A line current anomaly analysis method and system

Abstract

The application discloses a kind of line current anomaly analysis method and system, method includes: obtaining line current sequence and extracting target line current;Match historical current data to correct and obtain target initial line current;The instantaneous variation of each current and reference is calculated, whether the doubtful abnormal point of non-continuity is screened according to the threshold and the continuity of variation;The attitude angle of sensor is obtained at reference time and suspicious time, respectively reference current and suspicious current are decomposed into X axis, Y axis, Z axis three-dimensional component;By comparing whether each axis component deviation is over threshold, multidimensional anomaly determination is carried out.The application improves reference accuracy by historical data correction, effectively distinguishes transient disturbance and continuous fault by continuity judgment, and realizes current space three-dimensional decomposition and diagnosis by combining attitude angle, which significantly improves the accuracy, robustness and fault type recognition ability of anomaly detection.

Description

Technical Field

[0001] This invention belongs to the field of current measurement technology, and in particular relates to a method and system for analyzing abnormal line current. Background Technology

[0002] With the rapid development of smart grids and the Internet of Things for power, real-time and accurate sensing of the operating status of power systems has become crucial. Current, as one of the core parameters characterizing the operating status of a power system, directly affects the reliability and effectiveness of advanced applications such as relay protection, power quality analysis, equipment condition monitoring, and load control.

[0003] In actual current measurement processes, due to various factors such as electromagnetic interference, sensor transient saturation, transient errors in the data acquisition system, line faults, or drastic load changes, the measured current data inevitably contains outliers or distortions. If these abnormal data are not identified and processed in a timely manner, they will lead to misjudgments in subsequent analysis and decision-making, such as causing malfunctions or failures to operate protection devices, affecting the safe and stable operation of the system.

[0004] Currently, traditional current anomaly detection methods mostly focus on threshold judgment of current amplitude, such as setting a fixed upper and lower limit, and judging it as abnormal if it exceeds the range. Although such methods are simple and direct, they have obvious limitations: First, fixed thresholds are difficult to adapt to changes in system operation and dynamic range under different load conditions, and sensitivity and reliability are difficult to balance—if the threshold is set too high, it will easily lead to missed detection of real anomalies, and if it is set too low, it will easily produce false alarms. Second, relying solely on amplitude information cannot effectively identify anomalies where phase distortion occurs but amplitude changes are not obvious, such as certain types of harmonic pollution or phase shift. In addition, traditional methods lack an effective differentiation mechanism for brief pulse interference and continuous fault states, which may misjudge instantaneous noise interference as serious faults, or affect the judgment criteria for normal states due to too much continuous abnormal data.

[0005] Therefore, how to provide a current anomaly analysis method that can adapt to dynamic operating conditions, integrate multi-dimensional information, and effectively distinguish between transient interference and real faults has become a technical problem that urgently needs to be solved in this field. Summary of the Invention

[0006] This invention provides a method and system for analyzing abnormal line currents, which solves the above-mentioned technical problems.

[0007] In a first aspect, the present invention provides a method for analyzing abnormal line currents, comprising:

[0008] The line current of the line under test is obtained based on the preset target sensor within the current time period, and the line current sequence is obtained. The target line current is extracted from the line current sequence based on the preset current selection rule.

[0009] According to a preset current matching strategy, at least one historical initial line current corresponding to the target line current is matched in at least one preset historical line current sequence, and the target line current is corrected according to the at least one historical initial line current to obtain the target initial line current.

[0010] The difference between the other line currents in the line current sequence and the target initial line current is calculated to obtain at least one instantaneous current change, and the at least one instantaneous current change is sorted according to time order to obtain an instantaneous current change sequence.

[0011] Determine whether there is at least one non-continuous questionable instantaneous current change in the instantaneous current change sequence. A questionable instantaneous current change is an instantaneous current change that is greater than a preset threshold. A non-continuous questionable instantaneous current change is defined as a questionable instantaneous current change when the number of consecutively adjacent questionable instantaneous current changes is less than a preset number threshold.

[0012] The target sensor acquires the target attitude angle information at the target acquisition time. Based on the target initial line current and the target attitude angle information, anomaly analysis is performed on at least one suspicious instantaneous current change that is discontinuous, and anomaly analysis results are obtained. The target acquisition time is the acquisition time corresponding to the target initial line current.

[0013] In a second aspect, the present invention provides a line current anomaly analysis system, comprising:

[0014] The extraction module is configured to acquire the line current of the line under test within the current time period based on a preset target sensor, obtain a line current sequence, and extract the target line current from the line current sequence based on a preset current selection rule.

[0015] The correction module is configured to match at least one historical initial line current corresponding to the target line current in at least one historical line current sequence according to a preset current matching strategy, and to correct the target line current according to the at least one historical initial line current to obtain the target initial line current.

[0016] The sorting module is configured to subtract the other line currents in the line current sequence from the target initial line current to obtain at least one instantaneous current change, and sort the at least one instantaneous current change based on time order to obtain an instantaneous current change sequence.

[0017] The judgment module is configured to determine whether there is at least one non-continuous questionable instantaneous current change in the instantaneous current change sequence. The questionable instantaneous current change is an instantaneous current change that is greater than a preset threshold. A non-continuous questionable instantaneous current change is defined as a questionable instantaneous current change when the number of consecutively adjacent questionable instantaneous current changes is less than a preset number threshold.

[0018] The analysis module is configured to acquire the target attitude angle information of the target sensor at the target acquisition time, and to perform anomaly analysis on at least one suspicious instantaneous current change that is discontinuous based on the target initial line current and the target attitude angle information, so as to obtain anomaly analysis results, wherein the target acquisition time is the acquisition time corresponding to the target initial line current.

[0019] Thirdly, an electronic device is provided, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the steps of the line current anomaly analysis method of any embodiment of the present invention.

[0020] Fourthly, the present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein when the program instructions are executed by a processor, the processor performs the steps of the line current anomaly analysis method of any embodiment of the present invention.

[0021] The line current anomaly analysis method and system of this application firstly extracts the "target initial line current" from the original current sequence by introducing multi-dimensional stability assessment and adaptive correction driven by historical data. This target current has strong anti-interference capabilities and can truly reflect the steady-state operation characteristics of the system. This solves the problem of reference drift and misjudgment caused by fixed thresholds or single-moment references in traditional methods, and establishes a high-precision and high-reliability dynamic reference for subsequent analysis. Furthermore, by calculating instantaneous changes and implementing a continuous intelligent screening mechanism, random pulse interference is effectively removed, and suspected anomalies are accurately classified into instantaneous suspicious points and persistent faults. While ensuring the sensitivity of fault detection, the false alarm rate caused by noise is significantly reduced. For suspicious points, sensor attitude angle information is introduced to decompose the current vector into a physical three-dimensional space based on the conductor direction. This achieves a dimensional leap from traditional amplitude monitoring to spatial current state diagnosis, enabling anomalies such as ground leakage and unbalanced coupling, which are hidden in a single dimension, to be explicitly exposed and accurately located. Attached Figure Description

[0022] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 A flowchart of a line current anomaly analysis method provided in an embodiment of the present invention;

[0024] Figure 2 This is a structural block diagram of a line current anomaly analysis system provided in an embodiment of the present invention;

[0025] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0026] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0027] Please see Figure 1 The diagram shows a flowchart of a line current anomaly analysis method according to this application.

[0028] like Figure 1As shown, the method for analyzing abnormal line currents specifically includes the following steps:

[0029] Step S101: Based on the preset target sensor, obtain the line current of the line under test in the current time period to obtain the line current sequence, and extract the target line current from the line current sequence based on the preset current selection rule.

[0030] In this step, the line current sequence is divided into multiple consecutive time windows, each containing a preset number of line currents;

[0031] Calculate the multidimensional stability index of the current signal within each time window, and determine the window stability of each time window based on the multidimensional stability index. The multidimensional stability index includes the coefficient of variation of the current amplitude within the time window. The dispersion of the current phase angle within the time window Symmetry index of three-phase current within the time window and the total harmonic distortion of the current within the time window The expression for calculating window stability is:

[0032] ,

[0033] In the formula, For window stability, This is the weighting factor for the coefficient of variation of the current amplitude within the time window. This is a weighting coefficient for the dispersion of the current phase angle within the time window. The weighting coefficients represent the symmetry index of the three-phase currents within the time window. This is the weighting coefficient for the total harmonic distortion rate of the current within the time window. ;

[0034] Select the target time window with the highest stability value, and calculate the comprehensive confidence level of each line current within the target time window. The expression is as follows:

[0035] ,

[0036] In the formula, Let be the overall confidence level of the i-th line current within the target time window. Let be the current amplitude of the i-th line current. The average current amplitude within the target time window. The standard deviation of the current amplitude within the target time window. Let be the phase angle of the i-th line current within the target time window. The average phase angle within the target time window;

[0037] The line current with the highest overall confidence level is selected as the target line current.

[0038] In one specific embodiment, a target sensor (e.g., a Rogowski coil integrated with an inertial measurement unit (IMU) mounted on a transmission line or tower) acquires the instantaneous values ​​of the three-phase current of the line under test in real time at a fixed sampling frequency (e.g., 4 kHz). All current sampling points acquired within the current time period (e.g., the most recent 10 minutes) are arranged in chronological order to form a line current sequence, where each sampling point contains amplitude and phase information of the three-phase current.

[0039] Next, the line current sequence is divided into multiple consecutive time windows. The length of each time window is set based on the power frequency cycle, for example, 10 power frequency cycles (0.2 seconds for a 50Hz system), and includes a preset number of sampling points (e.g., 800 sampling points per window for a 4kHz sampling rate). Windows can slide with a 50% overlap rate to ensure the continuity of the analysis.

[0040] For each time window, calculate its multidimensional stability index, including:

[0041] coefficient of variation of current amplitude Calculate the mean of the current amplitude sequence within the window. with standard deviation ,but This indicator reflects the degree of fluctuation in current amplitude within a window; the smaller the value, the more stable the current.

[0042] current phase angle dispersion Extract the phase angle sequence of each sampling point within the window, and obtain the result by calculating the root mean square of the difference between each phase angle and the average phase angle. This indicator reflects the consistency and stability of the phase.

[0043] Symmetry index of three-phase current Perform symmetrical component transformation on the three-phase currents at each sampling point within the window to calculate the negative sequence current. With positive sequence current The amplitude ratio, i.e. This indicator reflects the degree of three-phase imbalance; the smaller the value, the more balanced the three phases are.

[0044] Total harmonic distortion of current Perform Fourier analysis on the current signal within the window to extract the fundamental amplitude. and the amplitude of each harmonic ,calculate , This index reflects the degree of waveform distortion, taking into account the highest harmonic order.

[0045] Using the above indicators, calculate the window stability score for each time window.

[0046] In summary, firstly, by dividing the time window and calculating its stability, the system can automatically identify the most stable and least disturbed period from dynamic operating data, avoiding reference selection deviations caused by random disturbances or transient processes. Secondly, by comprehensively considering the stability of multiple dimensions such as current amplitude, phase, three-phase balance, and harmonic distortion rate, the selected reference current not only has stable amplitude but also high waveform quality and good three-phase balance, providing a more comprehensive and reliable reference for subsequent anomaly analysis. Thirdly, by calculating the comprehensive confidence level of each sampling point, the current point closest to the ideal steady state is further refined within the stable window, further improving the accuracy and representativeness of the reference. This method overcomes the limitations of traditional methods that select references using a single threshold or fixed time period, significantly enhancing the adaptability and robustness of the anomaly detection algorithm to complex operating conditions, and laying a solid foundation for high-precision line current anomaly analysis.

[0047] Step S102: Match at least one historical initial line current corresponding to the target line current in at least one historical line current sequence according to a preset current matching strategy, and correct the target line current according to the at least one historical initial line current to obtain the target initial line current.

[0048] In this step, at least one historical line current sequence under the same season, weather conditions, and load type is extracted from a preset historical database. The sequence length of the historical line current sequence is equal to the sequence length of the line current sequence. Each line current in the line current sequence is set in a preset two-dimensional coordinate system to obtain each line current coordinate point. The coordinate points of each line current are connected sequentially to obtain the line current curve. The horizontal axis of the two-dimensional coordinate system represents the acquisition time, and the vertical axis represents the current value of the line current. At least one historical line current sequence is set in the preset two-dimensional coordinate system to obtain at least one historical line current curve. The curve similarity between the line current curve and each historical line current sub-curve is calculated, and the top K target historical line current sub-curves with the highest curve similarity are selected, where K is an integer greater than or equal to 1. For each selected target historical line current sub-curve, the historical line current that has the same relative position as the target line current in the line current sequence is extracted as a historical initial line current, resulting in K historical initial line currents.

[0049] It should be noted that the calculation of the curve similarity between the line current curve and each historical line current sub-curve is specifically as follows:

[0050] Key points, including current zero-crossing points and local extrema, are detected in both the line current curve and historical line current sub-curves. Based on the detected key point locations, a weight matrix is ​​generated corresponding to the sampling points of both curves, where the matrix elements corresponding to key points are assigned a preset weight value greater than 1. The Euclidean distance between each sampling point on the two curves is calculated, forming a local cost matrix. The weight matrix and the local cost matrix are combined, and a dynamic programming algorithm is used to calculate the minimum cumulative weighted cost from the start point to the end point of the curve, which serves as the weighted dynamic time warping distance. ;

[0051] Feature point sequences, consisting of zero-crossing points and peak points arranged in time sequence, are extracted from the line current curve and historical line current sub-curves, respectively. Each feature point is recorded for its type, time index, and amplitude. The time correspondence determined by dynamic time warping is used to align the feature point sequences of the two curves. For each aligned feature point pair, the sum of the absolute values ​​of their time index differences and the average value of their relative amplitude errors are calculated. Based on the sum of the time index differences and the average value of the relative amplitude errors, the matching degree of the feature point sequence is calculated using an exponential decay function. ;

[0052] The line current curve and historical line current sub-curves are divided into multiple overlapping short time periods, each lasting 2 to 3 power frequency cycles. A Discrete Fourier Transform is performed on each short time period to extract its fundamental amplitude, third harmonic amplitude, and fifth harmonic amplitude, and the percentages of the third and fifth harmonic amplitudes relative to the fundamental amplitude are calculated. For each pair of short time periods corresponding to the two curves, the cosine similarity between the eigenvectors formed by the fundamental amplitude and the harmonic percentages is calculated. The arithmetic mean of the cosine similarities for all pairs of short time periods is taken to obtain the short-time frequency domain feature consistency score. ;

[0053] Based on weighted dynamic time warping distance Matching degree and short-time frequency domain feature consistency score According to the formula The curve similarity is calculated, where The preset fusion weight coefficients, and .

[0054] In one specific embodiment, in step S102, at least one historical initial line current corresponding to the target line current is matched in at least one preset historical line current sequence according to a preset current matching strategy, and the target line current is corrected based on the at least one historical initial line current to obtain the target initial line current. This step aims to calibrate the currently selected reference current (target line current) using historical normal operation data, eliminating the influence of inherent sensor bias and slow environmental drift, thereby obtaining a more accurate and stable "target initial line current" as the final analysis benchmark. The specific implementation process is as follows:

[0055] The historical database stores the long-term current time-series data of the target line and synchronously records metadata such as the collection time, season, weather type (e.g., sunny, rainy, snowy) and total system load level for each data point.

[0056] Based on the acquisition time of the "target line current" obtained in step S101, retrieve all historical records in the historical database that meet the same season (e.g., all are summer), the same weather conditions (e.g., all are sunny), and similar load types (e.g., all are "normal daytime load").

[0057] From these records, historical data segments of the exact same length as the "line current sequence" used in the current analysis are extracted to form multiple "historical line current sequences".

[0058] The current line current sequence is plotted as a continuous line current curve with time as the horizontal axis and current amplitude as the vertical axis. Similarly, each retrieved historical line current sequence is plotted as a corresponding historical line current curve.

[0059] A composite similarity algorithm is used to accurately measure the degree of matching between the current curve and each historical curve, thereby obtaining the historical initial line current.

[0060] Furthermore, based on the similarity between at least one historical initial line current and the target line current, an adaptive weighting is applied to at least one historical initial line current to obtain a current weighting factor corresponding to the at least one historical initial line current. The expression for calculating the current weighting factor is as follows:

[0061] ,

[0062] ,

[0063] In the formula, In order to be with the first Current weighting factor corresponding to the initial historical line current. For the first Similarity between the initial historical line current and the target line current. For the first Similarity between the initial historical line current and the target line current. The timestamp for the target line current. For the first Timestamp of the initial historical line current. For time normalization factor, Let be the characteristic vector of the target line current. For the first The characteristic vector of the initial historical line current. The time decay constant;

[0064] The baseline historical initial line current is obtained by weighting and summing at least one historical initial line current according to each current weighting factor.

[0065] The target line current is obtained by fusing the target line current with the baseline historical initial line current.

[0066] In another specific embodiment, the similarity between the historical initial line current and the target line current is not obtained by directly calculating the two isolated current scalar values, but through an indirect mapping process.

[0067] The core logic is that the representativeness and reliability of a current point (historical initial line current) are determined by the similarity between its overall operating state (i.e., its historical line current curve) and the current target operating state (i.e., the line current curve where the target line current is located).

[0068] Therefore, the process of obtaining the similarity between the historical initial line current and the target line current is as follows:

[0069] The composite similarity algorithm is used to calculate the overall similarity between the current complete line current curve (including the target line current) and each candidate complete historical line current curve.

[0070] After the system selects the top K most similar historical curves based on curve similarity, it needs to extract a specific current value from each curve as the historical initial line current.

[0071] The extraction location is determined based on the time index or relative position of the target line current within the current curve sequence. For example, the target line current might be the nth sampling point in the entire current sequence.

[0072] The current value extracted from the same position n of the kth similar historical curve is identified as the kth historical initial line current. .

[0073] At this time, the The similarity between the initial historical line current and the target line current is directly defined as the first... The historical curve of the initial line current is similar to the current curve.

[0074] In summary, firstly, by introducing a composite similarity algorithm that integrates weighted DTW, key feature point alignment, and short-time frequency domain analysis, the accuracy and robustness of historical data matching are revolutionaryly improved. This method not only focuses on the overall waveform trend but also ensures accurate alignment of core electrical features such as zero-crossing points and peak values, as well as harmonic spectra. This results in a high degree of electrical consistency between the selected historical operating conditions and the current state, far exceeding traditional methods based on Euclidean distance or simple correlation coefficients. Secondly, the adaptive weighted correction strategy based on similarity and temporal proximity can intelligently integrate the most relevant and reliable historical information to dynamically generate an optimized benchmark that reflects both long-term operational statistical patterns and closely approximates the current instantaneous state. This effectively overcomes systematic errors caused by sensor calibration drift and changes in environmental temperature and humidity, significantly improving the accuracy, representativeness, and stability of the target initial line current as the benchmark for subsequent analysis. Finally, the target initial line current provides an extremely reliable and clean "ruler" for subsequent instantaneous anomaly detection and high-dimensional spatial analysis, fundamentally reducing the risk of misjudgment and missed judgment due to inaccurate benchmarks, and simultaneously enhancing the detection sensitivity and diagnostic specificity of the entire anomaly analysis.

[0075] Step S103: Subtract the other line currents in the line current sequence from the target initial line current to obtain at least one instantaneous current change, and sort the at least one instantaneous current change based on time order to obtain an instantaneous current change sequence.

[0076] In this step, the absolute difference between the amplitude of other line currents and the amplitude of the target initial line current is calculated to obtain the instantaneous current change. The other line currents are the line currents in the line current sequence excluding the target line current. Based on the acquisition time of each other line current, the corresponding instantaneous current changes are sorted to obtain the instantaneous current change sequence.

[0077] By comparing each sampling point (other line currents) with a high-precision, high-stability benchmark (target initial line current) intelligently corrected from historical data, the inherent fluctuations and load changes during normal line operation are effectively removed, highlighting subtle and genuine abnormal deviations. Secondly, the generated instantaneous current change sequence transforms complex current waveform information into a one-dimensional deviation time-series signal. This greatly simplifies the processing objects of subsequent algorithms, providing a direct and clear input for rapid threshold judgment and continuity analysis (such as identifying discontinuous and questionable instantaneous current changes in step S104).

[0078] Step S104: Determine whether there is at least one non-continuous questionable instantaneous current change in the instantaneous current change sequence. A questionable instantaneous current change is an instantaneous current change that is greater than a preset threshold. A non-continuous questionable instantaneous current change is defined as a questionable instantaneous current change when the number of consecutively adjacent questionable instantaneous current changes is less than a preset number threshold.

[0079] In one specific embodiment, after determining whether there is at least one non-continuous suspected instantaneous current change in the instantaneous current change sequence, if there is no at least one non-continuous suspected instantaneous current change, then it is determined whether there are M or more consecutive suspected instantaneous current changes in the instantaneous current change sequence.

[0080] If there are M or more consecutive questionable instantaneous current changes, it is determined that there is a persistent line fault in the line under test during the current time period.

[0081] If there are no consecutive M or more suspicious instantaneous current changes, then it is determined that there is no line fault in the line under test during the current time period.

[0082] In one specific embodiment, two key thresholds are preset: a change threshold and a consecutive number threshold (M). The change threshold is used to distinguish between normal fluctuations and suspected anomalies, and can be dynamically calculated based on the line's rated current and the statistical characteristics of historical operating data (such as mean and standard deviation). The consecutive number threshold M is usually set to an integer greater than 1 (e.g., M=5), representing the minimum number of consecutive anomalies required to determine a persistent fault.

[0083] The instantaneous current change sequence is scanned sequentially, and each element in the instantaneous current change sequence is compared with the change threshold.

[0084] If the instantaneous change in current exceeds the change threshold, then the instantaneous change in current at that moment is marked as a questionable instantaneous change in current.

[0085] Perform cluster analysis on the doubtful points in the record to identify which doubtful points appear continuously and find all continuous clusters of doubtful points. For example, if the indices [10, 11, 12, 15, 16] are doubtful points, two clusters are formed: Cluster A (continuous indices 10, 11, 12) and Cluster B (continuous indices 15, 16).

[0086] Define discontinuity: For each doubtful point, check the size of the cluster it belongs to (i.e., the number of continuous doubtful points in the cluster). If the size of the cluster is less than the preset continuous quantity threshold M, all the doubtful points within the cluster are defined as discontinuous doubtful instantaneous current change amounts. For example, if M = 3, all points in Cluster B (with a size of 2) are "discontinuous".

[0087] Define continuity: If the size of a cluster is greater than or equal to M, the cluster is marked as a persistent abnormal cluster, indicating the existence of an abnormal state that persists beyond the threshold.

[0088] There are discontinuous doubtful points: If at least one discontinuous doubtful instantaneous current change amount is detected (i.e., the size of at least one cluster < M), it is determined that there is a suspicion of instantaneous interference during the current time period. These isolated or transient abnormal points do not directly trigger a fault alarm. Instead, they are marked as "doubtful", and the corresponding original current data and timestamps are passed to step S105 for three-dimensional refined analysis based on the spatial attitude angle to finally confirm whether it is a real fault.

[0089] There are no discontinuous doubtful points, but there are persistent abnormal clusters: If no discontinuous doubtful points are detected, but there is at least one persistent abnormal cluster (cluster size M), it is directly determined that there is a persistent line fault (such as a short circuit, continuous overload, etc.) in the line under test during the current time period. At this time, a high-level fault alarm will be immediately triggered, and the complete fault recording data will be recorded. At the same time, corresponding protection control measures can be recommended or initiated.

[0090] There are no eligible doubtful points: If there are no continuous M or more doubtful instantaneous current change amounts, it is determined that there is no line fault in the line under test during the current time period, and the operating state is normal.

[0091] In summary, by setting a continuous quantity threshold M, isolated and transient out-of-threshold points caused by random electromagnetic noise, measurement pulse errors, etc., can be naturally filtered out and classified as "suspected" rather than "faults," thereby greatly reducing the false alarm rate and avoiding unnecessary malfunctions of protection devices. Secondly, this method establishes a clear and operable distinction between transient interference and real faults. The identification of persistent anomaly clusters can reliably capture the continuous current mutation characteristics exhibited by real faults such as short circuits and grounding, ensuring the sensitivity and timeliness of fault detection. Finally, the output of this step directly guides the optimized allocation of subsequent analysis resources: only suspected transient anomalies are sent to the computationally complex three-dimensional spatial analysis stage (step S105), while confirmed persistent faults directly enter the alarm and handling process. This hierarchical processing mechanism of "initial screening-precision judgment" significantly improves the real-time processing efficiency and resource utilization of the system while ensuring the overall analysis accuracy. It is especially suitable for online monitoring scenarios involving massive sampling data, laying a key technical foundation for building a highly reliable and adaptive intelligent line anomaly monitoring system.

[0092] Step S105: Obtain the target attitude angle information of the target sensor at the target acquisition time, and perform anomaly analysis on the at least one suspicious instantaneous current change that is discontinuous based on the target initial line current and the target attitude angle information to obtain the anomaly analysis result, wherein the target acquisition time is the acquisition time corresponding to the target initial line current.

[0093] In this step, the target sensor is equipped with attitude measurement units such as gyroscopes, which can measure the sensor's own spatial attitude angles (such as roll angle, pitch angle, and yaw angle) in real time. The target acquisition time is the moment when the target line current is acquired, and the target attitude angle information recorded at this time is used to establish the reference coordinate system for current decomposition.

[0094] Based on the target attitude angle information, the target initial line current is divided into X-axis initial current component, Y-axis initial current component and Z-axis initial current component.

[0095] Acquire a certain attitude angle information of the target sensor at a certain suspicious acquisition time, and divide the suspicious line current into X-axis suspicious current component, Y-axis suspicious current component and Z-axis suspicious current component according to the certain attitude angle information. The suspicious line current is the line current corresponding to a certain suspicious instantaneous current change, and the suspicious acquisition time is the acquisition time corresponding to a certain suspicious instantaneous current change.

[0096] Calculate the X-axis current deviation between the initial X-axis current component and the suspected X-axis current component, the Y-axis current deviation between the initial Y-axis current component and the suspected Y-axis current component, and the Z-axis current deviation between the initial Z-axis current component and the suspected Z-axis current component, respectively.

[0097] Determine whether the X-axis current deviation, Y-axis current deviation, and Z-axis current deviation are greater than preset deviation thresholds; if the X-axis current deviation, Y-axis current deviation, or Z-axis current deviation is greater than the preset deviation threshold, then it is determined that there is a line fault in the line under test at a certain suspicious acquisition time; if the X-axis current deviation, Y-axis current deviation, and Z-axis current deviation are all not greater than the preset deviation thresholds, then it is determined that there is no line fault in the line under test at a certain suspicious acquisition time.

[0098] In one specific embodiment, complete attitude angle information, including roll angle, is obtained from the target sensor (integrated with an IMU unit) at the target acquisition time (i.e., the acquisition time of the target line current in step S101). Pitch angle and yaw angle .

[0099] Based on this attitude angle and the pre-entered geographic direction vector of the traverse (e.g., the unit direction vector of the traverse in the northeast-northeast coordinate system obtained from the GIS system), a local orthogonal coordinate system (traverse coordinate system) is constructed with the traverse itself as the reference through coordinate transformation calculation. Wherein:

[0100] X-axis: Completely aligned with the direction of the conductor, representing the main direction of current transmission;

[0101] Y-axis: Perpendicular to the direction of the conductor in the horizontal plane;

[0102] Z-axis: Perpendicular to the plane formed by the X-axis and Y-axis, and generally points in the vertical direction.

[0103] Using the constructed conductor coordinate system transformation matrix, the target initial line current (a three-dimensional current vector) is decomposed into three orthogonal components in this coordinate system: the X-axis initial current component. Y-axis initial current component and Z-axis initial current component .

[0104] Iterate through each suspected instantaneous current change marked as discontinuous. For the j-th suspected instantaneous current change:

[0105] Obtain the corresponding original line current sampling value (the j-th suspected line current). ) and its collection time .

[0106] Extracting data from the sensor data stream at time [time] Recorded attitude angle information (at time...) Roll angle At any moment pitch angle At any moment Yaw angle ).

[0107] Because the sensor may sway due to wind vibrations, the attitude angle at the questionable moment differs from that at the reference moment. The system uses the attitude angle at this moment to recalculate the transformation relationship of the sensor relative to the traverse coordinate system at the current moment (even if the traverse direction is fixed, the rotation of the sensor body will change the correspondence between the measurement axis and the traverse coordinate system).

[0108] Using the recalculated transformation matrix at the current time, the j-th suspected line current is... Similarly, when decomposed into a conductor coordinate system, its three components are obtained: the X-axis suspected current component. Y-axis suspected current component and the suspected current component along the Z-axis .

[0109] Calculate the current deviation between the point of doubt and the reference point on each of the three coordinate axes:

[0110] ;

[0111] ;

[0112] ;

[0113] Each deviation value is compared with its corresponding preset deviation threshold. These thresholds can be set separately based on the statistical characteristics of each component during normal operation (such as mean and variance).

[0114] After performing the above analysis on all discontinuous suspicious points, the system summarizes the analysis results and generates an anomaly analysis report. The report clearly lists the time when the fault is finally confirmed, the corresponding abnormal component (such as "X-axis principal component overcurrent" or "Z-axis leakage current anomaly"), and the false alarm points that are judged as interference.

[0115] For points confirmed as faulty, the magnitude of the deviation and the component in which it is located can be further combined to help determine the type of fault (e.g., a grounding fault may be prominent on the Z-axis component).

[0116] This method extends traditional one-dimensional amplitude anomaly detection to the three-dimensional spatial domain by decomposing the current into three orthogonal directions—X (axial), Y (horizontal lateral), and Z (vertical lateral)—closely related to the physical structure of the line, greatly enriching the detection dimensions. For example, ground leakage current (mainly reflected in the Z-axis) or unbalanced circulating current caused by adjacent line coupling (mainly reflected in the Y-axis), which may be missed by traditional methods, are clearly exposed and evaluated separately in this method. Secondly, by dynamically compensating for changes in the sensor's own attitude, the uniformity and comparability of the current decomposition benchmark at different times are ensured, overcoming the problem of distortion of the physical meaning of measurement data due to installation point oscillation, making the three-dimensional decomposition results stable and reliable. Finally, based on a multi-dimensional independent threshold judgment mechanism, it can accurately judge the characterization characteristics of different types of anomalies on different components, significantly improving the recognition accuracy of complex anomalies (especially non-amplitude-dominated anomalies), while effectively suppressing misjudgments caused by transient pulse interference.

[0117] In summary, the method of this application firstly extracts the "target initial line current" from the original current sequence by introducing multi-dimensional stability assessment and historical data-driven adaptive correction. This target current is highly resistant to interference and accurately reflects the steady-state operating characteristics of the system. This solves the problem of reference drift and misjudgment caused by fixed thresholds or single-moment references in traditional methods, and establishes a high-precision and highly reliable dynamic reference for subsequent analysis. Secondly, by calculating instantaneous changes and implementing a continuous intelligent screening mechanism, random pulse interference is effectively eliminated, and suspected anomalies are accurately classified into instantaneous suspicious points and persistent faults. While ensuring fault detection sensitivity, the false alarm rate caused by noise is significantly reduced. For suspicious points, sensor attitude angle information is introduced to decompose the current vector into a physical three-dimensional space based on the conductor direction. This achieves a dimensional leap from traditional amplitude monitoring to spatial current state diagnosis, enabling anomalies such as ground leakage and unbalanced coupling, which are hidden in a single dimension, to be explicitly exposed and accurately located.

[0118] Please see Figure 2 The diagram shows a structural block diagram of a line current anomaly analysis system according to this application.

[0119] like Figure 2 As shown, the line current anomaly analysis system 200 includes an extraction module 210, a correction module 220, a sorting module 230, a judgment module 240, and an analysis module 250.

[0120] The extraction module 210 is configured to acquire the line current of the line under test within the current time period based on a preset target sensor, obtain a line current sequence, and extract a target line current from the line current sequence based on a preset current selection rule; the correction module 220 is configured to match at least one historical initial line current corresponding to the target line current in at least one preset historical line current sequence according to a preset current matching strategy, and correct the target line current based on the at least one historical initial line current to obtain a target initial line current; the sorting module 230 is configured to subtract the target initial line current from other line currents in the line current sequence to obtain at least one instantaneous current change, and sort the at least one instantaneous current change based on time order to obtain an instantaneous current change sequence; and the judgment module 230 is configured to... The interruption module 240 is configured to determine whether there is at least one non-continuous, questionable instantaneous current change in the instantaneous current change sequence. A questionable instantaneous current change is defined as an instantaneous current change greater than a preset threshold. A non-continuous questionable instantaneous current change is defined as a non-continuous one when the number of consecutively adjacent questionable instantaneous current changes is less than a preset threshold. The analysis module 250 is configured to acquire the target attitude angle information of the target sensor at the target acquisition time, and perform anomaly analysis on the at least one non-continuous, questionable instantaneous current change based on the target initial line current and the target attitude angle information to obtain anomaly analysis results. The target acquisition time is the acquisition time corresponding to the target initial line current.

[0121] It should be understood that Figure 2 The modules and references described in the document Figure 1 The steps described in the text correspond to those in the method described above. Therefore, the operations, features, and corresponding technical effects described above also apply to the method described in the text. Figure 2 The various modules in the document will not be described in detail here.

[0122] In other embodiments, the present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein when the program instructions are executed by a processor, the processor performs the line current anomaly analysis method in any of the above method embodiments.

[0123] In one embodiment, the computer-readable storage medium of the present invention stores computer-executable instructions, which are configured as follows:

[0124] The line current of the line under test is obtained based on the preset target sensor within the current time period, and the line current sequence is obtained. The target line current is extracted from the line current sequence based on the preset current selection rule.

[0125] According to a preset current matching strategy, at least one historical initial line current corresponding to the target line current is matched in at least one preset historical line current sequence, and the target line current is corrected according to the at least one historical initial line current to obtain the target initial line current.

[0126] The difference between the other line currents in the line current sequence and the target initial line current is calculated to obtain at least one instantaneous current change, and the at least one instantaneous current change is sorted according to time order to obtain an instantaneous current change sequence.

[0127] Determine whether there is at least one non-continuous questionable instantaneous current change in the instantaneous current change sequence. A questionable instantaneous current change is an instantaneous current change that is greater than a preset threshold. A non-continuous questionable instantaneous current change is defined as a questionable instantaneous current change when the number of consecutively adjacent questionable instantaneous current changes is less than a preset number threshold.

[0128] The target sensor acquires the target attitude angle information at the target acquisition time. Based on the target initial line current and the target attitude angle information, anomaly analysis is performed on at least one suspicious instantaneous current change that is discontinuous, and anomaly analysis results are obtained. The target acquisition time is the acquisition time corresponding to the target initial line current.

[0129] Computer-readable storage media may include a stored program area and a stored data area, wherein the stored program area may store an operating system and an application program required for at least one function; the stored data area may store data created based on the use of the line current anomaly analysis system, etc. Furthermore, the computer-readable storage medium may include high-speed random access memory, and may also include memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some embodiments, the computer-readable storage medium may optionally include memory remotely located relative to a processor, and this remote memory may be connected to the line current anomaly analysis system via a network. Examples of such networks include, but are not limited to, the Internet, corporate intranets, local area networks, mobile communication networks, and combinations thereof.

[0130] Figure 3 This is a schematic diagram of the structure of the electronic device provided in the embodiment of the present invention, such as... Figure 3As shown, the device includes a processor 310 and a memory 320. The electronic device may also include an input device 330 and an output device 340. The processor 310, memory 320, input device 330, and output device 340 can be connected via a bus or other means. Figure 3 Taking a bus connection as an example, the memory 320 is the computer-readable storage medium described above. The processor 310 executes various server functions and data processing by running non-volatile software programs, instructions, and modules stored in the memory 320, thereby implementing the line current anomaly analysis method described in the above embodiment. The input device 330 can receive input digital or character information and generate key signal inputs related to user settings and function control of the line current anomaly analysis system. The output device 340 may include a display screen or other display device.

[0131] The aforementioned electronic device can execute the method provided in the embodiments of the present invention, and has the corresponding functional modules and beneficial effects for executing the method. Technical details not described in detail in this embodiment can be found in the method provided in the embodiments of the present invention.

[0132] In one implementation, the above-described electronic device is used in a line current anomaly analysis system for a client application, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to:

[0133] The line current of the line under test is obtained based on the preset target sensor within the current time period, and the line current sequence is obtained. The target line current is extracted from the line current sequence based on the preset current selection rule.

[0134] According to a preset current matching strategy, at least one historical initial line current corresponding to the target line current is matched in at least one preset historical line current sequence, and the target line current is corrected according to the at least one historical initial line current to obtain the target initial line current.

[0135] The difference between the other line currents in the line current sequence and the target initial line current is calculated to obtain at least one instantaneous current change, and the at least one instantaneous current change is sorted according to time order to obtain an instantaneous current change sequence.

[0136] Determine whether there is at least one non-continuous questionable instantaneous current change in the instantaneous current change sequence. A questionable instantaneous current change is an instantaneous current change that is greater than a preset threshold. A non-continuous questionable instantaneous current change is defined as a questionable instantaneous current change when the number of consecutively adjacent questionable instantaneous current changes is less than a preset number threshold.

[0137] The target sensor acquires the target attitude angle information at the target acquisition time. Based on the target initial line current and the target attitude angle information, anomaly analysis is performed on at least one suspicious instantaneous current change that is discontinuous, and anomaly analysis results are obtained. The target acquisition time is the acquisition time corresponding to the target initial line current.

[0138] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of various embodiments or some parts of embodiments.

[0139] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for analyzing abnormal line currents, characterized in that, include: The line current of the line under test is obtained based on the preset target sensor within the current time period to obtain the line current sequence, and the target line current is extracted from the line current sequence based on the preset current selection rule. The line current sequence is divided into multiple consecutive time windows, and each time window contains a preset number of line currents. Calculate the multidimensional stability index of the current signal within each time window, and determine the window stability of each time window based on the multidimensional stability index. The multidimensional stability index includes the coefficient of variation of the current amplitude within the time window. The dispersion of the current phase angle within the time window Symmetry index of three-phase current within the time window and the total harmonic distortion of the current within the time window The expression for calculating the stability of the window is: ， In the formula, For window stability, This is the weighting factor for the coefficient of variation of the current amplitude within the time window. This is a weighting coefficient for the dispersion of the current phase angle within the time window. The weighting coefficients represent the symmetry index of the three-phase currents within the time window. This is the weighting coefficient for the total harmonic distortion rate of the current within the time window. ; Select the target time window with the highest window stability value, and calculate the comprehensive confidence level of each line current within the target time window. The expression is as follows: ， In the formula, Let be the overall confidence level of the i-th line current within the target time window. Let be the current amplitude of the i-th line current. The average current amplitude within the target time window. The standard deviation of the current amplitude within the target time window. Let be the phase angle of the i-th line current within the target time window. The average phase angle within the target time window; The line current with the highest overall confidence level is selected as the target line current. According to a preset current matching strategy, at least one historical initial line current corresponding to the target line current is matched in at least one preset historical line current sequence, and the target line current is corrected according to the at least one historical initial line current to obtain the target initial line current. The difference between the other line currents in the line current sequence and the target initial line current is calculated to obtain at least one instantaneous current change, and the at least one instantaneous current change is sorted according to time order to obtain an instantaneous current change sequence. Determine whether there is at least one non-continuous questionable instantaneous current change in the instantaneous current change sequence. A questionable instantaneous current change is an instantaneous current change that is greater than a preset threshold. A non-continuous questionable instantaneous current change is defined as a questionable instantaneous current change when the number of consecutively adjacent questionable instantaneous current changes is less than a preset number threshold. The target sensor acquires the target attitude angle information at the target acquisition time. Based on the target initial line current and the target attitude angle information, anomaly analysis is performed on at least one suspicious instantaneous current change that is discontinuous, and anomaly analysis results are obtained. The target acquisition time is the acquisition time corresponding to the target initial line current.

2. The method for analyzing abnormal line current according to claim 1, characterized in that, The step of matching at least one historical initial line current corresponding to the target line current in at least one preset historical line current sequence according to a preset current matching strategy includes: Extract at least one historical line current sequence under the same season, weather conditions, and load type from a preset historical database, wherein the sequence length of the historical line current sequence is equal to the sequence length of the line current sequence. Each line current in the line current sequence is set in a preset two-dimensional coordinate system to obtain each line current coordinate point, and each line current coordinate point is connected in sequence to obtain the line current curve. The horizontal axis of the two-dimensional coordinate system is the acquisition time, and the vertical axis of the two-dimensional coordinate system is the current value of the line current. The at least one historical line current sequence is set in a preset two-dimensional coordinate system to obtain at least one historical line current curve; Calculate the curve similarity between the line current curve and each historical line current sub-curve, and select the top K target historical line current sub-curves with the highest curve similarity, where K is an integer greater than or equal to 1; For each selected target historical line current sub-curve, extract the historical line current that has the same relative position as the target line current in the line current sequence, and use it as a historical initial line current to obtain K historical initial line currents.

3. The method for analyzing abnormal line current according to claim 2, characterized in that, The step of correcting the target line current based on the at least one historical initial line current to obtain the target initial line current includes: Based on the similarity between the at least one historical initial line current and the target line current, the at least one historical initial line current is adaptively weighted to obtain a current weighting factor corresponding to the at least one historical initial line current. The expression for calculating the current weighting factor is as follows: ， ， In the formula, In order to be with the first Current weighting factor corresponding to the initial historical line current. For the first Similarity between the initial historical line current and the target line current. For the first Similarity between the initial historical line current and the target line current. The timestamp for the target line current. For the first Timestamp of the initial historical line current. For time normalization factor, Let be the characteristic vector of the target line current. For the first The characteristic vector of the initial historical line current. The time decay constant; The reference historical initial line current is obtained by weighting and summing the at least one historical initial line current according to each current weighting factor. The target line current is obtained by fusing it with the reference historical initial line current.

4. The method for analyzing abnormal line current according to claim 1, characterized in that, After determining whether there is at least one discontinuous, questionable instantaneous current change in the sequence of instantaneous current changes, the method further includes: If there is no discontinuous at least one questionable instantaneous current change, then determine whether there are M or more consecutive questionable instantaneous current changes in the sequence of instantaneous current changes. If there are M or more consecutive suspicious instantaneous current changes, it is determined that the line under test has a persistent line fault during the current time period. If there are no consecutive M or more suspicious instantaneous current changes, then it is determined that there is no line fault in the line under test during the current time period.

5. The method for analyzing abnormal line current according to claim 1, characterized in that, The anomaly analysis of at least one suspicious instantaneous current change based on the target initial line current and the target attitude angle information yields the following anomaly analysis results: Based on the target attitude angle information, the target initial line current is divided into X-axis initial current component, Y-axis initial current component and Z-axis initial current component; Acquire a certain attitude angle information of the target sensor at a certain suspicious acquisition time, and divide a certain suspicious line current into an X-axis suspicious current component, a Y-axis suspicious current component, and a Z-axis suspicious current component based on the certain attitude angle information. The certain suspicious line current is the line current corresponding to a certain suspicious instantaneous current change, and the certain suspicious acquisition time is the acquisition time corresponding to a certain suspicious instantaneous current change. Calculate the X-axis current deviation between the initial X-axis current component and the suspected X-axis current component, the Y-axis current deviation between the initial Y-axis current component and the suspected Y-axis current component, and the Z-axis current deviation between the initial Z-axis current component and the suspected Z-axis current component, respectively. Determine whether the X-axis current deviation, the Y-axis current deviation, and the Z-axis current deviation are greater than a preset deviation threshold; If the X-axis current deviation, the Y-axis current deviation, or the Z-axis current deviation is greater than a preset deviation threshold, then it is determined that the line under test has a line fault at a certain suspicious acquisition time. If the X-axis current deviation, the Y-axis current deviation, and the Z-axis current deviation are all not greater than a preset deviation threshold, then it is determined that the line under test does not have a line fault at the suspected acquisition time.

6. A line current anomaly analysis system, characterized in that, include: The extraction module is configured to acquire the line current of the line under test within the current time period based on a preset target sensor, obtain a line current sequence, and extract the target line current from the line current sequence based on a preset current selection rule, wherein the line current sequence is divided into multiple consecutive time windows, and each time window contains a preset number of line currents. Calculate the multidimensional stability index of the current signal within each time window, and determine the window stability of each time window based on the multidimensional stability index. The multidimensional stability index includes the coefficient of variation of the current amplitude within the time window. The dispersion of the current phase angle within the time window Symmetry index of three-phase current within the time window and the total harmonic distortion of the current within the time window The expression for calculating the stability of the window is: ， In the formula, For window stability, This is the weighting factor for the coefficient of variation of the current amplitude within the time window. This is a weighting coefficient for the dispersion of the current phase angle within the time window. The weighting coefficients represent the symmetry index of the three-phase currents within the time window. This is the weighting coefficient for the total harmonic distortion rate of the current within the time window. ; Select the target time window with the highest window stability value, and calculate the comprehensive confidence level of each line current within the target time window. The expression is as follows: ， In the formula, Let be the overall confidence level of the i-th line current within the target time window. Let be the current amplitude of the i-th line current. The average current amplitude within the target time window. The standard deviation of the current amplitude within the target time window. Let be the phase angle of the i-th line current within the target time window. The average phase angle within the target time window; The line current with the highest overall confidence level is selected as the target line current. The correction module is configured to match at least one historical initial line current corresponding to the target line current in at least one historical line current sequence according to a preset current matching strategy, and to correct the target line current according to the at least one historical initial line current to obtain the target initial line current. The sorting module is configured to subtract the other line currents in the line current sequence from the target initial line current to obtain at least one instantaneous current change, and sort the at least one instantaneous current change based on time order to obtain an instantaneous current change sequence. The judgment module is configured to determine whether there is at least one non-continuous questionable instantaneous current change in the instantaneous current change sequence. The questionable instantaneous current change is an instantaneous current change that is greater than a preset threshold. A non-continuous questionable instantaneous current change is defined as a questionable instantaneous current change when the number of consecutively adjacent questionable instantaneous current changes is less than a preset number threshold. The analysis module is configured to acquire the target attitude angle information of the target sensor at the target acquisition time, and to perform anomaly analysis on at least one suspicious instantaneous current change that is discontinuous based on the target initial line current and the target attitude angle information, so as to obtain anomaly analysis results, wherein the target acquisition time is the acquisition time corresponding to the target initial line current.

7. An electronic device, characterized in that, include: At least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method according to any one of claims 1 to 5.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the method described in any one of claims 1 to 5.

Images