Power distribution network grounding fault positioning method based on bidirectional translation cross-correlation and waveform polarity identification

Abstract

The invention discloses a power distribution network grounding fault positioning method based on bidirectional translation cross-correlation and waveform polarity identification, and relates to the technical field of power system relay protection and power distribution network fault diagnosis. The invention aims to solve the problem that the positioning precision of an existing power distribution network grounding fault positioning method cannot meet the requirement. When the zero-sequence voltage leap occurs in the power distribution network, acquiring three-phase current data of each measurement node in the power distribution network to form an original data table, and reading a topological connection relation of the power distribution network; each element in the original data table is subjected to Carrenerus transform and wavelet denoising processing, and a pure transient zero-mode current sequence is obtained; performing bidirectional translation sliding on the pure transient zero-mode current sequences of the two adjacent measurement nodes to find an optimal time shift point, and obtaining waveform comprehensive characteristics of all measurement node pairs through energy effective values of the sequences corresponding to the two adjacent measurement nodes; and judging whether the link corresponding to each measurement node pair is a fault interval or not based on the waveform comprehensive characteristics and in combination with a preset fault criterion.

Description

Technical Field

[0001] This application belongs to the field of power system relay protection and distribution network fault diagnosis technology. Background Technology

[0002] With the deepening of smart grid construction, the demand for power supply reliability is increasing across society. However, high-resistance grounding faults occur frequently in distribution networks, which can easily lead to serious safety accidents such as equipment burnout and even forest fires, posing a huge threat to power grid safety and social stability. The traditional manual, blind line inspection and troubleshooting method can no longer meet the urgent needs of efficient operation in modern society.

[0003] Currently, the main method for locating grounding faults in distribution networks using monitoring data is to collect operational data from sensors deployed on the lines, and then analyze the data using feature extraction or algorithm models. The main methods currently employed include:

[0004] 1) Processing using single-point detection or traditional fault indicators. Because high-impedance fault signals are very weak and there are a large number of distributed capacitors in the distribution network, local detection is easily affected by severe noise interference, leading to missed detections. Moreover, it is very difficult to extract features from the complex traveling wave head reflection interference. Therefore, this processing mode can usually only determine the approximate branch segment, resulting in low positioning accuracy.

[0005] 2) The method using transient zero-mode current waveform similarity can be relatively effective in locating faults based on the differences between upstream and downstream waveforms. However, existing algorithms do not consider the case where the correlation coefficient is negative (i.e., polarity reversal), and only handle asynchrony issues through unidirectional signal shifting. While this method typically yields good theoretical results under ideal operating conditions, it fails to eliminate complex bidirectional communication synchronization errors in practical applications, leading to location failure. Especially in complex cases where the difference in capacitance to ground between upstream and downstream is small, both methods perform poorly because the current frequency and amplitude differences on both sides of the fault point are minimal, with only the polarities being opposite.

[0006] 3) Models that directly use the original correlation coefficient for logical judgment have the biggest problem of completely ignoring the objective impact of upstream and downstream signal amplitude attenuation and lacking a multi-dimensional correction mechanism for high-impedance grounding. They are easily misled by complex operating conditions, which can lead to misoperation or positioning failure, resulting in the final positioning accuracy not meeting the requirements. Summary of the Invention

[0007] This application aims to address the problem that existing methods for locating ground faults in distribution networks do not meet the required accuracy. It provides a ground fault location method for distribution networks based on bidirectional translational cross-correlation and waveform polarity identification. This method utilizes a wide-area high-precision synchronous measurement terminal to acquire transient information. Through bidirectional translational cross-correlation calculation and waveform polarity identification, it comprehensively considers the polarity of the correlation coefficient and amplitude weight correction to adapt to complex operating conditions such as high-resistance grounding in distribution networks and synchronization errors caused by communication delays, thereby achieving accurate location of ground faults in distribution networks.

[0008] The first aspect of this application provides a method for locating grounding faults in distribution networks based on bidirectional translational cross-correlation and waveform polarity identification, including:

[0009] When a zero-sequence voltage change occurs in the distribution network, the three-phase current data of each measuring node in the distribution network are collected to form an original data table, and the topological connectivity of the distribution network is read. Each element in the original data table includes a terminal number and a three-phase current array. The terminal number is used to indicate the topological location of the measuring node, and the three-phase current array is the three-phase high-frequency current data of the measuring node within the acquisition time window.

[0010] Each element in the original data table is subjected to Karen-Bauer transform and wavelet denoising to obtain a pure transient zero-mode current sequence.

[0011] The topological connectivity of the distribution network is traversed, and the optimal time shift point is found by bidirectional translation and sliding of the pure transient zero-mode current sequence of two adjacent measurement nodes. The waveform comprehensive characteristics of all measurement node pairs are obtained by using the energy effective value of the corresponding sequence of two adjacent measurement nodes.

[0012] Based on the waveform synthesis characteristics and combined with the preset fault criteria, it is determined whether the link corresponding to each measurement node is in a fault range.

[0013] In one possible design, performing Karen-Bauer transform and wavelet denoising on each element of the original data table to obtain a pure transient zero-mode current sequence includes:

[0014] The first in the original data table The three-phase high-frequency current data of each element are input into the Karen-Bauer transform and wavelet denoising module for Karen-Bauer transform and wavelet denoising processing to obtain the first element. The transient zero-mode current corresponding to each element;

[0015] Using the first data table The terminal number of each element and the transient zero-mode current are used to construct a pure transient zero-mode current element;

[0016] Repeat the above steps until all elements in the original data table have obtained the corresponding pure transient zero-mode current element;

[0017] A pure transient zero-mode current sequence is constructed using the pure transient zero-mode current elements corresponding to all elements in the original data table.

[0018] In one possible design, the Karenbauer transform and wavelet denoising module includes:

[0019] Karenbauer Transform Unit: Used to perform Karenbauer transformation on the three-phase high-frequency current data of each element in the original data table, and extract the initial transient zero-mode current sequence through the decoupling transformation matrix;

[0020] Resolution wavelet transform unit: used to decompose the initial transient zero-mode current sequence into discrete approximate components and detail components using preset wavelet basis functions.

[0021] Reconstruction unit: used to reconstruct the signal after setting the detailed components to zero, and to filter out high-frequency harmonics and noise.

[0022] In one possible design, the topological connectivity of the distribution network is traversed, and the optimal time shift point is found by bidirectional translation and sliding of the pure transient zero-mode current sequences of two adjacent measurement nodes. The waveform comprehensive characteristics of all measurement node pairs are obtained by using the effective energy values ​​of the corresponding sequences of two adjacent measurement nodes, including:

[0023] Read the topological connectivity of the power distribution network. The upstream terminal number and downstream terminal number of the link;

[0024] In the pure transient zero-mode current sequence, find the element whose terminal number is equal to the upstream terminal number and the downstream terminal number;

[0025] The two found elements are input into the waveform similarity cross-correlation calculation module for bidirectional translation and sliding calculation to obtain the first element. The waveform characteristics corresponding to each link;

[0026] Repeat the above steps until all links in the topological connectivity of the distribution network have obtained the corresponding waveform comprehensive characteristics.

[0027] In one possible design, the two found elements are input into the waveform similarity cross-correlation calculation module for bidirectional translation and sliding calculation to obtain the first... The waveform characteristics corresponding to each link include:

[0028] The two found elements are used as the upstream and downstream sequences, respectively. A bidirectional translation and sliding calculation is performed on the upstream and downstream sequences to obtain the cross-correlation coefficients under different time shifts.

[0029] The time shift point corresponding to the maximum absolute value of the cross-correlation coefficient is taken as the optimal time shift point. The cross-correlation coefficient corresponding to the optimal time shift point is extracted as the original cross-correlation coefficient, and then the polarity sign of the original cross-correlation coefficient is obtained.

[0030] Calculate the effective energy values ​​of the upstream and downstream sequences respectively;

[0031] The deviation between the effective energy values ​​of the upstream and downstream sequences is used as the amplitude correction coefficient.

[0032] The amplitude correction coefficient is used to correct the amplitude weight of the original cross-correlation coefficient to obtain an absolute comprehensive similarity coefficient that only considers the waveform shape approximation.

[0033] The polarity signs of the absolute comprehensive similarity coefficient and the original cross-correlation coefficient are used to construct waveform comprehensive features.

[0034] In one possible design, the bidirectional translation and sliding calculation of the upstream and downstream sequences to obtain the cross-correlation coefficients at different time shifts includes:

[0035] The upstream and downstream sequences are calculated using the following formula for bidirectional translation and sliding:

[0036] ;

[0037] in, Let be the translation step size, and , The maximum permissible number of time shift points for synchronization error. Indicates translation step size The cross-correlation coefficients under the following conditions;

[0038] and These represent the upstream sequence notation and the downstream sequence notation, respectively. , The number of sampling points for the found element. and It is decomposed into the average value of the upstream and downstream sequences.

[0039] In one possible design, the expression for the polarity sign of the original cross-correlation coefficient is:

[0040] ,

[0041] in, Indicates the polarity sign. Represents a symbolic function. This represents the original cross-correlation coefficients.

[0042] In one possible design, calculating the effective energy values ​​of the upstream and downstream sequences respectively includes:

[0043] The effective energy values ​​of the upstream and downstream sequences are calculated according to the following formula:

[0044] ,

[0045] ,

[0046] in, and They are respectively and The effective value of energy.

[0047] In one possible design, using the deviation between the effective energy values ​​of the upstream and downstream sequences as an amplitude correction coefficient includes:

[0048] The amplitude correction factor is calculated according to the following formula:

[0049] ,

[0050] in, This is the amplitude correction factor. and These represent taking the minimum and maximum values, respectively.

[0051] In one possible design, the step of using the amplitude correction coefficient to perform amplitude weight correction on the original cross-correlation coefficient to obtain an absolute comprehensive similarity coefficient that only considers waveform shape approximation includes:

[0052] The original cross-correlation coefficients are adjusted for magnitude weights according to the following formula:

[0053] ,

[0054] in, This is the absolute comprehensive similarity coefficient that only considers the approximation of waveform shape. This represents the original cross-correlation coefficients.

[0055] In one possible design, determining whether each measurement node's corresponding link is in a faulty interval, based on the waveform synthesis characteristics and a preset fault criterion, includes:

[0056] The preset fault criteria include:

[0057] If any of the following inequalities are satisfied, the link is considered to be abnormal:

[0058] Condition one: , This is a safety threshold;

[0059] Condition two: , Indicates the polarity sign.

[0060] The second aspect of this application provides a distribution network grounding fault location device based on bidirectional translational cross-correlation and waveform polarity identification. The distribution network grounding fault location device based on bidirectional translational cross-correlation and waveform polarity identification includes a processor and a memory. The memory stores at least one instruction, which is loaded and executed by the processor to implement the distribution network grounding fault location method based on bidirectional translational cross-correlation and waveform polarity identification as described above.

[0061] A third aspect of this application provides a computer storage medium storing at least one instruction, which is loaded and executed by a processor to implement the above-described method for locating grounding faults in a power distribution network based on bidirectional translational cross-correlation and waveform polarity identification.

[0062] The beneficial effects of this application are:

[0063] This application provides a method for locating ground faults in distribution networks based on bidirectional translational cross-correlation and waveform polarity identification. The method constructs a multi-resolution wavelet transform denoising module, a waveform similarity cross-correlation calculation module, and a fault interval discrimination module. These modules are used to establish the pure transient zero-mode current characteristics of wide-area nodes in a hierarchical manner and find the differences in polarity, amplitude, and similarity of the waveforms upstream and downstream of the fault point, thereby achieving accurate location of high-resistance ground faults in distribution networks.

[0064] This application utilizes a bidirectional translation calculation and amplitude polarity correction structure to describe the spatiotemporal evolution of transient zero-mode current in a distribution network across different sampling points. On one hand, this allows the algorithm to effectively eliminate the impact of sampling asynchrony caused by communication delays or hardware asynchrony through bidirectional sliding, ensuring accurate extraction of waveform comprehensive similarity under various complex distribution network topologies and capacitance distributions. On the other hand, by introducing polarity recording and amplitude weight correction mechanisms, the model no longer relies on specific capacitance difference environments, enabling a clearer expression of the essential differences in transient directions before and after the fault point, effectively solving the problem of blind spots caused by weak signals under high-resistance grounding conditions. This application enables more precise location of fault intervals, significantly reducing manual inspection time and substantially improving the management efficiency of power grid operation and maintenance. Attached Figure Description

[0065] Figure 1 This is a flowchart of a method for locating grounding faults in distribution networks based on bidirectional translational cross-correlation and waveform polarity identification. Detailed Implementation

[0066] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application. It should be noted that, unless otherwise specified, the embodiments and features in the embodiments of this application can be combined with each other.

[0067] Specific implementation method one: Refer to Figure 1 This embodiment specifically describes the distribution network grounding fault location method based on bidirectional translational cross-correlation and waveform polarity identification, which includes:

[0068] When a zero-sequence voltage change occurs in the distribution network, the three-phase current data of each measuring node in the distribution network are collected to form an original data table, and the topological connectivity of the distribution network is read. Each element in the original data table includes a terminal number and a three-phase current array. The terminal number is used to indicate the topological location of the measuring node, and the three-phase current array is the three-phase high-frequency current data of the measuring node within the acquisition time window.

[0069] Each element in the original data table is subjected to Karen-Bauer transform and wavelet denoising to obtain a pure transient zero-mode current sequence.

[0070] By traversing the topological connectivity of the distribution network, bidirectional translation and sliding calculations are performed on the pure transient zero-mode current sequences of two adjacent measurement nodes to obtain the waveform comprehensive characteristics of all measurement node pairs.

[0071] Based on the waveform synthesis characteristics and combined with the preset fault criteria, it is determined whether the link corresponding to each measurement node is in a fault range.

[0072] In one implementation, performing Karen-Bauer transform and wavelet denoising on each element in the original data table to obtain a pure transient zero-mode current sequence includes:

[0073] The first in the original data table The three-phase high-frequency current data of each element are input into the Karen-Bauer transform and wavelet denoising module for Karen-Bauer transform and wavelet denoising processing to obtain the first element. The transient zero-mode current corresponding to each element;

[0074] Using the first data table The terminal number of each element and the transient zero-mode current are used to construct a pure transient zero-mode current element;

[0075] Repeat the above steps until all elements in the original data table have obtained the corresponding pure transient zero-mode current element;

[0076] A pure transient zero-mode current sequence is constructed using the pure transient zero-mode current elements corresponding to all elements in the original data table.

[0077] In one implementation, the Karenbauer transform and wavelet denoising module includes:

[0078] Karenbauer Transform Unit: Used to perform Karenbauer transformation on the three-phase high-frequency current data of each element in the original data table, and extract the initial transient zero-mode current sequence through the decoupling transformation matrix;

[0079] Resolution wavelet transform unit: used to decompose the initial transient zero-mode current sequence into discrete approximate components and detail components using preset wavelet basis functions.

[0080] Reconstruction unit: used to reconstruct the signal after setting the detailed components to zero, and to filter out high-frequency harmonics and noise.

[0081] In one implementation, the step of traversing the topological connectivity of the distribution network and performing bidirectional translational sliding calculations on the pure transient zero-mode current sequences of two adjacent measurement nodes to obtain the waveform comprehensive characteristics of all measurement node pairs includes:

[0082] Read the topological connectivity of the power distribution network. The upstream terminal number and downstream terminal number of the link;

[0083] In the pure transient zero-mode current sequence, find the element whose terminal number is equal to the upstream terminal number and the downstream terminal number;

[0084] The two found elements are input into the waveform similarity cross-correlation calculation module for bidirectional translation and sliding calculation to obtain the first element. The waveform characteristics corresponding to each link;

[0085] Repeat the above steps until all links in the topological connectivity of the distribution network have obtained the corresponding waveform comprehensive characteristics.

[0086] In one implementation, the two found elements are input into the waveform similarity cross-correlation calculation module for bidirectional translation and sliding calculation to obtain the first element. The waveform characteristics corresponding to each link include:

[0087] The two found elements are used as the upstream and downstream sequences, respectively. A bidirectional translation and sliding calculation is performed on the upstream and downstream sequences to obtain the cross-correlation coefficients under different time shifts.

[0088] The time shift point corresponding to the maximum absolute value of the cross-correlation coefficient is taken as the optimal time shift point. The cross-correlation coefficient corresponding to the optimal time shift point is extracted as the original cross-correlation coefficient, and then the polarity sign of the original cross-correlation coefficient is obtained.

[0089] Calculate the effective energy values ​​of the upstream and downstream sequences respectively;

[0090] The deviation between the effective energy values ​​of the upstream and downstream sequences is used as the amplitude correction coefficient.

[0091] The amplitude correction coefficient is used to correct the amplitude weight of the original cross-correlation coefficient to obtain an absolute comprehensive similarity coefficient that only considers the waveform shape approximation.

[0092] The polarity signs of the absolute comprehensive similarity coefficient and the original cross-correlation coefficient are used to construct waveform comprehensive features.

[0093] In one implementation, the step of performing bidirectional translation and sliding calculations on the upstream and downstream sequences to obtain cross-correlation coefficients at different time shifts includes:

[0094] The upstream and downstream sequences are calculated using the following formula for bidirectional translation and sliding:

[0095] ;

[0096] in, Let be the translation step size, and , The maximum permissible number of time shift points for synchronization error. Indicates translation step size The cross-correlation coefficients under the following conditions;

[0097] and These represent the upstream sequence notation and the downstream sequence notation, respectively. , The number of sampling points for the found element. and It is decomposed into the average value of the upstream and downstream sequences.

[0098] In one implementation, the expression for the polarity sign of the original cross-correlation coefficient is:

[0099] ,

[0100] in, Indicates the polarity sign. Represents a symbolic function. This represents the original cross-correlation coefficients.

[0101] In one implementation, calculating the effective energy values ​​of the upstream and downstream sequences respectively includes:

[0102] The effective energy values ​​of the upstream and downstream sequences are calculated according to the following formula:

[0103] ,

[0104] ,

[0105] in, and They are respectively and The effective value of energy.

[0106] In one embodiment, using the deviation between the effective energy values ​​of the upstream and downstream sequences as an amplitude correction coefficient includes:

[0107] The amplitude correction factor is calculated according to the following formula:

[0108] ,

[0109] in, This is the amplitude correction factor. and These represent taking the minimum and maximum values, respectively.

[0110] In one implementation, the step of using the amplitude correction coefficient to perform amplitude weight correction on the original cross-correlation coefficient to obtain an absolute comprehensive similarity coefficient that only considers waveform shape approximation includes:

[0111] The original cross-correlation coefficients are adjusted for magnitude weights according to the following formula:

[0112] ,

[0113] in, This is the absolute comprehensive similarity coefficient that only considers the approximation of waveform shape. This represents the original cross-correlation coefficients.

[0114] In one implementation, determining whether the link corresponding to each measurement node is in a fault range based on the waveform synthesis characteristics and a preset fault criterion includes:

[0115] The preset fault criteria include:

[0116] If any of the following inequalities are satisfied, the link is considered to be abnormal:

[0117] Condition one: , This is a safety threshold;

[0118] Condition two: , Indicates the polarity sign.

[0119] To further illustrate the implementation scheme of this application, this embodiment provides a method for locating ground faults in distribution networks based on bidirectional translational cross-correlation and waveform polarity identification, including steps 1 to 6. The numbering of each step does not necessarily limit their execution order. This embodiment constructs a multi-resolution wavelet transform denoising module, a waveform similarity cross-correlation calculation module, and a dynamic threshold discrimination module. These modules are used to hierarchically establish the clean transient zero-mode current characteristics of wide-area nodes and find the differences in polarity, amplitude, and similarity of the waveforms upstream and downstream of the fault point, thereby achieving accurate location of high-resistance ground faults in the distribution network. Specifically, it includes:

[0120] S1. Deploy wide-area high-precision synchronous measurement terminals at key nodes in the distribution network substations and lines. When a zero-sequence voltage mutation is detected, collect the three-phase current data from each measurement terminal to form a raw data table, and read the distribution network topology connectivity list from the master station, as follows:

[0121] S101 involves installing high-precision synchronous measurement terminals on distribution network lines and at key nodes. Using the open-delta zero-sequence voltage mutation as the trigger criterion, and synchronously collecting three-phase current data from each node within a specific time window based on BeiDou / GPS high-precision time synchronization, a raw data table is generated. Each element in the raw data table contains two fields: terminal number and three-phase current array. The terminal number is an integer variable representing the topological location of the measurement node; the three-phase current array is a... floating-point arrays, It is a positive integer, recording the high-frequency current data of phases A, B, and C within the time window of the measurement node.

[0122] S102, obtain the number of elements in the original data table.

[0123] S103, establish a clean transient zero-mode current data list (initially an empty list) for subsequent storage of the denoised ground mode components.

[0124] S104: Read the static topology configuration file of the distribution network from the distribution network automation master station system or GIS database, and establish a list of distribution network topology connectivity relationships (initially an empty list).

[0125] S105, parse the static topology configuration file, and sequentially convert all adjacent terminal node pairs into elements of the distribution network topology connectivity list; each element of the distribution network topology connectivity list contains 3 fields, namely:

[0126] Link number: is an integer variable representing a specific segment of a physical link;

[0127] Upstream node: is an integer variable that corresponds to the terminal number on the power supply side of the line;

[0128] Downstream node: This is an integer variable that corresponds to the terminal number on the load side of the line.

[0129] S106, obtain the number of elements in the distribution network topology connectivity list.

[0130] S2, establish the Karen-Bauer transform and wavelet denoising module (TDM), whose input is a three-phase array and output is a clean zero-mode array, as follows:

[0131] S201, will a The floating-point three-phase array is used as the input array.

[0132] S202, perform a Karen-Bauer transform on the input array, and extract the initial transient zero-mode current sequence, i.e., the ground-mode component, through the decoupling transform matrix. The calculation formula is as follows:

[0133] ,

[0134] In the formula, These are three-phase high-frequency current sampling values. For line mode components, It is the zero-modulus component.

[0135] S203 inputs the initial transient zero-mode current sequence into the multi-resolution wavelet transform unit, uses the selected wavelet basis function (such as db4 wavelet basis) to decompose the signal into discrete approximate components and detail components. Based on the distribution characteristics that the transient zero-mode current is mainly concentrated in the mid-low frequency band, the high-frequency detail components are set to zero and then the signal is reconstructed to filter out high-frequency harmonics and noise interference.

[0136] S204, establish with The output array is a floating-point array of elements.

[0137] S205 stores the denoised results from S203 sequentially into the output array, serving as the output of the Karen-Bauer transform and wavelet denoising module.

[0138] S3. Process all contents of the original data table using TDM and store the results in the corresponding Pure Transient Zero Mode Current List (ZML) to obtain the Pure Zero Mode Current (PZMC) sequence, as follows:

[0139] S301, Initialize the count value of the data processing counter. .

[0140] S302, the first one in the original data table The three-phase current data of each element are used as the first temporary variable. .

[0141] S303, move the first temporary variable The input array is used as the TDM input array, and the output array is obtained through TDM processing, which is then used as the second temporary variable. .

[0142] S304, construct a new element Z in ZML, and set the two fields of Z as follows:

[0143] Terminal number is the first in the original data table The terminal number field of each element;

[0144] The transient zero-mode current is the second temporary variable. .

[0145] S305, add the new element Z to ZML.

[0146] S306, making .

[0147] S307, if If the number of elements is less than or equal to the number of elements in the original data table, proceed to S302; otherwise, proceed to S4.

[0148] S4. Establish a Waveform Correlation Module (WCM) for calculating waveform similarity cross-correlation. Its input is the PZMC (upstream and downstream sequences) of two adjacent measurement nodes, and its output is the corrected waveform composite features of adjacent nodes, as detailed below:

[0149] S401, the upstream sequence is denoted as The downstream sequence is denoted as Both are included A floating-point array of sampling points.

[0150] S402, let the translation step size be... ,in , This represents the maximum permissible time shift point for synchronization error. and Perform bidirectional translational sliding calculations to obtain cross-correlation coefficients at different time shifts. The formula is as follows:

[0151] ,

[0152] in, and It is decomposed into the average value of the upstream and downstream sequences.

[0153] Find the optimal time shift point corresponding to the maximum absolute value of the cross-correlation coefficient. Extract the optimal time shift point The corresponding cross-correlation coefficient is used as the original cross-correlation coefficient, i.e. .

[0154] S403, Extract and Record The polarity sign is denoted as .

[0155] Define a symbolic function to extract polarity:

[0156] .

[0157] S404, calculate respectively and The effective energy value is calculated using the following formula:

[0158] , ,

[0159] in, and They are respectively and The effective value of energy.

[0160] Seeking and The degree of deviation is used as the amplitude correction coefficient. :

[0161] .

[0162] For the original cross-correlation coefficients After adjusting the amplitude weights, an absolute comprehensive similarity coefficient is obtained that only considers the waveform shape similarity. .

[0163] S405, establishes waveform synthesis characteristics, including and Two fields are used to output the waveform synthesis features as the result of WCM.

[0164] S5, traverse the distribution network topology connectivity, calculate the similarity features of all adjacent intervals, and form a network-wide link similarity feature sequence, as follows:

[0165] S501, Initialize the list of network-wide link similarity features (an empty list).

[0166] S502, Initialize the link counter value .

[0167] S503, retrieve the first [item] from the distribution network topology connectivity list. Each element has an upstream terminal number and a downstream terminal number.

[0168] S504, find the element in ZML whose terminal number equals the upstream terminal number, and assign the pure transient zero-mode current sequence field of that element to the first temporary variable of the waveform. .

[0169] S505, find the element in ZML whose terminal number equals the downstream terminal number, and assign its PZMC field to the second temporary variable of the waveform. .

[0170] S506, As an upstream sequence, As a downstream sequence, the WCM module in S4 is called to process and obtain waveform synthesis features containing comprehensive similarity coefficients and polarity signs.

[0171] S507, a new element in establishing a list of network-wide link similarity features. Set its fields:

[0172] The link number is the link number of the element in the current distribution network topology connectivity list.

[0173] The overall similarity is the overall similarity field of the waveform's overall features.

[0174] The polarity symbol is the polarity symbol field of the waveform synthesis feature.

[0175] S508, with new elements Add it to the list of network-wide link similarity features.

[0176] S509, makes .

[0177] S510, if If the total number of links is less than or equal to the total number of links, proceed to S503; otherwise, proceed to S6.

[0178] S6, establish a fault interval discrimination module, traverse the entire network link similarity feature list to output accurate fault location, as detailed below:

[0179] S601, Set the correlation coefficient safety threshold (Typically, the value is in the range of 0.6 to 0.8); Establish a list of suspected fault link ranges for storing abnormal links (Fault Link List), which is an empty list; Determine the counter count value. ;

[0180] S602, extract the first element from the list of network-wide link similarity features. One element;

[0181] S603, Fault determination criteria: Combined with the comprehensive similarity coefficient obtained from S404 The polarity sign obtained with S403 The link is considered to be abnormal if any of the following inequalities are met:

[0182] Condition 1: Waveform distortion and amplitude attenuation: ,

[0183] Condition 2: Current polarity reversed: .

[0184] S604. If any of the abnormal conditions in S603 are met, the link number of the element is entered into the Fault Link List.

[0185] S605, makes ;

[0186] S606, if If the total number of links is less than or equal to the total number of links, proceed to S602; otherwise, proceed to S607.

[0187] S607 outputs the links contained in the list of suspected fault sections, and uses them as the precise location section of the final high-resistance grounding fault, triggering a network-wide GIS topology alarm and isolation and handling command.

[0188] Specific Implementation Method Two: The power grid grounding fault location device described in this implementation method includes a processor and a memory. The memory stores at least one instruction, which is loaded and executed by the processor to implement the power grid grounding fault location method as described in Specific Implementation Method One.

[0189] Specific Implementation Method 3: A computer storage medium as described in this embodiment stores at least one instruction, which is loaded and executed by a processor to implement the power distribution network grounding fault location method as described in Specific Implementation Method 1.

[0190] While specific embodiments of this application have been described herein with reference to them, it should be understood that these embodiments are merely examples of the principles and applications of this application. Therefore, it should be understood that many modifications can be made to the exemplary embodiments, and other arrangements can be designed without departing from the spirit and scope of this application as defined by the appended claims. It should be understood that different dependent claims and features described herein can be combined in ways different from those described in the original claims. It is also understood that features described in conjunction with individual embodiments can be used in other described embodiments.

Claims

1. A method for locating grounding faults in distribution networks based on bidirectional translational cross-correlation and waveform polarity identification, characterized in that, include: When a zero-sequence voltage change occurs in the distribution network, the three-phase current data of each measuring node in the distribution network are collected to form an original data table, and the topological connectivity of the distribution network is read. Each element in the original data table includes a terminal number and a three-phase current array. The terminal number is used to indicate the topological location of the measuring node, and the three-phase current array is the three-phase high-frequency current data of the measuring node within the acquisition time window. Each element in the original data table is subjected to Karen-Bauer transform and wavelet denoising to obtain a pure transient zero-mode current sequence. The topological connectivity of the distribution network is traversed, and the optimal time shift point is found by bidirectional translation and sliding of the pure transient zero-mode current sequence of two adjacent measurement nodes. The waveform comprehensive characteristics of all measurement node pairs are obtained by using the energy effective value of the corresponding sequence of two adjacent measurement nodes. Based on the waveform synthesis characteristics and combined with the preset fault criteria, it is determined whether the link corresponding to each measurement node is in a fault range.

2. The method for locating grounding faults in distribution networks based on bidirectional translational cross-correlation and waveform polarity identification according to claim 1, characterized in that, The step of performing Karen-Bauer transform and wavelet denoising on each element in the original data table to obtain a pure transient zero-mode current sequence includes: The first in the original data table The three-phase high-frequency current data of each element are input into the Karen-Bauer transform and wavelet denoising module for Karen-Bauer transform and wavelet denoising processing to obtain the first element. The transient zero-mode current corresponding to each element; Using the first data table The terminal number of each element and the transient zero-mode current are used to construct a pure transient zero-mode current element; Repeat the above steps until all elements in the original data table have obtained the corresponding pure transient zero-mode current element; A pure transient zero-mode current sequence is constructed using the pure transient zero-mode current elements corresponding to all elements in the original data table.

3. The method for locating grounding faults in distribution networks based on bidirectional translational cross-correlation and waveform polarity identification according to claim 2, characterized in that, The Karen-Bauer transform and wavelet denoising module includes: Karenbauer Transform Unit: Used to perform Karenbauer transformation on the three-phase high-frequency current data of each element in the original data table, and extract the initial transient zero-mode current sequence through the decoupling transformation matrix; Resolution wavelet transform unit: used to decompose the initial transient zero-mode current sequence into discrete approximate components and detail components using preset wavelet basis functions. Reconstruction unit: used to reconstruct the signal after setting the detailed components to zero, and to filter out high-frequency harmonics and noise.

4. The method for locating grounding faults in distribution networks based on bidirectional translational cross-correlation and waveform polarity identification according to claim 1, characterized in that, By traversing the topological connectivity of the distribution network, the optimal time shift point is found by bidirectional translation and sliding of the pure transient zero-mode current sequences of two adjacent measurement nodes, and the waveform comprehensive characteristics of all measurement node pairs are obtained by using the effective energy values ​​of the corresponding sequences of two adjacent measurement nodes, including: Read the topological connectivity of the power distribution network. The upstream terminal number and downstream terminal number of the link; In the pure transient zero-mode current sequence, find the element whose terminal number is equal to the upstream terminal number and the downstream terminal number; The two found elements are input into the waveform similarity cross-correlation calculation module for bidirectional translation and sliding calculation to obtain the first element. The waveform characteristics corresponding to each link; Repeat the above steps until all links in the topological connectivity of the distribution network have obtained the corresponding waveform comprehensive characteristics.

5. The method for locating grounding faults in distribution networks based on bidirectional translational cross-correlation and waveform polarity identification according to claim 4, characterized in that, The two found elements are input into the waveform similarity cross-correlation calculation module for bidirectional translation and sliding calculation to obtain the first element. The waveform characteristics corresponding to each link include: The two found elements are used as the upstream and downstream sequences, respectively. A bidirectional translation and sliding calculation is performed on the upstream and downstream sequences to obtain the cross-correlation coefficients under different time shifts. The time shift point corresponding to the maximum absolute value of the cross-correlation coefficient is taken as the optimal time shift point. The cross-correlation coefficient corresponding to the optimal time shift point is extracted as the original cross-correlation coefficient, and then the polarity sign of the original cross-correlation coefficient is obtained. Calculate the effective energy values ​​of the upstream and downstream sequences respectively; The deviation between the effective energy values ​​of the upstream and downstream sequences is used as the amplitude correction coefficient. The amplitude correction coefficient is used to correct the amplitude weight of the original cross-correlation coefficient to obtain an absolute comprehensive similarity coefficient that only considers the waveform shape approximation. The polarity signs of the absolute comprehensive similarity coefficient and the original cross-correlation coefficient are used to construct waveform comprehensive features.

6. The method for locating grounding faults in distribution networks based on bidirectional translational cross-correlation and waveform polarity identification according to claim 5, characterized in that, The step of performing bidirectional translation and sliding calculations on the upstream and downstream sequences to obtain cross-correlation coefficients at different time shifts includes: The upstream and downstream sequences are calculated using the following formula for bidirectional translation and sliding: ； in, Let be the translation step size, and , The maximum permissible number of time shift points for synchronization error. Indicates translation step size The cross-correlation coefficients under the following conditions; and These represent the upstream sequence notation and the downstream sequence notation, respectively. , The number of sampling points for the found element. and It is decomposed into the average value of the upstream and downstream sequences.

7. The method for locating grounding faults in distribution networks based on bidirectional translational cross-correlation and waveform polarity identification according to claim 5, characterized in that, The expression for the polarity sign of the original cross-correlation coefficient is: ， in, Indicates the polarity sign. Represents a symbolic function. This represents the original cross-correlation coefficients.

8. The method for locating grounding faults in distribution networks based on bidirectional translational cross-correlation and waveform polarity identification according to claim 6, characterized in that, The calculation of the effective energy values ​​of the upstream and downstream sequences respectively includes: The effective energy values ​​of the upstream and downstream sequences are calculated according to the following formula: ， ， in, and They are respectively and The effective value of energy.

9. The method for locating grounding faults in a distribution network based on bidirectional translational cross-correlation and waveform polarity identification according to claim 8, characterized in that, The amplitude correction factor is calculated according to the following formula: ， in, This is the amplitude correction factor. and These represent taking the minimum and maximum values, respectively. The original cross-correlation coefficients are adjusted for magnitude weights according to the following formula: ， in, This is the absolute comprehensive similarity coefficient that only considers the approximation of waveform shape. This represents the original cross-correlation coefficients.

10. The method for locating grounding faults in a distribution network based on bidirectional translational cross-correlation and waveform polarity identification according to claim 9, characterized in that, The step of determining whether each measurement node's corresponding link is in a faulty interval based on the waveform synthesis characteristics and a preset fault criterion includes: The preset fault criteria include: If any of the following inequalities are satisfied, the link is considered to be abnormal: Condition one: , This is a safety threshold; Condition two: , Indicates the polarity sign.

Images