Distribution feeder fault detection method considering transient frequency response of distributed power supply

Abstract

The invention discloses a power distribution feeder fault detection method considering transient frequency response of a distributed power supply, which relates to the technical field of power distribution network fault detection and comprises the following steps: acquiring a voltage traveling wave signal and performing wavelet transformation to obtain a multi-scale wavelet coefficient; calculating wavelet energy entropy and extracting local time-frequency energy characteristics; constructing a time-frequency energy matrix and calculating a norm of the time-frequency energy matrix to judge whether detection is started or not; and after a threshold condition is met, normalizing the energy matrix, screening out invalid data, calculating a non-zero norm, obtaining a distribution proportionality coefficient according to the non-zero norm, and finally completing fault identification. The method has better effects in the aspects of accuracy and applicability.

Description

Technical Field

[0001] This invention relates to the field of power distribution network fault detection technology, specifically to a power distribution feeder fault detection method that takes into account the transient frequency response of distributed generation. Background Technology

[0002] Single-phase grounding faults frequently occur in distribution networks, posing a threat to their safe operation. High-resistance grounding faults, in particular, often exhibit very subtle fault characteristics, which can even be mistaken for normal transient disturbances. Traditional overcurrent protection technologies cannot reliably detect these faults. With the continuous development of new energy technologies, distributed power sources are increasingly being integrated into distribution networks. The resulting transient frequency responses also impact fault detection technologies. High-resistance grounding faults can damage equipment insulation, disrupt the normal operation of the power grid, and potentially cause electric shock accidents and forest fires. Therefore, a fault detection method capable of quickly and accurately identifying faults and distinguishing them from disturbances is crucial for the normal and safe operation of distribution networks.

[0003] In distribution networks, steady-state quantities are often used for fault detection. However, this method is limited by the grounding method of the neutral point. As the transition resistance increases, the amplitude of the steady-state quantity decreases, making the extraction and identification of characteristic components difficult. Furthermore, the steady-state characteristic components detected during normal disturbances and faults can be confused. Although some scholars have proposed selecting fault feeders based on the polarity of transient quantities, this method is not only limited by the accuracy of current transformers, but also prone to false tripping due to the transient electrical quantities generated by normal transient disturbances in the distribution network. With the development of artificial intelligence algorithms, some scholars have proposed using artificial neural networks and support vector machines to identify fault characteristic components, thereby achieving fault detection. However, the physical meaning of these methods is not clear enough, and the accuracy of the detection results depends on the quality of the data. Installing a large number of communication and detection devices in the distribution network can acquire the real-time operating status of the grid and achieve fault detection. However, the cost of installing such a large number of devices and the challenges of grid renovation projects are highly controversial, and this method does not address the transient frequency response caused by distributed generation. Summary of the Invention

[0004] In view of the above-mentioned problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by this invention is that existing steady-state quantity detection methods are sensitive to the neutral point grounding method, have low detection sensitivity, and are weakly adaptable to distributed power sources. It is also important to understand how to quickly and accurately distinguish between high-resistance faults and disturbances.

[0006] To address the aforementioned technical problems, this invention provides the following technical solution: a method for detecting distribution feeder faults considering the transient frequency response of distributed power sources, comprising: acquiring voltage traveling wave signals in the distribution network; performing wavelet transform processing on the acquired traveling wave signals to obtain wavelet coefficients at multiple scales; calculating wavelet energy entropy based on energy distribution, and extracting local energy features in the time and frequency domains under a set sliding window and sliding step size; dividing the traveling wave signal into P segments in frequency and Q segments in time, constructing an energy matrix, calculating the L1 norm of the energy matrix to obtain the total energy E, and comparing it with a set threshold E. s A comparison is made to determine whether fault detection should be initiated; the energy matrix is ​​normalized, values ​​below a set threshold λ are filtered out, its non-zero norm L0 is calculated, and the distribution ratio coefficient ψ is obtained. Based on ψ and the set criterion Ψ... set The comparison results determine the test results.

[0007] As a preferred embodiment of the distribution feeder fault detection method considering the transient frequency response of distributed power sources described in this invention, the step of acquiring voltage traveling wave signals in the distribution network includes using a traveling wave detection device configured at the beginning of the line to acquire voltage traveling wave signals, with the sampling time set to 110μs and the sampling frequency set to 2kHz to 10kHz.

[0008] As a preferred embodiment of the distribution feeder fault detection method considering the transient frequency response of distributed power sources described in this invention, the step of performing wavelet transform processing on the acquired traveling wave signal to obtain wavelet coefficients at multiple scales includes decomposing the signal into wavelets at multiple scales, obtaining approximation coefficients and detail coefficients, and calculating the wavelet energy E at each scale. j Then calculate the energy ratio K. j Based on K j The wavelet energy entropy at each scale is obtained.

[0009] As a preferred embodiment of the distribution feeder fault detection method considering the transient frequency response of distributed power sources described in this invention, the extraction of local energy features in the time and frequency domains includes defining the sliding window width w and the sliding step size δ, calculating the wavelet energy entropy sequence by sliding the window in the signal, and forming local energy features at the time scale.

[0010] As a preferred embodiment of the distribution feeder fault detection method considering the transient frequency response of distributed power sources described in this invention, the energy matrix construction includes dividing the wavelet-transformed signal into P frequency bands, and further dividing each frequency band into Q time periods to obtain multiple time-frequency block energies E. aβ Combined to form a two-dimensional energy matrix E p Q.

[0011] As a preferred embodiment of the distribution feeder fault detection method considering the transient frequency response of distributed power sources described in this invention, wherein: the step of... s A comparison to determine whether to initiate fault detection involves calculating the L1 norm of the energy matrix to obtain the total energy E, and comparing E with a set threshold E0. s If E is greater than E, then... s If the signal is detected, fault detection will be initiated; otherwise, the system will return to re-acquire the signal.

[0012] As a preferred embodiment of the distribution feeder fault detection method considering the transient frequency response of distributed power sources described in this invention, the normalization process of the energy matrix includes normalizing the matrix elements to the [0,1] interval according to the maximum and minimum values, filtering out elements smaller than a threshold, calculating the non-zero norm L0, and obtaining the distribution ratio coefficient ψ accordingly. ψ is then compared with a preset criterion ψ... set The results are compared to determine the detection outcome.

[0013] Another objective of this invention is to provide a power distribution feeder fault detection system that considers the transient frequency response of distributed power sources, including a signal processing module, an energy entropy calculation module, an energy matrix construction module, and a fault identification module.

[0014] The signal processing module is used to acquire voltage traveling wave signals in the distribution network, perform wavelet transform processing on the acquired traveling wave signals, and obtain wavelet coefficients at multiple scales.

[0015] The energy entropy calculation module is used to calculate wavelet energy entropy based on energy distribution, and extract local energy features in the time and frequency domains under a set sliding window and sliding step size;

[0016] The energy matrix construction module is used to divide the traveling wave signal into P-segments in frequency and Q-segments in time, construct an energy matrix, calculate the L1 norm of the energy matrix to obtain the total energy E, and compare it with a set threshold E. s Compare the results to determine whether to initiate fault detection;

[0017] The fault identification module is used to normalize the energy matrix, filter out values ​​below a set threshold λ, calculate its non-zero norm L0, obtain the distribution ratio coefficient ψ, and then use ψ to determine the distribution ratio coefficient ψ based on a set criterion Ψ. set The comparison results determine the test results.

[0018] A computer device includes a memory and a processor, the memory storing a computer program, the processor executing the computer program to implement a method for detecting power distribution feeder faults that takes into account the transient frequency response of distributed power sources.

[0019] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of a power distribution feeder fault detection method that takes into account the transient frequency response of distributed power sources.

[0020] The beneficial effects of this invention are as follows: The power distribution feeder fault detection method considering the transient frequency response of distributed power sources provided by this invention is based on the acquisition and processing of traveling wave signals to obtain and identify fault characteristics for fault detection. It has the advantage of the traveling wave method being unaffected by the neutral point operating mode, and thus has strong applicability. By processing the traveling wave signal with wavelet energy entropy, an energy matrix is ​​established to reflect the distribution of signal energy in time and frequency. This reveals the difference in energy distribution between the fault traveling wave signal and the traveling wave signal under normal transient disturbance conditions. This enables the differentiation between fault and normal transient disturbance conditions, allowing for accurate fault identification. This invention achieves better results in terms of accuracy and applicability. Attached Figure Description

[0021] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 The first embodiment of the present invention provides an overall flowchart of a power distribution feeder fault detection method that considers the transient frequency response of distributed power sources. Detailed Implementation

[0023] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0024] Example 1, referring to Figure 1 As an embodiment of the present invention, a method for detecting distribution feeder faults considering the transient frequency response of distributed generation is provided, comprising:

[0025] S1: Acquire voltage traveling wave signals in the distribution network, perform wavelet transform processing on the acquired traveling wave signals, and obtain wavelet coefficients at multiple scales.

[0026] Furthermore, the acquisition of voltage traveling wave signals in the distribution network includes using a traveling wave detection device configured at the beginning of the line to acquire voltage traveling wave signals, with the sampling time set to 110μs and the sampling frequency set to 2kHz to 10kHz.

[0027] It should be noted that after detecting the traveling wave signal, wavelet transform is used to process it. In wavelet transform, the signal is decomposed into components of multiple scales or frequencies. These components are represented by different wavelet basis functions. For each level of wavelet transform, a set of wavelet coefficients is obtained, typically including low-frequency components (approximation coefficients) and high-frequency components (detail coefficients). These coefficients respectively contain the characteristics of the signal at different frequencies. Let the reconstruction coefficients of the traveling wave signal at decomposition scale j be y. j (k).

[0028] S2: Calculate wavelet energy entropy based on energy distribution, and extract local energy features in the time and frequency domains under the set sliding window and sliding step size.

[0029] Furthermore, we introduce the concept of wavelet energy entropy, which is derived from the combination of information entropy and wavelet transform. Let E be the energy of the sub-bands at different scales obtained after wavelet transform of the original traveling wave signal. j Then E1, E2, E3, ..., E n This involves dividing the signal energy at different scales; the sum of these energies represents the total signal energy E within that time window. Let K... j For partial energy E j The proportion of the total energy is used to calculate the wavelet energy entropy at different frequency scales. A sliding window is defined to slide across the signal to calculate the local features of the wavelet energy. The width of the sliding window is w, and the step size of the window sliding along the signal is δ, which is defined as the sliding factor δ. Each time the calculation is performed, the window will move by δ time points. With the help of the sliding time window, the wavelet energy entropy at different time scales can be calculated.

[0030] Decomposing the traveling wave signal into P frequency bands, and further decomposing the continuous wavelet coefficients in each frequency band into Q time intervals, yields several time-frequency blocks, thus allowing us to obtain the wavelet energy E in each time-frequency block. ab Thus, an energy matrix E that reflects the time-frequency distribution characteristics of energy can be constructed. PQ .

[0031] The total energy E of the energy matrix is ​​calculated using the m1 norm, taking into account signal interference in the distribution network. When E > E s When the fault detection is initiated, the fault detection process begins. If the fault is detected, the process continues; otherwise, data is collected again.

[0032] It should be noted that, firstly, the concept of wavelet energy entropy, derived from the combination of information entropy and wavelet transform, is introduced to quantitatively describe the characteristics of the traveling wave signal's variation in the time and frequency domains. Let E be the energy of the sub-bands at different scales obtained after the original traveling wave signal undergoes wavelet transform. j Then E1, E2, E3, ..., E n This involves dividing the signal energy at different scales. The sum of these divisions represents the total signal energy E within this time window, as shown in formula (1). Let K... j For partial energy E j The proportion of total energy, then K j The expression is shown below, thus yielding the wavelet energy entropy S. EE The definition is shown in formula (3).

[0033]

[0034] K j =E j / E (2)

[0035]

[0036] The larger the wavelet energy entropy value of a traveling wave signal, the richer and more uniformly distributed its frequency components in both the time and frequency domains. Conversely, a smaller value indicates a single frequency component and concentrated energy distribution within that time or frequency range.

[0037] Calculate the wavelet energy entropy of the voltage traveling wave signals generated by high-impedance grounding faults, capacitor switching, and load switching, respectively, and the wavelet energy entropy at different frequency scales as follows: Figure 1 As shown in the figure, the wavelet energy entropy is relatively large in the low-frequency band when the load switching and capacitor switching are switched, indicating that the energy distribution is relatively uniform. However, in the higher frequency band, the wavelet energy entropy is close to 0, indicating that the high-frequency component is simple and concentrated. The wavelet energy entropy of the fault does not change drastically in either the low-frequency or high-frequency band, indicating that the fault traveling wave signal has the characteristic of uniform energy distribution.

[0038] Based on the wavelet energy entropy calculation formula, a time window is introduced to obtain the change of wavelet energy entropy within that time window. In wavelet transform, the signal is decomposed into components of multiple scales or frequencies, which are represented by different wavelet basis functions. For each level of wavelet transform, a set of wavelet coefficients is obtained, which usually includes low-frequency components (approximation coefficients) and high-frequency components (detail coefficients). These coefficients respectively contain the characteristics of the signal at different frequencies.

[0039] Let the reconstruction coefficient of the traveling wave signal at decomposition scale j be y. j(k), where k represents the k-th sampling point, K = 1, 2, 3, ..., L, and L is the discrete time point of the entire signal. A sliding window is defined to slide across the signal to calculate the local features of wavelet energy. The width of the sliding window is w, and the step size of the window sliding along the signal is δ, defined as the sliding factor δ. Each time the calculation is performed, the window will move δ time points. Accordingly, the formula for calculating the wavelet energy entropy in time is shown in formula (4).

[0040]

[0041] In the formula, l = 1, 2, ..., N, N is the total number of window slides, and N = (Lw) / δ.

[0042] The time window sliding step was set to 10 μs, and a total of 11 steps were taken to obtain the change in wavelet energy entropy of the traveling wave signal under fault and transient disturbance in the time domain. In the initial stage of fault and normal transient disturbance conditions, the wavelet energy entropy increases, indicating an increase in the frequency components of the traveling wave signal and a more uniform energy distribution. As time progresses, the wavelet energy entropy begins to decrease, indicating a decrease in the frequency components and a more concentrated energy distribution. Combined with... Figure 1 The frequency domain distribution shows that the low-frequency components of normal transient disturbances decay slowly, while the high-frequency components decay quickly, and the wavelet energy entropy decays to 0 after 80μs. However, the wavelet energy entropy of faults decreases slowly over time and does not decay to 0. From the above analysis, we can conclude that: the low-frequency component content is high in normal transient disturbances, and the energy is concentrated in the low frequency. In the time domain, the amplitude of the energy entropy shows a decaying trend and eventually becomes 0. The high-frequency component is abundant in the fault traveling wave waveform, and the energy shows a uniform distribution in the time domain.

[0043] After analyzing the traveling wave waveforms of normal transient disturbances and faults in three dimensions of time, frequency, and amplitude, the differences between the two in the time and frequency domains can be obtained. Thus, the difference between the fault traveling wave signal and the normal transient disturbance signal can be intuitively reflected by wavelet energy entropy. Therefore, from the perspective of energy distribution difference, wavelet energy entropy can be used to construct a criterion for fault detection. The above analysis shows that the energy amplitude distribution of the fault traveling wave waveform in the time and frequency domains is significantly different from that of the normal transient disturbance condition.

[0044] S3: Divide the traveling wave signal into P-segments in frequency and Q-segments in time, construct an energy matrix, calculate the L1 norm of the energy matrix to obtain the total energy E, and compare it with a set threshold E. s A comparison is made to determine whether fault detection should be initiated.

[0045] Furthermore, the traveling wave signal is decomposed into P frequency bands, and the continuous wavelet coefficients in each frequency band are further decomposed into Q time-frequency blocks. This yields several time-frequency blocks, from which the wavelet energy E in each time-frequency block can be obtained. ab As shown in formula (5).

[0046]

[0047] Where 'a' refers to the 'a'-th frequency band, 'b' refers to the 'b'-th time period, and 'y'... a (k) refers to the wavelet coefficient at a certain sampling point k in the a frequency band.

[0048] Therefore, an energy matrix E that reflects the time-frequency distribution characteristics of energy can be constructed. PQ The distribution characteristics of energy in the time and frequency domains are quantitatively described by a matrix, as shown in formula (6).

[0049]

[0050] Under normal operating conditions of the distribution network, excluding transient disturbances such as load switching and capacitor switching, the traveling wave detection device cannot detect the traveling wave signal, thus the energy matrix E PQ The m1 norm should be 0, that is, the energy of each time-frequency block is 0 and their sum of energy is also 0. However, there are harmonics in the power system. These harmonic signals will be detected by the detection device, so that the m1 norm of the energy matrix is ​​not 0. The total energy E is calculated according to the m1 norm of the matrix as shown in formula (7). The traveling wave detection device in the distribution network has the ability to filter out harmonics. The total energy calculated will not be strictly equal to 0, but will fluctuate at a small value.

[0051]

[0052] This invention proposes a fault detection activation threshold, denoted as Es, to eliminate harmonic interference. When the total energy E is less than Es, the fault detection will not malfunction due to various noises in the power distribution network. When a ground fault or normal transient disturbance occurs, and the total energy E of the traveling wave signal is greater than Es, the fault detection is activated, thereby distinguishing between disturbances and faults and realizing reliable fault detection. Based on a large amount of simulation experimental data, Es is set to 25, and the fault activation criterion is shown in formula (8).

[0053] E>E s (8)

[0054] To eliminate the influence of dimensions on data analysis and thus make data processing more stable and efficient, matrix normalization is used to process the energy matrix, normalizing the value of each element in the energy matrix to the range [0, 1]. PQThe normalization process of each element is shown in formula (9).

[0055]

[0056] Among them, E max and E min These are the largest and smallest elements in the energy matrix. By using the normalization process of the matrix, we can obtain the characteristic matrix that reflects the difference between faults and transient disturbances, as shown in formula (10).

[0057]

[0058] Analysis of the energy distribution of faults and transient disturbances reveals that the energy distribution of transient disturbances is concentrated, while that of faults is uniform. For the energy matrix of a transient disturbance traveling wave, most elements will be close to 0. However, due to the uniform energy distribution, the energy matrix of a fault traveling wave has fewer elements close to 0. Therefore, calculating the total energy and using it as a detection criterion is insufficient. This is because it cannot be assumed that the sum of all non-zero elements in the energy matrix under disturbance conditions is necessarily less than that of a fault. Nevertheless, the difference in energy distribution characteristics between the two is well-established.

[0059] S4: Normalize the energy matrix, filter out values ​​below a set threshold λ, calculate its non-zero norm L0, obtain the distribution ratio coefficient ψ, and then apply ψ to the set criterion Ψ. set The comparison results determine the test results.

[0060] Furthermore, firstly, the elements in the normalized energy matrix are identified as valid. The non-zero norm L0 of the matrix is ​​introduced and extended. A threshold λ is set to filter out elements with very small values ​​but not zero. Only elements with values ​​greater than λ are considered as valid data. The expression for the non-zero norm of the energy matrix is ​​shown in formula (11).

[0061]

[0062] The ratio ψ is obtained by comparing L0 with the number of matrix elements Z. It is called the distribution coefficient, which quantitatively represents the distribution characteristics of energy. The expression of the distribution coefficient is shown in formula (12). Let ψset be used as the criterion for fault detection. When ψ>ψset, it is considered a fault, and otherwise it is a disturbance. Based on the distribution of energy amplitude in the time domain and frequency domain in the previous text, as well as a large amount of data analysis, the value is taken as 0.61.

[0063] The fault detection criterion proposed in this invention is shown in formula (13).

[0064]

[0065] ψ>ψset (13)

[0066] Example 2 is the second embodiment of the present invention, which differs from the previous embodiment in that:

[0067] If a function is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0068] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced 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 be any means that can contain, store, communicate, propagate, or transmit programs for use by, or in conjunction with, an instruction execution system, apparatus, or device.

[0069] More specific examples (a non-exhaustive list) of computer-readable media 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 programs can be printed, because programs can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpreting, or otherwise processing as necessary, and then stored in computer memory.

[0070] 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.

[0071] Example 3 is the third embodiment of the present invention. This embodiment provides a system for a power distribution feeder fault detection method that considers the transient frequency response of distributed power sources, including a signal processing module, an energy entropy calculation module, an energy matrix construction module, and a fault identification module.

[0072] The signal processing module is used to acquire voltage traveling wave signals in the power distribution network, perform wavelet transform processing on the acquired traveling wave signals, and obtain wavelet coefficients at multiple scales.

[0073] The energy entropy calculation module is used to calculate wavelet energy entropy based on energy distribution, and extract local energy features in the time and frequency domains under the set sliding window and sliding step size.

[0074] The energy matrix construction module is used to divide the traveling wave signal into P-segments in frequency and Q-segments in time, construct the energy matrix, calculate the L1 norm of the energy matrix to obtain the total energy E, and compare it with a set threshold E. s A comparison is made to determine whether fault detection should be initiated.

[0075] The fault identification module is used to normalize the energy matrix, filter out values ​​below a set threshold λ, calculate its non-zero norm L0, obtain the distribution ratio coefficient ψ, and then use ψ to determine the distribution ratio coefficient ψ based on the set criterion Ψ. set The comparison results determine the test results.

[0076] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A power distribution feeder fault detection method that considers the transient frequency response of distributed generation, characterized by, The method comprises: Collecting voltage traveling wave signals in a power distribution network, performing wavelet transform on the collected traveling wave signals to obtain wavelet coefficients at multiple scales; Calculating wavelet energy entropy based on energy distribution, and extracting local energy features in time domain and frequency domain under a set sliding window and sliding step; Dividing the traveling wave signals into P segments in frequency and Q segments in time, constructing an energy matrix, calculating the L1 norm of the energy matrix to obtain total energy, and comparing the total energy with a set threshold to determine whether to start fault detection; The energy matrix is normalized, values below a set threshold λ are filtered out, the non-zero norm L0 is calculated, a distribution proportion coefficient ψ is obtained, and a detection result is determined according to a comparison result of ψ and a set criterion Ψ set .

2. The method for distribution feeder fault detection considering transient frequency response of distributed generation as claimed in claim 1, wherein: The collecting of voltage traveling wave signals in a power distribution network comprises using a traveling wave detection device configured at the head of a line to collect voltage traveling wave signals, with a sampling time set to 110 μs and a sampling frequency set to 2 kHz to 10 kHz.

3. The method for distribution feeder fault detection considering transient frequency response of distributed generation of claim 2, wherein: The process of performing wavelet transform on the acquired traveling wave signal to obtain wavelet coefficients at multiple scales includes decomposing the signal into wavelet coefficients at multiple scales, obtaining approximation coefficients and detail coefficients, and calculating the wavelet energy E at each scale. j Then calculate the energy ratio K. j Based on K j The wavelet energy entropy at each scale is obtained.

4. The method for distribution feeder fault detection considering transient frequency response of distributed generation of claim 3, wherein: The extracting of local energy features in time domain and frequency domain comprises defining a sliding window width w and a sliding step δ, calculating a wavelet energy entropy sequence by sliding the window in the signal, and forming local energy features in time scale.

5. The method for distribution feeder fault detection considering transient frequency response of distributed generation as claimed in claim 4 wherein: The configuration energy matrix comprises dividing the wavelet-transformed signal into P frequency bands, each of which is further divided into Q time periods to obtain a plurality of time-frequency block energies E aβ , which are combined to form a two-dimensional energy matrix E p Q.

6. The method for distribution feeder fault detection considering transient frequency response of distributed generation as claimed in claim 5 wherein: the set threshold value E s comparing to determine whether to initiate fault detection includes calculating the LI norm of the energy matrix to obtain total energy E, and comparing E to a set threshold value E s If E is greater than E s , then fault detection is initiated; otherwise, return to reacquire the signal.

7. The method for distribution feeder fault detection considering transient frequency response of distributed generation as claimed in claim 6 wherein: The normalization of the energy matrix includes normalizing the matrix elements to the interval [0, 1] according to the maximum value and the minimum value, calculating the non-zero norm L0 after screening out elements less than a threshold value, and obtaining a distribution proportion coefficient ψ according to this, comparing ψ with a preset criterion ψ set to determine the detection result.

8. A system employing the power distribution feeder fault detection method considering transient frequency response of distributed generation according to any one of claims 1 to 7, characterized in that: The method comprises a signal processing module, an energy entropy calculation module, an energy matrix construction module, and a fault identification module; The signal processing module is used to collect voltage traveling wave signals in a power distribution network, perform wavelet transform on the collected traveling wave signals to obtain wavelet coefficients at multiple scales; The energy entropy calculation module is used to calculate wavelet energy entropy based on energy distribution, and extract local energy features in time domain and frequency domain under a set sliding window and sliding step; The energy matrix construction module is used for dividing the traveling wave signal into P segments in frequency and Q segments in time, constructing an energy matrix, calculating L1 norm of the energy matrix to obtain total energy E, and comparing the total energy E with a set threshold value E s to determine whether to start fault detection; The fault identification module is configured to normalize the energy matrix, filter out values lower than a set threshold λ, calculate a non-zero norm L0, obtain a distribution proportion coefficient ψ, and determine a detection result according to a comparison result of ψ and a set criterion Ψ set . 9.A computer device, comprising a memory and a processor, wherein the memory stores a computer program, and the computer device is configured to perform the method according to any one of claims 1-8 when the computer program is executed by the processor. The processor executes the computer program to realize the steps of the power distribution feeder fault detection method considering transient frequency response of distributed power sources according to any one of claims 1 to 7.

10. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to realize the steps of the power distribution feeder fault detection method considering transient frequency response of distributed power sources according to any one of claims 1 to 7.

Images