A short-circuit fault position analysis method for cable branch box

Abstract

This application provides a short-circuit fault location analysis method for cable branch boxes, comprising: obtaining the arrival time of the first wave and the reflection attenuation pattern information from the cable branch box, and screening out an initial fault branch candidate set; extracting feature vectors using a wavelet transform algorithm based on the reflection attenuation pattern of each branch in the initial fault branch candidate set, and constructing a transient current distribution map; identifying the grounding continuity of the shielding layer from the transient current distribution map, evaluating the reflection enhancement phenomenon appearing in the far-end branch, and obtaining the adjusted current distribution; converting the adjusted current distribution to the frequency domain using a Fourier transform algorithm, identifying the reflection peak caused by the impedance short-circuit point, and determining the preliminary fault distance location; and calculating the fault location coordinates based on the confirmed actual fault branch and its arrival time of the first wave and traveling wave propagation speed.

Description

Technical Field

[0001] This invention relates to the field of information technology, and in particular to a method for analyzing the location of short-circuit faults in cable branch boxes. Background Technology

[0002] In the field of power system operation and maintenance, fault location technology for cable branch boxes is particularly crucial, directly impacting power supply reliability and user safety. However, current fault location methods primarily rely on single-current intensity judgments, neglecting various interference factors in complex environments, such as differences in grounding conditions at branch joints. If the shielding layer of a branch has a high grounding impedance, indicating poor grounding conditions, the branch exhibits high-impedance termination characteristics at the node. According to traveling wave reflection theory, the voltage reflection coefficient of the incident wave at the high-impedance interface is positive, and the reflected wave and incident wave are superimposed with the same polarity, causing an abnormal increase in the signal voltage amplitude at that node. Conversely, if a branch has an actual short-circuit fault, the fault point forms a low-impedance path, allowing a large amount of incident wave energy to be transmitted to the fault point and absorbed and dissipated, resulting in a relatively weaker reflected signal amplitude. This difference leads to varying degrees of attenuation or enhancement of the current signal during transmission, rendering the seemingly simple judgment rule unreliable. Especially in multi-branch networks, complex signal reflections and interference often make it difficult for technicians to accurately pinpoint the specific location of the fault. The dynamic changes in signal reflection patterns within a branch box stem from the differences in physical characteristics of different branches, such as the state of the grounding shield, cable length, and contact quality at joints. These factors cause irregular attenuation or amplification of the signal during propagation. It is precisely because of the failure to effectively address these dynamic changes that fault location often falls into misjudgment. For example, in a multi-branch branch box, a distant branch may have a positive traveling wave reflection coefficient at its node due to a high grounding impedance of its shield. The superposition of reflected waves abnormally amplifies the signal amplitude, leading to the branch being mistakenly identified as the fault location. Conversely, the near-end fault branch, where a short circuit actually occurs, has a near-zero impedance at the short-circuit fault point, resulting in a voltage reflection coefficient of the traveling wave at that point that is close to zero. 1. The polarity of the reflected wave is opposite to that of the incident wave. The two cancel each other out at the branch node, resulting in a significantly lower voltage amplitude after superposition. At the same time, most of the incident wave energy is transmitted to the fault arc through the low-impedance path and dissipated by the arc resistance. The effective signal power returning to the node is already minimal. The double superposition of polarity reversal and energy dissipation ultimately causes the near-end fault branch to exhibit a reflected signal with low amplitude and inverted polarity at the node, which is difficult to identify as a fault feature. Summary of the Invention

[0003] This invention provides a short-circuit fault location analysis method for cable branch boxes, mainly including: The arrival time and reflection attenuation pattern of the first wave are obtained from the cable branch box, and the candidate set of initial fault branches is selected. Based on the reflection attenuation pattern of each branch in the initial fault branch candidate set, the wavelet transform algorithm is used to extract feature vectors and construct a transient current distribution map. The grounding continuity of the shielding layer is identified from the transient current distribution diagram, the enhanced reflection phenomenon in the far-end branch is evaluated, and the adjusted current distribution is obtained. Based on the adjusted current distribution, a Fourier transform algorithm is used to convert it to the frequency domain, identify the reflection peaks caused by impedance short circuit points, and determine the preliminary location of the fault distance. Based on the preliminary location of the fault distance, the reflected waveform of the corresponding spatial segment is captured, and the attenuation trend and polarity stability of the reflected waveform along the line direction are analyzed to identify the actual fault branch. Based on the confirmed actual fault branch, and combined with the arrival time of the first wave and the propagation speed of the traveling wave in the actual fault branch, the fault location coordinates are calculated. Compare the fault location coordinates with the physical topology of the branch to verify whether the reflected wave attenuation section covers the fault location coordinates. If they match, output the final fault location mark.

[0004] Furthermore, the step of obtaining the arrival time and reflection attenuation pattern information of the first wave from the cable branch box and filtering out the initial fault branch candidate set includes: Transient waveforms are collected from each branch node of the cable branch box. A bandpass filter is used to remove noise interference from the transient waveforms. The amplitude of the filtered traveling wave is detected along the time axis at the point where it first crosses the baseline. The time stamp of the point of the transition is used as the arrival time of the first wave of each branch. The waveform sequence within the time window after the first wave is extracted, and adjacent peaks in the waveform sequence are connected to form an envelope. The ratio of the amplitude of adjacent peaks and the falling slope of the envelope are extracted and combined to obtain the reflection attenuation pattern of each branch. The ratio of the amplitude of adjacent peaks and the descent slope are compared with the reference threshold obtained from the statistics of normally operating branches. If the ratio of the amplitudes deviates from the reference threshold and the first wave time falls within the near-end arrival window calculated according to the line length, then the branch is included in the initial fault branch candidate set.

[0005] Furthermore, the step of extracting feature vectors using wavelet transform algorithm and constructing a transient current distribution map based on the reflection attenuation pattern of each branch in the initial fault branch candidate set includes: The reflection attenuation pattern is decomposed into multiple layers using multi-scale wavelet transform. The time position, amplitude, and polarity sign of the modulus maxima are extracted from each decomposition layer. The modulus maxima parameters of each layer are spliced ​​and normalized along the frequency axis to obtain the feature vector of the branch. By comparing the timing of the first modulus maxima of different branches in the initial fault branch candidate set under the same frequency band, the arrival order of the traveling wave at the node of the cable branch box is determined, and the propagation direction of the incident wave is determined by the polarity sign, so as to obtain the arrival time, propagation direction and instantaneous current amplitude of each branch. A grid is arranged with the cable branch box node as the origin and the direction of each branch as the coordinate axis. The time trajectory of the maximum magnitude is drawn on the grid. Arrows and arrival sequence numbers are marked according to the propagation direction to obtain the transient current distribution map.

[0006] Furthermore, the step of identifying the grounding continuity of the shielding layer from the transient current distribution diagram, evaluating the enhanced reflection phenomenon occurring in the far-end branch, and obtaining the adjusted current distribution includes: Based on the amplitude trajectory and propagation direction markings of each branch in the transient current distribution diagram, the polarity of the incident wave component and the reflected wave component is compared, and the branches with the same polarity superposition are extracted as candidate branches for shielding grounding anomalies. The amplitude rise and spatial coordinates of the candidate branches for shielding grounding anomalies at the nodes are read to obtain the shielding layer grounding continuity performance index. Based on the grounding continuity performance index of the shielding layer, branches with spatial coordinates located in the far section are selected as far candidate branches. The amplitude rise of the far candidate branches is calculated according to the natural attenuation law of traveling waves along the line. The remaining amplitude after deducting attenuation is compared with the amplitude of the incident wave component of the far candidate branch to obtain the reflection enhancement coefficient. Based on the reflection enhancement coefficient, the amplitude trajectory of the corresponding far-end candidate branch in the transient current distribution diagram is scaled and corrected by the reciprocal of the reflection enhancement coefficient. The original amplitude trajectory is replaced with the corrected amplitude trajectory, and the amplitude trajectories of each branch are combined to obtain the adjusted current distribution.

[0007] Furthermore, the adjusted current distribution is converted to the frequency domain using a Fourier transform algorithm to identify reflection peaks caused by impedance short-circuit points and to determine the preliminary location of the fault distance, including: The amplitude trajectory of each branch is processed by Fast Fourier Transform to obtain the spectrum of the branch. For the reflection peaks that are equally spaced on the frequency axis in the spectrum, the frequency interval between adjacent reflection peaks is extracted to obtain the reflection peak feature sequence. The candidate distance is obtained by dividing the propagation speed of the traveling wave in the branch by twice the frequency interval. If the candidate distance falls within the physical length of the branch and the frequency interval repeats in the spectrum, the location of the impedance short circuit point is determined, and the preliminary fault distance is obtained.

[0008] Furthermore, based on the preliminary fault location, the reflected waveform of the corresponding spatial segment is captured, and the attenuation trend and polarity stability of the reflected waveform along the line direction are analyzed to confirm the actual fault branch, including: Based on the candidate distances of each candidate branch in the preliminary fault distance location, a spatial segment extending from the candidate distance as the center to both sides of a preset length is extracted from the amplitude trajectory of the candidate branch to obtain the reflected waveform sequence within the spatial segment; For the reflected waveform sequence, the amplitude and polarity sign are extracted point by point along the line direction to form an amplitude sequence and a polarity sequence arranged along the line direction; If the negative polarity sign in the polarity sequence appears continuously in the spatial segment, and the amplitude sequence shows a monotonically decreasing trend along the line direction, then the candidate branch is confirmed as the actual faulty branch.

[0009] Furthermore, the calculation of the fault location coordinates based on the confirmed actual fault branch, combined with the arrival time of the first wave and the propagation speed of the traveling wave in the actual fault branch, includes: The one-way propagation delay of the traveling wave is obtained by subtracting the arrival time of the first wave of the actual faulty branch from the fault occurrence time of the actual faulty branch. The fault distance is obtained by multiplying the traveling wave propagation speed by the one-way propagation delay of the traveling wave. The location point corresponding to the fault distance is marked along the actual fault branch from the node end to obtain the fault location coordinates.

[0010] Furthermore, the comparison of the fault location coordinates with the physical topology of the branch verifies whether the reflected wave attenuation section covers the fault location coordinates. If they match, the final fault location marker is output, including: The fault location coordinates are calculated by converting the branch direction in the physical topology of the branch to obtain the topological mapping point of the fault location coordinates on the branch. If the topology mapping point falls into the reflected wave attenuation section corresponding to the actual fault branch, then the final fault location marker is output.

[0011] The technical solutions provided by the embodiments of the present invention may include the following beneficial effects: This invention discloses a short-circuit fault location analysis method for cable branch boxes. It acquires the arrival time and reflection attenuation pattern of the first wave from the cable branch box using a traveling wave acquisition unit, filters out an initial candidate set of faulty branches, and extracts feature vectors based on the reflection attenuation pattern of each branch using a wavelet transform algorithm. A transient current distribution map is constructed, and the propagation direction, arrival order, and current amplitude changes of the electromagnetic transient in each branch after the fault are marked. From this, the grounding continuity of the shielding layer is identified, and the reflection enhancement phenomenon appearing in distant branches is evaluated, resulting in an adjusted current distribution. Fourier transform is then applied to the adjusted current distribution. The Fourier transform algorithm is used to convert the signal to the frequency domain, identify reflection peaks caused by impedance short circuits, and thus determine the initial fault location. Based on the initial location, the reflected waveform of the corresponding spatial segment is extracted, and its attenuation trend and polarity stability along the line direction are analyzed. If there is a continuous negative polarity reflection within the segment and the amplitude shows monotonically decreasing, the branch is confirmed as the actual fault branch. Finally, based on the confirmed actual fault branch, the fault location coordinates are calculated by combining the arrival time of the first wave and the propagation speed of the traveling wave. The physical topology is compared to verify whether the attenuation segment of the reflected wave covers the coordinate point. If they match, the final fault location marker is output. This invention organically integrates traveling wave morphology screening, wavelet feature extraction, transient current distribution map construction, Fourier reflection peak identification, and waveform attenuation trend verification, realizing accurate location of short-circuit faults in cable branch boxes and effectively improving the accuracy and reliability of fault diagnosis. Attached Figure Description

[0012] Figure 1 This is a flowchart of a short-circuit fault location analysis method for cable branch boxes according to the present invention.

[0013] Figure 2 This is a schematic diagram of a short-circuit fault location analysis method for cable branch boxes according to the present invention.

[0014] Figure 3 This is another schematic diagram of a short-circuit fault location analysis method for cable branch boxes according to the present invention. Detailed Implementation

[0015] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0016] like Figures 1-3 This embodiment of a short-circuit fault location analysis method for cable branch boxes may specifically include: Step S101: Obtain the arrival time and reflection attenuation pattern information of the first wave from the cable branch box, and filter out the initial fault branch candidate set.

[0017] Transient waveforms are collected from each branch node of the cable branch box. A bandpass filter is used to remove noise interference from the waveforms. The amplitude of the filtered traveling wave is detected along the time axis at the point where it first crosses the baseline, and the time stamp of this point is used as the arrival time of the first wave for each branch. Based on the arrival time of the first wave, a waveform sequence within the time window after the first wave is extracted. Adjacent peaks in the waveform sequence are connected to form an envelope. The ratio of the amplitudes of adjacent peaks and the descent slope of the envelope are extracted and combined to obtain the reflection attenuation pattern of each branch. For the reflection attenuation pattern, the amplitude ratio and descent slope are compared with reference thresholds obtained from statistics of normally operating branches. If the amplitude ratio deviates from the reference threshold and the arrival time of the first wave falls within the near-end arrival window calculated based on the line length, then the branch is included in the initial candidate set of faulty branches.

[0018] In one implementation, for fault location of cable branch boxes, the traveling wave acquisition typically covers the branch nodes of single-core cross-linked polyethylene cables in a 10 kV distribution line. The cable branch box generally has 2 to 5 outgoing branches connected inside, each grounded through a copper shielding layer. The transient traveling wave signal at the node reflects the reflection characteristics of the line impedance discontinuity point.

[0019] Specifically, the traveling wave acquisition unit consists of a Rogowski coil or a broadband current transformer, which is respectively installed on the cable sheath grounding wire or main cable core wire of each branch in the cable branch box. The traveling wave sampling rate is usually configured to be above 1 MHz to ensure that the details of the transient waveform at the steep leading edge are not lost. At the moment of fault occurrence, the traveling wave acquisition unit is synchronously triggered and records the transient voltage and transient current waveforms of each branch into the data buffer.

[0020] Understandably, the acquired raw waveform contains a 50 Hz power frequency component, power line carrier signal, and environmental electromagnetic noise, which need to be purified before feature analysis. The passband range of the bandpass filter is generally set to 10 kHz to 500 kHz, covering the main energy spectrum of the fault traveling wave. Power frequency and high-frequency white noise outside the cutoff frequency are suppressed, and the output of the filter is the denoised pure traveling wave.

[0021] Specifically, the process of locating the arrival time of the first wave of the filtered traveling wave along the time axis involves first calculating the mean and standard deviation of the traveling wave amplitude sequence for each branch within a static interval before the fault occurs. The mean is used as the baseline, and the mean is multiplied by a certain number of standard deviations to form the threshold for exceeding the limit. Scanning forward along the time axis, when the amplitude exceeds the threshold for the first time with multiple consecutive sampling points, the corresponding sampling time stamp is recorded as the arrival time of the first wave for that branch.

[0022] It should be noted that the design of continuous multi-point judgment can avoid single-point noise spikes being mistaken for the first wave, and the multiple of the threshold is generally taken as 3 to 5 times.

[0023] In one possible implementation, the threshold multiple is fine-tuned based on the grounding status of the branch shielding layer. Branches with higher grounding impedance have higher background noise, and the corresponding multiple is appropriately increased to reduce false triggering. Further, based on the arrival time of the first wave, a waveform sequence within a time window following the first wave is extracted. The length of the time window is generally tens to hundreds of microseconds after the arrival of the first wave, covering the arrival intervals of the secondary and multiple reflected waves. Local maxima are searched point-by-point within the time window to obtain a set of adjacent peaks.

[0024] For example, adjacent wave peaks are connected sequentially in chronological order to form a traveling wave envelope. The ratio of two adjacent peaks in the envelope reflects the degree of attenuation of the reflected wave relative to the incident wave; the descent slope of the envelope characterizes the overall energy dissipation rate of the waveform over time. In a multi-branch branch box, the ratio of adjacent peaks in the far-end high-impedance reflection branch is usually higher, and the descent slope of the envelope is gentler; while in the near-end low-impedance fault branch, due to energy transmission dissipation, the ratio of adjacent peaks is lower, and the descent slope of the envelope is steeper. The combination of the amplitude ratio and the descent slope constitutes the reflection attenuation pattern of the branch.

[0025] Preferably, the reference threshold is obtained through historical statistics of normally operating branches. In the waveform data set of daily inspections and normal operation of the cable branch box, the ratio of amplitude and the descent slope of each branch under fault-free conditions are statistically analyzed over a long period to obtain their respective mean and dispersion range. The interval defined by the mean plus or minus two standard deviations is used as the reference threshold interval. Simultaneously, the near-end arrival window is calculated based on the time difference between the first wave and the reflected wave, combined with the traveling wave propagation speed. The traveling wave propagation speed v is generally taken as 0.7 times the speed of light in a vacuum. The arrival time t1 of the first wave and the arrival time t2 of the first reflected wave at the branch measurement point are extracted, and the time difference t2 minus t1 is calculated. This time difference is multiplied by the speed v and divided by 2 to obtain the fault distance. If this fault distance is less than 0.3 times the total length of the corresponding branch, the branch is marked as near-end arrival. For each branch, the reflection attenuation pattern is compared with the reference threshold range, and the calculated fault distance is compared with the near-end arrival window. If the amplitude ratio deviates from the reference threshold range and the fault distance falls within the near-end arrival window, then the branch is included in the initial fault branch candidate set. The candidate set provides input for subsequent feature vector extraction and transient current distribution analysis, thereby narrowing the investigation scope in complex multi-branch reflection environments.

[0026] Step S102: Based on the reflection attenuation pattern of each branch in the initial fault branch candidate set, the wavelet transform algorithm is used to extract feature vectors and construct a transient current distribution map.

[0027] Based on the reflection attenuation patterns of each branch in the initial candidate set of fault branches, a multi-scale wavelet transform is used to decompose the reflection attenuation patterns into multiple layers. The time position, amplitude, and polarity sign of the modulus maxima are extracted from each decomposition layer. The modulus maxima parameters of each layer are concatenated and normalized along the frequency axis to obtain the feature vector of the branch. Based on the time position of the modulus maxima in the feature vector, the arrival order of the traveling wave at the cable branch box node is determined by comparing the first modulus maxima times of different branches in the candidate set within the same frequency band. The polarity sign is used to determine the propagation direction of the incident wave, thus obtaining the arrival time, propagation direction, and instantaneous current amplitude of each branch. Based on the arrival time, propagation direction, and instantaneous current amplitude, a grid is arranged with the cable branch box node as the origin and the branch directions as the coordinate axes. The time trajectory of the modulus maxima amplitude is plotted on the grid, and arrows are labeled according to the propagation direction and the arrival order numbering to obtain the transient current distribution map.

[0028] In one embodiment, the cable branch box is located within a ring main unit or switching station of a 10 kV distribution network, internally connecting 3 to 5 outgoing branches. An initial candidate set of faulty branches is obtained from the preceding processing stage and includes several branches suspected of being faulty. Using the candidate set as input, the time-frequency feature vector of each branch is extracted, and a transient current distribution map reflecting the transient propagation pattern after the fault is constructed.

[0029] Specifically, multi-scale wavelet transform is a method for decomposing time signals into different frequency bands. Compared with fixed-window-length spectral decomposition methods, wavelet transform has higher frequency resolution in the low-frequency band and higher time resolution in the high-frequency band, and can simultaneously capture the steep leading edge and slowly decaying components of traveling wave signals.

[0030] In one possible implementation, the selected wavelet basis is a compactly supported orthogonal wavelet with a fourth-order vanishing moment. The number of decomposition layers is set according to the sampling rate and the target frequency band. When the sampling rate is 2 MHz, the number of decomposition layers is 5 to 7.

[0031] It should be noted that the process of multi-layer decomposition of the reflection attenuation pattern involves applying high-pass and low-pass filters to the signal layer by layer, followed by a 2x downsampling, to obtain the detail coefficient sequence and approximation coefficient sequence for that layer. Each decomposition layer corresponds to a sub-frequency band, with the first layer's detail coefficients corresponding to the highest frequency band and the fifth layer's approximation coefficients corresponding to the lowest frequency band. For each layer's detail coefficient sequence, the signal is scanned point by point along the time axis to locate the sample point where the absolute value exhibits a local maximum and is greater than several times the standard deviation of the layer's mean. This sample point is the modulus maximum of that layer. The time position of the modulus maximum is calibrated as relative time, the amplitude is taken as the absolute value of the detail coefficient at that sample point, and the polarity sign is taken as the positive or negative sign of the coefficient itself.

[0032] In one embodiment, the modulus maxima parameters of each decomposition layer are sequentially concatenated from high frequency to low frequency according to the layer hierarchy. A maximum value normalization method is used, where only the amplitude parameters in the concatenated sequence are divided by the maximum amplitude value within the sequence, while the time position and polarity sign remain unchanged, resulting in the feature vector of the branch. The feature vector contains several time positions, normalized amplitudes, and polarity sign combinations along the frequency axis, reflecting the attenuation rhythm of the branch in different frequency bands. Further, based on the feature vector, the analysis stage of the traveling wave timing of each branch is entered. The goal of this stage is to read the arrival order and propagation direction of the traveling waves at the branch box nodes from the feature vector. In one embodiment, the time position of the modulus maxima in the feature vector is analyzed, and the moment of the first modulus maxima of each candidate branch under the first layer detail coefficient is extracted as the first wave arrival time at the node for that branch. The first wave arrival times of each branch in the candidate set are arranged in ascending order of value, and the branch at the top is determined to be the earliest arriving.

[0033] For example, if a branch box is connected to three branches, denoted as Branch 1, Branch 2, and Branch 3, and the first wave arrival times at the node are 100 microseconds, 103 microseconds, and 108 microseconds respectively, then the arrival order is Branch 1, Branch 2, and Branch 3.

[0034] Understandably, the polarity symbol reflects the phase relationship of the traveling wave at the node. A positive polarity symbol corresponds to the traveling wave propagating from the node to the far end of the branch, i.e., the incident wave propagates outward along the branch; a negative polarity symbol corresponds to the traveling wave returning from the far end of the branch to the node, i.e., the reflected wave. Simultaneously, the absolute value of the detail coefficient at the first modulus maxima position is read from the eigenvector and used as the instantaneous current amplitude of that branch at the node. After obtaining the arrival time, propagation direction, and instantaneous current amplitude of each branch, the process proceeds to the construction stage of the transient current distribution map. This map visually presents the propagation pattern of the electromagnetic transient in the multi-branch environment of the branch box after a fault.

[0035] Specifically, a grid is arranged with the cable branch box node as the origin of the coordinate system and the actual direction of each branch as the coordinate axis.

[0036] In one embodiment, the branch box connects four branches. Four rays extend eastward, southward, westward, and northward on a two-dimensional grid plane, serving as branch coordinate axes. The length of each ray is proportionally determined according to the actual cable length of the corresponding branch, with each graduation on the axis corresponding to a 50-meter cable distance. Along each branch coordinate axis, the trajectory of the modulus maximum amplitude over time is plotted as a broken line in the normal direction of the axis. The lateral position of the trajectory reflects the spatial distance reached by the traveling wave, and the height of the trajectory reflects the instantaneous current amplitude at that distance.

[0037] For example, arrows are drawn at the starting point of each branch's coordinate axis according to the propagation direction. The arrows point from the origin outwards from the axis, indicating the direction of incident wave propagation; the arrows point from the outwards from the axis back to the origin, indicating the direction of reflected wave return. The starting points of the trajectories on each axis are labeled with numbers 1, 2, 3, and 4 according to the arrival order. The transient current distribution map centrally presents the propagation direction, arrival order, and amplitude changes of each branch, allowing for the comparison and identification of abnormal amplitude increases caused by high-impedance reflections and amplitude reductions caused by low-impedance faults on the same graph.

[0038] Step S103: Identify the grounding continuity of the shielding layer from the transient current distribution diagram, evaluate the enhanced reflection phenomenon in the far-end branch, and obtain the adjusted current distribution.

[0039] Based on the amplitude trajectory and propagation direction markings of each branch in the transient current distribution diagram, the polarity of the incident wave component and the reflected wave component is compared. Branches exhibiting superposition of the same polarity are extracted as candidate branches for shielding grounding anomalies. The amplitude rise and spatial coordinates of the candidate branches at the nodes are read to obtain the shielding layer grounding continuity performance index. Based on the shielding layer grounding continuity performance index, branches with spatial coordinates located in the far-end section are selected as far-end candidate branches. The amplitude rise of the far-end candidate branches is calculated according to the natural attenuation law of traveling waves along the line. The ratio of the remaining amplitude after attenuation to the amplitude of the incident wave component of the far-end candidate branch is used to obtain the reflection enhancement coefficient. Based on the reflection enhancement coefficient, the amplitude trajectory of the corresponding far-end candidate branch in the transient current distribution diagram is scaled and corrected according to the reciprocal of the reflection enhancement coefficient. The corrected amplitude trajectory replaces the original amplitude trajectory, and the amplitude trajectories of each branch are combined to obtain the adjusted current distribution.

[0040] In one implementation, the pre-processing stage yields a transient current distribution map, which marks the propagation direction, arrival sequence, and current amplitude trajectory of each branch. Based on the distribution map, the spurious reflection enhancement caused by differences in shielding layer grounding conditions is further removed, thereby obtaining an adjusted current distribution that reflects the true fault characteristics.

[0041] Specifically, shielding continuity refers to the integrity of the conductive loop formed between the cable's metallic shielding layer and the ground via the grounding lead-out. In a power distribution network, the shielding layer of each branch at the branch box node is connected to a common ground busbar via a grounding copper braid. Under normal circumstances, the grounding impedance is generally less than 4 ohms. If the grounding copper braid of a branch is loose, corroded, or the grounding electrode has poor conductivity, the branch will exhibit high impedance termination characteristics at the node. Traveling waves will undergo positive reflection at this interface, and the reflected wave will superimpose with the incident wave of the same polarity.

[0042] It should be noted that the process of comparing the polarity of the incident wave component and the reflected wave component of each branch in the transient current distribution diagram is to scan several microsecond intervals along the time axis on the coordinate axis of each branch after the arrival of the first wave, and take the sign of the point with the largest absolute value of the amplitude in the interval as the polarity sign of the incident wave component. Then, the reflected wave component that is transmitted back after the incident wave first arrives at the other end is truncated according to the time window, and the sign of the point with the largest absolute value in the window is extracted as the polarity sign of the reflected wave component.

[0043] In one possible implementation, if the polarity signs of the incident wave and the reflected wave are both positive or both negative, the branch is determined to have polarity superposition at the node and is marked as a candidate branch for shielded grounding anomaly. The instantaneous current amplitude at the superposition peak of the candidate branch at the node location on the distribution map is read, and the steady-state current amplitude of the branch under normal conditions before the fault is subtracted; the difference is the amplitude rise. Simultaneously, the spatial coordinates of the candidate branch along the branch coordinate axis on the transient current distribution map are recorded.

[0044] In one embodiment, the branch box connects to four branches. Two of these branches exhibit overlapping polarities and significant amplitude increases, and are thus included in the candidate branch set for shielded grounding anomalies. The corresponding amplitude increase and spatial coordinates together constitute the shielded layer grounding continuity performance index. Furthermore, based on the shielded layer grounding continuity performance index, the reflection enhancement of the distant candidate branches is evaluated, with the goal of separating the false rise caused by high impedance reflection from the true fault characteristics.

[0045] It is understandable that the multiple branches connected by the branch box are physically divided into near-end and far-end sections. A far-end section generally refers to a section extending beyond half the total length of the line from the branch box node. For example, in a branch with a total length of 600 meters, the section more than 300 meters from the node is classified as far-end. The spatial coordinates of each candidate branch in the shielding layer grounding continuity performance index are checked. If the spatial coordinates fall within the far-end section, the candidate branch is selected as a far-end candidate branch. In one embodiment, the natural attenuation law of traveling waves along the line refers to the physical law that the amplitude of the traveling wave decreases exponentially with the propagation distance.

[0046] Specifically, the amplitude attenuation of a traveling wave in a uniform line follows an exponential relationship with an attenuation constant α as the parameter. This attenuation constant α is determined by the conductor resistance, insulation loss, and frequency of the line, and typically ranges from 0.05 to 0.15 nanops per kilometer for a 10 kV cross-linked polyethylene cable. The amplitude rise of the distant candidate branch is calculated by extrapolating the amplitude rise observed at the node to the distant end of the branch according to the exponential attenuation law, and then reconstructing it based on the actual incident wave component amplitude at that location.

[0047] Specifically, let A be the amplitude rise at the node, L be the propagation distance from the far end of the candidate branch to the node, and α be the attenuation constant. Then, after natural attenuation according to the exponential attenuation law, the remaining amplitude is A multiplied by e to the power of negative α multiplied by L. The ratio of the remaining amplitude to the amplitude of the incident wave component of the far-end candidate branch is defined as the reflection enhancement coefficient of that branch.

[0048] In one embodiment, the reflection enhancement coefficient is greater than 1, indicating that there is additional reflection superposition caused by high impedance termination at the node of the branch. Further, based on the reflection enhancement coefficient, a local correction step is performed on the transient current distribution map.

[0049] Specifically, the process of correcting the amplitude trajectory of the corresponding far-end candidate branch in the transient current distribution diagram is to divide the amplitude value at each scale of the original broken line trajectory on the branch coordinate axis by the reflection enhancement coefficient to obtain the corrected amplitude at that scale, and replace the broken line height at that scale with the corrected amplitude.

[0050] Preferably, the correction process only affects the magnitude of the amplitude and does not change the polarity direction or time coordinate of the broken line.

[0051] In one embodiment, the reflection enhancement coefficient is set to 1.6. Then, the height of the broken line at each scale on the branch coordinate axis is reduced by a multiple of 0.625, and the reflection superposition peak that was originally abnormally raised falls back to the amplitude level close to the real reflected wave.

[0052] For example, after all remote candidate branches have undergone amplitude scaling correction according to the aforementioned reflection enhancement coefficient, the corrected amplitude trajectory is reassembled with the amplitude trajectories of the remaining uncorrected branches according to the original branch coordinate axes to obtain the adjusted current distribution. The adjusted current distribution eliminates the false rise caused by high-impedance termination, allowing the amplitude reduction characteristics of faulty branches with genuine low-impedance short circuits to be preserved and clearly presented in the distribution map.

[0053] Step S104: For the adjusted current distribution, the Fourier transform algorithm is used to convert it to the frequency domain, identify the reflection peak caused by the impedance short circuit point, and determine the preliminary location of the fault distance.

[0054] Based on the adjusted current distribution, a Fast Fourier Transform (FFT) is applied to the amplitude trajectory of each branch to obtain the branch's spectrum. For the equally spaced reflection peaks on the frequency axis of the spectrum, the frequency intervals between adjacent reflection peaks are extracted to obtain a reflection peak feature sequence. Based on the frequency intervals in the reflection peak feature sequence, the propagation speed of the traveling wave in the branch divided by twice the frequency interval is used as a candidate distance. If the candidate distance falls within the physical length of the branch and the frequency interval repeats in the spectrum, the impedance short-circuit point location is determined, and the initial fault distance is obtained.

[0055] In one implementation, the adjusted current distribution has eliminated spurious reflection enhancement caused by differences in the grounding state of the shielding layer, wherein the amplitude trajectory of each branch reflects the true reflection superposition during the propagation of the traveling wave along that branch. Based on the adjusted current distribution, the distance coordinates of short-circuit impedance discontinuities are further located using frequency domain analysis.

[0056] Specifically, the fault traveling wave propagates along the line to the far end in the branch, encountering an impedance short-circuit point and generating a reflected wave that returns to the node. The reflected wave is then reflected again at the node and propagates to the far end, repeating this process multiple times to form superimposed reflections. In the time domain, this manifests as a periodic pulse train, and in the corresponding frequency domain spectrum, it appears as a set of equally spaced reflection peaks. The frequency interval between these reflection peaks is determined by the propagation time of the traveling wave along the line for one round trip; the shorter the propagation time, the larger the frequency interval.

[0057] In one possible implementation, the amplitude trajectory of each branch in the adjusted current distribution is sampled and truncated, with the truncated length covering the transient process interval of 10 to 50 milliseconds after the fault. The truncated amplitude trajectory is then processed using a Fast Fourier Transform to obtain the spectrum of the branch.

[0058] It should be noted that the identification process for reflection peaks evenly distributed along the frequency axis in the spectrum involves first scanning the spectrum along the frequency axis to locate the maximum points where the amplitude exceeds the local baseline threshold as candidate peak positions. The local baseline threshold is determined by setting a sliding window with a width of 50 MHz, calculating the average and standard deviation of the spectral amplitude within this window, and using the average plus three times the standard deviation as the local baseline threshold at the center frequency of the window. Subsequently, the candidate peak positions are sorted by frequency from low to high to form a candidate peak sequence. Next, the frequency difference between two adjacent candidate peaks in the candidate peak sequence is calculated, and all adjacent frequency differences are summarized into an adjacent difference set. From the adjacent difference set, frequency differences with an absolute difference less than 0.5 MHz that recur are selected; these recurring frequency differences are determined as the frequency interval between reflection peaks, and the corresponding candidate peak is confirmed as a reflection peak. Further, the reflection peak positions, reflection peak amplitudes, and frequency intervals of each branch are combined sequentially along the frequency axis to obtain the reflection peak feature sequence of that branch. The reflection peak feature sequence, with frequency interval as the core element, carries the response characteristics of the branch to short-circuit reflection in the frequency domain.

[0059] It is understandable that the propagation speed of the traveling wave in the branch is determined by the distributed inductance and distributed capacitance of the cable.

[0060] Specifically, for 10 kV cross-linked polyethylene cables, the propagation wave speed is generally taken as 0.5 to 0.7 times the speed of light in vacuum, which can be obtained in advance from the cable parameter table. Based on the frequency interval in the characteristic sequence of the reflection peaks, the candidate distance is calculated according to the formula L = v divided by 2 multiplied by Δf, where L represents the candidate distance, v represents the propagation wave speed of the traveling wave in the branch, and Δf represents the frequency interval.

[0061] For example, if the frequency interval of a branch is 100 kHz and the propagation wave speed is 2 x 10^8 meters per second, then the candidate distance L is 1000 meters.

[0062] In one embodiment, the candidate distance is verified. If the candidate distance falls within the physical length range of the branch and the frequency interval appears repeatedly in the spectrum, then the position corresponding to the candidate distance is determined to be the impedance short-circuit point position. If the candidate distance exceeds the physical length range of the branch, then the candidate distance is considered a false solution and discarded.

[0063] Preferably, the location of each verified impedance short-circuit point is marked on the corresponding scale of the branch coordinate axis, and the impedance short-circuit point locations on all branches are combined to obtain the preliminary fault distance location. The preliminary fault distance location exists in the form of a binary pair of candidate branches and candidate distances, carrying the fault spatial coordinates derived from the frequency domain reflection characteristics.

[0064] Step S105: Based on the initial fault location, capture the reflection waveform of the corresponding spatial segment and analyze its attenuation trend and polarity stability along the line direction. If there is a continuous negative polarity reflection in the segment and the amplitude is monotonically attenuating, then the branch is confirmed as a real fault branch.

[0065] Based on the candidate distances of each candidate branch in the preliminary fault location, a spatial segment extending a predetermined length from the amplitude trajectory of the candidate branch to both sides, centered on the candidate distance, is extracted to obtain the reflected waveform sequence within the spatial segment. For the reflected waveform sequence, the amplitude and polarity sign are extracted point-by-point along the line direction to form an amplitude sequence and a polarity sequence arranged along the line direction. Based on the amplitude sequence and the polarity sequence, if the negative polarity sign in the polarity sequence continuously appears within the spatial segment and the amplitude sequence shows a monotonically decreasing trend along the line direction, then the candidate branch is confirmed as a real fault branch.

[0066] In one implementation, the initial fault distance is obtained from the preceding processing stage, carrying the fault spatial coordinates derived from the frequency domain reflection characteristics, where each candidate branch corresponds to a candidate distance. Based on these candidate distances, the polarity and attenuation pattern of the reflected waveform are further verified along the line direction to eliminate misjudgments caused by spurious periodic peaks in the spectrum and confirm the true fault branch.

[0067] Specifically, for each candidate branch in the preliminary fault distance location, a local segment centered on the candidate distance is extracted from the amplitude trajectory of the candidate branch as a spatial segment.

[0068] In one possible implementation, the preset lengths extending on both sides of the spatial segment are calculated based on the cable wave velocity and the sampling interval. For example, the cable distance is extended by 20 to 50 meters on both sides of the candidate distance. The amplitude trajectory segment corresponding to the segment is the reflected waveform sequence.

[0069] It should be noted that the reflected waveform sequence is arranged point by point with spatial distance as the horizontal axis and instantaneous traveling wave current as the vertical axis. Each sampling point corresponds to a spatial position along the line direction within the spatial segment, and the amplitude and polarity sign of the traveling wave component at that position are the values ​​of that sampling point.

[0070] Specifically, the reflected waveform sequence is scanned point by point along the line direction from the candidate distance to both sides. For each sampling point, the absolute value of the traveling wave component amplitude at that point is taken to form an amplitude sequence, and the sign bit of the traveling wave component at that point is taken to form a polarity sequence. Both the amplitude sequence and the polarity sequence are arranged in spatial order along the line direction, and their lengths are consistent with the total number of sampling points in the reflected waveform sequence.

[0071] In one embodiment, 50 sampling points are taken on each side of the spatial segment, and the amplitude sequence and the polarity sequence each contain 101 elements.

[0072] It is understandable that the voltage reflection coefficient of the traveling wave at the low-impedance short-circuit fault point approaches -1, and the polarity of the reflected wave is opposite to that of the incident wave. Near the fault point, the reflected wave continuously exhibits a negative polarity sign along the line direction. Further away from the fault point, the traveling wave energy dissipates through the fault arc, and the remaining amplitude monotonically decreases along the line direction. The continuous negative polarity and monotonically decreasing amplitude are dual characteristics of a true low-impedance short-circuit fault in terms of waveform. Furthermore, the polarity sequence and amplitude sequence are jointly verified. The proportion of negative polarity signs is calculated along the line direction for the polarity sequence. If the proportion of negative polarity signs exceeds a preset threshold, it is determined that negative polarity signs continuously appear in the spatial segment. The preset threshold is preferably 0.8. The adjacent differences are calculated point-by-point along the line direction for the amplitude sequence. If all adjacent differences are less than or equal to zero, the amplitude sequence is determined to have a monotonically decreasing trend.

[0073] Preferably, a small amount of noise perturbation is allowed in the amplitude sequence, and branches with a proportion of adjacent differences less than zero exceeding a preset threshold are also judged as monotonically decreasing, where the threshold is 0.9.

[0074] In one embodiment, if the proportion of negative polarity signs in a candidate branch reaches 0.92 and the proportion of negative adjacent differences reaches 0.95, then both conditions are satisfied.

[0075] For example, for a branch box with three branches, if the negative polarity proportion of the polarity sequence of candidate branch one is 0.30 and the negative proportion of the adjacent difference of the amplitude sequence is 0.50, then the dual conditions are not met, and it is judged as a non-real fault branch; the two indicators of candidate branch two are 0.92 and 0.95 respectively, and both dual conditions are met. Based on the verification results of the dual conditions, a real fault branch identifier is output for each candidate branch.

[0076] In one embodiment, candidate branch two is identified as the real faulty branch, and other candidate branches are eliminated, thereby locking a single branch as the location of the real fault and avoiding misjudgment caused by high impedance reflection at a distance or frequency domain spurious peaks.

[0077] Step S106: Based on the confirmed actual fault branch, and combining its first wave arrival time and traveling wave propagation speed, calculate the fault location coordinates.

[0078] The one-way propagation delay of the traveling wave is obtained by subtracting the arrival time of the first wave from the actual faulty branch and the fault occurrence time corresponding to the actual faulty branch. The fault distance is obtained by multiplying the traveling wave propagation speed by the one-way propagation delay based on the one-way propagation delay. The location point corresponding to the fault distance is then marked along the actual faulty branch from the node end to obtain the fault location coordinates.

[0079] In one implementation, the actual fault branch has been verified for the polarity and attenuation pattern of the reflected waveform. Based on the actual fault branch, the spatial location coordinates of the fault are further given, which are obtained by jointly calculating the arrival time of the first wave and the propagation speed of the traveling wave.

[0080] Specifically, the time of the fault occurrence is determined by the start-up trigger time of the traveling wave acquisition unit in the cable branch box.

[0081] In one possible implementation, the triggering mechanism of the traveling wave acquisition unit adopts a current jump threshold method. When the absolute value of the instantaneous current on the actual faulty branch first exceeds a preset start threshold, waveform recording is triggered, and the triggering moment is latched as the fault occurrence moment. The difference between the arrival time of the first wave on the actual faulty branch and the fault occurrence moment is the one-way propagation delay of the traveling wave, which physically represents the time taken for the initial traveling wave generated at the fault point to propagate to the cable branch box node. Further, the traveling wave propagation speed is pre-calibrated by the distributed inductance and distributed capacitance of the cable. For a 10 kV cross-linked polyethylene cable, the traveling wave propagation speed is typically taken as 0.5 to 0.7 times the speed of light in a vacuum, for example, 1.7 x 10^8 meters per second. Multiplying the traveling wave propagation speed by the one-way propagation delay of the traveling wave yields the fault distance of the actual faulty branch.

[0082] For example, if the one-way propagation delay of the traveling wave is 3 microseconds and the propagation speed of the traveling wave is 1.7 x 10^8 meters per second, then the fault distance is 510 meters. Extending the line segment corresponding to the fault distance from the node end to the far end of the branch box along the actual route of the faulty branch, with the node end as the origin, the endpoint of the resulting line segment is the location point corresponding to the fault distance. Outputting the latitude and longitude values ​​of this location point in the geographic coordinate system yields the fault location coordinates.

[0083] Step S107: Compare the fault location coordinates with the physical topology of the branch to verify whether the reflected wave attenuation section covers the coordinate point. If they match, output the final fault location mark.

[0084] Based on the fault location coordinates and the pre-established branch physical topology, the fault location coordinates are converted according to the branch direction to obtain the topological mapping point of the fault location coordinates on the branch. Based on the topological mapping point and the reflected wave attenuation section corresponding to the actual fault branch, if the topological mapping point falls within the reflected wave attenuation section, the final fault location marker is output.

[0085] In one implementation, the fault location coordinates are obtained by the preceding processing stage and are calibrated using latitude and longitude values. These coordinates are then compared with the branch's physical topology and the reflected wave attenuation zone to complete the final verification of the fault location process.

[0086] Specifically, the physical topology of the branch is pre-established by the power distribution network operation and maintenance database. The physical topology of the branch records the actual laying direction, latitude and longitude of turning points, segment length and connection relationship of the cable branch box and its connected branches. Each branch is represented in the physical topology of the branch as a broken line segment pointing from the starting node to the far end.

[0087] In one possible implementation, the fault location coordinates are calculated according to the polygonal path of the actual faulty branch in the branch's physical topology. The calculation process involves projecting the fault location coordinates onto the nearest point on the polygonal segment, and then accumulating the cable distance value from the node end along the polygonal direction based on the segment length of the actual faulty branch. The polygonal segment location point corresponding to this cable distance value is the topology mapping point. Further, the reflected wave attenuation section is obtained from the preceding processing on the actual faulty branch, and is a continuous interval consisting of the starting cable distance and the ending cable distance.

[0088] It is understood that the reflected wave attenuation zone corresponds to a spatial range in which negative polarity reflection continues and its amplitude monotonically decreases. The cable distance value corresponding to the topology mapping point is compared with the starting cable distance and the ending cable distance. If the cable distance value is greater than or equal to the starting cable distance and less than or equal to the ending cable distance, then the topology mapping point is determined to fall within the reflected wave attenuation zone.

[0089] For example, the final fault location marker is output for the topology mapping point that has passed the determination.

[0090] In one embodiment, the final fault location marker is carried in the form of a triple, which includes the branch number of the actual faulty branch, the latitude and longitude values ​​of the topology mapping point, and the cable distance value. It is superimposed on the distribution network geographic information layer and displayed as a red positioning pin icon, so that maintenance personnel can go to the site for emergency repairs.

[0091] The above description of the embodiments is only for the purpose of helping to understand the technical solutions and core ideas of this application; those skilled in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or make equivalent substitutions for some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A short-circuit fault location analysis method for a cable branch box, characterized by, The method includes: The arrival time and reflection attenuation pattern of the first wave are obtained from the cable branch box, and the candidate set of initial fault branches is selected. Based on the reflection attenuation pattern of each branch in the initial fault branch candidate set, the wavelet transform algorithm is used to extract feature vectors and construct a transient current distribution map. The grounding continuity of the shielding layer is identified from the transient current distribution diagram, the enhanced reflection phenomenon in the far-end branch is evaluated, and the adjusted current distribution is obtained. Based on the adjusted current distribution, a Fourier transform algorithm is used to convert it to the frequency domain, identify the reflection peaks caused by impedance short circuit points, and determine the preliminary location of the fault distance. Based on the preliminary location of the fault distance, the reflected waveform of the corresponding spatial segment is captured, and the attenuation trend and polarity stability of the reflected waveform along the line direction are analyzed to identify the actual fault branch. Based on the confirmed actual fault branch, and combined with the arrival time of the first wave and the propagation speed of the traveling wave in the actual fault branch, the fault location coordinates are calculated. Compare the fault location coordinates with the physical topology of the branch to verify whether the reflected wave attenuation section covers the fault location coordinates. If they match, output the final fault location mark.

2. The cable branch box-oriented short-circuit fault location analysis method according to claim 1, characterized by, The process of obtaining the arrival time and reflection attenuation pattern information of the first wave from the cable branch box and filtering out the initial fault branch candidate set includes: Transient waveforms are collected from each branch node of the cable branch box. A bandpass filter is used to remove noise interference from the transient waveforms. The amplitude of the filtered traveling wave is detected along the time axis at the point where it first crosses the baseline. The time stamp of the point of the transition is used as the arrival time of the first wave of each branch. The waveform sequence within the time window after the first wave is extracted, and adjacent peaks in the waveform sequence are connected to form an envelope. The ratio of the amplitude of adjacent peaks and the falling slope of the envelope are extracted and combined to obtain the reflection attenuation pattern of each branch. The ratio of the amplitude of adjacent peaks and the descent slope are compared with the reference threshold obtained from the statistics of normally operating branches. If the ratio of the amplitudes deviates from the reference threshold and the first wave time falls within the near-end arrival window calculated according to the line length, then the branch is included in the initial fault branch candidate set.

3. The cable branch box-oriented short-circuit fault location analysis method according to claim 1, characterized by, The step of extracting feature vectors using wavelet transform algorithm based on the reflection attenuation pattern of each branch in the initial fault branch candidate set and constructing a transient current distribution map includes: The reflection attenuation pattern is decomposed into multiple layers using multi-scale wavelet transform. The time position, amplitude, and polarity sign of the modulus maxima are extracted from each decomposition layer. The modulus maxima parameters of each layer are spliced ​​and normalized along the frequency axis to obtain the feature vector of the branch. By comparing the timing of the first modulus maxima of different branches in the initial fault branch candidate set under the same frequency band, the arrival order of the traveling wave at the node of the cable branch box is determined, and the propagation direction of the incident wave is determined by the polarity sign, so as to obtain the arrival time, propagation direction and instantaneous current amplitude of each branch. A grid is arranged with the cable branch box node as the origin and the direction of each branch as the coordinate axis. The time trajectory of the maximum magnitude is drawn on the grid. Arrows and arrival sequence numbers are marked according to the propagation direction to obtain the transient current distribution map.

4. The cable branch box-oriented short-circuit fault location analysis method according to claim 1, characterized by, The process of identifying the grounding continuity of the shielding layer from the transient current distribution diagram, evaluating the enhanced reflection phenomenon in the far-end branch, and obtaining the adjusted current distribution includes: Based on the amplitude trajectory and propagation direction markings of each branch in the transient current distribution diagram, the polarity of the incident wave component and the reflected wave component is compared, and the branches with the same polarity superposition are extracted as candidate branches for shielding grounding anomalies. The amplitude rise and spatial coordinates of the candidate branches for shielding grounding anomalies at the nodes are read to obtain the shielding layer grounding continuity performance index. Based on the grounding continuity performance index of the shielding layer, branches with spatial coordinates located in the far section are selected as far candidate branches. The amplitude rise of the far candidate branches is calculated according to the natural attenuation law of traveling waves along the line. The remaining amplitude after deducting attenuation is compared with the amplitude of the incident wave component of the far candidate branch to obtain the reflection enhancement coefficient. Based on the reflection enhancement coefficient, the amplitude trajectory of the corresponding far-end candidate branch in the transient current distribution diagram is scaled and corrected by the reciprocal of the reflection enhancement coefficient. The original amplitude trajectory is replaced with the corrected amplitude trajectory, and the amplitude trajectories of each branch are combined to obtain the adjusted current distribution.

5. The cable branch box-oriented short-circuit fault location analysis method according to claim 1, characterized by, The adjusted current distribution is then converted to the frequency domain using a Fourier transform algorithm to identify reflection peaks caused by impedance short-circuit points and to determine the preliminary location of the fault distance, including: The amplitude trajectory of each branch is processed by Fast Fourier Transform to obtain the spectrum of the branch. For the reflection peaks that are equally spaced on the frequency axis in the spectrum, the frequency interval between adjacent reflection peaks is extracted to obtain the reflection peak feature sequence. The candidate distance is obtained by dividing the propagation speed of the traveling wave in the branch by twice the frequency interval. If the candidate distance falls within the physical length of the branch and the frequency interval repeats in the spectrum, the location of the impedance short circuit point is determined, and the preliminary fault distance is obtained.

6. The cable branch box-oriented short-circuit fault location analysis method according to claim 1, characterized by, The preliminary location of the fault based on the fault distance, the extraction of the reflected waveform of the corresponding spatial segment, and the analysis of the attenuation trend and polarity stability of the reflected waveform along the line direction to identify the actual fault branch include: Based on the candidate distances of each candidate branch in the preliminary fault distance location, a spatial segment extending from the candidate distance as the center to both sides of a preset length is extracted from the amplitude trajectory of the candidate branch to obtain the reflected waveform sequence within the spatial segment; For the reflected waveform sequence, the amplitude and polarity sign are extracted point by point along the line direction to form an amplitude sequence and a polarity sequence arranged along the line direction; If the negative polarity sign in the polarity sequence appears continuously in the spatial segment, and the amplitude sequence shows a monotonically decreasing trend along the line direction, then the candidate branch is confirmed as the actual faulty branch.

7. The cable branch box-oriented short-circuit fault location analysis method according to claim 1, characterized by, Based on the confirmed actual fault branch, and combining the arrival time of the first wave and the propagation speed of the traveling wave in the actual fault branch, the fault location coordinates are calculated, including: The one-way propagation delay of the traveling wave is obtained by subtracting the arrival time of the first wave of the actual faulty branch from the fault occurrence time of the actual faulty branch. The fault distance is obtained by multiplying the traveling wave propagation speed by the one-way propagation delay of the traveling wave. The location point corresponding to the fault distance is marked along the actual fault branch from the node end to obtain the fault location coordinates.

8. The cable branch box-oriented short-circuit fault location analysis method according to claim 1, characterized by, The process involves comparing the fault location coordinates with the physical topology of the branch to verify whether the reflected wave attenuation section covers the fault location coordinates. If they match, the final fault location marker is output, including: The fault location coordinates are calculated by converting the branch direction in the physical topology of the branch to obtain the topological mapping point of the fault location coordinates on the branch. If the topology mapping point falls into the reflected wave attenuation section corresponding to the actual fault branch, then the final fault location marker is output.

Images