sampled value sequence of each measuring point in the synchronous multi-point sample is achieved by the following formula: ， in, Indicates frequency index, Indicates the first The measuring point at the ... Frequency in group synchronized multi-point samples The range, This represents the number of sampling points within an observation window. Indicates the first The measuring point at the ... The first group of synchronous multi-point samples Each sample value, represents an imaginary number; The extraction of the fundamental frequency amplitude and multiple harmonic amplitudes is achieved through the following formula: ， in, This indicates the harmonic order index; higher harmonics include the 3rd, 5th, and 7th harmonics. Indicates the first The measuring point at the ... The first group of synchronous multi-point samples Second harmonic amplitude Indicates the first Frequency index of subharmonics Indicates the first The measuring point at the ... Frequency in group synchronized multi-point samples The range; The total harmonic distortion rate is calculated using the following formula: ， in, Indicates the first The measuring point at the ... Total harmonic distortion rate in a group of synchronized multi-point samples Represents the set of harmonic orders. , Indicates the first The measuring point at the ... The fundamental amplitude value in a group of synchronized multi-point samples; The expression for the multi-frequency harmonic anomaly characterization vector is as follows: ， in, Show the first The measuring point at the ... Multi-frequency harmonic anomaly representation vector in a group of synchronized multi-point samples. Indicates the first The measuring point at the ... The amplitude of the third harmonic in the group of synchronized multi-point samples. Indicates the first The measuring point at the ... The amplitude of the 5th harmonic in the group of synchronized multi-point samples, Indicates the first The measuring point at the ... The amplitude of the 7th harmonic in the group of synchronous multi-point samples, where T represents transpose.

5. The asynchronous multi-measurement point harmonic responsibility section identification method according to claim 1, characterized in that, The first-order change of the normalized multiple harmonic amplitude is obtained by the following formula: ， in, Indicates the first The measuring point at the ... The first group of synchronous multi-point samples Second harmonic amplitude Indicates the first The measuring point at the ... The first group of synchronous multi-point samples Second harmonic amplitude Indicates the first The measuring point at the ... The first group of synchronous multi-point samples The first-order change in the amplitude of the second harmonic; The second-order change of the normalized multiple harmonic amplitude is obtained by the following formula: ， in, Indicates the first The measuring point at the ... The first group of synchronous multi-point samples The second-order change in the amplitude of the subharmonic. Indicates the first The measuring point at the ... The first group of synchronous multi-point samples Second harmonic amplitude; The expression for the unified feature vector is as follows: ， in, Indicates the first The measuring point at the ... A unified feature vector in a group of synchronized multi-point samples. Indicates the first The measuring point at the ... Normalized third harmonic amplitude in a group of synchronized multi-point samples Indicates the first The measuring point at the ... Normalized 5th harmonic amplitude in a group of synchronized multi-point samples Indicates the first The measuring point at the ... Normalized 7th harmonic amplitude in a group of synchronized multi-point samples. Indicates the first The measuring point at the ... The first-order change in the normalized third harmonic amplitude in a group of synchronized multi-point samples. Indicates the first The measuring point at the ... The first-order change in the normalized fifth harmonic amplitude in a group of synchronized multi-point samples. Indicates the first The measuring point at the ... The first-order change in the normalized 7th harmonic amplitude in a group of synchronized multi-point samples. Indicates the first The measuring point at the ... The second-order change of the normalized third harmonic amplitude in a group of synchronized multi-point samples. Indicates the first The measuring point at the ... The second-order change of the normalized fifth harmonic amplitude in a group of synchronized multi-point samples. Indicates the first The measuring point at the ... The second-order change of the normalized 7th harmonic amplitude in a group of synchronized multi-point samples. Indicates the first The measuring point at the ... The normalized total harmonic distortion rate in a group of synchronous multi-point samples, where T represents the transpose.

6. The asynchronous multi-measurement point harmonic responsibility section identification method according to claim 1, characterized in that, The expression for the constrained optimization adaptive objective function is as follows: ， in, To maximize, Indicates the fusion weight. , , For preset weighting coefficients, Indicates based on fusion weights The calculated average branch divisibility, Indicates based on fusion weights The calculated average segment discrimination, Indicates based on fusion weights The calculated average common disturbance intensity, where T represents the transpose, This indicates the second normal form. Average branch indexing It can be obtained through the following formula: ， in, This represents the total number of samples from multiple points simultaneously. Indicates based on fusion weights The calculated first The average branch separability of a group of synchronized multi-point samples , Represents the set of unordered branch pairs. This represents the total number of unordered branch pairs. Indicates based on fusion weights Calculated branches and branch roads In the Branch separability of grouped synchronous multi-point samples , Indicates based on fusion weights Calculated branches In the The abnormal intensity of multiple synchronous samples in a group. , Indicates a branch The set of measurement points on the surface For preset weighting coefficients, Indicates based on fusion weights The calculated first The measuring point at the ... Abnormal intensity in multiple synchronized samples , Indicates the first The measuring point at the ... A unified feature vector in a group of synchronized multi-point samples. Indicates based on fusion weights Calculated branches In the The abnormal intensity of multiple synchronous samples in a group. Preset zero-prevention constant; Average segment discrimination It can be obtained through the following formula: ， in, Indicates based on fusion weights The calculated first The average segment discrimination of a group of synchronized multi-point samples, , Represents an unordered set of pairs of segments. Indicates the total number of unordered segments. Indicates based on fusion weights Calculated segment and section In the Segment discriminability of synchronized multi-point samples, , Indicates based on fusion weights Calculated segment In the The abnormal intensity of multiple synchronous samples in a group. , Indicates a section The set of measurement points on the surface Indicates based on fusion weights Calculated segment In the The abnormal intensity of multiple points of samples in a group; Average common disturbance intensity It can be obtained through the following formula: ， in, Indicates based on fusion weights The calculated first The common perturbation index of multiple synchronous samples in a group. , Represents the set of measurement points. Indicates the total number of measuring points. Indicates based on fusion weights The calculated first The measuring point at the ... Abnormal intensity in group synchronous multi-point samples.

7. The asynchronous multi-measurement point harmonic responsibility section identification method according to claim 1, characterized in that, The process involves determining responsibility and anomalies based on the anomaly intensity of each measuring point in each group of synchronous multi-point samples and the branch or section to which each measuring point belongs. The instantaneous responsibility label and anomaly flag are output for each group of synchronous multi-point samples, including: Based on the anomaly intensity of each measuring point in each synchronous multi-point sample and the branch or segment to which each measuring point belongs, calculate the anomaly intensity of each branch or segment in each synchronous multi-point sample, the branch separability of each disordered branch in each synchronous multi-point sample, the segment discriminability of each disordered segment in each synchronous multi-point sample, and the common disturbance index, responsibility confidence and overall anomaly intensity of each synchronous multi-point sample. Based on the anomaly intensity of each branch or segment in each synchronous multi-point sample, the branch separability of each disordered branch pair in each synchronous multi-point sample, the segment discrimination of each disordered segment pair in each synchronous multi-point sample, and the common disturbance index and responsibility confidence of each synchronous multi-point sample, the branch separability threshold, segment discrimination threshold, common disturbance threshold, event triggering threshold, and responsibility confidence threshold are constructed using the median and median absolute deviation. Based on the anomaly intensity of each branch or segment in each synchronous multi-point sample, the branch separability of each disordered branch pair in each synchronous multi-point sample, the segment discrimination of each disordered segment pair in each synchronous multi-point sample, and the common disturbance index and responsibility confidence of each synchronous multi-point sample, the instantaneous responsibility label of each synchronous multi-point sample is output as common side disturbance, branch side disturbance, segment side disturbance, or insufficient criteria, based on the branch separability threshold, segment discrimination threshold, common disturbance threshold, event triggering threshold, and responsibility confidence threshold. Synchronous multi-point samples whose overall anomaly intensity is greater than the event trigger threshold are marked as anomalies.

8. The asynchronous multi-measurement point harmonic responsibility section identification method according to claim 7, characterized in that, The branch separability threshold is constructed using the following formula: ， in, Indicates the branch separability threshold. Denotes the set of branches that are divisible. , Indicates a branch and branch roads In the Branch separability of grouped synchronous multi-point samples Represents the set of unordered branch pairs. This represents the total number of samples from multiple points simultaneously. This indicates the weight of the median absolute deviation of the predefined branch indexing. This represents the function for finding the median. This represents the function for calculating the absolute deviation of the median; The segment discrimination threshold is constructed using the following formula: ， in, Indicates the threshold for segment discrimination. Represents the set of segment discriminant values. , Indicates a section and section In the Segment discriminability of synchronized multi-point samples, Represents an unordered set of pairs of segments. This indicates the median absolute deviation weight of the discrimination of the preset segment; The common perturbation threshold is constructed using the following formula: ， in, Indicates the common perturbation threshold. This represents the set of common disturbance indicators. , Indicates the first The common perturbation index of multiple synchronous samples in a group. This indicates the weight of the median absolute deviation of the preset common disturbance index; The event trigger threshold is constructed using the following formula: ， in, Indicates the event trigger threshold. Represents the set of overall anomaly intensities. , Indicates the first The overall anomaly intensity of the group of synchronized multi-point samples, , Represents the set of measurement points. Indicates the total number of measuring points. Indicates the first The measuring point at the ... Abnormal intensity in multiple synchronized samples This represents the weight of the median absolute deviation of the preset total anomaly intensity; The responsibility confidence threshold is constructed using the following formula: in, Indicates the responsibility confidence threshold. This represents the set of responsibility confidence scores for all synchronized multi-point samples. , Indicates the first Responsibility confidence of group synchronous multi-point samples , Indicates the first The maximum value of the anomaly intensity among all measuring points in the synchronized multi-point sample group. Indicates the first The second largest anomaly intensity among all measuring points in the synchronized multi-point sample group. This represents the median absolute deviation weight of the preset responsibility confidence level.

9. The asynchronous multi-measurement point harmonic responsibility section identification method according to claim 8, characterized in that, The instantaneous responsibility label for each group of synchronous multi-point samples output is either a common-side disturbance, a branch-side disturbance, a segment-side disturbance, or there is insufficient criterion, including: If criterion 1 is met, the instantaneous responsibility label of the current synchronized multi-point sample is output as common side disturbance; where the criterion is that the common disturbance index of the current synchronized multi-point sample is not greater than the common disturbance threshold and the responsibility confidence is not greater than the confidence threshold. If criterion 1 is not satisfied but criterion 2 is satisfied, the responsible branch of the current synchronized multi-point sample is obtained; where criterion 2 is that there are at least two branches in the current synchronized multi-point sample with a branch separability greater than the branch separability threshold, and the responsible branch is the branch with the largest abnormal intensity among all branches in the current synchronized multi-point sample that satisfy criterion 2. If criterion 4 is satisfied, the responsible segment of the current synchronized multi-point sample is obtained, and the instantaneous responsibility label of the current synchronized multi-point sample is output as segment-side disturbance; where, criterion 4 is that there are multiple segments within the responsible branch and criterion 3 is satisfied, criterion 3 is that there are at least two segments whose segment discrimination is greater than the segment discrimination threshold, and the responsible segment is the segment with the largest abnormal intensity among all segments that satisfy criterion 3; If criterion 4 is not met, output the instantaneous responsibility label of the current synchronized multi-point sample as branch-side disturbance; If neither criterion 1 nor criterion 2 is satisfied, then the instantaneous responsibility label of the current synchronized multi-point sample is output as insufficient.

10. A system for implementing the asynchronous multi-point harmonic responsibility section identification method according to any one of claims 1-9, characterized in that, include: The synchronous multi-point sample construction module is used to reconstruct the timestamp of each local voltage waveform file based on the matching relationship between the local voltage waveform file of each measurement point and the update index record of the master station, and construct multiple sets of synchronous multi-point samples through the nearest neighbor time matching criterion. The multi-frequency harmonic anomaly characterization vector construction module is used to perform discrete Fourier transform on the sampled value sequence of each measuring point in each group of synchronous multi-point samples, extract the fundamental amplitude and multiple harmonic amplitudes respectively, calculate the total harmonic distortion rate, and construct the multi-frequency harmonic anomaly characterization vector of each measuring point in each group of synchronous multi-point samples based on the multiple harmonic amplitudes and the total harmonic distortion rate. The unified feature vector construction module is used to normalize the multi-frequency harmonic anomaly characterization vector of each measuring point in each group of synchronous multi-point samples, calculate the first-order and second-order changes of the normalized multiple harmonic amplitudes, and construct the unified feature vector of each measuring point in each group of synchronous multi-point samples. The module for optimizing the fusion weight and calculating the anomaly intensity is used to solve for the optimal fusion weight based on the unified feature vector of each measuring point in each group of synchronous multi-point samples and the branch or section to which each measuring point belongs, and to calculate the anomaly intensity of each measuring point in each group of synchronous multi-point samples based on the optimal fusion weight. The responsibility determination and anomaly determination module is used to determine responsibility and anomaly based on the anomaly intensity of each measuring point in each group of synchronous multi-point samples and the branch or section to which each measuring point belongs, and output the instantaneous responsibility label and whether it is abnormal for each group of synchronous multi-point samples. The abnormal event characterization module is used to merge synchronous multi-point samples with continuous and abnormal indices into abnormal events according to the index from smallest to largest. Within an abnormal event, a majority vote is performed based on the instantaneous responsibility labels of all synchronous multi-point samples to obtain the responsibility label of the abnormal event. The mean of the responsibility confidence of all synchronous multi-point samples is calculated to obtain the responsibility confidence of the abnormal event.

Images