Power distribution network multi-branch fault interval positioning method and system based on OMP algorithm

Abstract

The invention provides a power distribution network multi-branch fault interval positioning method and system based on an OMP algorithm, and the method comprises the steps: carrying out the subtraction of a reflection waveform in a fault state and a pre-stored normal reflection waveform in a normal state, and obtaining a waveform difference signal; first fault reflection wave heads in the waveform difference signals at the two ends are extracted respectively, fault distance estimation values corresponding to the two ends are obtained through calculation according to the fault reflection wave heads, and two groups of independent fault point candidate interval sets are generated; obtaining each candidate interval in an intersection of two groups of independent fault point candidate interval sets, constructing a local dictionary by using a fault sample waveform of a boundary node of the candidate interval, and performing sparse representation and reconstruction on the local dictionary by adopting an OMP algorithm; and the reconstruction error of each candidate interval is calculated, and the candidate interval with the minimum reconstruction error is selected as the final interval where the fault is located. According to the method, the real fault interval can be uniquely and accurately identified from a plurality of geometrically possible intervals.

Description

A Method and System for Locating Fault Sections in Distribution Networks Based on OMP Algorithm Technical Field

[0001] This invention relates to the field of fault location technology, and in particular to a method and system for locating multi-branch fault sections in a distribution network based on the OMP algorithm. Background Technology

[0002] With the acceleration of urbanization, power cables are increasingly widely used in power distribution networks due to their advantages such as small footprint and high reliability. However, the concealed nature and complex multi-branch topology of cable networks pose significant challenges to accurate fault location after failure.

[0003] Existing fault location methods mainly include impedance methods, artificial intelligence methods, and traveling wave methods. Impedance methods are easily affected by transition resistance and have low accuracy; artificial intelligence methods rely on a large amount of historical fault data for training and are at risk of getting trapped in local optima. Traveling wave methods have become mainstream due to their high accuracy and robustness, but single-ended traveling wave methods often cause ambiguity in the location interval due to multiple waveform reflections in multi-branch networks.

[0004] While the two-ended traveling wave method can eliminate some ambiguities, its performance is highly dependent on the traveling wave signal generated by natural faults, which is often weak and susceptible to noise interference. Furthermore, existing two-ended methods still struggle to uniquely determine the fault region when dealing with multi-branch structures. Summary of the Invention

[0005] The purpose of this invention is to provide a method and system for locating multi-branch fault sections in a distribution network based on the OMP algorithm, aiming to solve at least one of the problems in the background art.

[0006] In a first aspect, the present invention provides a method for locating multi-branch fault sections in a distribution network based on the OMP algorithm, the method comprising:

[0007] After a fault occurs, a self-modulated attenuated square wave signal is synchronously injected at both ends of the distribution network feeder, and the reflected waveforms at both ends are collected. The difference between the reflected waveform under the fault condition and the pre-stored normal reflected waveform under the normal condition is used to obtain the waveform difference signal.

[0008] The first fault reflection wavefront is extracted from the waveform difference signal at both ends, and the fault distance estimate corresponding to each end is calculated based on the fault reflection wavefront. With the two fault distance estimates as constraints, the depth-first search algorithm is started from the beginning and end of the line respectively to generate two independent sets of fault point candidate intervals.

[0009] For each candidate interval in the intersection of two independent sets of candidate fault point intervals, a local dictionary is constructed using the fault sample waveforms of its boundary nodes, and the OMP algorithm is used to perform sparse representation and reconstruction of the local dictionary.

[0010] Calculate the reconstruction error for each candidate interval, and select the candidate interval with the smallest reconstruction error as the final fault interval.

[0011] In some embodiments, the step of extracting the first fault reflection wavefront from the waveform difference signals at both ends includes:

[0012] Perform frequency slice wavelet transform on the waveform difference signal:

[0013]

[0014] Where W(t,ω,σ) is the time-frequency transformation result, σ is the scaling factor, and λ is the energy coefficient. for The conjugate function, The Fourier transform of the wavelet function p(t) Let u be the Fourier transform of the waveform difference signal, where u is the integral variable and ω is the angular frequency.

[0015] In some embodiments, the step of calculating the estimated fault distance values ​​corresponding to the two ends based on the fault reflected wavefront includes:

[0016] Obtain the arrival times τ1 and τ2 of the fault reflected wavefront at both ends, and calculate the fault distance estimate using the following formula:

[0017]

[0018] Where d1 and d2 are the estimated fault distances corresponding to the beginning and end points, respectively, L is the total route length, and v0 is the wave velocity;

[0019] The wave speed is calculated using the following formula:

[0020]

[0021] Where L0 and C0 represent the inductance and capacitance per unit length of the transmission line, and μ and ε are the permeability and dielectric constant of the core wire, respectively. r μ0, ε r ε0 and ε0 refer to the relative permeability, vacuum permeability, relative permittivity, and vacuum permittivity, respectively.

[0022] In some embodiments, the step of generating two independent sets of candidate fault point intervals by initiating a depth-first search algorithm from the beginning and end of the line, respectively, constrained by two fault distance estimates, includes:

[0023] Based on the adjacency matrix A and distance matrix D of the distribution network, with d1 and d2 as the maximum search radii respectively, depth-first search is performed simultaneously from the beginning and end, recording all line segments that satisfy the distance constraints, forming the beginning candidate interval set and the end candidate interval set respectively.

[0024] In some embodiments, the steps of obtaining each candidate interval in the intersection of two independent sets of candidate fault point intervals, constructing a local dictionary using the fault sample waveforms of its boundary nodes, and performing sparse representation and reconstruction of the local dictionary using the OMP algorithm include:

[0025] For each candidate interval (i,j) in the intersection Ω, the reflected wave sample signal x is pre-stored when the fault occurs at its first node i and last node j. i and x j The reflected wave sample signal is used as dictionary atoms to construct a local dictionary Φ. (i，j) =[x i x j ];

[0026] The measured fault reflection signal y is subjected to OMP sparse decomposition on the local dictionary to solve for the sparse coefficient vector:

[0027]

[0028] Where K is the sparsity. Here, θ represents the estimated values ​​of the sparse coefficients, and θ is the sparse coefficient vector.

[0029] By iteratively selecting the atom most relevant to the residual, the reconstructed signal of signal y is finally obtained.

[0030] In some embodiments, the step of calculating the reconstruction error of each candidate interval and selecting the candidate interval with the smallest reconstruction error as the final fault interval includes:

[0031] The reconstruction error for each candidate interval is calculated using the following formula:

[0032]

[0033] Where, ε (i，j) This represents the reconstruction error.

[0034] Secondly, this invention provides a multi-branch fault location system for a distribution network based on the OMP algorithm, the system comprising:

[0035] The waveform difference signal acquisition module is used to synchronously inject self-modulated attenuated square wave signals at both ends of the distribution network feeder after a fault occurs, and to collect the reflected waveforms at both ends. The difference between the reflected waveform under the fault state and the pre-stored normal reflected waveform under the normal state is used to obtain the waveform difference signal.

[0036] The fault distance calculation module is used to extract the first fault reflection wavefront from the waveform difference signal at both ends, and calculate the fault distance estimate corresponding to both ends based on the fault reflection wavefront. With the two fault distance estimates as constraints, the depth-first search algorithm is started from the beginning and end of the line respectively to generate two independent sets of fault point candidate intervals.

[0037] The candidate interval acquisition module is used to acquire each candidate interval in the intersection of two independent sets of fault point candidate intervals, construct a local dictionary using the fault sample waveforms of its boundary nodes, and perform sparse representation and reconstruction of the local dictionary using the OMP algorithm.

[0038] The reconstruction error calculation module is used to calculate the reconstruction error of each candidate interval and select the candidate interval with the smallest reconstruction error as the final fault interval.

[0039] Thirdly, the present invention provides a storage medium that stores one or more programs, which, when executed by a processor, implement the above-described method for locating multi-branch fault sections in a distribution network based on the OMP algorithm.

[0040] Fourthly, the present invention provides an electronic device, the electronic device comprising a memory and a processor, wherein:

[0041] The memory is used to store computer programs;

[0042] When the processor executes the computer program stored in the memory, it implements the above-mentioned method for locating multi-branch fault sections in a distribution network based on the OMP algorithm.

[0043] Compared with the prior art, the present invention has the following advantages:

[0044] 1. This invention proposes a novel fault zone localization method integrating dual-end injection, dual-end ranging, and the OMP algorithm. This method first utilizes dual-end synchronous injection and ranging to obtain a more reliable initial fault distance value. Then, dual-end depth-first search narrows the fault zone range. Finally, the OMP algorithm is innovatively introduced to sparsely fit the fault waveform. By minimizing the reconstruction error, the final fault zone is accurately determined, effectively improving the localization accuracy and robustness in complex multi-branch networks. Specifically, by employing dual-end active signal injection, the signal-to-noise ratio and detectability of fault feature signals are effectively improved. By combining dual-end ranging and depth-first search, the search range of fault points is efficiently narrowed, avoiding full network traversal. By introducing the OMP sparse reconstruction algorithm, based on the principle of minimum reconstruction error, the true fault zone can be uniquely and accurately identified from multiple geometrically possible intervals, solving the localization ambiguity problem in multi-branch networks. Attached Figure Description

[0045] Figure 1 is a flowchart of a method for locating multi-branch fault sections in a distribution network based on the OMP algorithm, according to an embodiment of the present invention.

[0046] Figure 2 is a schematic diagram of the waveform of the double-ended injection signal in this invention;

[0047] Figure 3 is a time-domain diagram of the original signal in this invention;

[0048] Figure 4 is a time-frequency domain analysis diagram of FSWT in this invention;

[0049] Figure 5 is a simulation diagram of a simple power distribution network in this invention;

[0050] Figure 6 shows the total internal reflection waveform at the head end when a fault occurs in this invention;

[0051] Figure 7 shows the difference waveform for the first and last faults in this invention;

[0052] Figure 8 is a comparison of OMP reconstruction errors in two possible intervals in this invention;

[0053] Figure 9 shows a comparison of OMP reconstruction errors under different transition resistances in this invention; Figure 10 is a schematic diagram of the structure of the adaptive control system of the power distribution system proposed in an embodiment of this invention.

[0054] The following detailed description, in conjunction with the accompanying drawings, will further illustrate the present invention. Detailed Implementation

[0055] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention. Unless otherwise defined, the technical or scientific terms used herein should have the ordinary meaning understood by those skilled in the art. The terms "comprising" and similar expressions used herein mean that the element or object preceding the word covers the element or object listed after the word and its equivalents, but does not exclude other elements or objects.

[0056] As shown in Figures 1 to 8, an embodiment of the present invention proposes a method for locating multi-branch fault sections in a distribution network based on the OMP algorithm. This method includes steps S101 to S104, wherein:

[0057] Step S101: After a fault occurs, a self-modulated attenuated square wave signal is synchronously injected at both ends of the distribution network feeder, and the reflected waveforms at both ends are collected. The difference between the reflected waveform under the fault state and the pre-stored normal reflected waveform under the normal state is calculated to obtain the waveform difference signal.

[0058] It should be noted that in this step, a GPS timing module is used to synchronize the time of the signal generators at both ends, with a synchronization error of no more than 1 microsecond; the phase of the injected signal at the beginning is set to 0°, and the phase of the injected signal at the end is set to 180°, so as to distinguish the source of the reflected wave.

[0059] In addition, it is necessary to pre-store a sample library of normal reflection waveforms of each node in the distribution network under fault-free conditions; after a fault occurs, the reflection waveforms at the beginning and end points collected in real time are subtracted from the corresponding fault-free sample waveforms to obtain the waveform difference signal that eliminates the reflection interference of normal nodes.

[0060] Step S102: Extract the first fault reflection wavefront from the waveform difference signal at both ends respectively, and calculate the fault distance estimate corresponding to both ends based on the fault reflection wavefront. Using the two fault distance estimates as constraints, start the depth-first search algorithm from the beginning and end of the line respectively to generate two independent sets of fault point candidate intervals.

[0061] In this step, frequency slice wavelet transform (FSWT) is performed on the waveform difference signal. The first transient event with the most concentrated energy is identified in the resulting time-frequency graph, and the corresponding time is the arrival time of the fault reflected wavefront. Combining the traveling wave propagation speed and the total length of the line, the distance d1 from the fault point to the beginning and the distance d2 from the end are calculated using the double-ended traveling wave ranging formula.

[0062] Specifically, frequency slice wavelet transform is performed on the waveform difference signal:

[0063]

[0064] Where W(t,ω,σ) is the time-frequency transformation result, σ is the scaling factor, and λ is the energy coefficient. for The conjugate function, The Fourier transform of the wavelet function p(t) Let u be the Fourier transform of the waveform difference signal, where u is the integral variable and ω is the angular frequency.

[0065] Let the scaling factor σ = ω / k, k > 0, then:

[0066]

[0067] Where k is the time-frequency resolution coefficient, used to adjust the time-frequency domain response sensitivity to the transform, and is independent of ω and u.

[0068] Furthermore, in some embodiments, since the first reflected waveform in the waveform difference data is the fault point reflected wavefront, by inputting complete one-dimensional waveform difference data and sampling frequency, the frequency-time relationship is constructed using frequency slice wavelet transform. The times τ1 and τ2 when the first frequency maximum appears are the times τ1 and τ2 when the fault point reflected wavefront reaches the beginning and end.

[0069] The voltage across the distributed capacitance and the current across the distributed inductance in a cable cannot change abruptly in a short time. The propagation process of a traveling wave requires a certain amount of time to establish. Its wave speed is the ratio of the cable's geometric length to the time required for propagation, which can be calculated using the following formula:

[0070]

[0071] Where L0 and C0 represent the inductance and capacitance per unit length of the transmission line, and μ and ε are the permeability and dielectric constant of the core wire, μ r μ0, ε r ε0 and ε0 refer to the relative permeability, vacuum permeability, relative permittivity, and vacuum permittivity, respectively.

[0072] Combining the arrival time difference of the wavefronts at both ends with the total route length L, the formula for calculating the distance from the fault point to both ends is:

[0073]

[0074] Where d1 and d2 are the estimated fault distances corresponding to the beginning and end points, respectively, L is the total route length, and v0 is the wave velocity.

[0075] In the process of generating the candidate interval set of fault points, it is necessary to perform a depth-first search simultaneously from the beginning and end of the distribution network based on the adjacency matrix A and distance matrix D, with d1 and d2 as the maximum search radii respectively, to record all line segments that meet the distance constraints, and form the candidate interval set at the beginning and the candidate interval set at the end respectively.

[0076] Step S103: Obtain each candidate interval in the intersection of two independent sets of candidate fault point intervals, construct a local dictionary using the fault sample waveforms of its boundary nodes, and use the OMP algorithm to perform sparse representation and reconstruction of the local dictionary;

[0077] In this step, for each candidate interval (i,j) in the intersection Ω, the reflected wave sample signal x is pre-stored when the fault occurs at its first node i and last node j. i and x j The reflected wave sample signal is used as dictionary atoms to construct a local dictionary Φ. (i,j) =[x i ,x j ];

[0078] The measured fault reflection signal y is subjected to OMP sparse decomposition on the local dictionary to solve for the sparse coefficient vector:

[0079]

[0080] Where K is the sparsity. Here, θ represents the estimated values ​​of the sparse coefficients, and θ is the sparse coefficient vector.

[0081] By iteratively selecting the atom most relevant to the residual, the reconstructed signal of signal y is finally obtained.

[0082] Step S104: Calculate the reconstruction error of each candidate interval, and select the candidate interval with the smallest reconstruction error as the final fault interval.

[0083] In some embodiments, the reconstruction error of each candidate interval is specifically calculated according to the following formula:

[0084]

[0085] Where, ε (i,j) This represents the reconstruction error.

[0086] For example, it should be noted that this embodiment uses a simple single-control power grid system as the test platform and performs simulation verification in the Matlab / Simulink environment. First, during signal injection and data acquisition, synchronization signal generators are configured at the beginning (S-end, corresponding to node 1) and end (E-end, corresponding to node 4) of the feeder. The injected signal is set to a self-modulated attenuated square wave with an amplitude of 1000V and a frequency of 2MHz. A GPS timing module ensures strict synchronization at both ends (error ≤ 1μs), with the signal phase at the S-end being 0° and at the E-end being 180°. Under fault-free conditions, the total reflection waveforms at the S-end and E-end are acquired and stored to construct a "fault-free sample library." Subsequently, an A-phase ground fault is set on the line between node 2 and node 3, with transition resistances set to 10Ω, 50Ω, and 100Ω respectively to simulate faults of different severity. After the fault occurs, signals are injected synchronously at both ends again, and the "fault reflection waveforms" at the S-end and E-end are acquired.

[0087] Then, the fault reflection waveform acquired at the S-terminal is subtracted point by point from its corresponding fault-free sample waveform to obtain the waveform difference signal at the S-terminal. Similarly, the data at the E-terminal is processed to obtain the waveform difference signal at the E-terminal. This operation effectively eliminates inherent reflection interference caused by normal line branches, joints, etc., so that the waveform difference signal mainly contains reflection characteristics from the fault point.

[0088] Next, frequency slice wavelet transform (FSWT) was performed on the waveform difference signals at the S and E ends respectively. In the resulting time-frequency diagram, the first transient event of energy concentration was clearly identified, and its corresponding times are the arrival times τ1 and τ2 of the fault reflected wavefront. Given that the total length L of the simulated line is 10km, and the traveling wave propagation velocity v is calculated based on the cable parameters, the distances d1 from the fault point at the S end and d2 from the fault point at the E end were calculated using the double-ended traveling wave ranging formula. With a 10Ω transition resistance, d1 = 4.998km and d2 = 5.002km were measured, which highly match the actual fault location (5km from the S end).

[0089] Next, based on the topology of a simple single-dominated power grid system, its adjacency matrix A and distance matrix D are constructed. With d1 = 4.998 km as a constraint, a depth-first search (DFS) is initiated from the S end (node ​​1), traversing all paths and recording all line segments whose cumulative distance does not exceed d1, resulting in the first candidate interval set Ω_S = {(1,2),(2,3)}. Similarly, with d2 = 5.002 km as a constraint, a reverse DFS is initiated from the E end (node ​​4), resulting in the last candidate interval set Ω_E = {(2,4),(2,3)}. The intersection of the two sets is then calculated to obtain the final set of intervals to be judged Ω = Ω_S ∩ Ω_E = {(2,3)}.

[0090] To fully verify the effectiveness of the OMP algorithm, this embodiment further constructs a more challenging scenario: assuming that due to measurement errors, d1' = 5.5km and d2' = 5.5km are measured. In this case, the DFS search yields Ω_S' = {(1,2),(2,3),(2,4)} and Ω_E' = {(2,4),(2,3),(1,2)}, whose intersection Ω' = {(1,2),(2,3),(2,4)} contains three possible candidate intervals.

[0091] For each interval in Ω', perform OMP discrimination:

[0092] - For the interval (2,3), retrieve the reflected wave samples x2(t) and x3(t) when the fault occurs at node 2 and node 3 from the pre-stored fault sample library to form a dictionary Φ_(2,3)=[x2,x3].

[0093] For the interval (2,4), retrieve samples x2(t) and x4(t) to form a dictionary Φ_(2,4)=[x2,x4].

[0094] Similarly, construct a dictionary Φ_(1,2) for the interval (1,2).

[0095] The actual measured fault reflection signal y is subjected to OMP sparse decomposition (sparseness K=2) on the above three dictionaries, and the reconstructed signal is calculated for each. And the reconstruction error ε. The calculation results are: ε(2,3)=0.675, ε(2,4)=3897.605, ε(1,2)=4125.331.

[0096] Finally, comparing the reconstruction errors of each candidate interval, the error of interval (2,3) is much smaller than that of the other two intervals. Therefore, the method of this invention accurately identifies the fault interval as (2,3), which is completely consistent with the preset fault location. The reconstruction error comparison chart intuitively demonstrates the powerful pattern matching and discrimination capabilities of the OMP algorithm. Furthermore, repeating the above experiments under high transition resistances of 50Ω and 100Ω, the OMP algorithm can still correctly identify the fault interval (2,3) through the principle of minimum reconstruction error, proving that the method of this invention has excellent robustness and adaptability to high-resistance faults.

[0097] In summary, this invention proposes a novel fault zone localization method that integrates dual-end injection, dual-end ranging, and the OMP algorithm. This method first utilizes dual-end synchronous injection and ranging to obtain a more reliable initial fault distance value. Then, it narrows the fault zone range through dual-end depth-first search. Finally, it innovatively introduces the OMP algorithm to sparsely fit the fault waveform, accurately determining the final fault zone by minimizing the reconstruction error, effectively improving the localization accuracy and robustness in complex multi-branch networks. Specifically, by employing dual-end active signal injection, the signal-to-noise ratio and detectability of fault feature signals are effectively improved. By combining dual-end ranging and depth-first search, the search range of fault points is efficiently narrowed, avoiding full network traversal. By introducing the OMP sparse reconstruction algorithm, and using the principle of minimum reconstruction error, the true fault zone can be uniquely and accurately identified from multiple geometrically possible intervals, solving the localization ambiguity problem in multi-branch networks.

[0098] As shown in Figure 9, an embodiment of the present invention also proposes a multi-branch fault location system for distribution networks based on the OMP algorithm, the system comprising:

[0099] The waveform difference signal acquisition module 10 is used to synchronously inject self-modulated attenuated square wave signals at both ends of the distribution network feeder after a fault occurs, and to collect the reflected waveforms at both ends. The waveform difference signal is obtained by subtracting the reflected waveform under the fault state from the pre-stored normal reflected waveform under the normal state.

[0100] The fault distance calculation module 20 is used to extract the first fault reflection wavefront from the waveform difference signal at both ends, and calculate the fault distance estimate corresponding to both ends based on the fault reflection wavefront. With the two fault distance estimates as constraints, the depth-first search algorithm is started from the beginning and end of the line respectively to generate two independent sets of fault point candidate intervals.

[0101] The candidate interval acquisition module 30 is used to acquire each candidate interval in the intersection of two independent sets of fault point candidate intervals, construct a local dictionary with the fault sample waveforms of its boundary nodes, and perform sparse representation and reconstruction of the local dictionary using the OMP algorithm.

[0102] The reconstruction error calculation module 40 is used to calculate the reconstruction error of each candidate interval and select the candidate interval with the smallest reconstruction error as the final fault interval.

[0103] In another aspect, the present invention also proposes a storage medium on which one or more programs are stored, which, when executed by a processor, implement the above-described method for locating multi-branch fault sections in a distribution network based on the OMP algorithm.

[0104] In another aspect, the present invention also proposes an electronic device, including a memory and a processor, wherein the memory is used to store a computer program, and the processor is used to execute the computer program stored in the memory to realize the above-mentioned method for locating multi-branch fault sections in a distribution network based on the OMP algorithm.

[0105] Those skilled in the art will understand that the logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a ordered list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus, or device (such as a computer-based system, a processor-included system, or other system that can fetch and execute instructions from, an instruction execution system, apparatus, or device). For the purposes of this specification, "computer-readable medium" can mean any means that can contain stored, communicated, propagated, or transmitted programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0106] More specific examples of computer-readable media (a non-exhaustive list) include: electrical connections (electronic devices) having one or more wires, portable computer disk drives (magnetic devices), random access memory (RAM), read-only memory (ROM), erasable and editable read-only memory (EPROM or flash memory), fiber optic devices, and portable optical disc read-only memory (CDROM). Furthermore, computer-readable media can even be paper or other suitable media on which the program can be printed, since the program can be obtained electronically, for example, by optically scanning the paper or other medium, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0107] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0108] While embodiments of the present invention have been described in detail above, it will be apparent to those skilled in the art that various modifications and variations can be made to these embodiments. However, it should be understood that such modifications and variations fall within the scope and spirit of the invention as set forth in the claims. Furthermore, the invention described herein may have other embodiments and can be implemented or carried out in various ways.

Claims

1. A method for locating multi-branch fault sections in a distribution network based on the OMP algorithm, characterized in that, The method includes: after a fault occurs, a self-modulated attenuated square wave signal is synchronously injected at both ends of the distribution network feeder, and the reflected waveforms at both ends are collected. The difference between the reflected waveform under the fault state and the pre-stored normal reflected waveform under the normal state is calculated to obtain a waveform difference signal. The first fault reflected wavefront is extracted from the waveform difference signal at both ends, and the fault distance estimate corresponding to each end is calculated based on the fault reflected wavefront. With the two fault distance estimates as constraints, a depth-first search algorithm is started from the beginning and end of the line to generate two independent sets of fault point candidate intervals. Each candidate interval in the intersection of the two independent sets of fault point candidate intervals is obtained, and a local dictionary is constructed using the fault sample waveforms of its boundary nodes. The OMP algorithm is used to perform sparse representation and reconstruction of the local dictionary. The reconstruction error of each candidate interval is calculated, and the candidate interval with the smallest reconstruction error is selected as the final fault location interval.

2. The method for locating multi-branch fault sections in a distribution network based on the OMP algorithm according to claim 1, characterized in that, The step of extracting the first fault reflection wavefront from the waveform difference signals at both ends includes: performing a frequency slice wavelet transform on the waveform difference signals. Where W(t, ω, σ) is the time-frequency transformation result, σ is the scaling factor, and λ is the energy coefficient. for The conjugate function, The Fourier transform of the wavelet function p(t) Let u be the Fourier transform of the waveform difference signal, where u is the integral variable and ω is the angular frequency.

3. The method for locating multi-branch fault sections in a distribution network based on the OMP algorithm according to claim 2, characterized in that, The step of calculating the estimated fault distance values ​​corresponding to the two ends based on the fault reflected wavefront includes: obtaining the arrival times τ1 and τ2 of the fault reflected wavefront at the beginning and end ends, and calculating the estimated fault distance values ​​according to the following formula: Where d1 and d2 are the estimated fault distances corresponding to the beginning and end points, respectively, L is the total route length, and v0 is the wave velocity; the wave velocity is calculated according to the following formula: Where L0 and C0 represent the inductance and capacitance per unit length of the transmission line, and μ and ε are the permeability and dielectric constant of the core wire, μ r μ0, ε r ε0 and ε0 refer to the relative permeability, vacuum permeability, relative permittivity, and vacuum permittivity, respectively.

4. The method for locating multi-branch fault sections in a distribution network based on the OMP algorithm according to claim 3, characterized in that, The steps of generating two independent candidate interval sets of fault points by starting a depth-first search algorithm from the beginning and end of the line with two fault distance estimates as constraints include: based on the adjacency matrix A and distance matrix D of the distribution network, using d1 and d2 as the maximum search radii respectively, performing a depth-first search simultaneously from the beginning and end of the line, recording all line segments that satisfy the distance constraints, and forming the candidate interval set at the beginning and the candidate interval set at the end respectively.

5. The method for locating multi-branch fault sections in a distribution network based on the OMP algorithm according to claim 4, characterized in that, The steps of obtaining each candidate interval in the intersection of two independent sets of candidate fault point intervals, constructing a local dictionary from the fault sample waveforms at its boundary nodes, and using the OMP algorithm to perform sparse representation and reconstruction of the local dictionary include: for each candidate interval (i, j) in the intersection Ω, pre-storing the reflected wave sample signal x when the fault occurs at its first end node i and last end node j. i and x j The reflected wave sample signal is used as dictionary atoms to construct a local dictionary Φ. (i，j) =[x i x j The measured fault reflection signal y is subjected to OMP sparse decomposition on the local dictionary to solve for the sparse coefficient vector. Where K is the sparsity. Here, θ represents the estimated sparse coefficients, and θ is the sparse coefficient vector. By iteratively selecting the atom most correlated with the residual, the reconstructed signal of signal y is finally obtained.

6. The method for locating multi-branch fault sections in a distribution network based on the OMP algorithm according to claim 5, characterized in that, The step of calculating the reconstruction error of each candidate interval and selecting the candidate interval with the smallest reconstruction error as the final fault location interval includes: calculating the reconstruction error of each candidate interval according to the following formula: Where, ε (i,j) This represents the reconstruction error.

7. A multi-branch fault location system for distribution networks based on the OMP algorithm, characterized in that, The system includes: a waveform difference signal acquisition module, used to synchronously inject self-modulated attenuated square wave signals at both ends of the distribution network feeder after a fault occurs, and to collect the reflected waveforms at both ends, and to subtract the reflected waveform under the fault state from the pre-stored normal reflected waveform under the normal state to obtain the waveform difference signal; a fault distance calculation module, used to extract the first fault reflected wavefront from the waveform difference signals at both ends, and to calculate the fault distance estimate corresponding to each end based on the fault reflected wavefront, and to use the two fault distance estimates as constraints to start a depth-first search algorithm from the beginning and end of the line respectively to generate two independent sets of fault point candidate intervals; a candidate interval acquisition module, used to acquire each candidate interval in the intersection of the two independent sets of fault point candidate intervals, to construct a local dictionary using the fault sample waveforms of its boundary nodes, and to perform sparse representation and reconstruction of the local dictionary using the OMP algorithm; and a reconstruction error calculation module, used to calculate the reconstruction error of each candidate interval, and to select the candidate interval with the smallest reconstruction error as the final fault location interval.

8. A storage medium, characterized in that, The storage medium stores one or more programs, which, when executed by a processor, implement the method for locating multi-branch fault sections in a distribution network based on the OMP algorithm as described in any one of claims 1-6.

9. An electronic device comprising a memory and a processor, wherein: The memory is used to store computer programs; when the processor executes the computer programs stored in the memory, it implements the method for locating multi-branch fault sections in a distribution network based on the OMP algorithm as described in any one of claims 1-6.

Images