Ground fault line selection method and system for distribution network

Abstract

Embodiments of the present invention provide a method and system for ground fault line selection in a distribution network, belonging to the field of ground fault line selection in distribution networks. The method includes collecting ground fault-related data and preprocessing the ground fault-related data; constructing an arc equivalent circuit model using the preprocessed ground fault-related data; establishing a multi-source dynamic verification feature cluster based on the output of the arc equivalent circuit model and the preprocessed ground fault-related data, wherein the multi-source dynamic verification feature cluster outputs a trusted dynamic feature fingerprint; and performing ground fault line selection based on the trusted dynamic feature fingerprint through cross-dimensional fusion and decision optimization. Through dynamic noise reduction, adaptive arc modeling, and multi-source feature fusion, the present invention improves the accuracy and anti-interference capability of ground fault detection, achieving reliable line selection under complex working conditions.

Description

Technical Field

[0001] The present invention relates to the technical field of ground fault line selection in a distribution network, and in particular to a ground fault line selection method and system for a distribution network. Background Art

[0002] As a vital component of the power system, the rapid and accurate detection of ground faults in distribution networks is crucial for ensuring power supply reliability. Due to the complex topology of distribution networks, the diverse neutral grounding methods (such as arc suppression coil grounding and low-resistance grounding), and the influence of fault characteristics on factors such as arc nonlinearity, environmental noise, and line parameter fluctuations, ground fault line selection remains a technical challenge in the industry.

[0003] Traditional line selection methods mainly rely on steady-state zero-sequence current amplitude, phase, or harmonic characteristics for judgment. However, in scenarios such as high-resistance grounding and intermittent arcing, the fault signal is weak and the characteristics are fuzzy. It is easily affected by the distributed capacitance of the line and load fluctuations, resulting in a high misjudgment rate. In recent years, methods such as high-frequency component analysis and traveling wave detection based on transient signals have been gradually applied. However, transient characteristics are easily affected by the randomness of the arc and the frequency response characteristics of the sensor, and lack the ability to model the dynamic evolution process of the fault and the relationship between spatial topology. In addition, existing arc models mostly use fixed resistance or simplified piecewise linearization models, which are difficult to reflect the time-varying nonlinear characteristics of the arc and the energy interaction process, resulting in insufficient feature extraction reliability. Summary of the Invention

[0004] The purpose of the embodiments of the present invention is to provide a ground fault line selection method and system for a distribution network, which improves the accuracy and anti-interference capability of ground fault detection through dynamic noise reduction, adaptive arc modeling and multi-source feature fusion, and realizes reliable line selection under complex working conditions.

[0005] In a first aspect, to achieve the above-mentioned objectives, an embodiment of the present invention provides a method for ground fault line selection in a distribution network, comprising: collecting ground fault-related data and preprocessing the ground fault-related data; constructing an arc equivalent circuit model using the preprocessed ground fault-related data; establishing a multi-source dynamic verification feature cluster based on the output of the arc equivalent circuit model and the preprocessed ground fault-related data, the multi-source dynamic verification feature cluster outputting a trusted dynamic feature fingerprint; and performing ground fault line selection based on the trusted dynamic feature fingerprint through cross-dimensional fusion and decision optimization.

[0006] Optionally, constructing the arc equivalent circuit model includes: defining the change of arc resistance over time and constructing a time-varying arc resistance model; establishing an arc equivalent circuit model including inductance, capacitance and nonlinear resistance based on the time-varying arc resistance model; and updating the initial arc resistance and line equivalent inductance in the arc equivalent circuit model in real time through the RLS algorithm.

[0007] Optionally, the establishment of a multi-source dynamic verification feature cluster includes: establishing a dynamic behavior feature group, a spatial topology feature group and a high-frequency fingerprint feature group, the dynamic behavior feature group includes the arc evolution trajectory and energy interaction characteristics, the spatial topology feature group includes the network structure fingerprint and the fault propagation path, and the high-frequency fingerprint feature group includes the spatiotemporal distribution law of high-frequency pulses and the phase relationship evolution of zero-sequence voltage and current.

[0008] Optionally, the establishment of a multi-source dynamic verification feature cluster also includes: calculating the time-varying box size, and calculating the fractal dimension based on the time-varying box size; calculating the dimensionality change rate of adjacent windows based on the fractal dimension; constructing a fractal dimension attenuation trajectory based on the dimensionality change rate of adjacent windows; and verifying the credibility of the data in the high-frequency fingerprint feature group based on the fractal dimension attenuation trajectory.

[0009] Optionally, verifying the credibility of the data in the high-frequency fingerprint feature group includes: in the first time period of arc triggering, not limiting the fractal dimension change rate, so as to completely capture the complexity mutation characteristics of the fault breakdown moment; in the second time period of arc triggering, limiting the fractal dimension change rate to a first interval, and if it exceeds the first interval, it is judged as abnormal; after the second time period of arc triggering, limiting the fractal dimension change rate to a second interval, and if it exceeds the second interval, it is judged as abnormal.

[0010] Optionally, the second interval is dynamically adjusted according to environmental noise, line length and sensor accuracy.

[0011] Optionally, the ground fault line selection through cross-dimensional fusion and decision optimization includes: multi-source feature time synchronization, spatial topology mapping, physical constraint verification, dynamic weight allocation, three-dimensional decision coordinate system construction and field strength aggregation.

[0012] Optionally, the multi-source feature time synchronization is based on the fault triggering moment, and interpolation and resampling of multi-source signals are performed; the spatial topology mapping abstracts the distribution network into a weighted graph structure, and quantifies the spatial attributes of the line nodes; the physical constraint verification is used to judge the physical rationality between features; the time axis of the three-dimensional decision coordinate system is the continuity evidence of the fault development stage, the frequency domain axis is the credibility accumulation of high-frequency transient features, and the space axis is the logical consistency verification of topological associations; the field strength aggregation is to convert the evidence of each dimension into energy density distribution in the field space, and locate the field strength extreme point through gradient tracking, so as to determine the fault line.

[0013] Optionally, the physical rationality between the judgment features includes: judging whether the high-frequency energy distribution conforms to the attenuation law determined by the line length, judging whether the arc dynamic parameters are within the value range allowed by the topological structure, and judging whether the phase distortion direction is spatially consistent with the fault location.

[0014] On the other hand, the present invention provides a ground fault line selection system for a distribution network, which is used to implement a ground fault line selection method for a distribution network. The system includes a control module, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the computer program to implement the ground fault line selection method for the distribution network.

[0015] This technical solution, by constructing a time-varying arc equivalent circuit model and introducing the RLS algorithm for real-time parameter updates, overcomes the linear simplification limitations of traditional arc models. It accurately characterizes the dynamic coupling characteristics of arc resistance, inductance, and capacitance, resolving feature distortion issues in high-resistance grounding and intermittent arcing scenarios. A multi-source dynamic verification feature cluster integrates dynamic behavioral feature groups (arc evolution trajectory, energy interaction), spatial topology feature groups (network structure fingerprint, fault propagation path), and high-frequency fingerprint feature groups (spatiotemporal distribution, phase evolution). Combined with a dynamic credibility verification mechanism based on fractal dimension attenuation trajectories, this approach enables adaptive screening and credibility quantification of high-frequency transient features in noisy environments, enhancing anti-interference capabilities. Through multi-source feature time synchronization, spatial topology mapping, three-dimensional decision coordinate system construction and field strength aggregation technology, the time domain continuity evidence, frequency domain credibility accumulation and spatial logic consistency verification are deeply integrated, which solves the limitations of traditional single-dimensional threshold criteria. In complex distribution networks containing distributed power sources and cable-overhead line hybrids, line selection is more accurate and positioning time is shorter, providing highly reliable and adaptable technical support for distribution network grounding fault detection.

[0016] Other features and advantages of the embodiments of the present invention will be described in detail in the subsequent detailed description. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] The accompanying drawings are used to provide a further understanding of the embodiments of the present invention and constitute a part of the specification. Together with the following detailed description, they are used to explain the embodiments of the present invention, but do not constitute a limitation of the embodiments of the present invention. In the accompanying drawings:

[0018] Figure 1 It is a flow chart of a ground fault line selection method for a distribution network.

[0019] Figure 2 It is a flow chart of high-frequency fingerprint feature verification. DETAILED DESCRIPTION

[0020] The following is combined with Figure 1 -Attached Figure 2 The specific implementation of the embodiment of the present invention is described in detail. It should be understood that the specific implementation described here is only used to illustrate and explain the embodiment of the present invention, and is not used to limit the embodiment of the present invention.

[0021] It should be noted that the acquisition, transmission, storage, use, and processing of data in the technical solution of this application are in compliance with the relevant provisions of national laws and regulations. In the embodiments of this application, certain software, components, models, and other existing solutions in the industry may be mentioned. These should be considered as exemplary. Their purpose is only to illustrate the feasibility of implementing the technical solution of this application, but it does not mean that the applicant has or will necessarily use such solutions.

[0022] In the process of realizing the present invention, the inventors of the present application found that the traditional method has problems such as the use of fixed threshold wavelet denoising during signal preprocessing, which easily leads to the loss of high-frequency features of transient arcs; the use of static arc models cannot accurately describe the time-varying characteristics of nonlinear impedances; the sole reliance on high-frequency fingerprints is susceptible to electromagnetic interference and causes misjudgment; and the lack of multi-dimensional feature fusion leads to insufficient accuracy in fault line selection.

[0023] Example 1

[0024] Reference Figure 1-Figure 2 , which is a first embodiment of the present invention, provides a ground fault line selection method for a distribution network, comprising:

[0025] S100: collecting ground fault related data and preprocessing the ground fault related data.

[0026] Specifically, the neutral point voltage, power oscillations of each feeder, environmental data, insulation aging data, etc. are collected; and high-frequency current sensors (bandwidth 2.5MHz) and voltage sensors are installed on each branch line of the distribution network to collect high-frequency transient characteristics, which is conducive to the detection of weak arc discharge signals.

[0027] Preferably, an 8th-order Butterworth low-pass filter with a cutoff frequency of 1.2 MHz is added before sampling to suppress high-frequency noise, avoid high-frequency components from being mixed into low frequencies, and ensure signal purity.

[0028] Furthermore, a trigger mechanism is set up. When the zero-sequence voltage mutation exceeds 5% of the rated voltage, the fault recording is triggered to record the 4-cycle data before and after the fault (sampling rate 2.5MHz).

[0029] Preferably, the collected ground fault related data is preprocessed to suppress noise and enhance fault characteristics, thereby laying a data foundation for high-precision modeling and identification.

[0030] Optimally, traditional wavelet threshold noise reduction (such as the Donoho threshold) oversmoothes transient arc signals, requiring dynamic threshold adjustment to preserve high-frequency features. This preprocessing solution includes a six-layer decomposition using the db4 wavelet basis. The number of wavelet decomposition layers is determined based on a balance between signal bandwidth coverage requirements and computational efficiency.

[0031] Furthermore, the wavelet dynamic threshold is adjusted based on the amplitude distribution of detail coefficients of each layer (max / median ratio). The wavelet dynamic threshold adjustment formula is as follows:

[0032]

[0033] in, represents the denoising threshold of the j-th layer wavelet coefficient, represents the standard deviation of the j-th layer wavelet coefficients, N represents the signal length, represents the kth wavelet coefficient of the jth layer, is the wavelet coefficient with the largest absolute value, representing mutation; is the median of the absolute values ​​of the coefficients in the jth layer, which is used to measure the degree of concentration.

[0034] Optimally, this formula takes into account the discreteness and non-Gaussianity of wavelet coefficients, and can adapt to the frequency domain distribution characteristics of different fault types. After noise reduction, the signal-to-noise ratio is improved by no less than 12dB, providing clear input for subsequent arc modeling and nonlinear feature extraction.

[0035] Preferably, the combined design of a high-frequency sensor (bandwidth 2.5MHz) and a low-pass filter (cut-off frequency 1.2MHz) effectively suppresses high-frequency noise aliasing, ensures signal purity, and accurately captures the transient characteristics of weak arcs. The wavelet denoising method (db4 wavelet basis + 6-layer decomposition) with dynamic threshold adjustment significantly improves the signal-to-noise ratio (≥12dB) while retaining high-frequency mutation signals (such as arc pulses), solving the problem of excessive smoothing of transient characteristics by traditional threshold denoising. Combined with the fault recording mechanism triggered by the zero-sequence voltage mutation (recording 4-cycle data before and after the fault), it provides a high-fidelity and high-timeliness data foundation for subsequent modeling and feature extraction.

[0036] S200: Constructing an arc equivalent circuit model using the pre-processed ground fault related data.

[0037] Specifically, the change of arc resistance over time is defined, and a time-varying arc resistance model is constructed. The expression of the time-varying arc resistance model is as follows:

[0038]

[0039] in, is the arc resistance at time t, is the initial arc resistance, is the arc current.

[0040] Preferably, this formula reflects the characteristic that arc impedance is affected by the current change rate, which is helpful for distinguishing between stable arcs and transient intermittent arcs.

[0041] Furthermore, considering the distributed parameters and nonlinear impedance characteristics of the distribution line, an arc equivalent circuit model including inductance, capacitance and nonlinear resistance is established. The formula of the arc equivalent circuit model is as follows:

[0042]

[0043] in, is the arc current; is the arc voltage; is the arc resistance at time t, which changes with time; is the equivalent inductive reactance of the line; is the line-to-ground capacitance; is the integral variable used to calculate the current integral, where t is the current moment.

[0044] Preferably, the arc equivalent circuit model combines the inductive and capacitive energy storage effects to reflect the response dynamics after arc triggering, and can accurately capture the current change trend during the initial arc ignition, stable arc burning, and extinction process.

[0045] Furthermore, a recursive least squares estimation (RLS) is performed every 5ms to update the arc model. and , realizing adaptive modeling of non-stationary arcs.

[0046] Preferably, nonlinear tools are used to characterize the complex behavior of the arc process and enhance the understanding of the nature of the fault.

[0047] Optimally, the arc equivalent circuit model can accurately characterize the dynamic characteristics of the arc from breakdown, arcing to extinction (such as nonlinear impedance mutation and periodic fluctuation), solving the defect of traditional linear model in describing transient process inadequately; the model parameters are updated in real time through recursive least squares (RLS) every 5ms. and , which enhances the adaptability to non-stationary arc conditions and provides physically explainable modeling support for distinguishing stable arcs from transient intermittent arcs.

[0048] S300: establishing a multi-source dynamic verification feature cluster according to the output of the arc equivalent circuit model and pre-processed ground fault related data, wherein the multi-source dynamic verification feature cluster outputs a credible dynamic feature fingerprint.

[0049] Specifically, the time-varying box size is calculated. The calculation formula for the time-varying box size is as follows:

[0050]

[0051] in, is the time-varying box size in the dynamic fractal dimension calculation, which is used to quantify the complexity change of the arc fault signal on the time axis; here t is the arc duration, which is calculated from the fault triggering moment; is the initial box size; is the box size decay coefficient, which is used to control the rate at which the box size decreases over time; is the segmentation threshold time (10ms); is the minimum box size.

[0052] Preferably, the initial box size and box size attenuation coefficient According to the current mutation trend output by the arc equivalent circuit model, the minimum box size Adapt to the collected line length, noise level and steady-state signal characteristics.

[0053] Preferably, in the initial stage (t≤10ms), the time-varying box size decreases linearly. , which can improve the resolution of the arc starting moment. In the steady state stage (t>10ms), the time-varying box size is fixed to , which can avoid the dimensionality drift caused by long-term calculation.

[0054] Preferably, the time-varying box size The smaller the size, the higher the time resolution, which can capture more subtle transient changes, such as arc reignition, short circuit breakdown and other sudden faults, and optimize the time-frequency focusing; the size of the time-varying box The larger the value, the larger the time window, which can effectively integrate the long-term energy distribution and trend changes of the signal, such as detecting slow-changing processes such as line temperature rise and insulation aging caused by arc faults (time scale > 1 second).

[0055] Arc fault transients (such as reignition and extinction) exhibit time-varying characteristics. Conventional methods using a fixed box size cannot adapt to these time-varying characteristics, and the choice of box size directly affects the quantization accuracy of signal complexity. This solution significantly improves arc fault detection's accuracy in analyzing complex transient characteristics and adaptability to operating conditions by adaptively adjusting the box size.

[0056] Further, Divide the time domain grid into the side length, count the number of grids containing signal points, and calculate the fractal dimension. The calculation formula of the fractal dimension is as follows:

[0057]

[0058] in, is the fractal dimension, is the time-varying box size, is the number of non-empty boxes.

[0059] Furthermore, the window is slid along the time axis to obtain the fractal dimension sequence ; Calculate the rate of change of dimensionality of adjacent windows:

[0060]

[0061] in, represents the rate of change of dimension, is the fractal dimension, where t represents the starting time of the current analysis window, is the step size of the sliding window.

[0062] Preferably, a fractal dimension attenuation trajectory is constructed based on the dimensionality change rate of adjacent windows. The fractal dimension attenuation trajectory is used for subsequent verification of the credibility of the high-frequency fingerprint.

[0063] Furthermore, based on the preprocessed data and the output of the arc equivalent circuit model, a multi-source dynamic verification feature cluster of "dynamic behavior-spatial topology-high-frequency fingerprint" was established. Through the physical correlation and information complementarity between features, a fault judgment system with strong explanatory power was obtained.

[0064] Specifically, the multi-source dynamic verification feature cluster includes a dynamic behavior feature group, a spatial topology feature group, and a high-frequency fingerprint feature group.

[0065] (1) The dynamic behavior characteristic group includes arc evolution trajectory and energy interaction characteristics.

[0066] The arc evolution trajectory is designed to capture the dynamics of the entire arc process from breakdown to stability, including the nonlinear impedance mutation in the initial breakdown stage, the periodic fluctuations in the stable arcing period, and the transient impact mode of the reignition / extinguishing events.

[0067] The energy interaction characteristics are used to analyze the energy exchange rules between the fault point and the system, including the release path of high-frequency transient energy, the coupling relationship between power frequency energy and high-frequency energy, and the attenuation gradient of energy propagation.

[0068] (2) The spatial topological feature group includes network structure fingerprint and fault propagation path.

[0069] The network structure fingerprint is the functional positioning of the line in the distribution network, including the branch level (trunk / branch / terminal), electrical distance (equivalent impedance to the power source point), and coupling strength of adjacent lines.

[0070] The fault propagation path is the fault impact range deduced based on electromagnetic coupling and topological connection, including the set of directly disturbed lines, the propagation boundary of secondary induced interference, and the chain reaction path of protection action.

[0071] (3) The high-frequency fingerprint feature group includes the spatiotemporal distribution of high-frequency pulses and the evolution of the phase relationship between zero-sequence voltage and current.

[0072] The temporal and spatial distribution laws of high-frequency pulses include the amplitude / polarity characteristics of the leading pulse, the resonant frequency components of the oscillation attenuation, and the statistical characteristics of the time intervals of the pulse clusters.

[0073] The evolution of the phase relationship between zero-sequence voltage and current includes the phase jump direction at the initial stage of the fault, the phase difference drift rate in the steady-state stage, and the phase oscillation mode caused by arc restrike.

[0074] Furthermore, high-frequency fingerprints (such as pulse amplitude and phase jump) are susceptible to electromagnetic interference contamination, and directly relying on them for fault location may lead to misjudgment.

[0075] Preferably, the fractal dimension decay trajectory is used to verify the credibility of the high-frequency fingerprint. After the arc is triggered, in the first time period (within the first 5 milliseconds), the fractal dimension change rate is not limited. , used to fully capture the complexity mutation characteristics of the fault breakdown moment; in the second time period of arc triggering (within 5 to 15 milliseconds), Limited to the first interval (-1.5≤ ≤+2.0), if it exceeds the first interval, it is considered abnormal; after the arc triggers the second time period, that is (after 15 milliseconds), Strictly limited to the second interval (-0.3≤ If the noise level is high (e.g. near a substation), the second interval can be widened by 1.5 times (-0.45≤ ≤+0.45); if the line length exceeds 10 kilometers or the sensor accuracy is high, the second interval is tightened to 0.8 times (-0.24≤ ≤+0.24).

[0076] Preferably, the transient process of arc fault (breakdown, restrike) will cause a sudden change in signal complexity. The complexity of random noise changes irregularly. Through dynamic interval adjustment, more than 90% of interference signals (such as false high-frequency fingerprints caused by equipment switching) can be eliminated and the false alarm rate can be reduced to below 5%.

[0077] Furthermore, through multi-source dynamic verification feature clusters, the fractal dimension attenuation trajectory is used to perform dynamic credibility verification on high-frequency fingerprints, and physical constraint verification is used to achieve cross-verification of multi-source features, and finally output a reliable dynamic feature fingerprint with strong anti-interference and physical interpretability.

[0078] Optimally, the fractal dimension calculation based on the time-varying box size optimizes the time-frequency focusing of transient details (such as arc reignition) and long-term trends (such as insulation aging), significantly improving the quantification accuracy of complex fault features. The credibility of high-frequency fingerprints is verified through the fractal dimension attenuation trajectory, and the physical correlation of multi-source feature groups is cross-verified to effectively eliminate false features caused by electromagnetic interference, reduce the false alarm rate, and solve the pain point that traditional single features are easily affected by noise.

[0079] S400: Based on trusted dynamic feature fingerprints, ground fault line selection is performed through cross-dimensional fusion and decision optimization.

[0080] Specifically, the time axis is synchronized, the fault triggering moment is used as the benchmark, a time scale system with millisecond-level accuracy is established, and multi-source signals are interpolated and resampled to eliminate acquisition synchronization errors.

[0081] Furthermore, spatial topological mapping is performed to abstract the distribution network into a weighted graph structure and quantify the spatial attributes of line nodes; an electrical correlation matrix between lines is constructed to describe the energy propagation path.

[0082] Furthermore, physical constraint checks are performed to verify the physical rationality of the features, including whether the high-frequency energy distribution conforms to the attenuation law determined by the line length, whether the arc dynamic parameters are within the value range allowed by the topological structure, and whether the phase distortion direction is spatially consistent with the fault location.

[0083] Preferably, a dynamic evaluation model of feature importance is constructed to dynamically adjust the weight distribution. In a low-noise environment, it focuses on detail resolution of high-frequency fingerprints; in complex topologies, it strengthens the correlation analysis of spatial propagation paths; in developmental faults, it tracks the evolution trend of dynamic behavior parameters.

[0084] Furthermore, a three-dimensional decision coordinate system is constructed, with the time axis representing continuity evidence of fault development stages; the frequency domain axis representing the credibility accumulation of high-frequency transient characteristics; and the spatial axis verifying the logical consistency of topological associations. This cross-dimensional fusion of time, frequency, and space is quantified into a unified criterion.

[0085] Furthermore, a field strength aggregation algorithm is used to convert evidence of each dimension into energy density distribution in the field space, and the extreme point of field strength is located through gradient tracking to determine the fault line.

[0086] Preferably, decision optimization is performed through dynamic weight allocation, physical rule verification and field strength aggregation algorithm, the most reliable fault judgment criteria are screened out from multi-source features, and finally high-precision line selection is achieved.

[0087] Preferably, a cross-dimensional fusion mechanism based on a three-dimensional decision coordinate system (time-frequency-space) realizes the spatiotemporal consistency verification of fault characteristics through millisecond-level time-scale synchronization, spatial topology mapping, and physical constraint verification; combined with the field strength aggregation algorithm, multi-dimensional evidence is converted into energy density distribution, and the field strength extreme points are located through gradient tracking, which significantly improves the line selection accuracy in complex noise environments and long line scenarios; the dynamic feature importance evaluation mechanism (focusing on high-frequency fingerprints, propagation paths or dynamic parameter evolution) further enhances the robustness and adaptability of the solution to high-noise, multi-distributed power grids.

[0088] The present invention also provides a ground fault line selection system for a distribution network, which is used to implement a ground fault line selection method for a distribution network. The system includes a control module, which includes a memory, a processor, and a computer program stored in the memory and runnable on the processor. The processor executes the computer program to implement the ground fault line selection method for the distribution network.

[0089] An embodiment of the present invention provides a storage medium storing a program, which, when executed by a processor, implements the ground fault line selection method for a power distribution network.

[0090] An embodiment of the present invention provides a processor, which is used to run a program, wherein the program executes the ground fault line selection method for a power distribution network when running.

[0091] An embodiment of the present invention provides a device comprising a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, it implements a method for selecting a ground fault line in a power distribution network. The device herein may be a server, a PC, a PAD, a mobile phone, or the like.

[0092] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing a ground fault line selection method for a power distribution network.

[0093] Those skilled in the art will appreciate that the embodiments of the present application may provide methods, systems, or computer program products. Therefore, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present application may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0094] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0095] These computer program instructions may also be stored in a computer readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0096] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0097] In a typical configuration, a computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0098] Memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. Memory is an example of a computer-readable medium.

[0099] Computer-readable media includes both permanent and non-permanent, removable and non-removable media that can be implemented using any method or technology for information storage. Information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change RAM (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic disk storage or other magnetic storage devices, or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media such as modulated data signals and carrier waves.

[0100] It should also be noted that the terms "comprises," "includes," or any other variations thereof are intended to encompass non-exclusive inclusion, such that a process, method, commodity, or apparatus that includes a series of elements includes not only those elements but also other elements not explicitly listed, or includes elements inherent to such process, method, commodity, or apparatus. In the absence of further limitations, an element defined by the phrase "comprises a ..." does not exclude the presence of other identical elements in the process, method, commodity, or apparatus that includes the element.

[0101] The above are merely embodiments of the present application and are not intended to limit the present application. For those skilled in the art, the present application may have various modifications and variations. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should be included within the scope of the claims of the present application.

Claims

1. A ground fault line selection method for a distribution network, characterized in that: include: Collecting ground fault related data and preprocessing the ground fault related data; constructing an arc equivalent circuit model using the pre-processed ground fault related data; A multi-source dynamic verification feature cluster is established based on the output of the arc equivalent circuit model and the pre-processed ground fault related data. The establishment of the multi-source dynamic verification feature cluster includes: Establish dynamic behavior feature groups, spatial topology feature groups and high-frequency fingerprint feature groups, The dynamic behavior feature group includes arc evolution trajectory and energy interaction characteristics, The spatial topology feature group includes network structure fingerprint and fault propagation path, The high-frequency fingerprint feature group includes the temporal and spatial distribution of high-frequency pulses and the phase relationship evolution of zero-sequence voltage and current; Using fractal dimension attenuation trajectories to perform dynamic credibility verification on the high-frequency fingerprints in the multi-source dynamic verification feature cluster, and achieving cross-validation of multi-source features through physical constraint verification to output a credible dynamic feature fingerprint; Based on the trusted dynamic feature fingerprint, grounding fault line selection is performed through cross-dimensional fusion and decision optimization.

2. The ground fault line selection method for distribution network according to claim 1, characterized in that: The constructing of the arc equivalent circuit model comprises: Define the change of arc resistance over time and build a time-varying arc resistance model; According to the time-varying arc resistance model, an arc equivalent circuit model including inductance, capacitance and nonlinear resistance is established; The initial arc resistance and line equivalent inductive reactance in the arc equivalent circuit model are updated in real time by the RLS algorithm.

3. The ground fault line selection method for distribution network according to claim 1, characterized in that: The establishing of a multi-source dynamic verification feature cluster further comprises: calculating a time-varying box size, and calculating a fractal dimension according to the time-varying box size; Calculating the dimension change rate of adjacent windows according to the fractal dimension; constructing a fractal dimension decay trajectory based on the dimensionality change rate of the adjacent windows; The credibility of the data in the high-frequency fingerprint feature group is verified according to the fractal dimension attenuation trajectory.

4. The ground fault line selection method for a distribution network according to claim 3, characterized in that: Verifying the credibility of the data in the high-frequency fingerprint feature group includes: In the first time period of arc triggering, the fractal dimension change rate is not restricted, so as to fully capture the complexity mutation characteristics at the moment of fault breakdown; In the second time period of arc triggering, the fractal dimension change rate is limited to the first interval, and if it exceeds the first interval, it is determined to be abnormal; After the arc is triggered for the second time period, the fractal dimension change rate is limited to the second interval, and if it exceeds the second interval, it is determined to be abnormal.

5. The ground fault line selection method for distribution network according to claim 4, characterized in that: The second interval is dynamically adjusted according to environmental noise, line length and sensor accuracy.

6. The ground fault line selection method for distribution network according to claim 1, characterized in that: The ground fault line selection through cross-dimensional fusion and decision optimization includes: multi-source feature time synchronization, spatial topology mapping, physical constraint verification, dynamic weight allocation, three-dimensional decision coordinate system construction and field strength aggregation.

7. The ground fault line selection method for a distribution network according to claim 6, characterized in that: The multi-source characteristic time synchronization is based on the fault triggering moment, and interpolation resampling is performed on the multi-source signals; The spatial topology mapping abstracts the distribution network into a weighted graph structure and quantifies the spatial attributes of line nodes; The physical constraint check is used to determine the physical rationality between features; The time axis of the three-dimensional decision coordinate system is the continuity evidence of the fault development stage, the frequency domain axis is the credibility accumulation of high-frequency transient characteristics, and the space axis is the logical consistency verification of topological association; The field strength aggregation is to convert the evidence of each dimension into the energy density distribution in the field space, locate the field strength extreme point through gradient tracking, and thus determine the fault line.

8. The ground fault line selection method for a distribution network according to claim 7, characterized in that: The physical rationality of the judgment features includes: judging whether the high-frequency energy distribution conforms to the attenuation law determined by the line length, judging whether the arc dynamic parameters are within the value range allowed by the topological structure, and judging whether the phase distortion direction is spatially consistent with the fault location.

9. A ground fault line selection system for a distribution network, characterized in that: The system includes a control module, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor. The processor executes the computer program to implement the ground fault line selection method for a distribution network according to any one of claims 1 to 8.

Images