Discharge fault location system for transmission line based on feature analysis

Abstract

This invention belongs to the field of discharge fault location technology. Specifically, it discloses a transmission line discharge fault location system based on feature analysis, comprising: a topology construction module, a hierarchical modeling module, a fault triggering module, a consistency verification module, and a precise location module. The topology construction module establishes a topology with towers as nodes and line segments as edges and deploys monitoring terminals. The hierarchical modeling module establishes a multi-wave velocity layered propagation model based on medium parameters. The fault triggering module collects high-frequency data, determines the fault when the criteria are exceeded, and obtains the wavefront polarity and center frequency. The consistency verification module uses polarity and frequency to identify the direct wave and arrival time, determines the affected road segment through graph search verification, and then integrates time difference, amplitude attenuation, and waveform similarity to determine the abnormal road segment. The precise location module traverses candidate points in the abnormal road segment, calculates the theoretical arrival time based on the hierarchical model to locate the fault, and improves the accuracy of fault location through the hierarchical wave velocity model.

Description

Technical Field

[0001] This invention belongs to the field of discharge fault location technology and relates to a transmission line discharge fault location system based on feature analysis. Background Technology

[0002] Transmission lines are exposed to the elements for extended periods, making them prone to discharge faults. Rapid and accurate fault location is crucial for power grid safety. Traveling wave fault location technology is widely used due to its high accuracy and minimal impact from line parameters. However, the actual propagation speed of traveling waves is affected by the medium, frequency dispersion, and other factors, resulting in differences between line and ground modes and frequency-dependent variations. Furthermore, in complex topologies with multiple triggering terminals, wavefront sources are mixed, making direct wave identification difficult. Existing technologies often employ fixed wave velocities or simple waveform similarity determination, which are insufficient to address these issues.

[0003] For example, Chinese invention patent CN113687192B discloses a method for acquiring and locating line discharge signals, including the following steps: S100, acquiring traveling wave sample data at the time of the fault; S200, defining the time length of the traveling wave sample data and extracting waveform segments at the time of the fault; S300, determining the waveform similarity between terminals based on the waveform segments uploaded by each monitoring terminal. If negatively similar, the waveform segment proceeds to S400. If positively similar, the waveform segment does not proceed to S400 and is marked outside the monitoring range. If dissimilar, it is directly discarded; S400, identifying the inflection point of the effective waveform segment; S500, acquiring the relative wave velocity; S600, accurately locating the cable fault. This method, by selecting waveforms of a specific time length for relevant determination, can effectively improve the reliability of data during fault diagnosis, and achieves inflection point identification of the traveling wave waveform by means of fitting curves, which is more accurate than the inflection point identification of traditional methods.

[0004] However, the aforementioned existing technologies still have the following shortcomings: First, this method only obtains the relative wave velocity through waveform similarity judgment and inflection point identification, without distinguishing between line mode and ground mode traveling waves, and without generating multiple wave velocity layers based on the center frequency of the discharge signal, frequency-velocity correction curve, and preset frequency bandwidth and step size. This results in a large deviation between the theoretical propagation time and the actual arrival time, reducing the accuracy of fault location.

[0005] Second, this method relies solely on waveform similarity to filter valid waveforms, failing to utilize wavefront polarity and center frequency to identify the direct wavefronts of each terminal, nor employing graph search algorithms to verify the consistency of polarity and frequency to determine the set of affected road segments. Furthermore, this method lacks a weighted fusion strategy based on time difference matching, amplitude attenuation error, and waveform similarity. Consequently, in complex topologies, it becomes difficult to accurately distinguish between direct and indirect waves, leading to deviations in abnormal road segment identification and impacting the reliability of final fault location. Summary of the Invention

[0006] In view of this, in order to solve the problems mentioned in the background technology, a transmission line discharge fault location system based on feature analysis is proposed.

[0007] The objective of this invention can be achieved through the following technical solution: This invention provides a transmission line discharge fault location system based on feature analysis, including: a topology construction module, which establishes a topology relationship with towers as nodes and line segments between adjacent towers as edges, and deploys traveling wave monitoring terminals on the towers.

[0008] The layered modeling module establishes a multi-wave velocity layered propagation model based on topological relationships and medium type parameters of each line segment.

[0009] The fault triggering module uses the terminal to collect high-frequency discharge data in real time. When the signal amplitude or rate of change exceeds the start-up criterion, it determines that a discharge fault has occurred and obtains the wavefront polarity and center frequency.

[0010] The consistency verification module identifies the direct wavefronts and arrival times of each terminal based on polarity and center frequency. Combining the hierarchical propagation model, it uses graph search to verify the consistency of polarity and frequency to determine the set of affected road segments. Based on the degree of agreement between the theoretical arrival time difference and the measured arrival time difference, amplitude attenuation, and waveform similarity, it identifies abnormal road segments.

[0011] The precise positioning module traverses candidate positions on abnormal line segments with a set step size, calculates the theoretical direct wavefront arrival time from each candidate position to each terminal based on the hierarchical propagation model, and thus determines the final fault location.

[0012] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) The present invention constructs a topology graph with towers as nodes and line segments as edges, and deploys traveling wave monitoring terminals on the towers, abstracting the line into a graph model, thereby effectively handling the multi-terminal traveling wave propagation relationship under complex topology, and avoiding positioning deviation caused by insufficient consideration of line structure.

[0013] This invention obtains the wave velocities of the line mode and ground mode from the dielectric parameters, and generates multiple wave velocity layers based on the center frequency of the discharge signal, the frequency-velocity correction curve, and the preset bandwidth and step size. It calculates the theoretical propagation time of each line segment under each wave velocity layer and stores them in layers to form a multi-wave velocity layer propagation model. This model matches the velocity differences caused by frequency dispersion in traveling wave propagation, thereby providing a more accurate time delay reference for fault location.

[0014] This invention obtains the wavefront polarity and center frequency. The consistency verification module uses a reference wavefront as a reference to identify the direct wavefront of each terminal using polarity and center frequency. Then, it combines graph search to perform consistency verification of polarity and frequency to determine the set of affected road segments. This effectively distinguishes direct waves from reflected and refracted waves, avoids wavefront mismatch, and achieves accurate location of faulty road segments in complex multi-terminal triggering scenarios.

[0015] This invention improves the reliability of abnormal road segment identification by using multi-dimensional feature weighting fusion based on the consistency between the arrival time difference (OTD) and the measured OTD, amplitude attenuation, and waveform similarity. It integrates the three dimensions of feature quantity—OTD consistency, amplitude attenuation error, and waveform similarity—for weighted fusion.

[0016] This invention reduces the computational complexity of the localization phase by traversing candidate fault points on abnormal line segments with a set step size, narrowing the search range from the global topology to a single abnormal line segment. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a schematic diagram showing the connections of the various modules in the system of the present invention.

[0019] Figure 2 This is a schematic diagram showing the steps involved in establishing the multi-wave velocity layered propagation model of the present invention.

[0020] Figure 3 This is a schematic diagram illustrating the steps for determining the candidate line segment set in this invention. Detailed Implementation

[0021] 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, and 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.

[0022] This invention achieves precise location of discharge faults in complex line topologies by establishing a multi-wave velocity layered propagation model, identifying direct wavefronts based on polarity and center frequency, performing graph search consistency verification, and fusing multi-dimensional features to determine abnormal road segments. Specifically, the system first establishes a topological relationship based on tower-line segments. Then, based on line dielectric parameters and the center frequency of the discharge signal, a layered propagation model is constructed, including line mode, ground mode, and multiple frequency-related wave velocity layers. Next, the direct wavefronts of each monitoring terminal are identified using wavefront polarity and center frequency, and consistency verification is performed through graph search to determine the set of affected road segments. Then, multi-dimensional features such as time difference, amplitude attenuation, and waveform similarity are fused to determine abnormal line segments. Finally, candidate points are traversed on the abnormal road segment with a step size to achieve precise location. This solves the problems of single wave velocity, difficulty in identifying direct waves, and inaccurate location of abnormal road segments in existing technologies.

[0023] Please see Figure 1 As shown, the present invention provides a transmission line discharge fault location system based on feature analysis. The system includes: a topology construction module, a hierarchical modeling module, a fault triggering module, a consistency verification module, and a precise location module.

[0024] In the above, the hierarchical modeling module is connected to the topology construction module, the fault triggering module, the consistency verification module, and the precise positioning module, respectively. The consistency verification module is also connected to the fault triggering module and the precise positioning module, respectively.

[0025] The topology construction module establishes a topology relationship with towers as nodes and line segments between adjacent towers as edges, and deploys traveling wave monitoring terminals on the towers.

[0026] For example, establishing the topological relationship includes: Obtain the geographical location information of each tower and the line length between adjacent towers, and determine the connection relationship between adjacent towers according to the tower numbering order.

[0027] The poles are used as nodes, and the line segments between adjacent poles are used as edges.

[0028] Configure corresponding traveling wave monitoring terminal identifiers for each node to form a topology for fault location.

[0029] The layered modeling module establishes a multi-wave velocity layered propagation model based on topological relationships and medium type parameters of each line segment. It should be noted that the wave velocity layer in this invention refers to a hierarchy divided according to different traveling wave propagation speed values, with each wave velocity layer corresponding to a propagation speed value and its associated frequency band range; a multi-wave velocity layer refers to a set of wave velocity layers composed of multiple different wave velocity values, used to characterize the velocity stratification phenomenon caused by frequency dispersion effects during traveling wave propagation.

[0030] Please see Figure 2 As shown, exemplarily, the establishment of the multi-wavelength layered propagation model includes: Q1. Obtain basic parameters: Obtain the earth resistivity, the average height of the conductor to the ground, and the relative permittivity of the medium surrounding the conductor from the medium type parameters of the line segment.

[0031] It should be noted that the above parameters can be obtained in the following ways: the average height of the conductor above the ground is calculated based on the suspension point height and sag data in the tower design drawings; the earth resistivity is obtained by referring to the recommended value of soil resistivity standards based on the geological conditions of the line corridor; and the relative permittivity... For overhead lines, the dielectric constant of air is taken as... The methods used to obtain the above parameters are all conventional techniques used by those skilled in the art.

[0032] Q2. Calculate the propagation velocity of the traveling wave in the wire mode: Calculate the propagation velocity of the traveling wave in the wire mode based on the relative permittivity of the medium surrounding the conductor. Specifically, use the formula... ,in For the propagation speed of the linear mode traveling wave, For the speed of light, we obtain .

[0033] Q3. Calculate the propagation velocity of the ground mode traveling wave: Obtain the center frequency of the discharge signal, and then calculate the propagation velocity of the ground mode traveling wave based on the average height of the conductor to the ground, the resistivity of the earth, and the center frequency of the discharge signal. Specifically, the following empirical formula is used: In the formula The resistivity of the earth. The center frequency of the discharge signal. This represents the average height of the conductor above the ground. The ground model velocity is lower than the conductor velocity, with a typical range of values. This formula is an empirical formula, where the constant 60 has dimensions of [missing information]. This constant is based on the unit conversion factor in the derivation process of Carson's theory. Substituting it into the standard units of the above parameters makes the second term in the denominator a dimensionless quantity.

[0034] The empirical formula is applicable under the condition that the center frequency of the discharge signal is: Earth resistivity Average height of the conductor above the ground This formula is simplified based on Carson's ground return impedance theory. Within the above range, the calculation error is less than 5%, which meets the engineering positioning accuracy requirements. When the range is exceeded, the skin effect and soil stratification have a significant impact. In this case, the ground mode wave velocity should be obtained by using the frequency-velocity correction curve established based on the JMarti model in step Q4 of this paper.

[0035] Q4. Generate multiple wave velocity layers: Determine the frequency-velocity correction curve corresponding to the center frequency of the discharge signal. Using this center frequency as the center and a preset frequency bandwidth as the interval, select multiple frequency points within the interval according to a preset frequency step size. Calculate the linear mode velocity or ground mode velocity corresponding to each frequency point, forming multiple wave velocity layers with different wave velocity values. Specifically, for overhead lines, calculate the ground mode velocity corresponding to each frequency point; for cable lines, calculate the linear mode velocity corresponding to each frequency point.

[0036] After forming multiple wave velocity layers, the frequency range of each wave velocity layer is determined with the frequency point at which the wave velocity layer is generated as the center and the preset frequency step size described in step Q4 as the radius. If the center frequency of a certain wavefront falls within the frequency range of two or more wave velocity layers, the wave velocity layer to which the wavefront belongs is the one with the smallest absolute value of the difference between the center frequency and the center frequencies of each wave velocity layer; if the differences are the same, the wave velocity layer with the lower wave velocity is selected.

[0037] It should be noted that the frequency-velocity correction curve is used to characterize the functional relationship between the propagation velocity of traveling waves and frequency. For overhead lines, the velocity of ground-mode traveling waves is significantly affected by the resistivity of the earth and frequency, while the velocity of line-mode traveling waves is approximately constant (i.e., the speed of light) across the entire frequency band. Therefore, in this embodiment, the frequency-velocity correction curve is only established for ground-mode traveling waves.

[0038] For example, this curve can be pre-established in the following way: First, using the center frequency of the discharge signal... Based on this, the target frequency band is determined as follows: (Lower limit not lower than 1kHz, upper limit not higher than 1MHz), take points at logarithmic intervals within this frequency band, taking points every ten octaves. points ( ), each frequency point Distribution, in the formula The lower limit of the target frequency band, Secondly, the JMarti frequency-varying circuit model in ATP-EMTP is used to sweep the frequency at each frequency point, based on the phase constant. according to Calculate the ground mode wave velocity at each point; finally, use cubic spline interpolation to fit the discrete points to form a continuous curve, and control the fitting error within ±2%.

[0039] Among them, the preset frequency bandwidth The selection principle is to ensure coverage of the main energy frequency band of the discharge signal, that is, based on the center frequency. an interval with radius However, the frequency range should not exceed the target frequency band range determined in step Q4. The principle for selecting the preset frequency step size is to select at least 5 frequency points within the bandwidth to ensure that the number of velocity layers is sufficient to reflect the velocity variation characteristics with frequency. The above bandwidth and step size values ​​can be appropriately adjusted according to the signal spectrum width and accuracy requirements in practical applications. If the bandwidth is too small, the number of velocity layers will be insufficient; if it is too large, frequency components unrelated to the signal will be introduced. The smaller the step size, the more velocity layers there will be and the higher the model accuracy, but the computational load will also increase accordingly.

[0040] Q5. Calculate the theoretical propagation time: Obtain the line length of each edge in the topology, divide the length of each edge by the wave velocity value of each wave velocity layer, and use the quotient as the theoretical propagation time of each edge under each wave velocity layer.

[0041] Q6. Layered storage: The theoretical propagation time of each side under each wave speed layer is stored in layers according to the wave speed value from large to small, forming a multi-wave speed layered propagation model.

[0042] The fault triggering module uses the terminal to collect high-frequency discharge data in real time. When the signal amplitude or rate of change exceeds the start-up criterion, it determines that a discharge fault has occurred and obtains the wavefront polarity and center frequency.

[0043] For example, obtaining the wavefront polarity and center frequency includes: Polarity determination: Locate the starting moment of the traveling wavefront, and read the amplitudes of the sampling points before and after that starting moment. If the amplitude of the later sampling point is greater than that of the earlier sampling point, it is determined to be positive polarity; otherwise, it is determined to be negative polarity. The principle is that the signal transition direction at the starting moment of the traveling wavefront directly reflects the polarity of the wavefront, and it can be accurately determined by comparing the amplitudes at the instants before and after the transition.

[0044] It should be noted that the threshold comparison method is used to locate the wavefront start time: when the signal amplitude of three consecutive sampling points all exceed the preset start threshold (such as 3 times the average background noise), and the amplitude after the first sampling point that exceeds the threshold shows a monotonically increasing or decreasing trend, then the sampling point is taken as the start time of the wavefront.

[0045] Center frequency extraction: Based on the arrival time of the wavefront, sampled data of a preset length are extracted forward and backward. Fourier transform is performed on the extracted sampled data to obtain a spectrum. The frequency corresponding to the maximum amplitude in the spectrum is taken as the center frequency of the wavefront.

[0046] The consistency verification module identifies the direct wavefronts and arrival times of each terminal based on polarity and center frequency. Combining the hierarchical propagation model, it uses graph search to verify the consistency of polarity and frequency to determine the set of affected road segments. Based on the degree of agreement between the theoretical arrival time difference and the measured arrival time difference, amplitude attenuation, and waveform similarity, it determines the abnormal road segments.

[0047] For example, identifying the direct wavefront of each terminal and its arrival time includes: Among all the triggered terminals, the wavefront with the earliest arrival time that meets the preset validity check conditions is taken as the reference wavefront, and the wave velocity layer corresponding to its polarity and center frequency is taken as the reference polarity and reference wave velocity layer, respectively.

[0048] Specifically, the preset validity verification conditions include at least the following: amplitude condition: the wavefront amplitude exceeds a preset trigger threshold (for example, the trigger threshold can be 3 times the average background noise) to ensure that the wavefront has a sufficient signal-to-noise ratio to distinguish it from noise interference; waveform characteristic condition: the duration of the rising edge of the wavefront is less than a preset time threshold (for example, based on the sampling rate of the monitoring terminal, when the sampling rate is 1MHz, it can be set to a preset time threshold). (When the sampling rate changes, it can be adjusted accordingly at intervals of 2 to 3 sampling points) to ensure that the wavefront has the steep rise characteristics of a traveling wavefront, distinguishing it from slowly changing interference signals or far-end reflected waves. The specific values ​​of the above thresholds can be adjusted by those skilled in the art based on actual line conditions and terminal parameters.

[0049] If multiple waveforms satisfying the above conditions exist simultaneously, the one with the earliest arrival time is taken as the reference waveform. If no waveform satisfying the above conditions exists among all triggered terminals, the trigger is deemed invalid, the system is reset, and it waits for the next trigger to avoid errors in subsequent positioning results due to noise-induced false triggering or abnormal waveforms.

[0050] For each triggered terminal, wavefronts within their recorded time period are examined sequentially. The first wavefront that simultaneously satisfies polarity consistency and a center frequency within the reference wave velocity layer frequency range is identified as the direct wavefront of that terminal, and the sampling point time corresponding to this wavefront is recorded as the arrival time of the direct wavefront of that terminal. Polarity consistency means that the polarity of the direct wavefront of the triggered terminal is the same as the reference polarity of the reference terminal. If the fault point is located between two adjacent terminals, the polarities of the direct wavefronts on both sides of the fault point are opposite. In this case, the side where the reference terminal is located is taken as the standard, and only the polarity of terminals on the same side is required to be consistent; the opposite polarity of terminals on opposite sides does not affect the consistency determination.

[0051] Preferably, if there is no wavefront that meets the above conditions among all the triggered terminals, the current trigger is determined to be invalid, and the system is reset and waits for the next trigger. More preferably, in order to avoid the system from looping infinitely due to invalid triggers, a maximum number of retries is set, and the counter is incremented by 1 each time an invalid trigger occurs. When the counter exceeds the maximum number of retries, the system exits the current positioning process and issues an alarm signal.

[0052] Please see Figure 3 As shown, exemplarily, the set of affected road segments includes: W1. Determine the reference terminal: Among all triggered terminals, the terminal with the smallest direct wavefront arrival time is taken as the reference terminal. Based on the matching of the shortest path theoretical propagation time and the measured time difference of the reference terminal, the candidate line segment set is determined.

[0053] Furthermore, the determination of the candidate line segment set includes: W1-1. Take the node where the reference terminal is located as the reference node, and take the wave velocity layer corresponding to the center frequency of its direct wave as the current search wave velocity layer.

[0054] W1-2. In the multi-wavelength layered propagation model, starting from the reference node, a graph search algorithm is used to calculate the theoretical propagation time of the shortest path from the reference node to each node. Specifically, the theoretical propagation time of the line segment between adjacent nodes is used as the edge weight, and Dijkstra's algorithm is executed with the reference node as the source node to iteratively update the cumulative shortest path time of each node, finally obtaining the theoretical propagation time of the shortest path from the reference node to each node.

[0055] W1-3. The difference between the measured arrival time of the direct wave of each triggered terminal and the arrival time of the direct wave of the reference terminal is taken as the measured time difference.

[0056] W1-4. Using the measured time difference as the center and the allowable time deviation as the radius, determine the theoretical propagation time range. The allowable time deviation is used to compensate for time measurement errors and model approximation errors during the traveling wave propagation process. The value is determined based on the sampling interval of the monitoring terminal. And time synchronization accuracy Determine, calculate using the following formula: ,in ( (sampling rate) The timing synchronization error of the terminal (usually 100%) For example, when the sampling rate is The time synchronization accuracy is hour, Those skilled in the art can calculate the above formula based on the sampling rate and timing accuracy of the actual terminal. value.

[0057] W1-5. Nodes whose shortest path propagation time falls within this range are considered valid nodes. All edges traversed by the shortest path from the reference node to each valid node are extracted, and the line segments corresponding to these edges are used to form a candidate line segment set.

[0058] W2. Determine the theoretical arrival order set: For each candidate line segment in the candidate line segment set, according to the multi-wave speed layer-by-layer propagation model, select multiple discretized hypothetical fault points on the candidate line segment according to the preset spatial step size, determine the theoretical arrival order of each triggered terminal under each hypothetical fault point, and merge them after deduplication to form the theoretical arrival order set of the candidate line segment.

[0059] Specifically, the determination of the arrival order set of the direct wave theory includes: W2-1, Discrete sampling fault points: Select multiple discrete points as hypothetical fault points on the current candidate line segment at a preset step size (e.g., 5% to 10% of the segment length). The discrete points must include at least the two endpoints of the line segment.

[0060] W2-2. Calculate the theoretical arrival time at each point: For each assumed fault point, based on the multi-wave speed layered propagation model, calculate the theoretical direct arrival time of the traveling wave from that point to each triggered terminal.

[0061] W2-3. Determine the theoretical arrival order: For each hypothetical fault point, sort the theoretical arrival times of each terminal in chronological order to obtain a theoretical arrival order (i.e., the sequence of terminals) corresponding to that point.

[0062] W2-4. Collect all sequences: After deduplication, merge the theoretical arrival sequences obtained from all hypothetical fault points on the current candidate line segment to form a set of theoretical arrival sequences for that candidate line segment.

[0063] W3. Obtain the measured arrival order: Sort the measured arrival times of the direct waves of each triggered terminal in chronological order to obtain the measured arrival order.

[0064] W4. Consistency condition judgment: Condition 1 is based on the theoretical arrival order set of the candidate line segment being the actual arrival order; Condition 2 is based on the consistent polarity of the direct wave of each triggered terminal; and Condition 3 is based on the fact that the center frequency of the direct wave of each triggered terminal falls within the frequency range of the same wave speed layer.

[0065] W5. Construct the set of affected road segments: The candidate road segments that meet conditions 1, 2 and 3 together constitute the set of affected road segments.

[0066] For example, the determination of abnormal line segments includes: S1. Time Difference Matching Degree Calculation: For each candidate road segment in the affected road segment set, multiple hypothetical fault points are selected on the road segment according to a set step size. The theoretical time difference (i.e., the theoretical time difference of the traveling wave propagating from that point to the beginning and end terminals) corresponding to each hypothetical fault point is calculated and compared with the measured time difference of the beginning and end triggering terminals. The hypothetical fault point with the smallest absolute difference is recorded as the matching fault location, and the absolute value of this difference is used as the time difference matching degree of the candidate road segment. The smaller this value, the higher the degree of matching between the candidate road segment and the measured time difference.

[0067] It should be noted that if a candidate road segment has a traveling wave monitoring terminal at only one end (such as the end of the line), the difference between the theoretical time difference and the measured time difference cannot be calculated. This candidate road segment will not participate in the time difference matching calculation and will directly proceed to the subsequent evaluation stages of amplitude attenuation error and waveform similarity. If the calculated matching fault location exceeds the actual length range of the candidate road segment (e.g., less than 0 or greater than the segment length), the nearest endpoint of the road segment will be taken as the matching fault location, and the absolute value of the difference between the theoretical time difference and the measured time difference corresponding to that endpoint will be recalculated as the time difference matching degree.

[0068] S2. Amplitude attenuation error calculation: Based on the medium type parameters and the center frequency of the traveling wave, determine the amplitude attenuation coefficient per unit distance, and then use the measured amplitude of the reference terminal as the benchmark amplitude to calculate the amplitude attenuation error of the candidate line segment.

[0069] Specifically, the unit distance amplitude attenuation coefficient Approximate calculation using the following formula: In the formula The center frequency of the traveling wave, This is a proportionality coefficient related to conductor type and earth resistivity, with dimensions of... This can be obtained by looking up a table. For example, for commonly used steel-cored aluminum stranded wire (LGJ series), when the earth resistivity... hour, The range of values ​​is ;when hour, The range of values ​​is ;when hour, The range of values ​​is .

[0070] Furthermore, the calculation of the amplitude attenuation error of the candidate line segment includes: S2-1. Calculate the propagation distance of the traveling wave from the location of the matched fault to each triggered terminal. Combined with the amplitude attenuation coefficient per unit distance of the candidate line segment (Unit is) ), to obtain the theoretical amplitude of each triggered terminal. For example, the theoretical amplitude of each triggered terminal is calculated using the following formula: In the formula The number of the triggered terminal , The total number of terminals that were triggered. The measured amplitude is used as a reference terminal.

[0071] S2-2. Calculate the absolute value of the difference between the theoretical amplitude and the measured amplitude of each triggered terminal and sum them to obtain the amplitude attenuation error of the candidate line segment.

[0072] S3. Waveform similarity calculation: Using the waveform of the reference terminal as the reference waveform, perform cosine similarity calculation on the waveforms of other triggered terminals, and take the average of all similarities as the waveform similarity.

[0073] Specifically, the waveform similarity calculation includes: taking the start time of the direct wavefront of the reference terminal as the reference point, and extracting sampled data of a preset length both forward and backward as reference waveform segments; for each of the other triggered terminals, taking the start time of their respective direct wavefronts as the reference point, and extracting waveform segments of the same length as waveform segments to be compared. The preset length is taken as 3 to 5 times the period corresponding to the center frequency of the direct wavefront, and the same window function is applied to each waveform segment.

[0074] The reference waveform segment is denoted as a vector, and the waveform segment to be compared of the triggered terminal is denoted as a vector. The cosine similarity is calculated, and then the average cosine similarity of all triggered terminals (excluding the reference terminal) is calculated as the waveform similarity of the candidate line segment.

[0075] S4. Weighted Fusion and Judgment: The time difference matching degree, amplitude attenuation error and waveform similarity are weighted and fused according to preset weights to calculate the comprehensive confidence of each candidate line segment, and then the candidate line segment with the highest comprehensive confidence is selected as the abnormal line segment.

[0076] It should be noted that the formula for calculating the overall confidence level is as follows: In the formula To assess the overall confidence level, The normalized time difference matching degree This is the normalized amplitude attenuation error. For waveform similarity, , , These are weights for time difference matching, amplitude attenuation error, and waveform similarity, used to quantify the contribution of different feature quantities to the identification of abnormal road sections, such as... , , The normalization process can employ extreme value normalization, which involves dividing the time difference matching degree and the amplitude attenuation error by their maximum values ​​in the candidate road segments, respectively, to ensure that their values ​​fall within a certain range. .

[0077] By using weighted fusion to calculate the overall confidence level of each candidate road segment, on the one hand, the weight allocation can reflect the actual weight of the influence of time difference matching degree, amplitude attenuation error and waveform similarity on different dimensions of fault segment identification, and reflect the difference in contribution of each feature quantity in confidence assessment; on the other hand, multiple feature quantities with different dimensions can be unified into a dimensionless overall confidence index, which makes it easier to intuitively compare the merits of each candidate road segment.

[0078] The weights can be set based on historical fault data, simulation samples, or engineering experience. For example, typical discharge fault waveforms and measured abnormal road segment data of multiple lines can be collected, and the time difference consistency, amplitude attenuation error, and waveform similarity of each candidate road segment can be calculated respectively. Using the actual abnormal road segment as a label, the optimal contribution of each feature quantity can be fitted by logistic regression or principal component analysis. After normalization, the weights are converted into weights, and the sum of the weights is 1. Those skilled in the art can flexibly adjust the weights according to the actual line conditions and accuracy requirements.

[0079] The precise positioning module traverses candidate positions on the abnormal line segment with a set step size, and calculates the theoretical direct wavefront arrival time from each candidate position to each terminal according to the hierarchical propagation model, so as to determine the final fault location.

[0080] For example, determining the final fault location includes: Y1. Generate candidate fault points: Select multiple candidate fault points in the abnormal line segment according to a preset step size. Preferably, the preset step size should not be less than the minimum resolution distance of the system. The sampling rate and timing synchronization accuracy of the monitoring terminal are jointly determined, and calculated using the following formula: .

[0081] Y2. Calculate the theoretical arrival time: For each candidate fault point, based on the multi-wave velocity layer propagation model, calculate the theoretical direct arrival time of the traveling wave from the candidate fault point to each triggered terminal. The wave velocity value used should be selected from the wave velocity layer corresponding to the current abnormal line segment, specifically the wave velocity value of the reference wave velocity layer.

[0082] It should be noted that the calculation process for the theoretical direct wavefront arrival time is as follows: Candidate fault points... Positioned on the edge Above, set Distance from node The distance is The side length is wave speed is Using the theoretical propagation time between nodes obtained in advance by the multi-wave speed layered propagation model, the node propagation time was determined. To the terminal node transmission time ,node arrive transmission time Then the traveling wave from arrive The theoretical propagation time is: This propagation time is the theoretical arrival time of the direct wavefront.

[0083] Y3. Calculate the time deviation: Subtract the theoretical arrival time of each direct wavefront at each candidate fault point from the actual measured time corresponding to each terminal to obtain the arrival time deviation of each terminal at each candidate fault point.

[0084] Y4. Select the optimal fault point: Sum the absolute values ​​of the arrival time deviations of all terminals under each candidate fault point to obtain the total time deviation of the candidate fault point, and then select the candidate fault point with the smallest total time deviation as the final fault location point.

[0085] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0086] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0087] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0088] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0089] Finally, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A transmission line discharge fault location system based on feature analysis, characterized in that: The system includes: The topology construction module establishes a topology relationship with towers as nodes and line segments between adjacent towers as edges, and deploys traveling wave monitoring terminals on the towers. The layered modeling module establishes a multi-wave velocity layered propagation model based on topological relationships and medium type parameters of each line segment. The fault triggering module uses the terminal to collect high-frequency discharge data in real time. When the signal amplitude or rate of change exceeds the start-up criterion, it determines that a discharge fault has occurred and obtains the wavefront polarity and center frequency. The consistency verification module identifies the direct wavefronts and arrival times of each terminal based on polarity and center frequency. Combined with the hierarchical propagation model, it uses graph search to verify the consistency of polarity and frequency to determine the set of affected road segments. Based on the degree of agreement between the theoretical arrival time difference and the measured arrival time difference, amplitude attenuation, and waveform similarity, it identifies abnormal road segments. The precise positioning module traverses candidate positions on abnormal line segments with a set step size, calculates the theoretical direct wavefront arrival time from each candidate position to each terminal based on the hierarchical propagation model, and thus determines the final fault location.

2. The transmission line discharge fault location system based on feature analysis according to claim 1, characterized in that: The establishment of topological relationships includes: Obtain the geographical location information of each tower and the line length between adjacent towers, and determine the connection relationship between adjacent towers according to the tower numbering order; Using towers as nodes and line segments between adjacent towers as edges; Configure corresponding traveling wave monitoring terminal identifiers for each node to form a topology for fault location.

3. The transmission line discharge fault location system based on feature analysis according to claim 1, characterized in that: The establishment of the multi-wave velocity layered propagation model includes: The earth resistivity, average height of the conductor to the ground, and relative permittivity of the medium surrounding the conductor are obtained from the medium type parameters of the line segment. Calculate the propagation speed of the traveling wave in the line mode based on the relative permittivity of the medium surrounding the conductor; The center frequency of the discharge signal is obtained, and then the propagation speed of the ground mode traveling wave is calculated based on the average height of the conductor to the ground, the resistivity of the earth, and the center frequency of the discharge signal. Determine the frequency-velocity correction curve corresponding to the center frequency of the discharge signal, and take the center frequency as the center and the preset frequency bandwidth as the interval. Within the interval, select multiple frequency points according to the preset frequency step size, and calculate the linear mode velocity or ground mode velocity corresponding to each frequency point to form multiple wave velocity layers with different wave velocity values. Obtain the line length of each edge in the topology, and divide the length of each edge by the wave velocity value of each wave velocity layer. The quotient is taken as the theoretical propagation time of each edge in each wave velocity layer. The theoretical propagation time of each side under each wave velocity layer is stored in layers according to the wave velocity value from large to small, forming a multi-wave velocity layer layered propagation model.

4. The transmission line discharge fault location system based on feature analysis according to claim 1, characterized in that: The acquisition of wavefront polarity and center frequency includes: To determine the starting time of the wavefront, read the amplitudes of the previous and next sampling points at that starting time. If the amplitude of the next sampling point is greater than that of the previous sampling point, it is determined to be positive polarity; otherwise, it is determined to be negative polarity. Based on the arrival time of the wavefront, sampled data of a preset length are extracted forward and backward. Fourier transform is performed on the extracted sampled data to obtain a spectrum. The frequency corresponding to the maximum amplitude in the spectrum is taken as the center frequency of the wavefront.

5. The transmission line discharge fault location system based on feature analysis according to claim 1, characterized in that: The identification of the direct wavefronts and arrival times of each terminal includes: Among all the triggered terminals, the wavefront with the earliest arrival time that meets the preset validity check conditions is taken as the reference wavefront, and the wave velocity layer corresponding to its polarity and center frequency is taken as the reference polarity and reference wave velocity layer, respectively. For each triggered terminal, the wavefronts within its recorded time period are examined sequentially in chronological order. The first wavefront that simultaneously satisfies the same polarity and whose center frequency is within the reference wave velocity layer frequency range is determined as the direct wavefront of that terminal, and the sampling point time corresponding to that wavefront is recorded as the arrival time of the direct wave of that terminal.

6. The transmission line discharge fault location system based on feature analysis according to claim 1, characterized in that: The set of affected road segments includes: The terminal with the smallest direct wavefront arrival time among all triggered terminals is taken as the reference terminal. Based on the matching of the shortest path theoretical propagation time and the measured time difference of the reference terminal, the candidate line segment set is determined. For each candidate line segment in the candidate line segment set, according to the multi-wave speed layer-by-layer propagation model, multiple discretized hypothetical fault points are selected on the candidate line segment according to the preset spatial step size. The theoretical arrival order of each triggered terminal under each hypothetical fault point is determined. After deduplication, they are merged to form the theoretical arrival order set of the candidate line segment. Arrival times of the measured direct waves from each triggered terminal are sorted in chronological order to obtain the measured arrival sequence. Condition 1 is the set of theoretical arrival sequences that the measured arrival sequence belongs to the candidate line segment; Condition 2 is the uniform polarity of the direct wave of each triggered terminal; and Condition 3 is that the center frequency of the direct wave of each triggered terminal falls within the frequency range of the same wave velocity layer. Candidate road segments that meet all three conditions (1, 2, and 3) constitute the set of affected road segments.

7. The transmission line discharge fault location system based on feature analysis according to claim 6, characterized in that: The set of candidate line segments includes: The node where the reference terminal is located is used as the reference node, and the wave velocity layer corresponding to the center frequency of its direct wave is used as the current search wave velocity layer. In the multi-wave speed layered propagation model, taking the reference node as the starting point, the graph search algorithm is used to calculate the theoretical propagation time of the shortest path from the reference node to each node. The difference between the measured arrival time of the direct wave of each triggered terminal and the arrival time of the direct wave of the reference terminal is taken as the measured time difference. The theoretical propagation time range is determined by taking the measured time difference as the center and the allowable time deviation as the radius; Nodes whose shortest path propagation time falls within this range are considered valid nodes. All edges traversed by the shortest path from the reference node to each valid node are extracted, and the line segments corresponding to these edges are used to form a candidate line segment set.

8. The transmission line discharge fault location system based on feature analysis according to claim 1, characterized in that: The identified abnormal line segments include: For each candidate road segment in the affected road segment set, multiple hypothetical fault points are selected on the road segment according to a set step size. The theoretical time difference corresponding to each hypothetical fault point is calculated and compared with the measured time difference of the first and last triggering terminals. The hypothetical fault point with the smallest absolute value of the difference is recorded as the matching fault location, and the absolute value of the difference is used as the time difference matching degree of the candidate road segment. Based on the medium type parameters and the center frequency of the traveling wave, the amplitude attenuation coefficient per unit distance is determined, and then the amplitude attenuation error of the candidate line segment is calculated using the measured amplitude of the reference terminal as the benchmark amplitude. Using the waveform of the reference terminal as the reference waveform, cosine similarity is calculated for the waveforms of other triggered terminals, and the average value of all similarities is taken as the waveform similarity. The time difference matching degree, amplitude attenuation error and waveform similarity are weighted and fused according to preset weights to calculate the comprehensive confidence of each candidate line segment, and then the candidate line segment with the highest comprehensive confidence is selected as the abnormal line segment.

9. The transmission line discharge fault location system based on feature analysis according to claim 8, characterized in that: The amplitude attenuation error of the candidate line segment is calculated as follows: Calculate the propagation distance of the traveling wave from the location of the matching fault to each triggered terminal, and combine it with the amplitude attenuation coefficient per unit distance of the candidate line segment to obtain the theoretical amplitude of each triggered terminal; Calculate the absolute value of the difference between the theoretical amplitude and the measured amplitude of each triggered terminal and sum them to obtain the amplitude attenuation error of the candidate line segment.

10. The transmission line discharge fault location system based on feature analysis according to claim 1, characterized in that: Determining the final fault location includes: Multiple candidate fault points are selected in abnormal line sections according to a preset step size; For each candidate fault point, the theoretical direct arrival time of the traveling wave from the candidate fault point to each triggered terminal is calculated based on the multi-wave speed layered propagation model. The arrival time deviation of each terminal under each candidate fault point is obtained by subtracting the theoretical arrival time of each direct wavefront at each candidate fault point from the actual measured time of each terminal. The absolute values ​​of the arrival time deviations of all terminals at each candidate fault point are summed to obtain the total time deviation of that candidate fault point. Then, the candidate fault point with the smallest total time deviation is selected as the final fault location point.

Images