Distribution network fault positioning method, system and device based on variational mode decomposition, and medium

Abstract

The invention discloses a distribution network fault positioning method, system and device based on variational mode decomposition and a medium, and relates to the technical field of power system fault positioning, and the method comprises the steps: collecting a traveling wave signal of a power distribution network, and carrying out the preprocessing of the collected traveling wave signal; performing modal decomposition on the preprocessed traveling wave signal, and outputting a plurality of modal components; based on the mode component obtained through decomposition, wave head arrival time is extracted; performing correction compensation based on the wave head arrival time difference, the phase difference and the propagation characteristics, and calculating the relative distance between the fault point and the known node; calculating the position coordinates of the fault point in the distribution network according to the geographic coordinates and the direction vectors of the nodes in combination with the relative distance; and constructing a visual interface, displaying the position of a fault point, and storing the traveling wave signal data and a fault positioning result into a database. According to the method, the adaptive correction function and the time difference compensation model are constructed, so that time difference errors caused by factors such as frequency change and phase nonlinearity can be corrected, and the accuracy of fault positioning is improved.

Description

Technical Field

[0001] This invention relates to the field of power system fault location technology, and in particular to a method, system, equipment and medium for distribution network fault location based on variational mode decomposition. Background Technology

[0002] As a crucial component of the power system, the distribution network's primary task is to transmit electrical energy from substations to end users. With increasing societal demand for electricity and the continuous integration of new energy sources, the operating environment and fault characteristics of the distribution network are becoming increasingly complex. This complexity demands higher fault location accuracy and faster response speeds to ensure the reliability of power supply. Traditional methods are often limited by factors such as sensor accuracy, noise interference from traveling wave signals, and variations in traveling wave propagation speed, making it difficult to guarantee fault location accuracy. Due to the large number of nodes and complex structure of the distribution network, existing fault location algorithms struggle to achieve efficient real-time calculations under multi-node conditions, impacting fault handling efficiency and the overall stability of the power grid.

[0003] Existing fault location technologies have certain limitations in practical applications. In the mode decomposition and feature extraction stages of traveling wave signals, it is difficult to adaptively adjust the signal characteristics. This results in unstable location accuracy under different fault types and complex working conditions. Due to the time asynchrony of sensors between different nodes, traditional time difference measurement is prone to deviation, which directly affects the accuracy of fault location. Summary of the Invention

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

[0005] Therefore, the problem to be solved by this invention is that it is difficult to adaptively adjust the signal characteristics in the mode decomposition and feature extraction stages of traveling wave signals. This leads to unstable positioning accuracy under different fault types and complex working conditions. Due to the asynchronous time of sensors between different nodes, traditional time difference measurement is prone to deviation, which directly affects the accuracy of fault location.

[0006] To address the aforementioned technical problems, this invention provides the following technical solution: a distribution network fault location method based on variational mode decomposition, comprising: acquiring traveling wave signals from the distribution network; obtaining frequency information using a frequency analyzer; preprocessing the acquired traveling wave signals; performing mode decomposition on the preprocessed traveling wave signals to output several mode components; extracting instantaneous phases and calculating instantaneous frequencies based on the decomposed mode components; analyzing the frequency variance through a sliding window to extract the wavefront arrival time; performing correction and compensation based on the wavefront arrival time difference, phase difference, and propagation characteristics; calculating the relative distance between the fault point and known nodes; calculating the location coordinates of the fault point in the distribution network based on the node's geographical coordinates and direction vector, combined with the relative distance; constructing a visualization interface to display the fault point location; storing the traveling wave signal data and fault location results in a database and archiving the data.

[0007] As a preferred embodiment of the distribution network fault location method based on variational mode decomposition described in this invention, the following steps are included: acquiring the traveling wave signal of the distribution network includes deploying current and voltage sensors at preset nodes of the distribution network to acquire the traveling wave signal; the preprocessing of the acquired traveling wave signal includes time synchronization based on a satellite synchronization system, removing noise interference using wavelet denoising technology, and standardizing the denoised signal.

[0008] As a preferred embodiment of the distribution network fault location method based on variational mode decomposition described in this invention, the mode decomposition of the preprocessed traveling wave signal includes adjusting the frequency and bandwidth of the modes gradually during the iteration process using the variational mode decomposition method until the set convergence criteria are met, and then outputting several mode components.

[0009] The beneficial effects of this preferred technical solution are as follows: by iteratively optimizing the frequency and bandwidth of the modal components and adaptively adjusting the center frequency, the extracted modes better match the intrinsic frequency characteristics of the original signal, thereby improving the accuracy of fault feature extraction.

[0010] As a preferred embodiment of the distribution network fault location method based on variational mode decomposition described in this invention, the extraction of wavefront arrival time includes: based on the obtained modal components, selecting frequency components with fault characteristics, performing Hilbert transform and constructing a complex signal, extracting its instantaneous phase and calculating the instantaneous frequency, analyzing the frequency variance through a sliding window, and extracting the wavefront arrival time.

[0011] The beneficial effects of this preferred technical solution are: the combination of Hilbert transform and complex signal reconstruction technology can process non-stationary traveling wave signals, effectively extract their instantaneous frequency information, and adapt to various complex power disturbance modes.

[0012] As a preferred embodiment of the distribution network fault location method based on variational mode decomposition described in this invention, the phase difference is calculated by: selecting a reference node A as the starting point for location and a node B as an auxiliary comparison; performing Fourier transform on the traveling wave signals of nodes A and B to obtain a complex signal in the frequency domain; calculating the magnitude of the complex signal in the frequency domain to obtain the amplitude spectrum; selecting the frequency component with the largest amplitude spectrum as the main frequency component; calculating the phase value of the main frequency component to obtain the phase of nodes A and B; and subtracting the phase of node B from the phase of node A to obtain the node phase difference.

[0013] The advantages of this preferred technical solution are: it adopts a nonlinear joint modeling method of phase difference and frequency, which avoids the failure of traditional methods under conditions such as multipath propagation and phase distortion, and is suitable for distribution networks with complex structures.

[0014] As a preferred embodiment of the distribution network fault location method based on variational mode decomposition described in this invention, the construction of the visualization interface includes: using a front-end framework to construct a visualization interface, including a distribution network topology map and a fault location display area; using a geographic information map to load the distribution network topology map, marking the fault location in the topology map, and displaying the node number; and allowing users who have passed real-name verification to view it.

[0015] As a preferred embodiment of the distribution network fault location method based on variational mode decomposition described in this invention, the step of storing traveling wave signal data and fault location results in the database includes storing the collected traveling wave signals and the distribution network fault locations generated by analysis in the database, setting up security access measures, backing up the stored data in the cloud, and performing integrity checks on the stored data and backup data according to a preset period. After the checks are completed, integrity check records are generated and synchronously stored in the database.

[0016] To address the aforementioned technical problems, this invention provides the following technical solution: a distribution network fault location system based on variational mode decomposition, comprising: a data acquisition and processing module, a mode decomposition module, a wavefront time extraction module, a fault location calculation module, and a visualization and storage module; the data acquisition and processing module acquires traveling wave signals from the distribution network, obtains frequency information using a frequency analyzer, and preprocesses the acquired traveling wave signals; the mode decomposition module performs mode decomposition on the preprocessed traveling wave signals, outputting several mode components; the wavefront time extraction module extracts the instantaneous phase and calculates the instantaneous frequency based on the decomposed mode components, analyzes the frequency variance through a sliding window, and extracts the wavefront arrival time; the fault location calculation module performs correction and compensation based on the wavefront arrival time difference, phase difference, and propagation characteristics, calculates the relative distance between the fault point and known nodes, and calculates the location coordinates of the fault point in the distribution network based on the node's geographical coordinates and direction vector, combined with the relative distance; the visualization and storage module constructs a visualization interface to display the fault location, and stores the traveling wave signal data and fault location results in a database for archiving.

[0017] A computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of a distribution network fault location method based on variational mode decomposition as described above.

[0018] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of a distribution network fault location method based on variational mode decomposition as described above.

[0019] The beneficial effects of this invention are as follows: This invention collects and preprocesses traveling wave signals, decomposes the preprocessed traveling wave signals using variational modes, calculates the wavefront arrival time based on the decomposition results, and calculates the location of the distribution network fault based on the wavefront arrival time difference; it improves the accuracy and anti-interference capability of fault location in complex environments and eliminates the influence of noise and error accumulation. Attached Figure Description

[0020] 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.

[0021] Figure 1 This is a flowchart of a distribution network fault location method based on variational mode decomposition in Example 1. Detailed Implementation

[0022] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0023] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0024] Example 1, referring to Figure 1 This is the first embodiment of the present invention, which provides a distribution network fault location method based on variational mode decomposition, including, as follows: Figure 1 As shown:

[0025] S1: Collect traveling wave signals from the power distribution network, obtain frequency information using a frequency analyzer, and preprocess the collected traveling wave signals.

[0026] S2: Perform mode decomposition on the preprocessed traveling wave signal and output several mode components.

[0027] S3: Based on the modal components obtained from the decomposition, extract the instantaneous phase and calculate the instantaneous frequency. Analyze the frequency variance through a sliding window and extract the wavefront arrival time.

[0028] S4: Based on the wavefront arrival time difference, phase difference, and propagation characteristics, correction and compensation are performed to calculate the relative distance between the fault point and the known node.

[0029] S5: Calculate the location coordinates of the fault point in the distribution network based on the node's geographical coordinates and direction vector, combined with the relative distance.

[0030] S6: Build a visual interface to display the location of the fault point, store the traveling wave signal data and fault location results in the database, and record and archive them.

[0031] It should be noted that existing fault location technologies have certain limitations in practical applications. In the mode decomposition and feature extraction stages of traveling wave signals, it is difficult to adaptively adjust the signal characteristics. This results in unstable location accuracy under different fault types and complex working conditions. Due to the asynchronous time of sensors between different nodes, traditional time difference measurement is prone to deviation, which directly affects the accuracy of fault location.

[0032] Therefore, to address the aforementioned problems, through steps S1-S6, traveling wave signals are collected and preprocessed at the distribution network nodes. Fault characteristic modes are extracted using variational mode decomposition, and wavefront arrival time is extracted by combining instantaneous frequency change detection. The fault point coordinates are calculated through phase difference correction and time difference compensation. Finally, the location results are displayed in a visualization interface. At the same time, secure data storage and cloud backup are achieved, improving the accuracy and response efficiency of fault location in complex environments.

[0033] Example 2 is the second embodiment of the present invention, which differs from the first embodiment in that: a distribution network fault location method based on variational mode decomposition further includes, in the embodiment of this application, step S1, collecting the traveling wave signal of the distribution network, obtaining frequency information using a frequency analyzer, and preprocessing the collected traveling wave signal.

[0034] Specifically, intelligent current and voltage sensors are installed at preset nodes in the distribution network to collect traveling wave signals, and a frequency analyzer is used to collect the frequency of the traveling wave signals and perform preprocessing.

[0035] The preprocessing of the acquired traveling wave signal includes time synchronization based on a satellite synchronization system, removal of noise interference from the acquired signal, and standardization of the denoised signal.

[0036] In an optional embodiment, wavelet denoising techniques can be used to remove noise interference, specifically including:

[0037] First, select a suitable wavelet basis based on the characteristics of the traveling wave signal. The selection criteria include the compactness, orthogonality, and smoothness of the wavelet function. Generally, when there are obvious signal abrupt changes, it is better to select a wavelet function with better time-domain localization ability.

[0038] Based on the sampling frequency and signal length, select an appropriate number of wavelet decomposition layers. The higher the number of decomposition layers, the more low-frequency information can be captured, but it may also lead to signal delay or resolution reduction. Usually, it is between 3 and 6 layers, and the adjustment is made according to experience and the frequency band characteristics of the fault signal.

[0039] The original traveling wave signal is decomposed into multiple wavelet coefficients (representing the low-frequency part) and detail coefficients (representing the high-frequency part) in several layers. The detail coefficients usually contain most of the noise information, while the approximation coefficients retain the main trend of the signal.

[0040] Thresholding is used to denoise detail coefficients. Common methods include "hard thresholding" and "soft thresholding":

[0041] Hard thresholding: Coefficients whose absolute value is less than the threshold are set to zero, while other coefficients are retained.

[0042] Soft thresholding: Not only are coefficients below the threshold set to zero, but coefficients above the threshold are also shrunk by the threshold to reduce abrupt changes.

[0043] The threshold can be selected based on statistical methods, such as automatically setting it using the noise standard deviation and signal length.

[0044] Reconstructing the signal: The wavelet coefficients of each layer after processing are inversely transformed and recombined into a denoised traveling wave signal. This step retains the main information in the original signal while significantly reducing high-frequency noise components.

[0045] Standardization and Verification: The denoised signal is standardized to make it suitable for subsequent operations such as mode decomposition and frequency analysis. Furthermore, the denoising effect can be verified by metrics such as signal-to-noise ratio (SNR) improvement and waveform smoothness.

[0046] In another optional embodiment, singular value decomposition (SVD) can also be used to remove noise interference, specifically including:

[0047] Constructing the phase space matrix: The original one-dimensional traveling wave signal is transformed into a two-dimensional matrix form through sliding recombination, that is, constructing an "extension matrix" or "delay embedding matrix". This process unfolds the local features of the signal into a matrix form, so that the principal component structure of the signal can be reflected.

[0048] Perform singular value decomposition on the matrix: decompose the matrix constructed above into several sets of singular values ​​and their corresponding singular vectors, representing the main structure and energy distribution of the signal. Larger singular values ​​correspond to the main structure in the signal, while smaller singular values ​​usually correspond to noise components.

[0049] Set a singular value screening strategy: Based on the energy proportion, singular value change trend or experience, select and retain the top few main singular values, and discard the parts corresponding to the remaining small singular values. For example, the top few singular values ​​with a cumulative energy proportion of 95% can be retained, and the rest are considered as noise contributions.

[0050] Reconstructing the noise reduction matrix: Using the selected singular values ​​and corresponding vectors, the signal matrix is ​​inversely transformed to reconstruct the two-dimensional matrix after noise removal.

[0051] Restoring to a one-dimensional signal: The denoised two-dimensional matrix is ​​restored to a one-dimensional traveling wave signal, thus restoring its original time series structure.

[0052] Post-processing and standardization: The amplitude of the denoised one-dimensional signal is standardized to improve the consistency of the signal comparison at different nodes or at different times, in preparation for subsequent processes such as mode decomposition and wavefront extraction.

[0053] It should be noted that the use of intelligent sensors improves the real-time performance and accuracy of data acquisition, and enables high-precision acquisition of traveling wave signals generated by different fault types. This provides a reliable data foundation for subsequent traveling wave analysis and localization. Using a frequency analyzer to deconstruct and analyze the frequency components of the traveling wave signal allows for better capture of the characteristic frequencies of the fault signal, thereby improving the accuracy of fault feature extraction. When a fault occurs, the frequency characteristics of the traveling wave signal are often quite significant; therefore, frequency analysis provides a basis for subsequent fault frequency discrimination. Through frequency decomposition, not only can the core frequency band of the fault signal be quickly identified, but noise signals can also be distinguished from fault signals, improving the reliability of the positioning system. Satellite synchronization systems can significantly reduce clock... The impact of drift on positioning accuracy can also achieve global synchronization between different nodes, providing reliable timing guarantees for the distribution network fault location system. By decomposing the traveling wave signal and selectively suppressing noise components, noise interference can be greatly reduced and the quality of the traveling wave signal can be improved, thereby enhancing the accuracy of subsequent fault feature extraction. Standardized processing improves the comparability of data, enabling traveling wave signals collected at different time points or different nodes to be analyzed under the same benchmark, thus ensuring the stability and accuracy of fault location. The preprocessing of the traveling wave signal ensures the efficiency of subsequent feature extraction and location, especially in handling multi-point faults or instantaneous faults in complex networks, enabling rapid identification of fault points and improving the stability and power supply reliability of the distribution network.

[0054] In the embodiments of this application, step S2 performs mode decomposition on the preprocessed traveling wave signal and outputs several mode components.

[0055] Specifically, variational mode decomposition (VMD) is performed on the preprocessed traveling wave signal to extract modal components of different frequency bands. The goal of VMD is to minimize the bandwidth of all modes, which is defined as the sum of the squares of the frequency change rates of the modal components, including...

[0056] The mode number K is set according to the empirical mode decomposition criterion, and the initial frequency ω is set. k It is 0.

[0057] Construct the bandwidth minimization objective function, the formula is:

[0058]

[0059] Where K is the total number of modal components, u k (t) represents the k-th modal component (IMF) at time t, indicating the portion of the signal in a specific frequency band, obtained through VMD decomposition, ω k Let be the center frequency of the k-th mode, and represent the main frequency components of each modal component. These are determined through initialization and optimization processes. Let be the differential operator at time t, representing the rate of change of the signal, used to calculate the bandwidth of each mode, j is the imaginary unit, t is time, and x(t) is the preprocessed traveling wave signal at time t.

[0060] By combining the bandwidth minimization objective function with signal reconstruction constraints, and introducing the Lagrange multiplier λ to achieve the signal reconstruction constraint, the decomposed modes can maintain the overall structure of the signal. The augmented Lagrange function L is constructed as follows:

[0061]

[0062] The Alternating Direction Multiplier Method (ADMM) iteratively updates the modal components and adjusts the Lagrange multipliers, as shown in the following formula:

[0063]

[0064] in, Let k be the component of the k-th mode in the (n+1)th iteration. Let λ be the component of the p-th mode at time t in the nth iteration (where p ≠ k). n+1 ρ is the Lagrange multiplier in the (n+1)th iteration, ρ is the adjustment coefficient used to control the adjustment rate, and n is the number of iterations.

[0065] The center frequency is adaptively adjusted based on the time characteristics of the current modal component. The formula for updating the center frequency is:

[0066]

[0067] in, Let be the center frequency of the k-th mode in the (n+1)th iteration. Let be the component of the k-th mode at time t in the nth iteration, and +∞ and -∞ be the upper and lower limits of the time integral, respectively.

[0068] The modal bandwidth energy estimation method is used to design a bandwidth convergence criterion formula, and the modal bandwidth B is calculated. k The formula is:

[0069]

[0070] Among them, B k Let be the bandwidth of the k-th mode, representing the degree of concentration of this mode in the frequency domain. Let u be the modal component at time t. k The first derivative of (t) reflects the rate of change of the signal and is used to measure the frequency change of the mode.

[0071] The bandwidth change is obtained by subtracting the absolute value of the nth iteration from the bandwidth at the (n+1)th iteration.

[0072] The modal bandwidth energy estimation method is used to set a convergence threshold. The bandwidth change is compared with the convergence threshold. When the bandwidth change is less than the convergence threshold, the iteration stops and the modal components, modal center frequencies, and modal bandwidths are output.

[0073] It should be noted that the decomposed modal components better match the true characteristics of traveling wave signals, helping to extract frequency components with fault characteristics and providing more reliable data support for fault location. By combining the bandwidth minimization objective function with signal reconstruction constraints and optimizing the modal bandwidth through the augmented Lagrangian function L, the decomposition accuracy of VMD is significantly improved. The introduction of the augmented Lagrangian function ensures the accuracy of signal reconstruction on the basis of bandwidth optimization, reduces the deviation in signal decomposition, and enhances the feature extraction effect of traveling wave signals. The optimization scheme is particularly suitable for complex and non-stationary fault signal processing, helping to improve the stability and accuracy of the fault location system. ADMM not only has advantages in convergence speed but also better adapts to the nonlinearity and multi-frequency bands of signals. The features greatly improve the decomposition efficiency and effectiveness of VMD. Adaptive adjustment can dynamically optimize the center frequency of modes, ensuring that the frequency characteristics of each mode match the actual frequency components of the signal source. The adaptive feature improves the decomposition accuracy and can better extract fault features from traveling wave signals, which helps to improve the accuracy of fault location. The introduction of the convergence judgment method greatly reduces redundant calculation time, making the signal decomposition process more efficient, especially suitable for real-time fault location scenarios. The dynamic monitoring method based on bandwidth changes effectively improves the convergence accuracy of VMD, can ensure the stability of decomposition results, and helps to accurately extract fault signals. The output results lay a solid foundation for the judgment of fault type and accurate location, improving the response speed and accuracy of fault diagnosis.

[0074] In this embodiment of the application, in step S3, based on the modal components obtained by decomposition, the instantaneous phase is extracted and the instantaneous frequency is calculated. The frequency variance is analyzed by sliding window, and the wavefront arrival time is extracted.

[0075] Based on the obtained modal components, frequency components with fault characteristics are screened out, Hilbert transform is performed to construct complex signals, instantaneous phase is extracted and instantaneous frequency is calculated, frequency variance is analyzed through sliding window, and wavefront arrival time is extracted.

[0076] Specifically, based on the obtained modal components, they are screened, and modal components that conform to the frequency range of fault characteristics are retained.

[0077] In an optional embodiment, the results of variational mode decomposition can be filtered using a frequency bandwidth filtering method, specifically including:

[0078] Obtain all modal components and their center frequency and bandwidth information: After the signal has undergone variational mode decomposition (VMD), multiple modal components are output, each corresponding to a center frequency and bandwidth. This information can usually be obtained directly from the VMD output, reflecting the frequency domain distribution characteristics of each component.

[0079] Set the fault characteristic frequency range: Based on the historical operating experience of the actual distribution network or the known typical fault characteristics, set a target frequency range, such as tens of hertz to thousands of hertz.

[0080] This frequency range typically includes the traveling wave response bands caused by common short circuits, grounding, electric arcs, or switching operations.

[0081] Determine if the modal center frequency falls within the target range: Iterate through the center frequencies of each modal component in turn and determine whether they are within the set characteristic frequency range.

[0082] If a mode falls within this range, it is considered to contain modal components that may contain valid fault characteristics.

[0083] Assess modal bandwidth stability: For each candidate mode, further analyze whether its frequency bandwidth is too wide or deviates from the concentration. If the modal bandwidth is too wide, it may indicate that it contains other non-faulty components. If the bandwidth is too narrow, it may indicate that the mode is too localized and the signal energy is insufficient. Generally, retain the modal components with medium bandwidth and high frequency concentration.

[0084] Calculate modal energy or variance: Perform further energy estimation on modal components that fall within the target frequency range (e.g., calculate the mean square value or energy density of their waveforms), and remove modal components with extremely low energy and insufficient information content.

[0085] Retain valid modes and discard irrelevant modes: The modal components that pass the above screening are marked as "valid modes" and used for subsequent Hilbert transform, instantaneous frequency calculation and wavefront time extraction steps; the modal components that do not pass the screening are directly discarded and do not participate in subsequent analysis.

[0086] In another optional embodiment, the results of variational mode decomposition can also be screened using the energy entropy screening method, specifically including:

[0087] Extracting the signals of each modal component: For multiple modal components after variational mode decomposition (VMD), their original time-domain waveforms are retained for subsequent energy characteristic analysis.

[0088] Divide the modal signal into multiple time windows: set a fixed-length time sliding window for each modal component (e.g., one window for every 100 sampling points) to observe its local energy changes.

[0089] Calculate the energy value for each window: sum the squares of the signal amplitudes of the modal components in each window to obtain the energy estimate for that window. This gives the "local energy distribution" of the modal component over the entire time domain.

[0090] Normalize local energy and calculate energy entropy: Normalize the energy values ​​of each time window (i.e. convert them into energy proportions), and then calculate the energy entropy of the modal component according to the information entropy formula. The larger the energy entropy, the more uniform the energy distribution and the more "chaotic" the signal; the smaller the entropy value, the more concentrated the energy in the signal, the clearer the structure, and the more likely it is a fault feature.

[0091] Setting an energy entropy threshold for screening: Based on experience, an upper limit threshold for energy entropy is set. Usually, modal components with lower energy entropy (i.e., signals with concentrated energy and significant structure) are retained, while modal components with excessively high energy entropy are considered to be mainly composed of noise and should be removed.

[0092] Output filtering results: The modal components that are determined to be valid by energy entropy are used as inputs for subsequent processing, such as wavefront extraction, phase calculation, and fault location.

[0093] For the filtered modal components u k '(t) performs H(u) transform k The imaginary part of the signal is obtained by '(t), and the formula is:

[0094]

[0095] Wherein, H(u) k '(t)) represents the Hilbert transform result of the filtered modal components, ∞ and -∞ are the upper and lower limits of the principal value integral, respectively, and τ is the integration variable.

[0096] The selected modal components are combined with the Hilbert transform result to form a complex signal z. k (t), the formula is:

[0097] z k (t)=u k '(t)+jH(u k '(t)),

[0098] Through complex signal z k The instantaneous phase θ(t) is extracted using the following formula:

[0099] θ(t)=arg(z k (t)),

[0100] Taking the higher-order derivative of the instantaneous phase θ(t), we obtain the instantaneous frequency f. inst (t), the formula is:

[0101]

[0102] Among them, f inst (t) represents the instantaneous frequency at time t, where t is time. The first derivative of the instantaneous phase represents the rate of change of the phase at time t. The second derivative of the instantaneous phase at time t is used to capture the acceleration characteristics of the phase and increase the sensitivity to signal abrupt changes. ln(1+…) is the natural logarithm function, which performs nonlinear compression on the whole and controls the range of values, so that the frequency response has good performance under both stationary and abrupt signals.

[0103] It should be noted that the calculation of instantaneous frequency depends on the phase change over time. Therefore, the first derivative is the most fundamental element for calculating instantaneous frequency. In existing technologies, instantaneous frequency is usually calculated directly using the first derivative of the phase. However, relying solely on the first derivative is susceptible to noise interference and cannot effectively capture changes during signal abrupt changes. Therefore, while the first derivative serves as a foundation, further enhancement processing is needed. The squared term of the second derivative captures the acceleration information of the signal, especially significantly improving the sensitivity to abrupt changes. By introducing the squared term of the second derivative, abrupt changes that the first derivative cannot accurately detect can be identified. Since abrupt changes in instantaneous frequency often... Since the phase acceleration changes significantly, the second derivative plays a core role in abrupt change detection. Traditional instantaneous frequency calculations mostly use only the first derivative to estimate the frequency, which has a weak response to abrupt changes. Introducing the square term of the second derivative not only enhances the sensitivity to abrupt changes but also makes the algorithm more stable in detecting abrupt change points. It is particularly suitable for extracting abrupt change features in traveling wave signals. Through nonlinear compression, it ensures that the instantaneous frequency will not have a maximum value within the normal range, and the abrupt change signal can be significantly amplified. Existing methods mostly use linear weighting, simply adding or superimposing the derivatives. Linear processing can easily lead to insufficient difference in response between stationary signals and abrupt change signals. The nonlinear properties of the natural logarithm function enable the mutation response to grow smoothly under large signal changes, while remaining normal under stationary signals. This makes the algorithm more adaptable to various types of signals. By introducing the square term of the second derivative, the formula's ability to respond to instantaneous frequency mutations is significantly improved, overcoming the problem of insufficient mutation response caused by using only the first derivative. Using the natural logarithm function for nonlinear transformation allows the formula to significantly amplify the response to mutation signals while maintaining smoothness for stationary signals, thus improving the algorithm's adaptability to different signal types. The combination of the square term of the second derivative and the nonlinear function effectively suppresses the influence of high-frequency noise on mutation detection, making it suitable for instantaneous frequency analysis in noisy environments.

[0104] The window size W was set using an experimental method, and the instantaneous frequency f was... inst(t) Using a sliding window of size W, calculate the variance σ of the instantaneous frequencies within the window. 2 (t), the formula is:

[0105]

[0106] Where, σ 2 (t) represents the variance of the instantaneous frequency within the sliding window at time t, indicating the fluctuation amplitude of the instantaneous frequency. W is the size of the sliding window, the number of sampling points within the window, and f inst (i) represents the instantaneous frequency value at time i. This represents the average instantaneous frequency within the window.

[0107] The variance threshold is set using the standard deviation multiple method. The variance of the instantaneous frequency within the window is compared with the variance threshold. The instantaneous frequency time point corresponding to the variance of the first instantaneous frequency within the window that is greater than the variance threshold is recorded as the wavefront arrival time.

[0108] It should be noted that the screening method ensures that the final analyzed signal characteristics match the actual fault characteristics, significantly improving the accuracy of signal processing. In complex power systems, signals often contain a large amount of background noise unrelated to the fault. The frequency bandwidth screening method can significantly improve the accuracy of fault location, providing reliable basic data for subsequent signal analysis. The application of Hilbert transform ensures the representation of the decomposed signal in the complex frequency domain, making the extraction of instantaneous frequencies more accurate. Through the analysis of instantaneous frequencies, the onset time and propagation path of the fault can be identified, providing strong support for accurate fault location. In transient signals, instantaneous frequencies reflect the transient changes of the signal. By analyzing instantaneous frequencies, the abrupt change point of the fault can be discovered, thereby accurately determining the onset of the fault. The higher-order derivative method ensures that the signal can accurately reflect frequency characteristics even under slight disturbances, providing important data support for precise location of wavefront arrival time. The calculation of window variance can capture subtle changes in the signal, thereby improving the accuracy of instantaneous frequency calculation. This dynamic calculation method can reduce misjudgments caused by signal instability and enhance the stability and reliability of fault location. Compared with a fixed threshold, the standard deviation multiple method can adaptively adjust to the fluctuation characteristics under different signal environments, which helps to improve the robustness of the fault location system and ensures the complete capture of fault information in the signal in time and space. By extracting the wavefront arrival time, high-precision fault location can be achieved, thereby reducing the troubleshooting time of power system faults and improving the system's response efficiency.

[0109] In this embodiment of the application, step S4 is performed to correct and compensate based on the wavefront arrival time difference, phase difference and propagation characteristics, and the relative distance between the fault point and the known node is calculated.

[0110] Specifically, a reference node A is selected as the starting point for positioning, and a node B is selected as an auxiliary comparison. The wavefront arrival time difference Δt between nodes A and B is calculated using the following formula:

[0111] Δt=|t arrA -t arrB |,

[0112] Among them, t arrA Let t be the arrival time of the wavefront at node A. arrB This represents the arrival time of the wavefront at node B.

[0113] Perform Fourier transform on the traveling wave signals at nodes A and B to obtain complex signals in the frequency domain. Calculate the magnitude of the complex signals in the frequency domain to obtain the amplitude spectrum. Select the frequency component with the largest amplitude spectrum as the main frequency component. Calculate the phase value of the main frequency component to obtain the phase of nodes A and B.

[0114] Subtracting the phase of node B from the phase of node A yields the node phase difference Δφ. AB .

[0115] An adaptive correction function G(r,Δφ) is constructed using the phases of nodes A and B. AB This is used to correct the nonlinear effect of phase difference on time difference under different frequency and phase conditions. The formula is:

[0116]

[0117] Where r is the frequency of the current traveling wave signal.

[0118] Calculate the time difference compensation value Δt using the time difference compensation value formula. opt The formula is:

[0119]

[0120] Among them, tan -1 (Δφ AB The arctangent of the phase difference is used as a compensation term for the time difference to smoothly adjust any small errors that may occur in the phase difference.

[0121] The wavefront arrival time difference Δt and the time difference compensation value Δt opt Adding them together, we get the positioning time difference Δt. ' .

[0122] It should be noted that in traditional time difference compensation formulas, commonly used calculation methods typically only consider a simple linear phase-to-time difference conversion. However, in practical applications, due to the nonlinear effects of traveling wave signals during propagation (such as multipath interference and attenuation), simple linear formulas can lead to inaccurate compensation results. The introduction of an adaptive function allows the time difference compensation value to be adaptively adjusted according to changes in frequency and phase difference through dynamic correction. The dynamic correction method is suitable for signal compensation needs in complex environments and avoids the introduction of additional multivariable external interference sources while maintaining computational stability. The phase difference typically fluctuates within the range of -π to π; however, changes within this range are affected by a truncation effect during the phase-to-time conversion (i.e., time compensation may become distorted when the phase approaches an extreme value). By introducing an arctangent operation, the conversion can be smoother under large phase differences, thus avoiding distortion near extreme points. The nonlinear response of the function allows the time difference compensation formula to more accurately adapt to a wide range of phase changes. Furthermore, the choice of this function ensures that its range approaches a constant, thus maintaining both computational accuracy and output convergence. Utilizing the nonlinear characteristics of the signal's frequency and phase difference, adaptive adjustment of the compensation value is achieved, enabling the time difference to more accurately reflect changes in actual signal propagation. The improvements enhance the dynamic response capability and accuracy of the compensation value. The introduction of compensation mitigates the truncation effect that easily occurs in traditional formulas over large phase difference ranges, allowing the formula to be applied over a wider phase range and improving the robustness and adaptability of time difference compensation. By combining multiple complex nonlinear functions, computational instability caused by multivariate interference is avoided. Especially when the traveling wave frequency or phase difference changes, it still provides accurate and stable time difference compensation values. This optimization allows the formula to adapt to complex and variable real-world environments, improving overall application reliability.

[0123] In this embodiment of the application, in step S5, the location coordinates of the fault point in the distribution network are calculated based on the node's geographical coordinates and direction vector, combined with the relative distance.

[0124] Based on the relative positioning formula, and using the positioning time difference Δt ' The formula for calculating the distance D from the fault point to node A is:

[0125]

[0126] Where v is the propagation speed of the traveling wave.

[0127] Calculate the Euclidean distance d between node A and node B using the Euclidean distance formula. AB ; Calculate the direction vector from node A to node B.

[0128] Normalizing the direction vector using Euclidean distance yields the unit direction vector (c). x ,c y Using the coordinates of node A, the unit direction vector, and the distance D, calculate the coordinates (x, y) of the fault point. D ,y D The formula is:

[0129] x D =x A +D·c x y D =y A +D·c y .

[0130] It should be noted that in power systems, fault signals propagate rapidly, and even small changes in the time difference can significantly affect the fault location. Therefore, accurately calculating the wavefront arrival time difference is crucial for improving location accuracy. Performing a Fourier transform on the traveling wave signals at nodes A and B converts the time-domain signal into a frequency-domain complex signal. The main frequency components are identified through their amplitude spectra. The Fourier transform more clearly reveals the essential characteristics of the fault signal, providing accurate data for subsequent location calculations. Compared to methods relying solely on time difference, phase difference provides information in frequency space, which is more helpful in accurately identifying fault points in complex power networks. The introduction of a correction function allows the fault location system to adaptively adjust according to the real-time frequency characteristics of the signal, thereby improving fault location accuracy in environments with varying frequencies. Through nonlinear adaptive adjustment, the system eliminates errors caused by frequency changes, effectively enhancing fault location accuracy. The robustness of the fault detection system is enhanced by the introduction of compensation values. The positioning system can eliminate uncertainties caused by measurement errors between nodes, improving the reliability of wavefront arrival time difference. The calculation of these compensation values ​​adjusts the errors in the phase difference, thereby ensuring the positioning accuracy of the fault point. This is particularly effective in situations with multi-path signal transmission, significantly optimizing the positioning effect. The normalization of the direction vector ensures that the calculation of the fault point not only depends on the distance but also considers the signal propagation direction, thus better aligning with the actual distribution network structure and improving the actual positioning accuracy. By combining the coordinates of node A, the unit direction vector, and the distance D, the coordinates of the fault point are calculated using a formula. This fault positioning method based on a planar coordinate system not only achieves spatial positioning of the fault point but also reduces errors through comprehensive consideration of multiple factors. This ensures that the location of the fault point is accurately reflected in the actual distribution network layout, providing a clear spatial reference for fault diagnosis and significantly shortening the fault repair time.

[0131] In this embodiment of the application, step S6 involves constructing a visualization interface to display the location of the fault point, storing the traveling wave signal data and fault location results in the database, and recording and archiving them.

[0132] Specifically, a front-end framework is used to build a visual interface, including a distribution network topology diagram and a fault location display area. A geographic information map is used to load the distribution network topology diagram, mark the fault location in the topology diagram, and display the node number, allowing users who have been verified by real name to view it.

[0133] Furthermore, the collected traveling wave signals and the resulting distribution network fault locations are stored in a database, and secure access measures are set up. The database will back up the stored data to the cloud and periodically perform integrity checks on the stored data and backup data. After the checks are completed, integrity check records are generated and synchronously stored in the database.

[0134] It should be noted that the introduction of the database not only improves the efficiency of data management but also facilitates the rapid retrieval of large-scale data. It helps in the long-term preservation and management of traveling wave signal data, laying the foundation for fault recording and analysis in power systems. By setting up secure access measures, unauthorized access can be effectively prevented, ensuring data privacy and security. This security measure makes the storage of distribution network data more secure and reliable, effectively guaranteeing the integrity and confidentiality of system data. Cloud backups are geographically distributed across different servers, so even if the local database fails or data is lost, data can still be recovered through cloud backups, avoiding the risk of data irrecoverability due to a single hardware failure. The cloud backup mechanism is particularly suitable for fault management in distribution network systems, effectively ensuring the continuity of traveling wave signals and fault location data, providing crucial support for data recovery. Regular integrity checks also improve the efficiency of data auditing, ensuring that data remains highly reliable even when a fault occurs. The combination of cloud and local database backups provides comprehensive protection for distribution network data, ensuring the stability and continuity of traveling wave signals and fault location information. Long-term data accumulation can also support the application of big data analytics, providing a scientific basis for intelligent maintenance of power distribution systems by mining patterns in historical data.

[0135] Example 3, the third embodiment of the present invention, differs from the previous two embodiments in that it is a distribution network fault location system based on variational mode decomposition, including a data acquisition and processing module, a mode decomposition module, a wavefront time extraction module, a fault location calculation module, and a visualization and storage module. The data acquisition and processing module acquires traveling wave signals from the distribution network, obtains frequency information using a frequency analyzer, and preprocesses the acquired traveling wave signals. The mode decomposition module performs mode decomposition on the preprocessed traveling wave signals, outputting several mode components. The wavefront time extraction module extracts the instantaneous phase and calculates the instantaneous frequency based on the decomposed mode components, analyzes the frequency variance through a sliding window, and extracts the wavefront arrival time. The fault location calculation module performs correction and compensation based on the wavefront arrival time difference, phase difference, and propagation characteristics, calculates the relative distance between the fault point and known nodes, and calculates the location coordinates of the fault point in the distribution network based on the node's geographical coordinates and direction vector, combined with the relative distance. The visualization and storage module constructs a visualization interface to display the fault location and stores the traveling wave signal data and fault location results in a database for archiving.

[0136] If the aforementioned functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present 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 described in the various embodiments of the present 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.

[0137] 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.

[0138] 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, because 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.

[0139] 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 in combination with any 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.

[0140] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended 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 method for fault location in distribution networks based on variational mode decomposition, characterized in that: include, The traveling wave signal of the power distribution network is collected, the frequency information is obtained using a frequency analyzer, and the collected traveling wave signal is preprocessed. Modal decomposition is performed on the preprocessed traveling wave signal to output several modal components; Based on the modal components obtained from the decomposition, the instantaneous phase is extracted and the instantaneous frequency is calculated. The frequency variance is analyzed by a sliding window, and the wavefront arrival time is extracted. Based on the wavefront arrival time difference, phase difference, and propagation characteristics, corrections and compensations are performed to calculate the relative distance between the fault point and known nodes; Based on the node's geographical coordinates and direction vector, combined with the relative distance, the location coordinates of the fault point in the distribution network are calculated; A visual interface is built to display the location of the fault point. Traveling wave signal data and fault location results are stored in the database and archived.

2. The distribution network fault location method based on variational mode decomposition as described in claim 1, characterized in that: The acquisition of traveling wave signals in the distribution network includes deploying current and voltage sensors at preset nodes in the distribution network to acquire traveling wave signals. The preprocessing of the acquired traveling wave signal includes time synchronization based on a satellite synchronization system, removal of noise interference from the acquired signal, and standardization of the denoised signal.

3. The distribution network fault location method based on variational mode decomposition as described in claim 2, characterized in that: The modal decomposition of the preprocessed traveling wave signal includes using a variational modal decomposition method to gradually adjust the frequency and bandwidth of the modes during the iteration process until the set convergence criteria are met, and then outputting several modal components.

4. The distribution network fault location method based on variational mode decomposition as described in claim 3, characterized in that: The extraction of wavefront arrival time includes: based on the obtained modal components, screening out frequency components with fault characteristics, performing Hilbert transform and constructing a complex signal, extracting its instantaneous phase and calculating the instantaneous frequency, analyzing the frequency variance through a sliding window, and extracting the wavefront arrival time.

5. The distribution network fault location method based on variational mode decomposition as described in claim 4, characterized in that: The phase difference calculation method includes selecting a reference node A as the positioning starting point and a node B as an auxiliary comparison, performing Fourier transform on the traveling wave signals of nodes A and B to obtain a complex signal in the frequency domain, calculating the magnitude of the complex signal in the frequency domain to obtain the amplitude spectrum, selecting the frequency component with the largest amplitude spectrum as the main frequency component, calculating the phase value of the main frequency component to obtain the phase of nodes A and B, and subtracting the phase of node B from the phase of node A to obtain the node phase difference.

6. The distribution network fault location method based on variational mode decomposition as described in claim 5, characterized in that: The construction of the visualization interface includes using a front-end framework to build the visualization interface, including a distribution network topology diagram and a fault location display area; Use a geographic information map to load the topology of the distribution network, mark the fault location in the topology map, and display the node number; Users who have passed real-name verification are allowed to view this information.

7. The distribution network fault location method based on variational mode decomposition as described in claim 6, characterized in that: The step of storing traveling wave signal data and fault location results in the database includes storing the collected traveling wave signals and the distribution network fault locations generated by analysis in the database, setting up security access measures, backing up the stored data in the cloud, and performing integrity checks on the stored data and backup data at preset intervals. After the checks are completed, integrity check records are generated and synchronously stored in the database.

8. A distribution network fault location system based on variational mode decomposition, employing the distribution network fault location method based on variational mode decomposition as described in any one of claims 1 to 7, characterized in that: It includes a data acquisition and processing module, a modal decomposition module, a wavefront time extraction module, a fault location calculation module, and a visualization and storage module; The data acquisition and processing module acquires traveling wave signals from the power distribution network, obtains frequency information using a frequency analyzer, and preprocesses the acquired traveling wave signals. The mode decomposition module performs mode decomposition on the preprocessed traveling wave signal and outputs several mode components; The wavefront time extraction module extracts the instantaneous phase and calculates the instantaneous frequency based on the decomposed modal components, and extracts the wavefront arrival time by analyzing the frequency variance through a sliding window. The fault location calculation module performs correction and compensation based on wavefront arrival time difference, phase difference and propagation characteristics, calculates the relative distance between the fault point and known nodes, and calculates the location coordinates of the fault point in the distribution network based on the node's geographical coordinates and direction vector, combined with the relative distance. The visualization and storage module constructs a visualization interface to display the location of the fault point and stores the traveling wave signal data and fault location results in the database for record archiving.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the distribution network fault location method based on variational mode decomposition as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the distribution network fault location method based on variational mode decomposition as described in any one of claims 1 to 7.

Images