High-resistance ground fault line selection method and system for neutral point low-resistance grounding distribution network

Abstract

This invention relates to the field of distribution network relay protection and fault location technology, and discloses a method and system for selecting fault lines in high-resistance grounding faults in distribution networks with neutral points grounded through low resistance. The method includes: determining candidate fault feeders, a set of suspected fault feeders, and a set of healthy feeders based on threshold comparisons of all passive discrimination scores; selecting the continuous frequency band that maximizes the difference in residual energy between the set of suspected fault feeders and the set of healthy feeders as candidate detection sub-bands; selecting corresponding grid-connected equipment as active detection sources based on all average zero-sequence transmission gains; calculating the active detection indication quantity of the corresponding feeder based at least on the zero-sequence current generated by each feeder in the target distribution network during injection; and determining the target fault feeder based on all active detection indication quantities and candidate fault feeders. This invention is suitable for distribution networks with neutral points grounded through low resistance and a large number of grid-connected equipment, and can improve the robustness and reliability of fault selection results in high-resistance grounding fault scenarios.

Description

Technical Field

[0001] This invention relates to the field of distribution network relay protection and fault location technology, and in particular to a method and system for locating high-resistance grounding faults in distribution networks with neutral points grounded by small resistances. Background Technology

[0002] Neutral-point grounded distribution networks with low resistance are widely used in medium-voltage distribution systems. In such systems, when a single-phase ground fault occurs, the bus zero-sequence voltage and feeder zero-sequence current change, and protection devices typically use this information to detect the fault and identify the faulty feeder. For ground faults with high fault resistance, traditional passive fault location methods are prone to problems such as insufficient fault location sensitivity, increased false alarm rate, and reduced distinguishability between faulty and healthy feeders.

[0003] With the large-scale integration of grid-connected power electronic equipment such as photovoltaic inverters, wind power converters, and energy storage converters into the distribution network, the zero-sequence network characteristics of the system have changed significantly compared to traditional distribution networks. On the one hand, the switching frequency, modulation frequency, and sideband components of grid-connected power electronic equipment introduce harmonic disturbances and non-stationary spectral components into the zero-sequence current. On the other hand, changes in the penetration rate of new energy sources, the proportion of grid-connected equipment, topology switching status, and operating conditions cause dynamic changes in the system's equivalent zero-sequence impedance and fault propagation characteristics. These changes weaken the applicability of traditional passive line selection methods.

[0004] Existing active fault location methods typically improve the identification capability of high-resistance grounding faults by comparing the responses of each feeder through manually applied signals, signals injected by arc suppression devices, signals injected by zero-sequence signal sources, or characteristic disturbances emitted by existing controllable equipment within the system. However, when the active detection frequency band overlaps with the switching frequency sideband of grid-connected power electronic equipment, the active detection signal is easily contaminated by background harmonics; when the zero-sequence coupling between the active detection source and the feeder to be identified is weak, the injected signal is unlikely to form sufficient separation on the faulty feeder, thus affecting the fault location effect. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for selecting high-resistance grounding faults in a distribution network with a neutral point grounded by a small resistor, in order to solve the problems of weak high-resistance grounding fault characteristics, easy interference of fixed frequency bands by switch sidebands, unreasonable selection of active detection sources, and difficulty in forming a stable closed-loop verification between active and passive criteria in scenarios with high penetration of new energy and widespread access of grid-connected power electronic equipment.

[0006] In a first aspect, embodiments of the present invention provide a method for selecting the fault location of a high-resistance grounding fault in a distribution network with a neutral point grounded by a small resistor, including: Acquire zero-sequence electrical data, grid-connected equipment operating status data, network topology data, and zero-sequence parameter data of the target distribution network; The passive discrimination score of each feeder in the target distribution network is obtained by multi-feature fusion calculation on the zero-sequence electrical data, and the candidate fault feeder, the set of suspected fault feeders and the set of sound feeders are determined by threshold comparison based on all the passive discrimination scores. Based on the operating status data of the grid-connected equipment, an interference frequency band set is constructed, and within the available frequency bands excluding the interference frequency band set, a continuous frequency band that makes the difference in residual energy between the suspected faulty feeder set and the healthy feeder set the largest is selected as a candidate detection sub-band. Based on the network topology data and the zero-sequence parameter data, the average zero-sequence transmission gain from each grid-connected device to each suspected fault feeder in the target distribution network is calculated in the candidate detection subband. Based on all the average zero-sequence transmission gains, the corresponding grid-connected device is selected as the active detection source using the minimum gain maximum criterion. The active detection source is controlled to inject a zero-sequence detection signal into the target distribution network. The active detection indication quantity corresponding to the feeder is calculated based at least on the zero-sequence current generated by each feeder in the target distribution network during the injection. The target fault feeder is determined based on all the active detection indication quantities and the candidate fault feeders.

[0007] Preferably, the step of performing multi-feature fusion calculation on the zero-sequence electrical data to obtain the passive discrimination score for each feeder in the target distribution network, and determining the candidate fault feeder, the set of suspected fault feeders, and the set of healthy feeders based on threshold comparison of all the passive discrimination scores, includes: The power frequency phasor characteristics, harmonic anomaly characteristics, and transient attenuation characteristics of each feeder in the target distribution network are extracted from the obtained bus zero-sequence voltage and each feeder zero-sequence current. For each feeder, the power frequency phasor characteristics, harmonic anomaly characteristics, and transient attenuation characteristics are respectively normalized and weighted to calculate the passive discrimination score of the corresponding feeder. Determine the maximum and second-largest values ​​among all the passive discrimination scores, and obtain the passive separation degree by calculating the difference between the maximum and the second-largest values; If the maximum value is not less than the first preset action threshold and the passive separation degree is not less than the preset separation threshold, then the feeder corresponding to the maximum value is determined as a candidate fault feeder. If the maximum value is less than the first preset action threshold or the passive separation degree is less than the preset separation threshold, then each feeder whose passive discrimination score is not less than a preset proportion of the maximum value is determined as a suspected fault feeder. A set of suspected faulty feeders is formed based on each of the suspected faulty feeders, and a set of healthy feeders is formed based on each remaining feeder after excluding the set of suspected faulty feeders.

[0008] Preferably, the step of constructing an interference frequency band set based on the grid-connected equipment operating status data, and then selecting, within the available frequency bands excluding the interference frequency band set, a continuous frequency band that maximizes the difference in residual energy between the suspected faulty feeder set and the healthy feeder set as a candidate detection sub-band, includes: Extract the harmonic interference frequency generated by each grid-connected device from the grid-connected device's operating status data, and construct an interference frequency band set based on each of the harmonic interference frequencies; The available frequency bands are obtained by removing the set of interference frequency bands from the frequency range of the zero-sequence electrical signals of the target distribution network. With the goal of maximizing the residual energy difference between the suspected faulty feeder set and the healthy feeder set, continuous frequency bands within the available frequency band are selected as candidate detection subbands.

[0009] Preferably, the residual energy difference between the set of suspected faulty feeders and the set of healthy feeders is obtained by summing the absolute values ​​of the residual energy differences between each pair of suspected faulty feeders and each healthy feeder, wherein the residual energy is the integral energy of the square of the modulus of the zero-sequence current spectrum increment of the feeder after the fault relative to before the fault within the available frequency band.

[0010] Preferably, the step of calculating the average zero-sequence transmission gain from each grid-connected device to each suspected fault feeder in the target distribution network within the candidate probe subband based on the network topology data and the zero-sequence parameter data, and selecting the corresponding grid-connected device as the active probe source based on the minimum gain maximum criterion according to all the average zero-sequence transmission gains, includes: Based on the network topology data and the zero-sequence parameter data, calculate the zero-sequence transfer function of each grid-connected device in the target distribution network to each suspected fault feeder at each frequency point in the candidate detection sub-band; The zero-order transfer function is integrated within the candidate probe subband, and the integral result is divided by the bandwidth of the candidate probe subband to obtain the average zero-order transfer gain. Determine the minimum average zero-sequence transmission gain of each of the grid-connected devices to each of the suspected fault feeders within the candidate probe subband; The grid-connected device corresponding to the maximum value among all the minimum values ​​is taken as the active detection source.

[0011] Preferably, the step of calculating the zero-sequence transfer function of each grid-connected device in the target distribution network to each suspected fault feeder at each frequency point in the candidate detection subband based on the network topology data and the zero-sequence parameter data includes: Based on the network topology data and the zero-sequence parameter data, a zero-sequence network model of the target distribution network is established; The ratio of the predicted zero-sequence response amplitude calculated according to the zero-sequence network model to the amplitude of the injected zero-sequence detection signal is used to obtain the zero-sequence transfer function of each grid-connected device in the target distribution network to each suspected fault feeder at each frequency point in the candidate detection subband.

[0012] Preferably, the step of controlling the active detection source to inject a zero-sequence detection signal into the target distribution network, calculating the active detection indication corresponding to each feeder based at least on the zero-sequence current generated by each feeder in the target distribution network during the injection, and determining the target fault feeder based on all the active detection indications and the candidate fault feeders, includes: The active detection source is controlled to inject a zero-sequence detection signal with a center frequency located within the candidate detection sub-band into the target power distribution network. The band-limited incremental response of each feeder in the target distribution network is calculated based on the zero-sequence current generated during the injection period. The active detection indication of the feeder is calculated based on the correlation coefficient between the band-limited incremental response and the zero-sequence detection signal and the amplitude increment of the band-limited response. Determine the maximum and second largest values ​​among all the active detection indications, and obtain the active separation degree by calculating the difference between the maximum and the second largest values; The feed line corresponding to the maximum value is taken as the active candidate feed line. If the active candidate feed line is the same as the candidate fault feed line and the active separation degree is not less than the second preset action threshold, then the active candidate feed line is determined as the target fault feed line.

[0013] Preferably, the step of calculating the band-limited incremental response of each feeder based on the zero-sequence current generated during injection in the target distribution network, and calculating the active detection indication of the feeder based on the correlation coefficient between the band-limited incremental response and the zero-sequence detection signal, and the amplitude increment of the band-limited response, includes: For each feeder in the target distribution network, the zero-sequence current generated during the injection and the baseline zero-sequence current obtained before the injection are respectively bandpass filtered with the candidate probe subband as the passband, and the obtained band-limited zero-sequence current signal is differentially processed to obtain the corresponding band-limited incremental response of the feeder. Normalized cross-correlation calculations are performed on the band-limited incremental response and the zero-sequence probe signal of each feeder to obtain the correlation coefficient of the corresponding feeder. For each feeder, the zero-sequence current generated during injection and the baseline zero-sequence current obtained before injection are respectively bandpass filtered with the candidate probe subband as the passband, and the obtained band-limited zero-sequence current effective value is differentially calculated to obtain the band-limited response amplitude increment of the corresponding feeder. The correlation coefficient and the band-limited response amplitude increment of each feeder are respectively normalized and weighted to obtain the active detection indication of the corresponding feeder.

[0014] Preferably, the zero-sequence electrical data includes the bus zero-sequence voltage and the zero-sequence current of each feeder; the grid-connected equipment operating status data includes the grid-connected status of the equipment, switching frequency and modulation frequency; the network topology data includes the bus connection relationship, feeder switch status, equipment access node and network branch connection relationship; and the zero-sequence parameter data includes the zero-sequence resistance, zero-sequence inductance, and zero-sequence capacitance of each feeder, as well as the neutral point grounding resistance and transformer zero-sequence parameters.

[0015] Secondly, embodiments of the present invention provide a high-resistance grounding fault location system for a distribution network with a neutral point grounded by a small resistor, comprising: The data acquisition module is used to acquire zero-sequence electrical data, grid-connected equipment operating status data, network topology data, and zero-sequence parameter data of the target distribution network. The passive discrimination module is used to perform multi-feature fusion calculation on the zero-sequence electrical data to obtain the passive discrimination score of each feeder in the target distribution network, and to determine the candidate fault feeder, the set of suspected fault feeders and the set of sound feeders based on threshold comparison of all the passive discrimination scores. The frequency band selection module is used to construct an interference frequency band set based on the grid-connected equipment operating status data, and within the available frequency bands excluding the interference frequency band set, to select the continuous frequency bands that make the difference in residual energy between the suspected faulty feeder set and the healthy feeder set the largest as candidate detection sub-bands. The detection source selection module is used to calculate the average zero-sequence transmission gain of each grid-connected device to each suspected fault feeder in the candidate detection sub-band based on the network topology data and the zero-sequence parameter data, and select the corresponding grid-connected device as the active detection source according to the minimum gain maximum criterion based on all the average zero-sequence transmission gains. An active detection and closed-loop decision module is used to control the active detection source to inject zero-sequence detection signals into the target distribution network, calculate the active detection indication quantity corresponding to each feeder based at least on the zero-sequence current generated by each feeder in the target distribution network during the injection, and determine the target fault feeder based on all the active detection indication quantities and the candidate fault feeders.

[0016] Compared with the prior art, the method and system for selecting high-resistance grounding faults in a distribution network with a neutral point grounded by a small resistor, as described in this invention, have the following advantages at least one point: First, by performing multi-feature fusion calculation on zero-sequence electrical data to obtain passive discrimination scores, it is possible to quickly complete the judgment of candidate fault feeders and narrow down the set of suspected fault feeders. In the early stage of a fault, the first round of screening can be completed without starting active detection, effectively reducing the frequency of active detection and improving the efficiency of fault selection. Second, based on the operating status data of grid-connected equipment, an interference frequency band set is constructed. Within the available frequency bands after removing interference, the continuous frequency bands with the largest residual energy difference are selected as candidate detection sub-bands. This can effectively avoid the interference of grid-connected equipment switching harmonics on the detection process, and maximize the feature differentiation between suspected faulty feeders and healthy feeders, thereby improving the signal-to-noise ratio and identification accuracy of active detection. Third, calculate the average zero-sequence transmission gain from each network-connected device to the suspected fault feeder based on network topology data and zero-sequence parameter data. Use the minimum gain maximum criterion to select active detection sources to ensure that the excitation effect on all suspected fault feeders is balanced and sufficient, and avoid excitation failure, fault omission or misjudgment due to improper selection of detection sources. Fourth, by injecting zero-sequence detection signals through active detection sources and calculating the active detection indication of each feeder, the active detection results are combined with the candidate fault feeders obtained by passive discrimination for collaborative decision-making, thus constructing a closed-loop feeder selection mechanism of passive initial selection and active verification, which improves the robustness and reliability of feeder selection results in high-resistance grounding fault scenarios. Attached Figure Description

[0017] Figure 1 This is a topology diagram of a distribution network with a neutral point grounded through a small resistor in an embodiment of the present invention; Figure 2 This is a flowchart illustrating a method for selecting a high-resistance grounding fault in a distribution network with a neutral point grounded by a small resistor, according to an embodiment of the present invention. Figure 3 This is a schematic diagram of the process for determining candidate faulty feeders, a set of suspected faulty feeders, and a set of healthy feeders according to an embodiment of the present invention; Figure 4 This is a schematic diagram of the process for determining candidate probe subbands according to an embodiment of the present invention; Figure 5 This is a schematic diagram of the process for determining an active detection source according to an embodiment of the present invention; Figure 6 This is a schematic diagram of the process for determining the target fault feeder according to an embodiment of the present invention; Figure 7 This is a schematic diagram of a high-resistance grounding fault selection system for a distribution network with a neutral point grounded by a small resistor, according to an embodiment of the present invention. Figure label: Neutral point grounding resistance; Feeders; SVG, static var generator; In the feeder The fault point on; In the feeder The fault points are: 01. Data acquisition module; 02. Passive discrimination module; 03. Frequency band selection module; 04. Detection source selection module; 05. Active detection and closed-loop decision module. Detailed Implementation

[0018] The specific embodiments of the present invention will be described in further detail below with reference to the accompanying drawings and examples. The following examples are for illustrative purposes only and are not intended to limit the scope of the invention.

[0019] In the description of this invention, it should be understood that the terms "first" and "second," etc., are used to distinguish different objects, rather than to describe a specific order.

[0020] In the description of this invention, it should be noted that, unless otherwise defined, all technical and scientific terms used in this invention have the same meaning as commonly understood by those skilled in the art. The terminology used in this specification is for the purpose of describing specific embodiments only and is not intended to limit the invention. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0021] A distribution network with a neutral point grounded through a small resistor is a medium-voltage distribution network in which the neutral point is connected to the ground through a low-value resistor. For example... Figure 1 As shown, this is a topology diagram of a distribution network with the neutral point grounded through a small resistor in an embodiment of the present invention. (Refer to...) Figure 1 In this embodiment of the invention, the neutral point grounded distribution network uses the busbar as the core of the power distribution, and is equipped with a neutral point grounding resistor. Multiple feeders ( This constitutes the basic power distribution architecture, with multiple grid-connected power electronic devices configured on the feeders; these power electronic devices include feeders. Wind power and SVG (Static Var Generator), feeders Photovoltaics and energy storage, feeders For charging piles on the road, devices lacking direct zero-sequence injection capability can achieve equivalent zero-sequence detection signal injection through a matching grounding transformer, auxiliary injection branch, or zero-sequence control interface. Among these, Indicated in the feeder The fault point on Indicated in the feeder The fault point on it.

[0022] like Figure 2 The diagram shown is a flowchart illustrating a method for selecting a high-resistance grounding fault in a distribution network with a neutral point grounded through a small resistor, according to an embodiment of the present invention. (Refer to...) Figure 2This invention provides a method for locating high-resistance grounding faults in a distribution network with a neutral point grounded through a small resistor, comprising the following steps: S1. Obtain zero-sequence electrical data, grid-connected equipment operating status data, network topology data, and zero-sequence parameter data of the target distribution network; Target distribution network Figure 1 The diagram shows a distribution network with a neutral point grounded through a small resistor. Zero-sequence electrical data includes bus zero-sequence voltage and zero-sequence current for each feeder. Grid-connected equipment operating status data includes the grid connection status, switching frequency, and modulation frequency of the equipment. Network topology data includes bus connection relationships, feeder switch status, equipment access nodes, and network branch connection relationships. Zero-sequence parameter data includes the zero-sequence resistance, zero-sequence inductance, and zero-sequence capacitance of each feeder, as well as the neutral point grounding resistance and transformer zero-sequence parameters.

[0023] Specifically, voltage sampling units and current sampling units installed at the busbar and each feeder are used to collect the three-phase voltage of the busbar. , , and the Three-phase current of the feeder , , .

[0024] Bus zero-sequence voltage and feeder zero-sequence current are defined as follows: in, Represents the zero-sequence voltage of the bus. Indicates the first Zero-sequence current of the feeder.

[0025] The sampling frequency should be no less than 10kHz; the lengths of the observation window before and after the fault can be 1 to 5 power frequency cycles and 1 to 10 power frequency cycles, respectively, to balance the stability of power frequency estimation and the requirements for transient feature extraction.

[0026] Simultaneously, operational status data of grid-connected power electronic equipment is collected, including but not limited to equipment switching frequency, modulation frequency, control mode, output, equipment availability, and equipment access nodes. Network topology data can be provided by station control systems, distribution network automation systems, or protection communication networks, while zero-sequence parameter data can be obtained from equipment nameplate parameters, line ledger parameters, or online identification results.

[0027] S2. Perform multi-feature fusion calculation on zero-sequence electrical data to obtain the passive discrimination score of each feeder in the target distribution network, and determine the candidate fault feeder, suspected fault feeder set and sound feeder set based on threshold comparison of all passive discrimination scores. like Figure 3 As shown, this is a flowchart illustrating step S2. (Refer to...) Figure 3 Step S2 includes: S201. Extract the power frequency phasor characteristics, harmonic anomaly characteristics, and transient attenuation characteristics of each feeder in the target distribution network from the obtained bus zero-sequence voltage and each feeder zero-sequence current. The construction process of power frequency phasor characteristics, harmonic anomaly characteristics, and transient attenuation characteristics is explained in detail below: 1) Power frequency phasor characteristics: The power frequency phasor of the zero-sequence voltage of the bus is obtained through discrete Fourier transform, sliding Fourier algorithm, or phase-locked loop estimation method. and the Zero-sequence current power frequency phasor of the feeder .

[0028] Specifically, power frequency phasor characteristics It can be constructed as follows: in, and Indicates non-negative weights. This represents the maximum value of the power frequency phasor amplitude of the zero-sequence current of all feeders. This indicates the smallest positive number whose denominator is zero. Indicates the phase angle sign. This indicates that the reference phase difference is preset according to the grounding method and polarity convention of the distribution network. For implementation methods that only use amplitude information, only the first term in the above formula can be used.

[0029] It should be noted that weight and These components are used to characterize the power frequency amplitude component and phase relationship component in relation to the power frequency phasor features, respectively. The relative contributions. Since both the power frequency amplitude component and the phase relationship component are used to characterize the degree of fault suspicion of the feeder at the power frequency level, and both play a positive role in fault discrimination, the weights are taken as non-negative values. At the same time, in order to keep the fusion result within the same dimension range of the normalized components and to facilitate the subsequent unified setting of thresholds, the two are constrained to a convex combination with a sum of 1. Specifically, and The range of values ​​for are respectively , In this embodiment The value is 0.6. The value is set to 0.4. When the system's power frequency amplitude characteristics are more stable, the value can be appropriately increased. When the phase relationship is more sensitive to fault detection, it can be appropriately increased. .

[0030] 2) Harmonic anomaly characteristics: set up Indicates the first observation window before the fault. The spectrum of zero-sequence current in a feeder. The spectrum corresponding to the observation window after the fault is represented, and the harmonic anomaly increment spectrum is defined. for: In the preset harmonic analysis frequency band Internal harmonic anomaly characteristics It can be constructed as follows: in, This represents the maximum value of the energy integral of the harmonic anomaly increment spectrum within the preset harmonic analysis frequency band for all feeders.

[0031] 3) Transient decay characteristics: Transient band filtering is performed on the zero-sequence current after the fault, and the filtering result is denoted as... Transient decay characteristics It can be constructed as follows: in, Indicates the time when the fault occurred. This represents the time window for integrating transient features. This represents the attenuation weighting coefficient, which is determined based on the fault transient attenuation characteristics and the system zero-sequence parameters. Indicates the first Zero-sequence current after transient band filtering following a feeder fault.

[0032] S202. Normalize and weightedly fuse the power frequency phasor characteristics, harmonic anomaly characteristics and transient attenuation characteristics of each feeder to obtain the passive discrimination score of the corresponding feeder. After normalizing the power frequency phasor characteristics, harmonic anomaly characteristics, and transient attenuation characteristics of each feeder, the passive discrimination score of the corresponding feeder is obtained by weighted summation. The weights corresponding to the normalized power frequency phasor characteristics, harmonic anomaly characteristics, and transient attenuation characteristics are as follows: , , .

[0033] It should be noted that the power frequency phasor characteristics reflect the steady-state power frequency offset, the harmonic anomaly characteristics reflect the frequency domain abnormal energy increment, and the transient attenuation characteristics reflect the transient response differences after the fault. All three are positive discrimination information for the faulty feeder; therefore, their corresponding weights are non-negative. Furthermore, to ensure that the fusion result remains within the unified dimension range of the normalized characteristics and to facilitate consistent threshold setting under different operating conditions, the weights of the three are constrained to a convex combination with a sum of 1. Specifically, , , The range of values ​​for are respectively , , In this embodiment The value is 0.4. The value is 0.3. The value is set to 0.3. When the steady-state quantity at the power frequency is more reliable, it can be appropriately increased. When the system harmonic anomalies are more pronounced, the voltage can be appropriately increased. When the transient decay difference is more significant, the value can be appropriately increased. .

[0034] S203. Determine the maximum and second largest values ​​among all passive discrimination scores, and obtain the passive separation degree by calculating the difference between the maximum and second largest values; Specifically, the passive separation degree is calculated using the following formula: in, Indicates passive separation degree. This represents the maximum value among all passive discrimination scores. This represents the second largest value among all passive discrimination scores.

[0035] S204. If the maximum value is not less than the first preset action threshold and the passive separation degree is not less than the preset separation threshold, then the feeder corresponding to the maximum value is determined as the candidate fault feeder. Specifically, the current disturbance level is determined based on the new energy penetration rate, the proportion of grid-type equipment, and the system equivalent zero-sequence impedance change coefficient, and the first preset action threshold and the preset separation threshold are determined according to the current disturbance level.

[0036] The penetration rate of new energy sources can be defined as: in, Indicates the penetration rate of new energy sources. It represents a collection of new energy power sources. Indicates the first The rated active power or current active power output of a new energy power source. This indicates the total system load power or the reference load power.

[0037] The proportion of network-type devices can be defined as: in, This indicates the proportion of network-type devices. This represents a collection of grid-type power electronic devices. Indicates the first The rated capacity or currently available capacity of each network-type device. This represents the set of power electronic devices currently connected to the grid. Indicates the first The rated capacity or currently available capacity of each grid-connected power electronic device.

[0038] The system's equivalent zero-sequence impedance variation coefficient can be defined as: in, This represents the coefficient of change of the system's equivalent zero-sequence impedance. This represents the system's equivalent zero-sequence impedance under the current operating conditions. It represents the system's equivalent zero-sequence impedance under the reference operating condition.

[0039] The state vector is denoted as: A disturbance level library is established by clustering historical fault data, real-time operation data, and simulation data.

[0040] Specifically, the K-means clustering algorithm is used to divide the samples into... There are several perturbation level clusters, each corresponding to a cluster center. and a set of threshold parameters For the current operating conditions, the current disturbance level is selected according to the principle of minimum Euclidean distance: Then let: in, This indicates the first preset action threshold. This indicates the preset separation threshold. Indicates the first The passive action threshold corresponding to each disturbance level Indicates the first The separation threshold corresponding to each disturbance level.

[0041] Furthermore, if the maximum value is not less than the first preset action threshold and the passive separation degree is not less than the preset separation threshold, then the feeder corresponding to the maximum value is determined as a candidate fault feeder.

[0042] S205. If the maximum value is less than the first preset action threshold or the passive separation degree is less than the preset separation threshold, then each feeder with a preset ratio value that is not less than the maximum value is identified as a suspected fault feeder. Specifically, the preset ratio of the maximum value is... , ( (A preset coefficient between 0 and 1) In this embodiment The value range is 0.75 to 0.95.

[0043] Furthermore, the score for each passive discrimination is not less than The feeder was identified as a suspected faulty feeder. This method allows the few feeders most likely to fail to be reserved for the active detection phase.

[0044] S206. Based on each suspected faulty feeder, a set of suspected faulty feeders is formed, and based on each remaining feeder after excluding the set of suspected faulty feeders, a set of healthy feeders is formed.

[0045] Each suspected faulty feeder is combined to obtain a set of suspected faulty feeders, and each remaining feeder after excluding the set of suspected faulty feeders is combined to obtain a set of healthy feeders.

[0046] S3. Construct an interference frequency band set based on the grid-connected equipment operation status data, and within the available frequency bands excluding the interference frequency band set, select the continuous frequency bands that maximize the difference in residual energy between the suspected faulty feeder set and the healthy feeder set as candidate detection sub-bands. like Figure 4 As shown, this is a flowchart illustrating step S3. (Refer to...) Figure 4 Step S3 includes: S301. Extract the harmonic interference frequency generated by each grid-connected device from the grid-connected device operation status data, and construct an interference frequency band set based on each harmonic interference frequency. Based on the operating status data of grid-connected equipment, the switching frequency and modulation frequency of each device are extracted, and the first step of each device is calculated sequentially. The center frequency of the first-order switching frequency sideband is calculated using the following formula: in, Indicates the first The first device First-order switching frequency, the first The center frequency of the sideband corresponding to the first modulation frequency. , The highest analysis order indicating the switching frequency. , Indicates the highest analysis order of the modulation frequency. Indicates the first The switching frequency of each device Indicates the first The modulation frequency of each device.

[0047] Based on the center frequency of each sideband, according to the preset guard band width A harmonic interference avoidance zone is constructed by taking the union of the avoidance zones corresponding to all grid-connected devices, ultimately forming a set of interference frequency bands used to avoid harmonic interference. The preset protection frequency band width is determined based on the harmonic diffusion range of the switching sidebands of the grid-connected devices and the system's anti-interference requirements.

[0048] S302. Eliminate the set of interference frequency bands from the frequency range of the zero-sequence electrical signal of the target distribution network to obtain the usable frequency band; Pre-set the total analysis frequency band of the target distribution network zero-sequence electrical signals ( This indicates the lower limit of the total analysis bandwidth. This indicates the upper limit of the total analysis frequency band (the total analysis frequency band is set according to the effective analysis range of the zero-sequence signal in the distribution network, the sampling frequency, and the frequency band distribution of fault characteristics). Within this total analysis frequency band, the interference frequency band set is removed, and the remaining continuous frequency band without harmonic interference is the usable frequency band. Its mathematical expression is: in, This represents the set of interference frequency bands, ensuring that there are no harmonic disturbances caused by the switching sidebands of grid-connected equipment within the available frequency band.

[0049] S303. With the goal of maximizing the residual energy difference between the set of suspected faulty feeders and the set of healthy feeders, a continuous frequency band is selected as a candidate detection subband within the available frequency band.

[0050] The difference in residual energy between the set of suspected faulty feeders and the set of healthy feeders is obtained by summing the absolute values ​​of the residual energy differences between each pair of suspected faulty feeders and each healthy feeder. Here, residual energy is the integral energy of the squared magnitude of the zero-sequence current spectrum increment of the feeder after the fault relative to before the fault, over the available frequency band.

[0051] Specifically, define the first feeder in frequency band Residual energy inside for: in, Indicates the first The increase in the zero-sequence current spectrum after a feeder fault relative to before the fault.

[0052] The reason for using residual energy as a frequency band evaluation metric is that, in high-resistance grounding fault scenarios, faulty feeders and healthy feeders will exhibit different degrees of spectral energy increments in certain frequency bands, and this difference is the basis for subsequent active detection to distinguish them.

[0053] Considering that the switching frequency and sidebands of grid-connected power electronic equipment will introduce strong background interference in the corresponding frequency band, we first construct an interference frequency band set, and then search for candidate probe subbands in the available frequency band.

[0054] For any frequency band If the difference in residual energy between feeders in the suspected faulty feeder set and feeders in the healthy feeder set is greater in a given frequency band, then that frequency band is more conducive to distinguishing the two types of feeders. Based on this idea, a frequency band discrimination objective function is constructed: in, Indicates frequency band Overall ability to distinguish between suspected faulty feeders and healthy feeders This represents a set of feeders suspected of having faults. Indicates the suspected faulty feeder number. This indicates a complete feeder number.

[0055] This function characterizes the frequency band difference between suspected faulty feeders and healthy feeders. The total difference in residual energy within. When A larger value indicates that the overall difference in frequency domain residual energy between the suspected faulty feeder and the healthy feeder is more pronounced within that frequency band; conversely, when... When the value is relatively small, it indicates that the frequency band has a weak ability to separate the two types of feeders, which is not conducive to subsequent active detection.

[0056] Therefore, with To maximize the optimization objective, all continuous candidate frequency bands that meet the preset bandwidth constraints are traversed within the available frequency band. The discrimination value of each candidate frequency band is calculated, and the continuous frequency band that maximizes the objective function is selected as the final candidate detection sub-band.

[0057] Specifically, candidate detector subbands are selected using the following formula: in, This indicates a candidate probe subband.

[0058] In practical operation, the total analysis bandwidth will be... Discretization is performed according to a preset frequency step size, and a preset candidate bandwidth set is also performed. Iterate through all candidate center frequencies and candidate bandwidth combinations to construct continuous candidate frequency bands. Only the available frequency bands are reserved. For each candidate frequency band, calculate the frequency band discrimination corresponding to each valid candidate frequency band. The final selection made The frequency band that achieves the maximum value is used as a candidate detector subband. .

[0059] S4. Based on network topology data and zero-sequence parameter data, calculate the average zero-sequence transmission gain from each grid-connected device to each suspected fault feeder in the candidate detection sub-band, and select the corresponding grid-connected device as the active detection source according to the minimum gain maximum criterion based on all average zero-sequence transmission gains. like Figure 5 As shown, this is a flowchart illustrating step S4. (Refer to...) Figure 5 Step S4 includes: S401. Based on network topology data and zero-sequence parameter data, calculate the zero-sequence transfer function of each grid-connected device in the target distribution network to each suspected fault feeder at each frequency point in the candidate detection sub-band. Specifically, step S401 includes: 1) Based on network topology data and zero-sequence parameter data, establish a zero-sequence network model of the target distribution network; Based on network topology data such as bus connection relationships, feeder switch status, and equipment access nodes, and combined with zero-sequence parameter data such as line zero-sequence resistance, zero-sequence inductance, zero-sequence capacitance, neutral point grounding resistance, and transformer zero-sequence parameters, a zero-sequence equivalent network model of the target distribution network is constructed by building a node admittance matrix or a zero-sequence impedance matrix, and based on Kirchhoff's laws and the principle of circuit equivalent transformation.

[0060] 2) The zero-sequence transfer function of each grid-connected device to each suspected fault feeder in the target distribution network at each frequency point in the candidate detection subband is obtained by using the ratio of the predicted zero-sequence response amplitude calculated according to the zero-sequence network model to the injected zero-sequence detection signal amplitude.

[0061] Specifically, for the first When a device injects a unit amplitude zero-sequence detection signal into the target distribution network, the first... feeder at frequency The ratio of the predicted zero-sequence response amplitude to the injected zero-sequence probe signal amplitude is defined as the zero-sequence transfer function. : in, Indicates the first When the zero-sequence detection signal is injected into the device, the first... A suspected faulty feeder at frequency The predicted zero-sequence current amplitude at that location. Indicates the first The device is at frequency The amplitude of the zero-sequence probe signal injected at the location.

[0062] S402. Integrate the zero-order transfer function within the candidate probe subband, and divide the integral result by the bandwidth of the candidate probe subband to obtain the average zero-order transfer gain. Specifically, no. The device to the first A suspected faulty feeder in the candidate detection subband Average zero-sequence transmission gain within Defined as: in, This indicates the bandwidth of the candidate probe subband.

[0063] S403. Determine the minimum average zero-sequence transmission gain in the candidate probe subband for each grid-connected device to each suspected fault feeder. S404. Use the grid-connected device corresponding to the maximum value among all minimum values ​​as the active detection source.

[0064] It should be noted that if only the criterion of "maximizing the gain of a certain feeder" is adopted, the following problem may easily occur: the selected equipment may provide strong excitation for a suspected faulty feeder, but weak excitation for another suspected faulty feeder. Furthermore, since the actual faulty feeder has not yet been determined, this selection will reduce the overall robustness of the active detection phase. Therefore, steps S403 to S404 adopt the minimum gain maximization criterion.

[0065] For each candidate probe source Step S403 examines the minimum of the average zero-sequence transmission gain over all suspected faulty feeders: in, Indicates when the device When selected as an active detection source, the ability to ensure the excitation of the feeder that is the most difficult to excite among all suspected faulty feeders.

[0066] like A larger value indicates that even under the most unfavorable conditions, the device can still provide an average excitation intensity of no less than that value to any one of the suspected faulty feeders; if If the value is small, it indicates that the device has insufficient excitation capability for at least one suspected faulty feeder, which is not conducive to fault diagnosis.

[0067] Therefore, step S404 will actively detect the source. Defined as the device that maximizes the aforementioned minimum gain, i.e.: When the actual faulty feeder is unknown, priority is given to selecting a detection source that is not excessively weak for all suspected faulty feeders, thereby improving the detectability and stability of active detection at the most unfavorable fault location.

[0068] In practice, node admittance matrices or zero-order equivalent networks can be established based on the current network topology and zero-order parameters, and for each candidate device... and every suspected faulty feeder calculate This forms a gain matrix. Then, by taking the minimum value row by row or device by device, and then taking the maximum value among these minimum values, the active detection source can be determined. .

[0069] S5. Control the active detection source to inject zero-sequence detection signals into the target distribution network, calculate the active detection indication of the corresponding feeder based on the zero-sequence current generated by each feeder in the target distribution network during the injection, and determine the target fault feeder based on all active detection indications and candidate fault feeders.

[0070] like Figure 6 As shown, this is a flowchart illustrating step S5. (Refer to...) Figure 6 Step S5 includes: S501, Control the active detection source to inject a zero-sequence detection signal with a center frequency located within the candidate detection subband into the target distribution network; Control of active detection sources The injection center frequency is located in the candidate probe subband Short-time zero-sequence detection signal within Specifically, in this embodiment, the detection signal uses a windowed single-frequency sinusoidal signal, which is expressed as follows: in, Indicates the amplitude of the detected signal. Represents the window function. This indicates the center frequency of the candidate probe subband. Indicates the initial phase. This indicates the duration of the zero-sequence probe signal injection.

[0071] S502. Calculate the band-limited incremental response of each feeder based on the zero-sequence current generated during the injection period in the target distribution network, and calculate the active detection indication of the corresponding feeder based on the correlation coefficient between the band-limited incremental response and the zero-sequence detection signal and the amplitude increment of the band-limited response. Specifically, step S502 includes: 1) For each feeder in the target distribution network, the zero-sequence current generated during the injection period and the baseline zero-sequence current obtained before the injection are bandpass filtered with the candidate probe subband as the passband, and the obtained band-limited zero-sequence current signals are differentially processed to obtain the band-limited incremental response of the corresponding feeder. set up Indicates the first The zero-sequence current of the feeder during injection. This represents the baseline zero-sequence current before injection, and the two are compared around the candidate probe subband. After bandpass filtering, the first... Band-limited incremental response of the feeder: in, Indicates the first The feeder line in the candidate detector sub-band Band-limited incremental response within, This represents a bandpass filtering operation with the candidate probe subband as the passband.

[0072] 2) Perform normalized cross-correlation calculations on the band-limited incremental response and zero-sequence probe signal of each feeder to obtain the correlation coefficient of the corresponding feeder; Specifically, the following formula is used to calculate the first... correlation coefficient of feeder : 3) For each feeder, the zero-sequence current generated during the injection period and the baseline zero-sequence current obtained before the injection are bandpass filtered with the candidate probe subband as the passband, and the obtained band-limited zero-sequence current effective value is differentially calculated to obtain the band-limited response amplitude increment of the corresponding feeder. Specifically, no. Band-limited response amplitude increment of the feeder It can be calculated using the following formula: 4) Perform normalized weighted fusion calculation on the correlation coefficient and band-limited response amplitude increment of each feeder to obtain the active detection indication of the corresponding feeder.

[0073] Will and Normalized to and Then, the active detection indication can be obtained by weighted summation: in, and Indicates non-negative weights. Indicates the first Active detection indication of the feeder.

[0074] It should be noted that, This mainly reflects the consistency between the feeder response and the probe signal in terms of waveform shape and timing. This primarily reflects the strength of the feeder's energy response to active injection. Both are used to characterize the significance of the feeder's response to the injected signal during the active detection phase; therefore, their weights are... and All values ​​are non-negative. Meanwhile, to maintain the active detection indication... To ensure the consistency of normalized dimensions and facilitate subsequent unified threshold setting, the two are constrained to a convex combination whose sum is 1. Specifically, and The range of values ​​for are respectively , In this embodiment The value is 0.6. The value is set to 0.4. When the correlation between the injected response waveform and the probe signal becomes more stable, the value can be appropriately increased. When the amplitude increment better reflects the strength of the faulty feeder response, it can be appropriately increased. .

[0075] S503. Determine the maximum and second largest values ​​among all active detection indications, and obtain the active separation degree by calculating the difference between the maximum and second largest values; Specifically, the active separation degree is calculated using the following formula: in, Indicates the degree of active separation. This represents the maximum value among all active detection indicators. This represents the second largest value among all active detection indications.

[0076] S504. The feeder corresponding to the maximum value is taken as the active candidate feeder. If the active candidate feeder is the same as the candidate fault feeder and the active separation degree is not less than the second preset action threshold, then the active candidate feeder is determined as the target fault feeder.

[0077] Specifically, the current disturbance level is determined based on the penetration rate of new energy sources, the proportion of grid-type equipment, and the system's equivalent zero-sequence impedance change coefficient, and a second preset action threshold is determined according to the current disturbance level.

[0078] The process of determining the second preset action threshold can be found in step S204. After selecting the current disturbance level according to the minimum Euclidean distance principle for the current working condition, let: in, This indicates the second preset action threshold. Indicates the first The active action threshold corresponding to each disturbance level.

[0079] Furthermore, if the active candidate feeder is the same as the candidate fault feeder and the active separation degree is not less than the second preset action threshold, then the active candidate feeder is determined as the target fault feeder.

[0080] It should be noted that if the active-passive closed-loop consistency check condition is enabled, then when the passive separation degree is greater than zero, the following must also be satisfied: in, and Indicates the current disturbance level The corresponding dynamic threshold is used. The fault feeder determination result is only output when this condition is met, thereby improving decision consistency.

[0081] Specifically, and According to the first The ratio statistic is obtained from the known correct samples at each perturbation level. For example, when for mean When setting its standard deviation, you can set: in, This represents a safety factor greater than zero, used to adjust the width of the consistency interval. The value can be determined based on the coverage requirements of the validation sample set and the trade-off between the false positive and false negative rates; its range is as follows. In this embodiment The value is set to 2.0. A smaller value can be used when the system is operating relatively stably. To improve the strictness of consistency decisions; when the system fluctuates greatly or is noisy, a larger value can be taken. To improve fault tolerance.

[0082] This invention provides a method for selecting fault lines in high-resistance grounding faults in distribution networks with neutral points grounded by low resistance. By performing multi-feature fusion calculations on zero-sequence electrical data to obtain a passive discrimination score, it can quickly complete the identification of candidate faulty feeders and narrow down the set of suspected faulty feeders. The first round of screening can be completed without initiating active detection in the early stages of a fault, effectively reducing the frequency of active detection and improving fault selection efficiency. Based on the operating status data of grid-connected equipment, an interference frequency band set is constructed. Within the available frequency bands after removing interference, the continuous frequency bands with the largest residual energy differences are selected as candidate detection sub-bands. This effectively avoids interference from switching harmonics of grid-connected equipment on the detection process and maximizes the characteristic areas between suspected faulty feeders and healthy feeders. The system employs a gradation method to improve the signal-to-noise ratio and recognition accuracy of active detection. Based on network topology data and zero-sequence parameter data, it calculates the average zero-sequence transmission gain from each grid-connected device to the suspected faulty feeder. The minimum gain maximum criterion is used to select the active detection source, ensuring a balanced and sufficient excitation effect on all suspected faulty feeders. This avoids excitation failure, missed fault detection, or misjudgment due to improper detection source selection. By injecting zero-sequence detection signals into the active detection source and calculating the active detection indication for each feeder, the active detection results are combined with the candidate faulty feeders obtained through passive discrimination for collaborative decision-making. This constructs a closed-loop line selection mechanism of passive initial selection and active verification, improving the robustness and reliability of line selection results in high-resistance grounding fault scenarios.

[0083] like Figure 7 As shown, this is a schematic diagram of a high-resistance grounding fault location system for a distribution network with a neutral point grounded by a small resistor, according to an embodiment of the present invention. (Refer to...) Figure 7 An embodiment of the present invention provides a high-resistance grounding fault location system for a distribution network with a neutral point grounded by a small resistor, comprising: Data acquisition module 01 is used to acquire zero-sequence electrical data, grid-connected equipment operating status data, network topology data and zero-sequence parameter data of the target distribution network; The passive discrimination module 02 is used to perform multi-feature fusion calculation on zero-sequence electrical data to obtain the passive discrimination score of each feeder in the target distribution network, and to determine the candidate fault feeder, the set of suspected fault feeders and the set of sound feeders based on threshold comparison of all passive discrimination scores. The frequency band selection module 03 is used to construct an interference frequency band set based on the grid-connected equipment operation status data, and within the available frequency bands that exclude the interference frequency band set, select the continuous frequency band that makes the difference in residual energy between the suspected faulty feeder set and the healthy feeder set the largest as a candidate detection sub-band. The detection source selection module 04 is used to calculate the average zero-sequence transmission gain from each grid-connected device to each suspected fault feeder in the target distribution network based on network topology data and zero-sequence parameter data, and select the corresponding grid-connected device as the active detection source based on the minimum gain maximum criterion according to all average zero-sequence transmission gains. The active detection and closed-loop decision module 05 is used to control the active detection source to inject zero-sequence detection signals into the target distribution network. It calculates the active detection indication of the corresponding feeder based on the zero-sequence current generated by each feeder in the target distribution network during the injection, and determines the target fault feeder based on all active detection indications and candidate fault feeders.

[0084] It should be noted that each module in the above-mentioned high-resistance grounding fault location system for a neutral-point-grounded distribution network with low resistance can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the corresponding operations of each module. For specific limitations regarding the high-resistance grounding fault location system for a neutral-point-grounded distribution network with low resistance, please refer to the limitations regarding the high-resistance grounding fault location method for a neutral-point-grounded distribution network with low resistance mentioned above; both have the same function and role, and will not be repeated here.

[0085] In summary, the present invention provides a method and system for selecting high-resistance grounding faults in a distribution network with a neutral point grounded by a small resistance. By performing multi-feature fusion calculations on zero-sequence electrical data to obtain a passive discrimination score, it can quickly complete the identification of candidate fault feeders and narrow down the set of suspected fault feeders. The first round of screening can be completed without initiating active detection in the early stages of a fault, effectively reducing the frequency of active detection and improving fault selection efficiency. Based on the operating status data of grid-connected equipment, an interference frequency band set is constructed. Within the available frequency bands after removing interference, the continuous frequency bands with the largest residual energy differences are selected as candidate detection sub-bands. This effectively avoids interference from switching harmonics of grid-connected equipment on the detection process and maximizes the comparison between suspected fault feeders and healthy feeders. The feature discrimination is improved, enhancing the signal-to-noise ratio and recognition accuracy of active detection. The average zero-sequence transmission gain from each grid-connected device to the suspected fault feeder is calculated based on network topology data and zero-sequence parameter data. The minimum gain maximum criterion is used to select the active detection source, ensuring a balanced and sufficient excitation effect on all suspected fault feeders, avoiding excitation failure, missed fault detection, or misjudgment due to improper detection source selection. Zero-sequence detection signals are injected through the active detection source, and the active detection indication of each feeder is calculated. The active detection results are combined with the candidate fault feeders obtained from passive discrimination for collaborative decision-making, constructing a closed-loop line selection mechanism of passive initial selection and active verification, improving the robustness and reliability of line selection results in high-resistance grounding fault scenarios.

[0086] The various embodiments in this specification are described in a progressive manner. For directly identical or similar parts of the embodiments, refer to each other. Each embodiment focuses on its differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions in the method embodiments. It should be noted that the technical features of the above embodiments can be combined arbitrarily. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as the combination of these technical features does not contradict each other, it should be considered within the scope of this specification.

[0087] The above description is only a preferred embodiment of the present invention. It should be noted that for those skilled in the art, several improvements and substitutions can be made without departing from the technical principles of the present invention, and these improvements and substitutions should also be considered within the scope of protection of the present invention.

Claims

1. A method for selecting the fault location of a high-resistance grounding fault in a distribution network with a neutral point grounded through a small resistor, characterized in that, include: Acquire zero-sequence electrical data, grid-connected equipment operating status data, network topology data, and zero-sequence parameter data of the target distribution network; The passive discrimination score of each feeder in the target distribution network is obtained by multi-feature fusion calculation on the zero-sequence electrical data, and the candidate fault feeder, the set of suspected fault feeders and the set of sound feeders are determined by threshold comparison based on all the passive discrimination scores. Based on the operating status data of the grid-connected equipment, an interference frequency band set is constructed, and within the available frequency bands excluding the interference frequency band set, a continuous frequency band that makes the difference in residual energy between the suspected faulty feeder set and the healthy feeder set the largest is selected as a candidate detection sub-band. Based on the network topology data and the zero-sequence parameter data, the average zero-sequence transmission gain from each grid-connected device to each suspected fault feeder in the target distribution network is calculated in the candidate detection subband. Based on all the average zero-sequence transmission gains, the corresponding grid-connected device is selected as the active detection source using the minimum gain maximum criterion. The active detection source is controlled to inject a zero-sequence detection signal into the target distribution network. The active detection indication quantity corresponding to the feeder is calculated based at least on the zero-sequence current generated by each feeder in the target distribution network during the injection. The target fault feeder is determined based on all the active detection indication quantities and the candidate fault feeders.

2. The method for selecting the fault location of a high-resistance grounding fault in a distribution network with a neutral point grounded by a small resistor according to claim 1, characterized in that, The process of performing multi-feature fusion calculation on the zero-sequence electrical data to obtain the passive discrimination score for each feeder in the target distribution network, and determining the candidate fault feeder, the set of suspected fault feeders, and the set of healthy feeders based on threshold comparison of all the passive discrimination scores, includes: The power frequency phasor characteristics, harmonic anomaly characteristics, and transient attenuation characteristics of each feeder in the target distribution network are extracted from the obtained bus zero-sequence voltage and each feeder zero-sequence current. For each feeder, the power frequency phasor characteristics, harmonic anomaly characteristics, and transient attenuation characteristics are respectively normalized and weighted to calculate the passive discrimination score of the corresponding feeder. Determine the maximum and second-largest values ​​among all the passive discrimination scores, and obtain the passive separation degree by calculating the difference between the maximum and the second-largest values; If the maximum value is not less than the first preset action threshold and the passive separation degree is not less than the preset separation threshold, then the feeder corresponding to the maximum value is determined as a candidate fault feeder. If the maximum value is less than the first preset action threshold or the passive separation degree is less than the preset separation threshold, then each feeder whose passive discrimination score is not less than a preset proportion of the maximum value is determined as a suspected fault feeder. A set of suspected faulty feeders is formed based on each of the suspected faulty feeders, and a set of healthy feeders is formed based on each remaining feeder after excluding the set of suspected faulty feeders.

3. The method for selecting the fault location of a high-resistance grounding fault in a distribution network with a neutral point grounded by a small resistor according to claim 1, characterized in that, The process of constructing an interference frequency band set based on the grid-connected equipment operating status data, and then selecting, within the available frequency bands excluding the interference frequency band set, a continuous frequency band that maximizes the difference in residual energy between the suspected faulty feeder set and the healthy feeder set as a candidate detection sub-band, includes: Extract the harmonic interference frequency generated by each grid-connected device from the grid-connected device's operating status data, and construct an interference frequency band set based on each of the harmonic interference frequencies; The available frequency bands are obtained by removing the set of interference frequency bands from the frequency range of the zero-sequence electrical signals of the target distribution network. With the goal of maximizing the residual energy difference between the suspected faulty feeder set and the healthy feeder set, continuous frequency bands within the available frequency band are selected as candidate detection subbands.

4. The method for selecting the fault location of a high-resistance grounding fault in a distribution network with a neutral point grounded by a small resistor according to claim 3, is characterized in that, The residual energy difference between the set of suspected faulty feeders and the set of healthy feeders is obtained by summing the absolute values ​​of the residual energy differences between each pair of suspected faulty feeders and each healthy feeder, wherein the residual energy is the integral energy of the square of the modulus of the zero-sequence current spectrum increment of the feeder after the fault relative to before the fault within the available frequency band.

5. The method for selecting the fault location of a high-resistance grounding fault in a distribution network with a neutral point grounded by a small resistor according to claim 1, characterized in that, The step of calculating the average zero-sequence transmission gain from each grid-connected device to each suspected fault feeder in the target distribution network within the candidate probe subband based on the network topology data and the zero-sequence parameter data, and selecting the corresponding grid-connected device as the active probe source based on the minimum gain maximum criterion according to all the average zero-sequence transmission gains, includes: Based on the network topology data and the zero-sequence parameter data, calculate the zero-sequence transfer function of each grid-connected device in the target distribution network to each suspected fault feeder at each frequency point in the candidate detection sub-band; The zero-order transfer function is integrated within the candidate probe subband, and the integral result is divided by the bandwidth of the candidate probe subband to obtain the average zero-order transfer gain. Determine the minimum average zero-sequence transmission gain of each of the grid-connected devices to each of the suspected fault feeders within the candidate probe subband; The grid-connected device corresponding to the maximum value among all the minimum values ​​is taken as the active detection source.

6. The method for selecting the fault location of a high-resistance grounding fault in a distribution network with a neutral point grounded by a small resistor according to claim 5, is characterized in that, The step of calculating the zero-order transfer function of each grid-connected device in the target distribution network to each suspected fault feeder at each frequency point in the candidate detection subband based on the network topology data and the zero-order parameter data includes: Based on the network topology data and the zero-sequence parameter data, a zero-sequence network model of the target distribution network is established; The ratio of the predicted zero-sequence response amplitude calculated according to the zero-sequence network model to the amplitude of the injected zero-sequence detection signal is used to obtain the zero-sequence transfer function of each grid-connected device in the target distribution network to each suspected fault feeder at each frequency point in the candidate detection subband.

7. The method for selecting the fault location of a high-resistance grounding fault in a distribution network with a neutral point grounded by a small resistor according to claim 1, characterized in that, The step of controlling the active detection source to inject a zero-sequence detection signal into the target distribution network, calculating the active detection indication corresponding to each feeder based at least on the zero-sequence current generated by each feeder in the target distribution network during the injection, and determining the target fault feeder based on all the active detection indications and the candidate fault feeders, includes: The active detection source is controlled to inject a zero-sequence detection signal with a center frequency located within the candidate detection sub-band into the target power distribution network. The band-limited incremental response of each feeder in the target distribution network is calculated based on the zero-sequence current generated during the injection period. The active detection indication of the feeder is calculated based on the correlation coefficient between the band-limited incremental response and the zero-sequence detection signal and the amplitude increment of the band-limited response. Determine the maximum and second largest values ​​among all the active detection indications, and obtain the active separation degree by calculating the difference between the maximum and the second largest values; The feed line corresponding to the maximum value is taken as the active candidate feed line. If the active candidate feed line is the same as the candidate fault feed line and the active separation degree is not less than the second preset action threshold, then the active candidate feed line is determined as the target fault feed line.

8. The method for selecting the fault location of a high-resistance grounding fault in a distribution network with a neutral point grounded by a small resistor according to claim 7, characterized in that, The step of calculating the band-limited incremental response of each feeder in the target distribution network based on the zero-sequence current generated during injection, and calculating the active detection indication of the feeder based on the correlation coefficient between the band-limited incremental response and the zero-sequence detection signal, and the amplitude increment of the band-limited response, includes: For each feeder in the target distribution network, the zero-sequence current generated during the injection and the baseline zero-sequence current obtained before the injection are respectively bandpass filtered with the candidate probe subband as the passband, and the obtained band-limited zero-sequence current signal is differentially processed to obtain the corresponding band-limited incremental response of the feeder. Normalized cross-correlation calculations are performed on the band-limited incremental response and the zero-sequence probe signal of each feeder to obtain the correlation coefficient of the corresponding feeder. For each feeder, the zero-sequence current generated during injection and the baseline zero-sequence current obtained before injection are respectively bandpass filtered with the candidate probe subband as the passband, and the obtained band-limited zero-sequence current effective value is differentially calculated to obtain the band-limited response amplitude increment of the corresponding feeder. The correlation coefficient and the band-limited response amplitude increment of each feeder are respectively normalized and weighted to obtain the active detection indication of the corresponding feeder.

9. The method for selecting the fault location of a high-resistance grounding fault in a distribution network with a neutral point grounded by a small resistor according to claim 1, characterized in that, The zero-sequence electrical data includes the bus zero-sequence voltage and the zero-sequence current of each feeder. The grid-connected equipment operating status data includes the grid-connected status, switching frequency, and modulation frequency of the equipment. The network topology data includes the bus connection relationship, feeder switch status, equipment access node, and network branch connection relationship. The zero-sequence parameter data includes the zero-sequence resistance, zero-sequence inductance, and zero-sequence capacitance of each feeder, as well as the neutral point grounding resistance and transformer zero-sequence parameters.

10. A fault location system for high-resistance grounding faults in a distribution network with a neutral point grounded by a small resistor, characterized in that, include: The data acquisition module is used to acquire zero-sequence electrical data, grid-connected equipment operating status data, network topology data, and zero-sequence parameter data of the target distribution network. The passive discrimination module is used to perform multi-feature fusion calculation on the zero-sequence electrical data to obtain the passive discrimination score of each feeder in the target distribution network, and to determine the candidate fault feeder, the set of suspected fault feeders and the set of sound feeders based on threshold comparison of all the passive discrimination scores. The frequency band selection module is used to construct an interference frequency band set based on the grid-connected equipment operating status data, and within the available frequency bands excluding the interference frequency band set, to select the continuous frequency bands that make the difference in residual energy between the suspected faulty feeder set and the healthy feeder set the largest as candidate detection sub-bands. The detection source selection module is used to calculate the average zero-sequence transmission gain of each grid-connected device to each suspected fault feeder in the candidate detection sub-band based on the network topology data and the zero-sequence parameter data, and select the corresponding grid-connected device as the active detection source according to the minimum gain maximum criterion based on all the average zero-sequence transmission gains. An active detection and closed-loop decision module is used to control the active detection source to inject zero-sequence detection signals into the target distribution network, calculate the active detection indication quantity corresponding to each feeder based at least on the zero-sequence current generated by each feeder in the target distribution network during the injection, and determine the target fault feeder based on all the active detection indication quantities and the candidate fault feeders.

Images