Power distribution network fault positioning method based on traveling wave detection

Abstract

The invention relates to the technical field of power distribution network fault positioning, in particular to a power distribution network fault positioning method based on traveling wave detection, and the method comprises the following steps: S1, distributed detection network construction and signal collection: collecting a three-phase current traveling wave signal in real time; s2, time-frequency joint analysis and feature extraction: extracting time-frequency energy distribution features; s3, intelligent screening of effective fault wave heads: screening out the effective fault wave heads; s4, initial arrival time parameter determination: determining the initial arrival time of each detection point; s5, according to topology self-adaptive error correction, transmission path wave velocity deviation is corrected through line parameter dynamic compensation, and clock errors of the asynchronous detection device are eliminated through adjacent node clock difference estimation; and S6, multi-source data fusion and fault positioning: resolving space coordinates of fault points. According to the invention, the method improves the recognition accuracy of fault features, and solves the problems of poor anti-interference capability and high misjudgment rate of a single feature in a conventional method.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution network fault location, and particularly to a distribution network fault location method based on traveling wave detection. Background Art

[0002] With the intelligent development of the distribution network, the operation and maintenance of the distribution network face higher requirements. Distribution network fault location is a key link to ensure the safe and stable operation of the power system. Traditional fault location methods mostly rely on high-precision synchronous clock devices, such as GPS systems, or perform single-point measurements through local detection devices. However, with the increasing complexity and diversity of the distribution network, especially in complex multi-branch and long-distance environments, traditional fault location methods are difficult to meet the requirements of accuracy and efficiency.

[0003] There are some obvious deficiencies in the existing distribution network fault location technologies. First of all, most traditional methods rely on synchronous clocks, which require the clocks of each detection device to be highly synchronized, increasing the cost of equipment and the difficulty of installation and maintenance. Secondly, existing methods usually adopt a fixed wave speed assumption, ignoring the dynamic influence of line type, conductor material and environmental factors on the wave speed, resulting in large positioning errors. In addition, traditional methods still rely on a single criterion in feature extraction and fault wavehead identification, with poor anti-interference ability and being easily affected by noise interference and signal loss, thus reducing the accuracy and reliability of fault location.

[0004] The purpose of the present invention is to provide a distribution network fault location method based on traveling wave detection, which can accurately identify effective fault waveheads and dynamically compensate for wave speed and clock errors, significantly improving the fault location accuracy and reliability in complex distribution network environments. At the same time, through the combination of multi-source data fusion and optimization algorithms, the accuracy and calculation efficiency of fault point location are optimized. Summary of the Invention

[0005] The present invention provides a distribution network fault location method based on traveling wave detection.

[0006] A distribution network fault location method based on traveling wave detection includes the following steps: S1, construction of a distributed detection network and signal acquisition: Configure asynchronous traveling wave detection devices at each branch node of the distribution network to collect three-phase current traveling wave signals in real time; S2, time-frequency joint analysis and feature extraction: Perform time-frequency joint analysis on the collected traveling wave signals, extract time-frequency energy distribution features, and generate a multi-dimensional feature matrix including frequency components, energy gradients and time correlations; S3, intelligent screening of effective fault waveheads: Establish a wavehead effectiveness discrimination model based on the multi-dimensional feature matrix, and screen out effective fault waveheads through the joint criterion of time-frequency energy mutation coefficient, waveform similarity index and polarity feature matching degree; S4. Initial arrival time parameter determination: Based on the time-frequency characteristic parameters of the effective fault wavefront and in combination with the signal propagation path length and the reference wave velocity, determine the initial arrival time of each detection point; S5. Topology adaptive error correction: Construct a network topology adaptive correction model, dynamically compensate and correct the wave velocity deviation of the transmission path through line parameter compensation, and estimate and eliminate the clock error of the asynchronous detection device by using the clock difference between adjacent nodes; S6. Multi-source data fusion and fault location: Fuse the corrected arrival times of multiple detection points, and in combination with the line segment wave velocity distribution model, establish a spatio-temporal constraint equation set, and solve the spatial coordinates of the fault point through an optimization algorithm.

[0007] Optionally, the distributed detection network construction and signal acquisition in S1 include: S11. Topology node classification and detection level division: Based on the distribution network topology structure, divide the nodes into backbone nodes, primary branch nodes, and secondary branch nodes, where; Configure main detection devices at backbone nodes at intervals of 3 - 5 km, with a sampling rate of not less than 1 MHz; Configure secondary detection devices at the beginning of the branch line of primary branch nodes, with a sampling rate of 500 kHz; Optionally configure micro detection devices at secondary branch nodes, with a sampling rate of 200 kHz; S12. Deployment of asynchronous detection devices: Install traveling wave detection devices with independent clock sources at the selected nodes; S13. Signal acquisition: Real-time collect three-phase current traveling wave signals through the deployed traveling wave detection devices.

[0008] Optionally, the time-frequency joint analysis and feature extraction in S2 include: S21. Signal adaptive decomposition and preprocessing: Use the improved Hilbert-Huang transform (HHT) to perform intrinsic mode decomposition on the traveling wave signal to obtain IMF components that meet the narrowband conditions; S22. Time-frequency energy ridge tracking: Calculate the instantaneous frequency and energy amplitude for each IMF component, and extract the time-frequency energy ridge that satisfies , where is the energy mutation threshold; S23. Multidimensional feature matrix construction: Generate a multidimensional feature matrix based on the energy ridge, including frequency components , energy gradient , and time correlation ; ; S24. Feature matrix standardization: Standardize the multidimensional feature matrix Perform normalization processing.

[0009] Optionally, the improved Hilbert-Huang transform (HHT) in S21 includes: S211, Adaptive noise injection and initial decomposition: Inject times of adaptive white noise into the original traveling wave signal to generate a noisy signal ; S212, Intrinsic mode function decomposition and residual update: Iteratively decompose each noisy signal ; S213, Dynamic stopping criterion control: Define the similarity of adjacent IMF components and determine whether the mode decomposition terminates; S214, Ensemble average and mode reconstruction: Perform ensemble average on the decomposition results of times of noise injection.

[0010] Optionally, the intelligent screening of effective fault waveheads in S3 includes: S31, Calculation of time-frequency energy mutation coefficient: Calculate the time-frequency energy mutation coefficient for the th ridge line ; S32, Calculation of waveform similarity index: Calculate the similarity index between the th ridge line and the reference waveform ; S33, Calculation of polarity feature matching degree: Extract the three-phase polarity combination , and calculate the matching degree ; S34, Comprehensive discrimination using combined criterion: Use a weighted scoring model to screen effective waveheads, expressed as: ; If and only if it satisfies: ; where , , are weight coefficients, is the comprehensive discrimination threshold, and is the comprehensive score.

[0011] Optionally, the determination of the initial wave arrival time parameter in S4 includes: S41, Extraction of time-frequency features of effective waveheads: Extract time-frequency feature parameters from the screened effective fault waveheads, including the wavehead arrival time , instantaneous frequency ; S42, Calculation of signal propagation path length: Calculate the propagation path length from the fault point to the detection point according to the distribution network topology ; S43, Dynamic calculation of reference wave velocity: Based on the instantaneous frequency calculate the reference wave velocity ; S44, Initial arrival time calculation: Calculate the initial arrival time according to the propagation path length and the reference wave velocity, expressed as: ; where, is the initial arrival time of the th wavefront.

[0012] Optionally, the topology adaptive error correction in S5 includes: S51, Dynamic compensation of line parameters to correct the wave velocity deviation of the transmission path: Build a wave velocity dynamic compensation model based on the line type, conductor material and environmental parameters (temperature and humidity), and correct the reference wave velocity; S52, Estimation of clock difference between adjacent nodes to eliminate asynchronous error: Use the signal transmission time difference between adjacent detection devices to inversely deduce the clock deviation, calculate the clock error of the asynchronous detection device by comparing the initial arrival time of adjacent nodes with the theoretical transmission time, and correct the initial arrival time.

[0013] Optionally, the dynamic compensation of line parameters to correct the wave velocity deviation of the transmission path in S51 includes: S511, Calculation of line parameter influence coefficient: Calculate the basic wave velocity correction coefficient according to the line type and conductor material ; S512, Calculation of environmental temperature and humidity compensation coefficient: Calculate the environmental compensation coefficient based on the temperature and humidity data in the real-time environment ; S513, Construction of comprehensive correction model: Superimpose the basic wave velocity correction coefficient and the environmental compensation coefficient to generate the total wave velocity correction coefficient, expressed as: ; ; where, is the total correction coefficient of the th section of the line, is the corrected wave velocity, is the path selection indication variable, is the total number of line segments.

[0014] Optionally, the estimation of clock difference between adjacent nodes to eliminate asynchronous error in S52 includes: S521, Calculation of signal transmission time difference between adjacent nodes: Calculate the signal transmission time difference between adjacent detection devices ; S522, Theoretical transmission time calculation: Calculate the theoretical transmission time based on the line length and the corrected wave velocity ; S523, Clock deviation estimation and correction: Calculate the clock deviation and correct the time of arrival of the wave by comparing the signal transmission time difference with the theoretical transmission time, expressed as: ; ; wherein, is the clock deviation of the th detection point, is the corrected time of arrival of the wave.

[0015] Optionally, the multi-source data fusion and fault location in S6 include: S61, Construction of space-time constraint equations: Based on the corrected time of arrival of the wave at multiple detection points and the wave velocity of line segments , establish space-time constraint equations, expressed as: ; wherein, is the spatial coordinate of the fault point, is the coordinate of the th detection point, is the corrected time of arrival of the wave at the th detection point, is the fault occurrence time, is the time residual of the th detection point, is the number of effective detection points ; S62, Definition of multi-objective optimization function: Construct an objective function that minimizes the weighted sum of squared residuals, expressed as: ; wherein, is the weight coefficient of the th detection point; S63, Solve using the improved particle swarm optimization algorithm: Use the adaptive inertia weight particle swarm algorithm to solve the optimal fault coordinates, specifically including: Particle state update: ; Dynamic inertia weight adjustment: ; wherein, is the velocity of the particle in dimension at the th iteration, is the position of the particle in dimension at the th iteration, is the dynamic inertia weight, which controls the update of the particle velocity. , are the learning factors, which respectively control the weights of the particle moving towards the individual optimal and global optimal directions. , are random numbers uniformly distributed, which are used to introduce randomness. is the individual historical optimal position of the particle in dimension . is the global historical optimal position of the population in dimension . is the maximum number of iterations. is the current number of iterations. is the upper limit value of the dynamic inertia weight. is the lower limit value of the dynamic inertia weight.

[0016] Advantages of the present invention: In the present invention, a distributed asynchronous detection network is constructed, which combines time-frequency joint analysis technology and intelligent wavehead screening algorithm, breaks through the dependence on synchronous clocks of traditional distribution network fault location methods, and realizes high-precision fault location in a complex distribution network environment without GPS signals. This method uses a multi-dimensional feature matrix to accurately extract fault features, and adopts a weighted scoring model to optimize the screening of wavehead effectiveness, significantly improving the recognition accuracy of fault features and solving the problems of poor anti-interference ability of single features and high misjudgment rate existing in previous methods.

[0017] In the present invention, by introducing topology adaptive error correction technology, aiming at the positioning error problem caused by the fixed wave velocity in traditional methods, the influence of factors such as line parameters, environmental temperature and humidity on the wave velocity is dynamically compensated, and the wave velocity deviation is accurately corrected, further improving the accuracy and reliability of fault location. At the same time, using the clock difference estimation technology between adjacent detection points, the clock deviation is inversely deduced through the time difference between asynchronous detection devices, effectively eliminating the clock synchronization error, avoiding the dependence on high-precision synchronous clock devices, and reducing the equipment cost.

[0018] In the present invention, high-precision fault location is realized through an improved particle swarm optimization algorithm. Combining the corrected time of arrival of the wave at the detection point and the wave velocity information of line segments, a spatio-temporal constraint equation is constructed and the positioning result is optimized by minimizing the weighted sum of squared residuals objective function. This method can dynamically adjust the weights of detection points, quickly converge to the optimal fault point coordinates, and significantly improve the positioning accuracy and response speed in a complex distribution network environment. Description of the Drawings

[0019] To more clearly illustrate the technical solutions in the present invention or the prior art, the following will briefly introduce the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only those of the present invention. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can also be obtained based on these drawings.

[0020] Figure 1 Schematic flow diagram of the positioning method according to an embodiment of the present invention; Figure 2 Schematic diagram of intelligent screening of effective fault wave heads according to an embodiment of the present invention. Detailed implementation manners

[0021] The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments. At the same time, it should be noted here that in order to make the embodiments more detailed, the following embodiments are the best and preferred embodiments. For some well-known technologies, those skilled in the art can also adopt other alternative ways to implement them; and the accompanying drawings are only for more specific description of the embodiments, and are not intended to specifically limit the present invention.

[0022] It should be noted that in the specification, when referring to "an embodiment", "embodiments", "exemplary embodiments", "some embodiments", etc., it indicates that the described embodiments may include specific features, structures or characteristics, but not necessarily every embodiment includes such specific features, structures or characteristics. In addition, when combining embodiments to describe specific features, structures or characteristics, implementing such features, structures or characteristics in combination with other embodiments (whether explicitly described or not) should be within the knowledge of those skilled in the relevant art.

[0023] Generally, terms can be understood at least in part from their use in the context. For example, at least in part depending on the context, the term "one or more" used herein can be used to describe any feature, structure or characteristic in a singular sense, or can be used to describe a combination of features, structures or characteristics in a plural sense. In addition, the term "based on" can be understood as not necessarily intended to convey a set of exclusive factors, but rather, at least in part depending on the context, allowing for the existence of other factors that may not be explicitly described.

[0024] As Figure 1 - Figure 2 shown, a distribution network fault location method based on traveling wave detection includes the following steps: S1, construction of a distributed detection network and signal acquisition: Configure asynchronous traveling wave detection devices at each branch node of the distribution network to collect three-phase current traveling wave signals in real time; S2, time-frequency joint analysis and feature extraction: Perform time-frequency joint analysis on the collected traveling wave signals, extract time-frequency energy distribution features, and generate a multi-dimensional feature matrix including frequency components, energy gradients, and time correlations; S3, Intelligent Screening of Effective Fault Wavefronts: Based on the multi-dimensional feature matrix, establish a discriminant model for wavefront effectiveness. Through the combined criterion of time-frequency energy mutation coefficient, waveform similarity index, and polarity feature matching degree, screen out the effective fault wavefronts; S4, Determination of Initial Wave Arrival Time Parameters: According to the time-frequency characteristic parameters of the effective fault wavefronts, combined with the signal propagation path length and the reference wave velocity, determine the initial wave arrival time of each detection point; S5, Topology Adaptive Error Correction: Construct a network topology adaptive correction model. Dynamically compensate and correct the wave velocity deviation of the transmission path through line parameter dynamic compensation, and estimate and eliminate the clock error of asynchronous detection devices using the clock difference between adjacent nodes; S6, Multi-source Data Fusion and Fault Location: Fuse the corrected wave arrival times of multiple detection points, combine with the line segment wave velocity distribution model, establish a spatio-temporal constraint equation set, and solve the spatial coordinates of the fault point through an optimization algorithm; Through the above content, construct a distributed asynchronous detection network, accurately extract fault features by combining time-frequency joint analysis technology, intelligently screen effective wavefronts using multi-dimensional criteria, break through the limitation of traditional dependence on synchronous clocks, and further dynamically compensate wave velocity deviation and clock error through the topology adaptive correction model, significantly improving the positioning accuracy in complex distribution network environments.

[0025] The construction of the distributed detection network and signal acquisition in S1 include: S11, Topology Node Classification and Detection Hierarchy Division: Based on the distribution network topology structure, divide the nodes into main nodes, first-level branch nodes, and second-level branch nodes, where; Configure main detection devices at intervals of 3 - 5 km for main nodes, with a sampling rate of not less than 1 MHz; Configure secondary detection devices at the head of the branch line for first-level branch nodes, with a sampling rate of 500 kHz; Selectively configure micro detection devices for second-level branch nodes, with a sampling rate of 200 kHz; S12, Deployment of Asynchronous Detection Devices: Install traveling wave detection devices with independent clock sources at the selected nodes; S13, Signal Acquisition: Real-time collect three-phase current traveling wave signals through the deployed traveling wave detection devices; Through the above content, hierarchical topology node classification realizes differential configuration of detection resources. Main nodes use high-sampling-rate main detection devices to ensure the signal capture accuracy of key paths, branch nodes deploy secondary / micro devices as needed to reduce hardware costs, and asynchronous detection devices rely on independent clock sources to get rid of GPS synchronization dependence. Combined with the selective deployment strategy, while ensuring the monitoring accuracy of the main line, it significantly reduces the equipment investment in multi-branch scenarios.

[0026] The time-frequency joint analysis and feature extraction in S2 include: S21, Signal Adaptive Decomposition and Preprocessing: Use the improved Hilbert-Huang Transform (HHT) to perform intrinsic mode decomposition on the traveling wave signal to obtain IMF components that meet the narrowband condition; S22, Time-Frequency Energy Ridge Tracking: Calculate the instantaneous frequency and energy amplitude for each IMF component, and extract the time-frequency energy ridge that meets , where is the energy mutation threshold, expressed as: ; ; where, is the layer IMF component, is the instantaneous phase angle of the IMF component, is the derivative with respect to time; S23, Multidimensional Feature Matrix Construction: Generate a multidimensional feature matrix including frequency components, energy gradient and time correlation , expressed as: ; where, is the average frequency component of the th ridge line, is the analysis time window length, is the starting time when the traveling wave reaches the detection device, is the th instantaneous frequency of the ridge line, is the energy gradient, is the th instantaneous energy amplitude of the ridge line, is the th ending time of the ridge line, is the th ending time of the ridge line, is the time correlation between ridge lines when calculated), is the th instantaneous energy amplitude of the ridge line, is the average value of the instantaneous energy amplitudes of the th ridge line, is the average value of the instantaneous energy amplitudes of the th ridge line, is the standard deviation of the instantaneous energy amplitudes of the th ridge line, is the Standard deviation of the instantaneous energy amplitude of the strip ridge line S24, Feature matrix standardization: For the multi-dimensional feature matrix Perform normalization processing, expressed as: ; ; ; Among them, Are the mean and standard deviation of the frequency components respectively, , , Are the normalized frequency components, energy gradients, and time correlations respectively; Through the above content, the improved Hilbert-Huang transform (HHT) adaptively decomposes the traveling wave signal, combines the time-frequency energy ridge line tracking technology to accurately capture the fault characteristics, significantly improves the detection ability of weak signals, constructs a multi-dimensional feature matrix to comprehensively characterize the time-frequency characteristics of the traveling wave signal, solves the problems of single feature and poor anti-interference ability in traditional methods, and eliminates the dimension difference and enhances the comparability and separability of features through feature matrix standardization processing.

[0027] The improved Hilbert-Huang transform (HHT) in S21 includes: S211, Adaptive noise injection and initial decomposition: Inject Into the original traveling wave signal Times of adaptive white noise to generate a noisy signal , expressed as: ; Among them, Is the noise amplitude coefficient, Is a white noise signal with a mean of zero and a variance of 1, Is the number of noise injection times; S212, Intrinsic mode function decomposition and residual update: Perform iterative decomposition on each noisy signal , specifically including: The Layer IMF extraction: ; Among them, the empirical mode decomposition The narrowband condition for the IMF needs to be that the difference between the number of extreme points and the number of zero-crossing points does not exceed 1 and the local mean is determined by the average of the upper and lower envelope lines; Residual update: ; Initial condition: ; S213, Dynamic stopping criterion control: Define the similarity of adjacent IMF components, and determine whether the mode decomposition terminates, which is expressed as: ; where is the similarity of adjacent IMF components, used to determine whether the mode decomposition terminates. When the decomposition is terminated, and the final IMF set is output , is the -th layer intrinsic mode function (IMF), is the -th layer intrinsic mode function (IMF); S214, Ensemble average and mode reconstruction: Perform ensemble averaging on the decomposition results of times of noise injection, which is expressed as: ; where is the final -th layer IMF component, which is the ensemble average result, is the -th layer IMF component obtained by decomposition after the -th noise injection; Through the above content, adaptive noise injection effectively suppresses the mode mixing problem of traditional EMD, significantly improves the decomposition accuracy of traveling wave signals, optimizes the decomposition depth through dynamic stopping criteria, avoids the generation of redundant modes, improves the calculation efficiency, and combines ensemble averaging technology to further enhance the physical interpretability and anti-noise ability of IMF components.

[0028] The effective fault wavefront intelligent screening in S3 includes: S31, Calculation of time-frequency energy mutation coefficient: Calculate the time-frequency energy mutation coefficient for the -th ridge line, which is expressed as: ; where is the maximum value of the energy gradients of all ridge lines, is the reference frequency, taking the 20th harmonic value of the power grid power frequency (50 Hz); S32, Calculation of waveform similarity index: Calculate the waveform similarity index between the -th ridge line and the reference waveform, which is expressed as: ; where is the instantaneous energy amplitude of the -th ridge line, is the The mean value of the instantaneous energy amplitude of the strip ridge line, is the pre-stored energy amplitude of the fault waveform, and is the mean value of the reference waveform; S33, calculation of the polarity feature matching degree: Extract the three-phase polarity combination , and calculate the matching degree , which is expressed as: ; Among them, is the phase traveling wave polarity (+1 represents positive polarity, -1 represents negative polarity), is the reference polarity in the pre-stored fault type library, is the Kronecker function, which is expressed as ; S34, comprehensive discrimination of the combined criterion: Use a weighted scoring model to screen effective waveheads, which is expressed as: ; When and only when it satisfies: ; Among them, , , are the weight coefficients, satisfying , is the comprehensive discrimination threshold, with a value of 0.75, is the comprehensive score, indicating the comprehensive confidence level of the th strip ridge line as an effective fault wavehead; Through the above content, the accuracy and reliability of fault feature identification are significantly improved. The weighted scoring model is used to dynamically evaluate the effectiveness of waveheads, solving the problems of high misjudgment rate and poor adaptability of traditional single-criterion methods. Through the hard constraint of the energy mutation coefficient, it is ensured that the screening results have clear physical meaning and fault correlation.

[0029] The determination of the initial wave arrival time parameter in S4 includes: S41, extraction of the time-frequency characteristics of effective waveheads: Extract the time-frequency characteristic parameters of the screened effective fault waveheads, including the wavehead arrival time and the instantaneous frequency , which is expressed as: ; ; Among them, is the wavehead energy amplitude of the th strip ridge line, is the instantaneous frequency of the th strip ridge line; S42, Signal propagation path length calculation: Calculate the propagation path length from the fault point to the detection point according to the distribution network topology structure , expressed as: ; Wherein, is the length of the th line segment, is the path selection indication variable, indicating that the th line segment is in the propagation path, is the total number of line segments; S43, Dynamic calculation of reference wave velocity: Calculate the reference wave velocity based on the instantaneous frequency , expressed as: ; Wherein, is the reference wave velocity of the th wavefront, , indicating the reference wave velocity at power frequency , , indicating the reference frequency (the 20th harmonic of power frequency); S44, Initial wave arrival time calculation: Calculate the initial wave arrival time according to the propagation path length and the reference wave velocity, expressed as: ; Wherein, is the initial wave arrival time of the th wavefront; Through the above content, combined with the calculation of the signal propagation path length and the dynamic reference wave velocity, the accuracy and reliability of the wave arrival time calculation are significantly improved. By dynamically adjusting the reference wave velocity through frequency, the positioning error problem caused by the fixed wave velocity is solved, and the path length based on the distribution network topology structure.

[0030] The topology adaptive error correction in S5 includes: S51, Dynamic compensation of line parameters to correct the wave velocity deviation of the transmission path: Construct a wave velocity dynamic compensation model based on the line type, conductor material and environmental parameters (temperature and humidity) to correct the reference wave velocity; S52, Estimation of clock difference between adjacent nodes to eliminate asynchronous error: Use the signal transmission time difference between adjacent detection devices to reverse infer the clock deviation, calculate the clock error of the asynchronous detection device by comparing the initial wave arrival time of adjacent nodes with the theoretical transmission time, and correct the initial wave arrival time; Through the above content, the reference wave velocity is accurately corrected in combination with the line type, conductor material, and environmental parameters, solving the positioning error problem caused by the fixed wave velocity. At the same time, by using the adjacent node clock difference estimation technology, the clock deviation is inversely deduced by comparing the signal transmission time difference of adjacent detection devices, effectively eliminating the clock error of asynchronous detection devices.

[0031] The dynamic compensation of line parameters in S51 to correct the wave velocity deviation of the transmission path includes: S511, calculation of line parameter influence coefficient: Calculate the basic wave velocity correction coefficient according to the line type and conductor material , expressed as: ; Among them, is the basic wave velocity correction coefficient of the th section of the line, represents the line type weight coefficient (0.6 for overhead lines and 0.4 for cables), is the line type correction value (+0.05 for overhead lines and -0.03 for cables), is the conductor material weight coefficient, is the material correction value (-0.02 for aluminum conductors and +0.01 for copper conductors); S512, calculation of environmental temperature and humidity compensation coefficient: Calculate the environmental compensation coefficient based on the temperature and humidity data in the real-time environment , expressed as: ; Among them, is the environmental compensation coefficient of the th section of the line, is the temperature compensation coefficient, is the real-time temperature, is the reference temperature ( ), is the humidity compensation coefficient, is the real-time humidity, is the reference humidity (60); S513, construction of comprehensive correction model: Superimpose the basic wave velocity correction coefficient and the environmental compensation coefficient to generate the total wave velocity correction coefficient, expressed as: ; ; Among them, is the total correction coefficient of the th section of the line, is the corrected wave velocity, is the path selection indicator variable, represents that the th section of the line is in the propagation path, is the total number of line segments; Through the above content, the accuracy and adaptability of wave velocity correction are significantly improved. By accumulating the correction coefficients of each line segment, it supports accurate wave velocity calculation in complex distribution network scenarios, and the real-time environmental parameter compensation mechanism further enhances the dynamic response ability of the model.

[0032] The estimation of the clock difference between adjacent nodes in S52 to eliminate asynchronous errors includes: S521, calculation of the signal transmission time difference between adjacent nodes: Calculate the signal transmission time difference between adjacent detection devices , expressed as: ; where is the initial arrival time difference between the th detection point and the adjacent th detection point, is the initial arrival time of the th detection point, is the initial arrival time of the th adjacent detection point; S522, calculation of the theoretical transmission time: Calculate the theoretical transmission time based on the line length and the corrected wave velocity , expressed as: ; where is the theoretical time for the signal to be transmitted from the th detection point to the th detection point, is the line length between adjacent detection points, obtained from the topological structure database; S523, clock deviation estimation and correction: By comparing the signal transmission time difference with the theoretical transmission time, calculate the clock deviation and correct the arrival time, expressed as: ; ; where is the clock deviation of the th detection point, is the corrected arrival time; Through the above content, by comparing the signal transmission time difference and the theoretical transmission time of adjacent detection devices, the clock deviation of asynchronous detection devices is accurately deduced and the initial arrival time is corrected, significantly improving the accuracy and reliability of fault location. Based on the calculation of the theoretical transmission time with the corrected wave velocity and combined with real-time topological structure data, the accuracy of clock deviation estimation is ensured; there is no need to rely on high-precision synchronous clock devices.

[0033] Multi-source data fusion and fault location in S6 include: S61, construction of spatio-temporal constraint equations: Based on the time of arrival (TOA) corrected at multiple detection points and the wave velocity of line segments , establish spatio-temporal constraint equations, expressed as: ; Among them, is the spatial coordinate of the fault point, is the coordinate of the th detection point, is the corrected TOA of the th detection point, is the fault occurrence time, is the time residual of the th detection point, is the number of effective detection points ; S62, definition of multi-objective optimization function: Construct an objective function to minimize the weighted sum of squared residuals, expressed as: ; Among them, is the weight coefficient of the th detection point, , is the standard deviation of time measurement of the th detection point, is the maximum value of the corrected wave velocities of all detection points; S63, solution using improved particle swarm optimization algorithm: Use an adaptive inertia weight particle swarm algorithm to solve the optimal fault coordinates, specifically including: Update of particle state: ; Dynamic adjustment of inertia weight: ; Among them, is the velocity of the particle in dimension at the th iteration, is the position of the particle in dimension at the th iteration, is the dynamic inertia weight, controlling the update of particle velocity, , are learning factors, respectively controlling the weights of the particle moving towards the individual optimal and global optimal directions, , are uniformly distributed random numbers, used to introduce randomness, is the individual historical optimal position of the particle in dimension , is the global optimal position of the swarm in dimension The globally historically optimal position is the maximum number of iterations is the current number of iterations is the upper limit value of the dynamic inertia weight is the lower limit value of the dynamic inertia weight; Through the above content, high-precision fusion of the time of arrival of the multi-detection point correction wave is achieved. The weighted residual sum of squares minimization objective function is adopted to dynamically allocate the weights of the detection points, significantly improving the positioning accuracy in complex distribution network scenarios. By improving the particle swarm optimization algorithm and combining the dynamic inertia weight adjustment strategy, it quickly converges to the optimal fault point coordinates.

[0034] This invention covers any substitutions, modifications, equivalent methods, and solutions made within the essence and scope of this invention. To enable the public to have a thorough understanding of this invention, specific details are described in detail in the following preferred embodiments of this invention. However, those skilled in the art can fully understand this invention without these detailed descriptions. Additionally, well-known methods, processes, procedures, components, and circuits, etc., are not described in detail to avoid unnecessary confusion to the essence of this invention.

[0035] The above description is only a preferred embodiment of this invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of this invention, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of this invention.

Claims

1. A method for fault location in a distribution network based on traveling wave detection, characterized in that It includes the following steps: S1. Construction of distributed detection network and signal acquisition: Configure asynchronous traveling wave detection devices at each branch node of the distribution network to collect three-phase current traveling wave signals in real time; S2. Time-frequency joint analysis and feature extraction: Conduct time-frequency joint analysis on the collected traveling wave signals, extract time-frequency energy distribution features, and generate a multi-dimensional feature matrix including frequency components, energy gradients, and time correlations; S3. Intelligent screening of effective fault wavefronts: Establish a wavefront effectiveness discrimination model based on the multi-dimensional feature matrix, and screen out effective fault wavefronts through a joint criterion of time-frequency energy mutation coefficient, waveform similarity index, and polarity feature matching degree; S4. Determination of initial wave arrival time parameters: According to the time-frequency characteristic parameters of the effective fault wavefronts, combined with the signal propagation path length and the reference wave velocity, determine the initial wave arrival time at each detection point; S5. Topology adaptive error correction: Construct a network topology adaptive correction model, dynamically compensate and correct the wave velocity deviation of the transmission path through line parameter compensation, and estimate and eliminate the clock error of the asynchronous detection device using the clock difference between adjacent nodes; S6. Multi-source data fusion and fault location: Fusion the corrected wave arrival times of multiple detection points, combined with the line segment wave velocity distribution model, establish a spatio-temporal constraint equation set, and solve the spatial coordinates of the fault point through an optimization algorithm.

2. The method for locating faults in a distribution network based on traveling wave detection according to claim 1, wherein The construction of the distributed detection network and signal acquisition in S1 includes: S11. Topology node classification and detection level division: Based on the topology structure of the distribution network, divide the nodes into main nodes, primary branch nodes, and secondary branch nodes, where; Configure main detection devices at main nodes at intervals of 3 - 5 km, with a sampling rate of not less than 1 MHz; Configure secondary detection devices at the beginning of the branch line of primary branch nodes, with a sampling rate of 500 kHz; Selectively configure micro detection devices at secondary branch nodes, with a sampling rate of 200 kHz; S12. Deployment of asynchronous detection devices: Install traveling wave detection devices with independent clock sources at the selected nodes; S13. Signal acquisition: Real-time collect three-phase current traveling wave signals through the deployed traveling wave detection devices.

3. A method for fault location in a distribution network based on traveling wave detection according to claim 1, characterized in that, The time-frequency joint analysis and feature extraction in S2 includes: S21. Signal adaptive decomposition and preprocessing: Use the improved Hilbert-Huang transform to perform intrinsic mode decomposition on the traveling wave signals to obtain IMF components that meet the narrowband conditions; S22, Time-frequency energy ridge tracking: Calculate the instantaneous frequency for each IMF component and the energy amplitude , and extract the time-frequency energy ridge that satisfies , where is the energy mutation threshold; S23, Multidimensional Feature Matrix Construction: Based on the energy ridge line, generate a multidimensional feature matrix including frequency components , energy gradient , and time correlation ; ; S24, Feature matrix standardization: Normalize the multi-dimensional feature matrix by performing normalization processing.

4. A method for fault location in a distribution network based on traveling wave detection according to claim 3, characterized in that, The improved Hilbert-Huang transform in S21 includes: S211, Adaptive Noise Injection and Initial Decomposition: Inject into times of adaptive white noise into the original traveling wave signal to generate a noisy signal ; S212, Intrinsic mode decomposition and residual update: Iteratively decompose each noisy signal for iterative decomposition; S213. Dynamic stop criterion control: Define the similarity of adjacent IMF components to judge whether the mode decomposition terminates; S214, Ensemble Averaging and Modal Reconstruction: Perform ensemble averaging on the decomposition results of the sub-noise injection.

5. A method for fault location in a distribution network based on traveling wave detection according to claim 4, characterized in that, The intelligent screening of effective fault wavefronts in S3 includes: S31, Calculation of time-frequency energy mutation coefficient: Calculate the time-frequency energy mutation coefficient for the th ridge line ; S32, Waveform similarity index calculation: Calculate the similarity index between the th ridge line and the reference waveform ; S33, Calculation of polarity feature matching degree: Extract the three-phase polarity combination , calculate the matching degree ; S34. Comprehensive discrimination of joint criteria: Use a weighted scoring model to screen effective wavefronts, expressed as: ； If and only if the following conditions are met: ; Among them, , , are weight coefficients, is the comprehensive discrimination threshold, is the comprehensive score.

6. A method for fault location in a distribution network based on traveling wave detection according to claim 1, characterized in that, The determination of initial wave arrival time parameters in S4 includes: S41, Extraction of time-frequency characteristics of effective waveheads: Extract time-frequency characteristic parameters from the selected effective fault waveheads, including the arrival time of the waveheads , instantaneous frequency ; S42, Signal propagation path length calculation: Calculate the propagation path length from the fault point to the detection point according to the distribution network topology structure ; S43, Dynamic calculation of reference wave velocity: Based on instantaneous frequency Calculate the reference wave velocity ; S44. Calculation of initial wave arrival time: Calculate the initial wave arrival time according to the propagation path length and the reference wave velocity, expressed as: ； Among them, is the initial arrival time of the th wavefront.

7. A method for fault location in a distribution network based on traveling wave detection according to claim 1, characterized in that, The topology adaptive error correction in S5 includes: S51. Dynamic compensation of line parameters to correct the wave velocity deviation of the transmission path: Construct a wave velocity dynamic compensation model based on line type, conductor material, and environmental parameters to correct the reference wave velocity; S52. Eliminate the asynchronous error by estimating the clock difference between adjacent nodes: Use the signal transmission time difference of adjacent detection devices to inversely deduce the clock deviation, calculate the clock error of the asynchronous detection device by comparing the initial arrival time and the theoretical transmission time of adjacent nodes, and correct the initial arrival time.

8. A method for fault location in a distribution network based on traveling wave detection according to claim 7, characterized in that The dynamic compensation of line parameters in S51 to correct the wave velocity deviation of the transmission path includes: S511, Calculation of line parameter influence coefficient: Calculate the fundamental wave velocity correction coefficient according to the line type and conductor material ; S512, Ambient Temperature and Humidity Compensation Coefficient Calculation: Calculate the ambient compensation coefficient based on the temperature and humidity data in the real-time environment ; S513. Construct the comprehensive correction model: Superimpose the basic wave velocity correction coefficient and the environmental compensation coefficient to generate the total wave velocity correction coefficient, expressed as: ； ； Among them, is the total correction coefficient of the th section of the line, is the wave velocity after correction, is the path selection indication variable, is the total number of line segments.

9. A method for fault location in a distribution network based on traveling wave detection according to claim 8, characterized in that, The elimination of asynchronous error by estimating the clock difference between adjacent nodes in S52 includes: S521, Calculation of signal transmission time difference between adjacent nodes: Calculate the signal transmission time difference between adjacent detection devices ; S522, Theoretical Transmission Time Calculation: Calculate the theoretical transmission time based on the line length and the corrected wave velocity ; S523. Estimate and correct the clock deviation: Calculate the clock deviation and correct the arrival time by comparing the signal transmission time difference with the theoretical transmission time, expressed as: ； ； Among them, is the clock deviation of the th detection point, is the corrected time of arrival.

10. A method for fault location in a distribution network based on traveling wave detection according to claim 9, characterized in that, The multi-source data fusion and fault location in S6 includes: S61, Construction of Spatiotemporal Constraint Equation System: Based on the Time of Arrival (TOA) Corrected by Multiple Detection Points and the wave velocity of line segments , establish the spatiotemporal constraint equation, expressed as: ； Among them, is the spatial coordinate of the fault point, is the coordinate of the th detection point, is the corrected time of arrival of the th detection point, is the fault occurrence time, is the time residual of the th detection point, is the number of effective detection points; S62. Define the multi-objective optimization function: Construct the objective function of minimizing the weighted sum of squared residuals, expressed as: ； Among them, is the weight coefficient of the th detection point; S63. Solve using the improved particle swarm optimization algorithm: Use the adaptive inertia weight particle swarm algorithm to solve the optimal fault coordinates, specifically including: Particle state update: ; Dynamic inertia weight adjustment: ; Wherein, is the velocity of the particle in dimension at the -th iteration, is the position of the particle in dimension at the -th iteration, is the dynamic inertia weight, which controls the update of the particle velocity, , are the learning factors, which respectively control the weights of the particle moving towards the individual optimal and global optimal directions, , are uniformly distributed random numbers, which are used to introduce randomness, is the individual historical optimal position of the particle in dimension , is the global historical optimal position of the population in dimension , is the maximum number of iterations, is the current number of iterations, is the upper limit value of the dynamic inertia weight, is the lower limit value of the dynamic inertia weight.

Images