Power distribution network fault section locating method based on beidou satellite timing

Abstract

The application belongs to the technical field of fault positioning, and discloses a power distribution network fault section positioning method based on Beidou satellite timing, which comprises the following steps: collecting three-phase currents, and synchronously aligning the three-phase currents by using the Beidou satellite to obtain a three-phase current synchronous data sequence; performing modulus transformation on the three-phase current synchronous data sequence to obtain zero-mode current components and line-mode current components; performing adaptive waveform decomposition on the zero-mode current components and the line-mode current components to obtain intrinsic modal components, and performing signal reconstruction on target components to obtain an enhanced fault traveling wave signal; performing waveform curvature mutation detection on the enhanced fault traveling wave signal to obtain a wave head arrival time and a time difference observation sequence; performing geometric analysis on a topological structure based on the time difference observation sequence and a traveling wave propagation speed to obtain a suspected fault section; and comprehensively judging the suspected fault section to obtain an actual fault section. The application can improve the efficiency of power distribution network fault positioning.

Description

Technical Field

[0001] This invention belongs to the field of fault location technology, and in particular relates to a method for locating fault sections in distribution networks based on BeiDou satellite timing. Background Technology

[0002] In the field of fault location in distribution networks, traditional methods mainly rely on local clocks to collect data from multiple monitoring points and use the traveling wave time difference method to locate faults. However, the local clocks used in traditional methods lack a highly stable timing reference, making it difficult to achieve accurate phase synchronization of three-phase current data from different monitoring points. This results in microsecond-level or even millisecond-level time deviations in the collected current data. At the same time, existing technologies perform relatively crude modulus transformation processing on three-phase current data, making it difficult to effectively suppress non-ideal noise components and complete mode purification. Furthermore, the lack of a scientific mode screening mechanism makes it impossible to accurately identify key components characterizing high-frequency transient features, resulting in poor enhancement of the fault traveling wave signal and limited accuracy in detecting the arrival time of the wavefront.

[0003] The shortcomings of the aforementioned traditional methods directly lead to a decrease in the reliability of fault location. Insufficient clock synchronization accuracy makes subsequent time difference calculations based on traveling wave propagation characteristics prone to errors, resulting in geometric analysis errors of the fault point. Furthermore, weaknesses in the signal processing stage lead to incomplete fault feature extraction, further affecting the accuracy of suspected fault location assessments. Ultimately, the efficiency and reliability of fault location fail to meet the actual needs of rapid operation and maintenance in distribution networks, extending fault investigation time and expanding the power supply impact area. Therefore, improving the clock synchronization accuracy and signal processing quality for fault location in distribution networks has become an urgent technical problem to be solved. Summary of the Invention

[0004] To address the shortcomings of existing technologies, the present invention aims to provide a method for locating fault sections in distribution networks based on BeiDou satellite timing, which can improve the efficiency of fault location in distribution networks.

[0005] To achieve the above objectives, the present invention provides a method for locating fault sections in a distribution network based on BeiDou satellite timing, comprising:

[0006] S1. Real-time acquisition of the three-phase current at monitoring points on the distribution network, and using the clock signal output by the Beidou satellite to perform phase synchronization and alignment of the three-phase current to obtain the three-phase current synchronization data sequence of the distribution network.

[0007] S2. Perform modulus transformation on the three-phase current synchronization data sequence to obtain the zero-mode current component and line-mode current component of the distribution network.

[0008] S3. Perform adaptive waveform decomposition on the zero-mode current component and the line-mode current component to obtain the intrinsic mode components of the distribution network. Select the target component that characterizes the high-frequency transient characteristics from the intrinsic mode components and reconstruct the signal of the target component to obtain the enhanced fault traveling wave signal of the distribution network.

[0009] S4. Detect waveform curvature change in the enhanced fault traveling wave signal to obtain the arrival time of the wavefront of the enhanced fault traveling wave signal, and obtain the time difference observation sequence of the distribution network based on the arrival time of the wavefront and the preset time reference.

[0010] S5. Based on the time difference observation sequence and the traveling wave propagation speed of the distribution network, the topology of the distribution network is geometrically analyzed to obtain the suspected fault sections of the distribution network.

[0011] S6. Based on the wavefront polarity and energy attenuation amplitude characteristics of the enhanced fault traveling wave signal, a comprehensive assessment is conducted on the suspected fault sections to obtain the actual fault sections of the distribution network.

[0012] Preferably, in S1, the three-phase current synchronization data sequence of the distribution network is obtained, including:

[0013] Raw three-phase current sampling data are collected at monitoring points in the distribution network, and initial time stamps based on local clocks are marked for the monitoring points.

[0014] It receives the standard time signal transmitted by the Beidou satellite and, based on the standard time signal, synchronizes and calibrates the local clock of the monitoring point to obtain the calibrated time tag of the monitoring point;

[0015] A unified reference time axis for the power distribution network is established based on the calibrated time stamps.

[0016] The calibrated time stamps and the original three-phase current sampling data are mapped onto a unified reference time axis to obtain the three-phase current synchronization data sequence of the distribution network.

[0017] Preferably, in S2, the zero-mode current component and line-mode current component of the distribution network are obtained, including:

[0018] Baseline calibration is performed on the three-phase current synchronization data sequence to obtain standard three-phase current data for the distribution network;

[0019] The standard three-phase current data is mapped from the phase domain to the mode domain to obtain the original zero-mode and line-mode components of the distribution network.

[0020] Analyze the signal quality of the zero-mode original component and the line-mode original component, suppress the non-ideal noise components in the zero-mode original component and the line-mode original component, and obtain the intermediate components of the zero-mode current and the line-mode current of the distribution network.

[0021] Modal purification processing is performed on the intermediate components of zero-mode current and line-mode current to obtain the zero-mode current components and line-mode current components of the distribution network.

[0022] Preferably, in S3, the target component characterizing the high-frequency transient features is selected from the intrinsic mode components, including:

[0023] Waveform decomposition of zero-mode current components and line-mode current components yields the intrinsic mode components of the distribution network.

[0024] Extract the instantaneous frequency sequence and marginal energy spectrum of the intrinsic mode components;

[0025] Based on the instantaneous frequency sequence and marginal energy spectrum, evaluate the high-frequency marginal energy and total marginal energy of the intrinsic mode components;

[0026] Based on the high-frequency marginal energy, total marginal energy, and instantaneous frequency sequence, the intrinsic modal components are comprehensively evaluated to obtain the modal significance evaluation value of the intrinsic modal components;

[0027] Based on the modal significance evaluation value, the intrinsic modal components are sorted, and the intrinsic modal components whose modal significance evaluation value exceeds the preset threshold are taken as target components.

[0028] Preferably, the formula for calculating the modal significance evaluation value is:

[0029] ;

[0030] Where S represents the modal significance evaluation value, and E h E represents the marginal energy in the high-frequency band. t Represents the total marginal energy. This represents the variance of the instantaneous frequency sequence. This indicates the preset weighting coefficients.

[0031] Preferably, in S3, the enhanced fault traveling wave signal of the distribution network is obtained, including:

[0032] Analyze the phase consistency of the target component to obtain the degree of phase consistency of the target component;

[0033] Based on the modal saliency evaluation value and phase consistency of the target component, the corresponding fusion weight is assigned to the target component;

[0034] Based on the fusion weight, the target components are phase-aligned and superimposed to obtain the superimposed signal of the distribution network;

[0035] The superimposed signals are subjected to waveform shaping to obtain the enhanced fault traveling wave signal of the distribution network.

[0036] Preferably, in S4, the time difference observation sequence of the distribution network is obtained, including:

[0037] The enhanced fault traveling wave signal is segmented by a sliding window to obtain the signal analysis segment of the enhanced fault traveling wave signal;

[0038] A continuous and smooth curve for enhancing the fault traveling wave signal is obtained by performing cubic spline interpolation fitting on the signal analysis segment.

[0039] By analyzing the curvature value of the continuous smooth curve at the center point of the sliding window, the waveform curvature sequence of the enhanced fault traveling wave signal is obtained;

[0040] Multi-scale difference analysis was performed on the waveform curvature sequence to obtain the location of the abrupt change point in the enhanced fault traveling wave signal;

[0041] The time point at which the first abrupt change occurs is determined as the arrival time of the wavefront of the enhanced fault traveling wave signal;

[0042] By comparing and analyzing the arrival time of the wavefront with the time reference generated by the BeiDou satellite, the absolute time difference of the monitoring point is obtained;

[0043] The absolute time differences are summarized and arranged in the order of the monitoring point numbers to obtain the time difference observation sequence of the distribution network.

[0044] Preferably, in S5, the suspected fault sections of the distribution network are obtained, including:

[0045] Based on the traveling wave propagation speed of the distribution network, the absolute time difference in the time difference observation sequence is mapped to the corresponding traveling wave propagation distance difference;

[0046] Based on the topology of the distribution network, construct the electrical connection diagram of the distribution network;

[0047] Spatiotemporal matching analysis was performed between the traveling wave propagation distance difference and the electrical connection diagram. In the electrical connection diagram, the spatial location interval with the highest matching degree with the traveling wave propagation distance difference combination was located.

[0048] Based on the line segment to which the spatial location interval belongs in the electrical connection diagram, the initial candidate fault segments of the distribution network are generated.

[0049] The initial candidate fault sections are optimized and screened to obtain the suspected fault sections of the distribution network.

[0050] Preferably, the optimization and screening of the initial candidate fault sections specifically involves:

[0051] Based on the electrical connection diagram, fault deduction is performed on the initial candidate fault sections to obtain the theoretical arrival time difference relationship of the traveling wave in the distribution network.

[0052] By comparing the arrival time difference relationship of traveling wave theory with the time difference observation sequence, the degree of time difference agreement of the distribution network is obtained;

[0053] Based on the degree of time difference matching, a multi-dimensional consistency judgment is made on the initial candidate fault segments to obtain the spatial exclusion relationship between the initial candidate fault segments.

[0054] Based on spatial exclusion relationships, the initial candidate fault sections are sorted and screened to obtain the suspected fault sections of the distribution network.

[0055] Preferably, in S6, the actual fault sections of the distribution network are obtained, including:

[0056] Extract the wavefront polarity direction features associated with the suspected fault section from the enhanced fault traveling wave signal;

[0057] Based on the topology of the distribution network, the consistency between the theoretical propagation direction and the polarity direction of the wavefront of the enhanced fault traveling wave signal is determined, and the directional consistency index of the distribution network is obtained.

[0058] Extract the energy attenuation amplitude characteristics associated with suspected fault sections from the enhanced fault traveling wave signal;

[0059] The degree of matching between the electrical distance between monitoring points and suspected fault sections and the energy attenuation amplitude characteristics is evaluated to obtain the attenuation matching index of the distribution network.

[0060] By integrating the directional consistency index and the attenuation matching index, a comprehensive reliability score for the distribution network is obtained.

[0061] The suspected fault section with the highest comprehensive credibility score is taken as the actual fault section of the distribution network.

[0062] The present invention has the following beneficial effects:

[0063] This invention utilizes a high-stability clock signal output from the BeiDou satellite to achieve precise phase synchronization of the three-phase current at monitoring points in the power distribution network. Through scientific modulus transformation and mode purification, it can effectively suppress non-ideal noise components. Furthermore, through adaptive waveform decomposition, it accurately filters target components that characterize high-frequency transient features and completes signal reconstruction, which can significantly improve the clarity and reliability of fault traveling wave signals, providing high-quality and highly reliable core data support for fault location.

[0064] This invention accurately captures the arrival time of wavefronts by detecting abrupt changes in waveform curvature. It efficiently locates suspected fault sections by combining the traveling wave propagation speed with the geometric analysis of the distribution network topology. Furthermore, by comprehensively judging the polarity direction of the wavefront and the amplitude characteristics of energy attenuation, the accuracy of fault section location can be greatly improved. At the same time, the data processing and judgment logic in the location process are optimized, significantly improving the overall efficiency of fault section location in the distribution network, helping to quickly troubleshoot faults and reduce the impact on power supply. Attached Figure Description

[0065] Figure 1 This is a schematic flowchart of the method of the present invention;

[0066] Figure 2 This is a diagram showing the data acquisition results of the three-phase current synchronization in the power distribution network in an embodiment of the present invention;

[0067] Figure 3 This is a diagram of the zero-mode and line-mode current components after modulus transformation in an embodiment of the present invention;

[0068] Figure 4 This is a diagram showing the enhanced fault traveling wave signal and wavefront arrival time detection in an embodiment of the present invention. Detailed Implementation

[0069] The embodiments of the present invention will be further described below with reference to the accompanying drawings:

[0070] like Figure 1 As shown, the method for locating fault sections in a distribution network based on BeiDou satellite timing includes the following steps:

[0071] S1. Real-time acquisition of three-phase current at monitoring points on the distribution network, and using the high-stability clock signal output by the Beidou satellite to perform phase synchronization and alignment of the three-phase current, thereby obtaining the three-phase current synchronization data sequence of the distribution network.

[0072] S2. Perform modulus transformation on the three-phase current synchronization data sequence to obtain the zero-mode current component and line-mode current component of the distribution network.

[0073] S3. Perform adaptive waveform decomposition on the zero-mode current component and the line-mode current component to obtain the intrinsic mode components of the distribution network. Select the target component that characterizes the high-frequency transient characteristics from the intrinsic mode components and reconstruct the signal of the target component to obtain the enhanced fault traveling wave signal of the distribution network.

[0074] S4. Detect waveform curvature change in the enhanced fault traveling wave signal to obtain the arrival time of the wavefront of the enhanced fault traveling wave signal, and obtain the time difference observation sequence of the distribution network based on the arrival time of the wavefront and the preset time reference.

[0075] S5. Based on the time difference observation sequence and the traveling wave propagation speed of the distribution network, the topology of the distribution network is geometrically analyzed to obtain the suspected fault sections of the distribution network.

[0076] S6. Based on the wavefront polarity and energy attenuation amplitude characteristics of the enhanced fault traveling wave signal, a comprehensive assessment is conducted on the suspected fault sections to obtain the actual fault sections of the distribution network.

[0077] In S1, the three-phase current synchronization data sequence of the distribution network is obtained, including:

[0078] Raw three-phase current sampling data are collected at monitoring points in the distribution network, and initial time stamps based on local clocks are marked for the monitoring points.

[0079] It receives the standard time signal transmitted by the Beidou satellite and, based on the standard time signal, synchronizes and calibrates the local clock of the monitoring point to obtain the calibrated time tag of the monitoring point;

[0080] A unified reference time axis for the power distribution network is established based on the calibrated time stamps.

[0081] The original three-phase current sampling data is mapped onto a unified reference time axis according to the calibrated time label to obtain the three-phase current synchronization data sequence of the distribution network.

[0082] Current acquisition devices are deployed at various pre-set monitoring points in the distribution network. These devices continuously capture real-time changes in the three-phase current at each monitoring point, recording current values ​​at fixed time intervals to form continuous raw three-phase current sampling data. Each monitoring point is equipped with an independent local clock module, which records the current time in real time. Upon acquiring each set of raw three-phase current sampling data, the time information displayed on the local clock module is directly associated with the corresponding data set, completing the initial time tagging based on the local clock and ensuring that each set of raw three-phase current sampling data has a unique initial time tag.

[0083] Each monitoring point's clock calibration module continuously receives the standard time signal (containing precise time reference information) broadcast by the BeiDou satellite via a dedicated signal receiving antenna. The clock calibration module decodes the received standard time signal, extracts the standard time data, and directly compares this standard time data with the time currently displayed on the monitoring point's local clock module to find the time difference. Based on the found time difference, the operating parameters of the local clock module are adjusted to ensure that the time displayed on the local clock module is completely consistent with the standard time broadcast by the BeiDou satellite, completing the local clock synchronization calibration. The time displayed on the local clock module at this point is the calibrated time. This calibrated time is then associated with the corresponding original three-phase current sampling data to form a calibrated time tag for the monitoring point.

[0084] Collect calibrated time stamps from all monitoring points in the distribution network, extract the calibrated time information contained within, and organize and summarize all calibrated time information to determine the start and end points of the time range. Based on the time reference in the standard time signal transmitted by the BeiDou satellite, construct a continuous and unified time axis according to fixed time scale division rules. The time scale of this time axis is consistent with the time unit of the calibrated time information, ensuring that the calibrated time of all monitoring points in the distribution network can accurately correspond to a specific position on this time axis, forming a unified reference time axis for the distribution network.

[0085] Each set of raw three-phase current sampling data and its corresponding calibrated timestamp are extracted from each monitoring point. Based on the time information in the calibrated timestamps, a time node that perfectly matches the data is found on a unified reference time axis. This set of raw three-phase current sampling data is directly associated with the corresponding time node on the unified reference time axis. The raw three-phase current sampling data of all monitoring points are arranged sequentially according to the time order of the unified reference time axis, so that all data are presented in an orderly manner under the same time base, ultimately forming a three-phase current synchronization data sequence of the distribution network.

[0086] By accurately collecting raw three-phase current sampling data and marking initial time tags, and relying on the BeiDou satellite standard time signal to complete local clock synchronization calibration, a unified reference time axis is established and the data and time tags are accurately mapped, ensuring that the three-phase current data of all monitoring points are on the same time reference, effectively eliminating time deviations between different monitoring points, and providing accurate and synchronous basic data support for various data processing links in the subsequent location of fault sections in the distribution network.

[0087] In S2, the zero-mode current component and line-mode current component of the distribution network are obtained, including:

[0088] Baseline calibration is performed on the three-phase current synchronization data sequence to obtain standard three-phase current data for the distribution network;

[0089] The standard three-phase current data is mapped from the phase domain to the mode domain to obtain the original zero-mode and line-mode components of the distribution network.

[0090] Analyze the signal quality of the zero-mode original component and the line-mode original component, suppress the non-ideal noise components in the zero-mode original component and the line-mode original component, and obtain the intermediate components of the zero-mode current and the line-mode current of the distribution network.

[0091] Modal purification processing is performed on the intermediate components of zero-mode current and line-mode current to obtain the zero-mode current components and line-mode current components of the distribution network.

[0092] Using the data from the stable operation phase without fault interference in the three-phase current synchronization data sequence as a benchmark, the average value of the current data of each phase in this phase is calculated. The average value of the corresponding phase is subtracted from each data point in the three-phase current synchronization data sequence to eliminate the DC offset and baseline drift in the data, ensuring that the data is at a unified benchmark level, and thus obtaining the standard three-phase current data of the distribution network.

[0093] Using a preset modulus transformation rule, a specific linear combination operation is performed on the A-phase, B-phase, and C-phase current signals in the standard three-phase current data. Through this combination method, the three-phase current signals that were originally in the phase domain are transformed into the modulus domain space, and the components that can reflect the common change characteristics of the three-phase currents and the components that reflect the relative change characteristics between phases are separated out, which are respectively used as the zero-mode original components and line-mode original components of the distribution network.

[0094] A comprehensive observation and analysis of the waveform morphology and variation trend of the zero-mode original component and the line-mode original component is conducted to identify the disorderly fluctuations that do not follow the normal signal variation law and have no clear physical meaning. These parts are non-ideal noise components. Through targeted signal filtering processing, these disorderly fluctuations are removed from the zero-mode original component and the line-mode original component, and the effective signal components that meet the expectations are retained to obtain the intermediate components of the zero-mode current and the line-mode current of the distribution network.

[0095] The modal characteristics of the intermediate components of zero-mode current and line-mode current are further identified. Stray components that are not matched with the core modal characteristics of zero-mode or line-mode current are checked one by one and completely removed so that the intermediate components of zero-mode current retain only the pure signal characteristics related to zero-mode and the intermediate components of line-mode current retain only the pure signal characteristics related to line-mode, thus obtaining the zero-mode current components and line-mode current components of the distribution network.

[0096] Baseline calibration ensures the uniformity of the data benchmark, modulus mapping enables effective conversion from the phase domain to the mode domain and mode separation, noise suppression eliminates invalid interference components, and modal purification enhances the purity of each modal signal. The four-step operation is progressive and provides accurate and pure modal current data support for subsequent high-frequency transient feature extraction and fault traveling wave signal processing, ensuring the accuracy of subsequent fault location-related analysis.

[0097] In S3, target components characterizing high-frequency transient features are selected from the intrinsic mode components, including:

[0098] Waveform decomposition of zero-mode current components and line-mode current components yields the intrinsic mode components of the distribution network.

[0099] Extract the instantaneous frequency sequence and marginal energy spectrum of the intrinsic mode components;

[0100] Based on the instantaneous frequency sequence and marginal energy spectrum, evaluate the high-frequency marginal energy and total marginal energy of the intrinsic mode components;

[0101] Based on the high-frequency marginal energy, total marginal energy, and instantaneous frequency sequence, the intrinsic modal components are comprehensively evaluated to obtain the modal significance evaluation value of the intrinsic modal components;

[0102] Based on the modal significance evaluation value, the intrinsic modal components are sorted, and the intrinsic modal components whose modal significance evaluation value exceeds the preset threshold are taken as target components.

[0103] The formula for calculating the modal significance evaluation value is:

[0104] ;

[0105] Where S represents the modal significance evaluation value, and E h E represents the marginal energy in the high-frequency band. t Represents the total marginal energy. This represents the variance of the instantaneous frequency sequence. This represents the preset weighting coefficients, and .

[0106] Signal reconstruction is performed on the target components to obtain the enhanced fault traveling wave signal of the distribution network, including:

[0107] Analyze the phase consistency of the target component to obtain the degree of phase consistency of the target component;

[0108] Based on the modal saliency evaluation value and phase consistency of the target component, the corresponding fusion weight is assigned to the target component;

[0109] Based on the fusion weight, the target components are phase-aligned and superimposed to obtain the superimposed signal of the distribution network;

[0110] The superimposed signals are subjected to waveform shaping to obtain the enhanced fault traveling wave signal of the distribution network.

[0111] The zero-mode current component and the line-mode current component are separated step by step according to their respective oscillation characteristics. The complex signal is decomposed into multiple oscillation components with non-overlapping frequency ranges that can independently reflect the local characteristics of the signal. These oscillation components are the intrinsic mode components of the distribution network.

[0112] The frequency changes of each intrinsic mode component at different times are tracked, and the frequency values ​​corresponding to each time node are recorded one by one to form a complete instantaneous frequency sequence. At the same time, the energy distribution of each intrinsic mode component in each frequency range is statistically analyzed. By sorting out the correspondence between energy and frequency, a marginal energy spectrum reflecting the law of energy change with frequency is drawn.

[0113] The frequency range of the high-frequency band is defined based on the instantaneous frequency sequence. The energy distribution area corresponding to the high-frequency band is locked in the marginal energy spectrum. The total energy in this area is calculated to obtain the marginal energy of the high-frequency band of the intrinsic mode components. The total marginal energy of the intrinsic mode components is obtained by calculating the cumulative sum of the energy in all frequency intervals in the marginal energy spectrum.

[0114] By combining the ratio of marginal energy in the high-frequency band to the total marginal energy, and observing the stability of the instantaneous frequency sequence, these two factors are comprehensively weighed in a fixed proportion to form a value that can fully reflect the prominence of the high-frequency transient characteristics of the intrinsic mode components. This value is the modal significance evaluation value of the intrinsic mode components of the distribution network.

[0115] The high-frequency marginal energy is obtained by analyzing the marginal energy spectrum of the intrinsic mode components. The energy part corresponding to the high-frequency band is separated from the marginal energy spectrum of the intrinsic mode components.

[0116] The total marginal energy is calculated by summing the marginal energy spectra of the intrinsic mode components across the entire frequency band.

[0117] The variance of the instantaneous frequency sequence is obtained by first extracting the instantaneous frequency sequence of the intrinsic mode components, and then calculating the average of the squares of the differences between each value in the sequence and the sequence mean.

[0118] The preset weighting coefficient is a fixed value set between 0 and 1 based on the high-frequency transient characteristics extraction requirements of the traveling wave signal of the distribution network fault.

[0119] The formula for calculating the modal significance evaluation value comprehensively evaluates the characteristics of intrinsic modal components by weighted summation of two parts. The first part reflects the proportion of high-frequency components in the intrinsic modal components through the ratio of high-frequency marginal energy to total marginal energy. The second part reflects the stability of the instantaneous frequency sequence by calculating the sum of 1 divided by 1 and the variance of the instantaneous frequency sequence; the smaller the variance, the larger the result, indicating greater frequency stability.

[0120] By adjusting the influence of the two components in the comprehensive evaluation through preset weighting coefficients, the final value can quantify the ability of intrinsic mode components to characterize the high-frequency transient characteristics of distribution network faults.

[0121] When the marginal energy of the high-frequency band increases while the total marginal energy remains constant, the ratio of the marginal energy of the high-frequency band to the total marginal energy will increase, leading to an increase in the first part of the value, and consequently, an increase in the overall value. When the variance of the instantaneous frequency sequence decreases, the sum of 1 divided by 1 and the variance will increase, leading to an increase in the second part of the value, and consequently, an increase in the overall value.

[0122] When the preset weight coefficient increases, the influence weight of the first part in the comprehensive evaluation increases, while the influence weight of the second part decreases. If the value of the first part is greater than the value of the second part, the comprehensive value will increase, and vice versa.

[0123] All intrinsic modal components are arranged in descending order of their modal significance evaluation values. A fixed judgment criterion is pre-set as a preset threshold. The evaluation value of each intrinsic modal component is compared with the preset threshold one by one. Intrinsic modal components with evaluation values ​​exceeding the preset threshold are identified as target components characterizing high-frequency transient features.

[0124] By comparing the phase states of all target components at the same time point, we can observe whether the phase change trends of each target component are consistent and whether the phase values ​​at the same moment are similar. Through this direct comparison and analysis, we can determine the phase matching situation between all target components and obtain the phase consistency of the target components.

[0125] Based on the modal saliency evaluation value of the target component, the higher the evaluation value, the more significant its high-frequency transient characteristics, and the higher the weight is assigned accordingly. At the same time, the phase consistency is taken into account. The more consistent the phase of the target component, the more stable its contribution to signal reconstruction, and the higher its weight ratio will be. Combining these two factors, a unique fusion weight is determined for each target component.

[0126] Select the phase of one of the target components as the reference phase, adjust the phase state of all other target components to make the phase of all target components completely consistent at each same time node, and complete the phase alignment of the target components; then, according to their respective assigned fusion weights, perform superposition calculation on all phase-aligned target components to obtain the superimposed signal of the distribution network.

[0127] The waveform of the superimposed signal is smoothed to remove minor spikes and irregular fluctuations generated during the superposition process. At the same time, the characteristic parts of the waveform related to the fault traveling wave are enhanced to make the peak, rising edge and falling edge of the waveform clearer and more distinct, and to make the signal characteristics corresponding to the fault traveling wave more prominent, so as to obtain the enhanced fault traveling wave signal of the distribution network.

[0128] The intrinsic mode components are accurately extracted through adaptive waveform decomposition. The core target components are selected by combining multi-dimensional feature extraction and comprehensive evaluation. Then, the feature recognition of the fault traveling wave signal is effectively enhanced by phase alignment superposition and waveform shaping. The purity and clarity of the signal are improved, providing a high-quality signal foundation for subsequent fault traveling wave front detection and time difference calculation, and ensuring the accuracy of fault location-related analysis.

[0129] In S4, the time difference observation sequence of the distribution network is obtained, including:

[0130] The enhanced fault traveling wave signal is segmented by a sliding window to obtain the signal analysis segment of the enhanced fault traveling wave signal;

[0131] A continuous and smooth curve for enhancing the fault traveling wave signal is obtained by performing cubic spline interpolation fitting on the signal analysis segment.

[0132] By analyzing the curvature value of the continuous smooth curve at the center point of the sliding window, the waveform curvature sequence of the enhanced fault traveling wave signal is obtained;

[0133] Multi-scale difference analysis was performed on the waveform curvature sequence to obtain the location of the abrupt change point in the enhanced fault traveling wave signal;

[0134] The time point at which the first abrupt change occurs is determined as the arrival time of the wavefront of the enhanced fault traveling wave signal;

[0135] By comparing and analyzing the arrival time of the wavefront with the time reference generated by the BeiDou satellite, the absolute time difference of the monitoring point is obtained;

[0136] The absolute time differences are summarized and arranged in the order of the monitoring point numbers to obtain the time difference observation sequence of the distribution network.

[0137] By setting a fixed-length sliding window and a uniform sliding step size, the enhanced fault traveling wave signal is sequentially truncated according to the range of the sliding window. Each truncation yields a continuous signal segment, and all the truncated continuous signal segments are the signal analysis segments of the enhanced fault traveling wave signal.

[0138] For each signal analysis segment, all signal data points within that segment are selected as interpolation nodes. Based on the rules of cubic spline interpolation, a piecewise cubic polynomial interpolation function is constructed, ensuring that the function satisfies the condition of smoothness and continuity between each adjacent interpolation node. This interpolation function is used to supplement data and fit curves in the signal analysis segment, ultimately forming a continuous and smooth curve that enhances the fault traveling wave signal.

[0139] For each continuous smooth curve, the position of the center point of the corresponding sliding window is determined. By calculating the first and second derivatives of the curve at the center point, the curvature at the center point is calculated according to the core definition logic of curvature. According to the sliding order of the sliding window, the curvature value corresponding to the center point of each window is recorded in sequence. These curvature values ​​arranged in sequence together constitute the waveform curvature sequence of the enhanced fault traveling wave signal.

[0140] Multiple different scale parameters are set, each corresponding to a specific difference window length. For the waveform curvature sequence, the difference between adjacent data points is calculated at each scale to obtain the difference sequence at different scales. The data changes in the difference sequence at each scale are compared and analyzed to find the location where the values ​​in all difference sequences change drastically. This location is the abrupt change point of the enhanced fault traveling wave signal.

[0141] The time information corresponding to all mutation points is sorted out, and these mutation points are arranged in chronological order. The earliest mutation point is selected and the time point corresponding to the mutation point is directly determined as the arrival time of the wavefront of the enhanced fault traveling wave signal.

[0142] A fixed time reference is extracted from the standard time signal transmitted by the BeiDou satellite. This time reference is a unified and accurate time reference. The arrival time of the wavefront of the enhanced fault traveling wave signal corresponding to each monitoring point is directly compared with the time reference, and the time difference between the two is calculated. This time difference is the absolute time difference of the monitoring point.

[0143] Collect the absolute time difference corresponding to each monitoring point in the distribution network. Arrange all absolute time differences in an orderly manner according to the unique numbering sequence of the monitoring points in advance, ensuring that each absolute time difference corresponds one-to-one with the corresponding monitoring point number, forming an ordered dataset, which is the time difference observation sequence of the distribution network.

[0144] The continuity and smoothness of the signal were ensured by sliding window segmentation and cubic spline interpolation fitting. The accurate identification of abrupt change points was achieved based on center point curvature calculation and multi-scale difference analysis, ensuring the accurate determination of wavefront arrival time. After comparison with the BeiDou satellite time reference and orderly arrangement, a precise and standardized time difference observation sequence was obtained, which provided high-precision time data support for subsequent time difference-based geometric analysis of fault sections, improving the accuracy and reliability of fault location.

[0145] In S5, suspected fault sections of the distribution network are obtained, including:

[0146] Based on the traveling wave propagation speed of the distribution network, the absolute time difference in the time difference observation sequence is mapped to the corresponding traveling wave propagation distance difference;

[0147] Based on the topology of the distribution network, construct the electrical connection diagram of the distribution network;

[0148] Spatiotemporal matching analysis was performed between the traveling wave propagation distance difference and the electrical connection diagram. In the electrical connection diagram, the spatial location interval with the highest matching degree with the traveling wave propagation distance difference combination was located.

[0149] Based on the line segment to which the spatial location interval belongs in the electrical connection diagram, the initial candidate fault segments of the distribution network are generated.

[0150] The initial candidate fault sections are optimized and screened to obtain the suspected fault sections of the distribution network, including:

[0151] Based on the electrical connection diagram, fault deduction is performed on the initial candidate fault sections to obtain the theoretical arrival time difference relationship of the traveling wave in the distribution network.

[0152] By comparing the arrival time difference relationship of traveling wave theory with the time difference observation sequence, the degree of time difference agreement of the distribution network is obtained;

[0153] Based on the degree of time difference matching, a multi-dimensional consistency judgment is made on the initial candidate fault segments to obtain the spatial exclusion relationship between the initial candidate fault segments.

[0154] Based on spatial exclusion relationships, the initial candidate fault sections are sorted and screened to obtain the suspected fault sections of the distribution network.

[0155] In S6, the actual fault sections of the distribution network are obtained, including:

[0156] Extract the wavefront polarity direction features associated with the suspected fault section from the enhanced fault traveling wave signal;

[0157] Based on the topology of the distribution network, the consistency between the theoretical propagation direction and the polarity direction of the wavefront of the enhanced fault traveling wave signal is determined, and the directional consistency index of the distribution network is obtained.

[0158] Extract the energy attenuation amplitude characteristics associated with suspected fault sections from the enhanced fault traveling wave signal;

[0159] The degree of matching between the electrical distance between monitoring points and suspected fault sections and the energy attenuation amplitude characteristics is evaluated to obtain the attenuation matching index of the distribution network.

[0160] By integrating the directional consistency index and the attenuation matching index, a comprehensive reliability score for the distribution network is obtained.

[0161] The suspected fault section with the highest comprehensive credibility score is taken as the actual fault section of the distribution network.

[0162] The traveling wave propagation speed of the distribution network line is determined. This speed is an inherent physical parameter of the distribution network line. The absolute time difference of each monitoring point in the time difference observation sequence is multiplied by the traveling wave propagation speed. Through this operation, the absolute time difference in the time dimension is directly converted into the distance difference in the spatial dimension. Each absolute time difference corresponds to a unique traveling wave propagation distance difference.

[0163] The topological information of the power distribution network, including line layout, node distribution, monitoring point installation location, and connection methods between lines, is comprehensively sorted out. This information is presented in a graphical way, clearly marking the start and end nodes, length, and specific location of monitoring points on each line. This accurately reflects the electrical connection logic between different parts of the power distribution network and constructs a complete electrical connection diagram of the power distribution network.

[0164] Using the electrical connection diagram as a spatial reference, all traveling wave propagation distance differences are grouped together as a whole. The degree of fit between this combination and the distance difference combination corresponding to different spatial location intervals in the electrical connection diagram is calculated one by one. During the comparison process, the spatial range with the highest numerical and logical fit with the traveling wave propagation distance difference combination is found based on the positional relationship of each monitoring point and the route of the line. This spatial range is the spatial location interval that has been located.

[0165] The electrical connection diagram clearly defines the specific line segments corresponding to the spatial location intervals mentioned above. These line segments are the parts of the distribution network with clear start and end ranges and electrical attributes. The line segments are directly identified as the initial candidate fault sections of the distribution network to ensure that the initial candidate fault sections completely correspond to the spatial location intervals.

[0166] Based on the constructed electrical connection diagram, assuming that each initial candidate fault section has a fault, the theoretical time for the traveling wave to travel from the fault section to each monitoring point is calculated by combining the location of the section, the line length and the traveling wave propagation speed. Then, by calculating the difference in theoretical arrival time between different monitoring points, the theoretical arrival time difference relationship of the traveling wave corresponding to each group of initial candidate fault sections is obtained.

[0167] The theoretical arrival time difference relationship of each initial candidate fault segment is compared with the previously obtained distribution network time difference observation sequence one by one. The numerical fit between the theoretical time difference and the observed absolute time difference is verified pair by pair. The degree of fit between the two is obtained through comprehensive numerical comparison. This degree of fit is the time difference matching degree of the distribution network.

[0168] Using the degree of time difference matching as the core criterion, and combining factors such as the spatial location of each initial candidate fault segment in the electrical connection diagram and the line connection logic, the consistency of all initial candidate fault segments is verified. If the time difference matching of two initial candidate fault segments contradicts each other and cannot simultaneously meet the logical requirements of fault location, then it is determined that there is a spatial exclusion relationship between the two.

[0169] The initial candidate fault sections are sorted from highest to lowest according to the degree of time difference matching. Sections with a high degree of time difference matching are retained first. At the same time, according to the spatial exclusion relationship, sections that are excluded from the high matching section are eliminated. After sorting and conflict elimination, the remaining initial candidate fault sections are the suspected fault sections of the distribution network.

[0170] By accurately mapping the propagation speed of traveling waves to the absolute time difference, the conversion from time to space is achieved. Spatiotemporal matching and positioning are completed based on the electrical connection diagram. Then, through fault deduction, time difference comparison and spatial exclusion relationship analysis, optimization and screening are carried out to narrow down the possible range of faults step by step, ensuring the accuracy and reliability of suspected fault sections, and laying a solid foundation for the accurate judgment of actual fault sections in the future.

[0171] set up The noise suppression ratio is 42.2 dB, the BeiDou timing accuracy is ±25 ns, and the traveling wave propagation speed is 2.96 × 10⁻⁶. 8 Based on core parameters such as m / s, modal purity of 95.3%, sliding window length of 5ms, and wavefront time difference of -6.482ms, a fault location experiment was conducted in the distribution network, yielding the following results: Figures 2-4 The data shown includes three-phase current synchronization data, modulus transformation results, and enhanced fault traveling wave signal and wavefront detection results.

[0172] Figure 2 The results of synchronous acquisition of three-phase current at multiple monitoring points based on BeiDou satellite timing were demonstrated. Figure 2 The three-phase currents (blue, green, and red curves) of phases A, B, and C exhibit standard power frequency sinusoidal fluctuation characteristics within a 0-200ms time window, with amplitudes ranging from approximately 0.8 to 1.6A, and a 120-degree phase difference between the three phases. Through a high-stability clock signal output from the BeiDou satellite (time accuracy ±25ns), the local clocks at each monitoring point are synchronized, and the original three-phase current sampling data are mapped to a unified reference time axis. This eliminates time deviations between different monitoring points and provides a precise and synchronized data foundation for subsequent fault transient feature extraction.

[0173] Figure 3 The figure presents the zero-mode current component (purple) and the line-mode current component (orange) after modulus transformation. Through phase-domain to mode-domain mapping, the zero-mode current component stabilizes at a DC level of approximately 2.0 A, reflecting the common variation characteristics of the three-phase currents; the line-mode current component exhibits periodic fluctuations of approximately ±0.3 A, characterizing the relative variation characteristics between phases. The high-frequency noise of the zero-mode component is effectively suppressed, while the line-mode component retains clear power frequency characteristics, demonstrating the effectiveness of baseline calibration, noise suppression, and modal purification processing. The modal purity reaches 95.3%, providing clean modal data for adaptive waveform decomposition and fault traveling wave extraction.

[0174] Figure 4 The enhanced fault traveling wave signal (red curve) after adaptive waveform decomposition and signal reconstruction is shown, along with the wavefront identification results based on waveform curvature abrupt change detection. Through sliding window segmentation, cubic spline interpolation fitting, and multi-scale difference analysis, the algorithm accurately captured the wavefront arrival time at 93.49 ms (marked by the red dashed line). The enhanced traveling wave signal has an amplitude of 0.85 A, a noise suppression ratio of 42.2 dB, and clear and sharp wavefront features. Combined with the BeiDou satellite time reference, the wavefront time difference was calculated to be -6.482 ms. This high-precision time difference observation will be used for subsequent topological geometric analysis and fault segment location.

[0175] The algorithm involved in this embodiment can be executed by an electronic device, which includes a memory, a processor, and a computer program stored in the memory and capable of running on the processor. The processor executes the software to implement the above-mentioned algorithm calculation.

Claims

1. A method for locating fault sections in a distribution network based on BeiDou satellite timing, characterized in that, The methods include: S1. Real-time acquisition of the three-phase current at monitoring points on the distribution network, and using the clock signal output by the Beidou satellite to perform phase synchronization and alignment of the three-phase current to obtain the three-phase current synchronization data sequence of the distribution network. S2. Perform modulus transformation on the three-phase current synchronization data sequence to obtain the zero-mode current component and line-mode current component of the distribution network. S3. Perform adaptive waveform decomposition on the zero-mode current component and the line-mode current component to obtain the intrinsic mode components of the distribution network. Select the target component that characterizes the high-frequency transient characteristics from the intrinsic mode components and reconstruct the signal of the target component to obtain the enhanced fault traveling wave signal of the distribution network. S4. Detect waveform curvature change in the enhanced fault traveling wave signal to obtain the arrival time of the wavefront of the enhanced fault traveling wave signal, and obtain the time difference observation sequence of the distribution network based on the arrival time of the wavefront and the preset time reference. S5. Based on the time difference observation sequence and the traveling wave propagation speed of the distribution network, the topology of the distribution network is geometrically analyzed to obtain the suspected fault sections of the distribution network. S6. Based on the wavefront polarity and energy attenuation amplitude characteristics of the enhanced fault traveling wave signal, a comprehensive analysis is conducted on the suspected fault sections to obtain the actual fault sections of the distribution network. In S3, target components characterizing high-frequency transient features are selected from the intrinsic mode components, including: Waveform decomposition of zero-mode current components and line-mode current components yields the intrinsic mode components of the distribution network. Extract the instantaneous frequency sequence and marginal energy spectrum of the intrinsic mode components; Based on the instantaneous frequency sequence and marginal energy spectrum, evaluate the high-frequency marginal energy and total marginal energy of the intrinsic mode components; Based on the high-frequency marginal energy, total marginal energy, and instantaneous frequency sequence, the intrinsic modal components are comprehensively evaluated to obtain the modal significance evaluation value of the intrinsic modal components; Based on the modal significance evaluation value, the intrinsic modal components are sorted, and the intrinsic modal components whose modal significance evaluation value exceeds the preset threshold are taken as target components. The formula for calculating the modal significance evaluation value is: ； Where S represents the modal significance evaluation value, and E h E represents the marginal energy in the high-frequency band. t Represents the total marginal energy. This represents the variance of the instantaneous frequency sequence. This indicates the preset weighting coefficients.

2. The method for locating fault sections in a distribution network based on BeiDou satellite timing as described in claim 1, characterized in that, In S1, the three-phase current synchronization data sequence of the distribution network is obtained, including: Raw three-phase current sampling data are collected at monitoring points in the distribution network, and initial time stamps based on local clocks are marked for the monitoring points. It receives the standard time signal transmitted by the Beidou satellite and, based on the standard time signal, synchronizes and calibrates the local clock of the monitoring point to obtain the calibrated time tag of the monitoring point; A unified reference time axis for the power distribution network is established based on the calibrated time stamps. The calibrated time stamps and the original three-phase current sampling data are mapped onto a unified reference time axis to obtain the three-phase current synchronization data sequence of the distribution network.

3. The method for locating fault sections in a distribution network based on BeiDou satellite timing as described in claim 1, characterized in that, In S2, the zero-mode current component and line-mode current component of the distribution network are obtained, including: Baseline calibration is performed on the three-phase current synchronization data sequence to obtain standard three-phase current data for the distribution network; The standard three-phase current data is mapped from the phase domain to the mode domain to obtain the original zero-mode and line-mode components of the distribution network. Analyze the signal quality of the zero-mode original component and the line-mode original component, suppress the non-ideal noise components in the zero-mode original component and the line-mode original component, and obtain the intermediate components of the zero-mode current and the line-mode current of the distribution network. Modal purification processing is performed on the intermediate components of zero-mode current and line-mode current to obtain the zero-mode current components and line-mode current components of the distribution network.

4. The method for locating fault sections in a distribution network based on BeiDou satellite timing as described in claim 1, characterized in that, In S3, the enhanced fault traveling wave signal of the distribution network is obtained, including: Analyze the phase consistency of the target component to obtain the degree of phase consistency of the target component; Based on the modal saliency evaluation value and phase consistency of the target component, the corresponding fusion weight is assigned to the target component; Based on the fusion weight, the target components are phase-aligned and superimposed to obtain the superimposed signal of the distribution network; The superimposed signals are subjected to waveform shaping to obtain the enhanced fault traveling wave signal of the distribution network.

5. The method for locating fault sections in a distribution network based on BeiDou satellite timing as described in claim 1, characterized in that, In S4, the time difference observation sequence of the distribution network is obtained, including: The enhanced fault traveling wave signal is segmented by a sliding window to obtain the signal analysis segment of the enhanced fault traveling wave signal; A continuous and smooth curve for enhancing the fault traveling wave signal is obtained by performing cubic spline interpolation fitting on the signal analysis segment. By analyzing the curvature value of the continuous smooth curve at the center point of the sliding window, the waveform curvature sequence of the enhanced fault traveling wave signal is obtained; Multi-scale difference analysis was performed on the waveform curvature sequence to obtain the location of the abrupt change point in the enhanced fault traveling wave signal; The time point at which the first abrupt change occurs is determined as the arrival time of the wavefront of the enhanced fault traveling wave signal; By comparing and analyzing the arrival time of the wavefront with the time reference generated by the BeiDou satellite, the absolute time difference of the monitoring point is obtained; The absolute time differences are summarized and arranged in the order of the monitoring point numbers to obtain the time difference observation sequence of the distribution network.

6. The method for locating fault sections in a distribution network based on BeiDou satellite timing as described in claim 1, characterized in that, In S5, suspected fault sections of the distribution network are obtained, including: Based on the traveling wave propagation speed of the distribution network, the absolute time difference in the time difference observation sequence is mapped to the corresponding traveling wave propagation distance difference; Based on the topology of the distribution network, construct the electrical connection diagram of the distribution network; Spatiotemporal matching analysis was performed between the traveling wave propagation distance difference and the electrical connection diagram. In the electrical connection diagram, the spatial location interval with the highest matching degree with the traveling wave propagation distance difference combination was located. Based on the line segment to which the spatial location interval belongs in the electrical connection diagram, the initial candidate fault segments of the distribution network are generated. The initial candidate fault sections are optimized and screened to obtain the suspected fault sections of the distribution network.

7. The method for locating fault sections in a distribution network based on BeiDou satellite timing as described in claim 6, characterized in that, The optimization and screening of the initial candidate fault sections is as follows: Based on the electrical connection diagram, fault deduction is performed on the initial candidate fault sections to obtain the theoretical arrival time difference relationship of the traveling wave in the distribution network. By comparing the arrival time difference relationship of traveling wave theory with the time difference observation sequence, the degree of time difference agreement of the distribution network is obtained; Based on the degree of time difference matching, a multi-dimensional consistency judgment is made on the initial candidate fault segments to obtain the spatial exclusion relationship between the initial candidate fault segments. Based on spatial exclusion relationships, the initial candidate fault sections are sorted and screened to obtain the suspected fault sections of the distribution network.

8. The method for locating fault sections in a distribution network based on BeiDou satellite timing as described in claim 1, characterized in that, In S6, the actual fault sections of the distribution network are obtained, including: Extract the wavefront polarity direction features associated with the suspected fault section from the enhanced fault traveling wave signal; Based on the topology of the distribution network, the consistency between the theoretical propagation direction and the polarity direction of the wavefront of the enhanced fault traveling wave signal is determined, and the directional consistency index of the distribution network is obtained. Extract the energy attenuation amplitude characteristics associated with suspected fault sections from the enhanced fault traveling wave signal; The degree of matching between the electrical distance between monitoring points and suspected fault sections and the energy attenuation amplitude characteristics is evaluated to obtain the attenuation matching index of the distribution network. By integrating the directional consistency index and the attenuation matching index, a comprehensive reliability score for the distribution network is obtained. The suspected fault section with the highest comprehensive credibility score is taken as the actual fault section of the distribution network.

Images