Line fault detection method based on beat frequency attenuation principle and high-frequency signal information entropy
Through the detection method based on the beat frequency attenuation principle and high-frequency signal information entropy, high-frequency fault characteristics are extracted and relationship models are constructed, and the problems of insufficient line fault detection sensitivity and weak anti-interference ability in the existing technology are solved, achieving more accurate Fault location and detection effects of lower false alarm rates.
Patent Information
- Application Number
- CN202410939728.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-15
- Publication Date
- 2024-11-05
AI Technical Summary
The existing line fault detection methods are insufficient in high-impedance fault detection and complex network environments, have weak anti-interference capabilities, and rely on fundamental frequency signals, lack theoretical foundations, and have high implementation costs.
Using the detection method based on beat frequency attenuation principle and high-frequency signal information entropy, high-frequency fault components are extracted through wavelet transformation and high-pass filtering, Hilbert transform and envelope spectrum analysis are carried out, and the relationship model between the attenuation coefficient and the fault distance is constructed. And extract the information entropy change characteristics and combine it with neural network to make fault judgments and positioning.
It improves the sensitivity and anti-interference ability of fault detection, enhances the adaptability and generalization ability of the algorithm, achieves faster and more accurate fault identification and positioning, reduces the false alarm rate, and is suitable for detection of complex fault types.
Smart Images

Figure CN118897152A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of line fault detection, and in particular to a line fault detection method based on the beat frequency attenuation principle and high-frequency signal information entropy. Background Art
[0002] The safe and stable operation of the power system is crucial to modern society. As an important part of the power system, the fault detection and location of the transmission line has always been a hot topic in the field of power engineering. Line fault detection methods aim to quickly and accurately identify and locate various faults in the transmission line, such as short circuit, line break, grounding, etc. The importance of these methods is reflected in the ability to detect and handle faults in a timely manner, minimize power outage time, ensure the continuity and reliability of power supply, and reduce the risk of equipment damage and maintenance costs. Traditional line fault detection methods mainly include distance protection based on impedance measurement, traveling wave method, and current differential protection. The working principles of these methods usually rely on the measurement and analysis of electrical parameters such as voltage and current, and use specific algorithms to determine the occurrence and location of the fault. For example, distance protection estimates the fault distance by calculating the impedance of the fault point, and the traveling wave method uses the propagation characteristics of the high-frequency electromagnetic waves generated by the fault on the line to locate the fault. However, these existing technologies have some problems, such as the measurement accuracy is greatly affected by line parameters and load conditions, the detection sensitivity of certain types of faults (such as high-resistance faults) is insufficient, and the positioning accuracy is reduced in complex network topologies.
[0003] In order to solve the above problems, researchers have proposed a variety of improvement methods. For example, artificial intelligence technologies such as neural networks and fuzzy logic are used to improve the accuracy and adaptability of fault detection; high-precision and synchronized data provided by synchronized phasor measurement units (PMUs) are used to improve fault location; new fault detection algorithms based on wideband signal analysis are developed, etc. These methods have indeed made progress in some aspects, but compared with line fault detection methods based on the beat frequency attenuation principle and high-frequency signal information entropy, there are still some defects. First, many improved methods still rely mainly on baseband signals and make insufficient use of high-frequency transient information, while high-frequency signals often contain richer fault feature information. Secondly, some methods, such as artificial intelligence-based technologies, although they perform well in some scenarios, lack theoretical support, and their generalization ability and interpretability need to be improved. In addition, some methods, such as PMU-based technologies, require a lot of hardware investment and are costly to implement. Finally, many improved methods are still not ideal when dealing with complex fault types or multiple faults.
[0004] Therefore, there is an urgent need for a technical solution that can improve the sensitivity of fault detection and enhance the anti-interference ability. Summary of the invention
[0005] In order to solve the deficiencies of the prior art, the embodiments of the present application provide a line fault detection method, system and device based on the beat frequency attenuation principle and high-frequency signal information entropy. The present application solves the technical problems of low sensitivity and weak anti-interference ability of the prior art fault detection.
[0006] The embodiment of the present application provides a line fault detection method based on the beat frequency attenuation principle and high-frequency signal information entropy, including: collecting high-frequency transient characteristic signals at target nodes on the line based on a preset sampling frequency; extracting high-frequency fault components in the high-frequency transient characteristic signals using wavelet transform combined with high-pass filtering; performing Hilbert transform on the high-frequency fault components, and performing envelope spectrum analysis on the transformed signals to extract beat frequency features; fitting the attenuation characteristics of the beat frequency features to construct a relationship model between the attenuation coefficient and the fault distance; extracting the information entropy change characteristics of the high-frequency fault components to obtain a target information entropy sequence; judging the line fault based on the relationship model and the target information entropy sequence, and obtaining the fault location.
[0007] In a possible implementation, wavelet transform is combined with high-pass filtering to extract high-frequency fault components in high-frequency transient characteristic signals, including: using a target wavelet basis function to decompose the high-frequency transient characteristic signal into detail coefficients and approximate coefficients of different frequency bands; based on a preset number of decomposition layers, retaining the decomposed target high-frequency detail coefficients to reconstruct the high-frequency transient characteristic signal; using a high-pass filter with a preset adjustable cutoff frequency, and according to the lower limit of the frequency band of the target layer decomposition, adaptively filtering the reconstructed signal to obtain a high-frequency fault component.
[0008] In a possible implementation, a Hilbert transform is performed on a high-frequency fault component, and an envelope spectrum analysis is performed on the transformed signal to extract beat frequency characteristics, including: performing a Hilbert transform on the signal of the high-frequency fault component to construct an analytical signal; extracting the instantaneous amplitude and instantaneous phase through the analytical signal to obtain a target envelope signal; detecting the mutation point of the target envelope signal to divide the envelope signal into multiple quasi-steady-state segments; obtaining the corresponding local envelope spectrum by estimating the periodogram of each signal segment; finding the peak value of each local envelope spectrum within a preset frequency range, and obtaining a comprehensive beat frequency component through fusion calculation to extract the beat frequency characteristics.
[0009] In a possible implementation, the attenuation characteristics of the beat frequency feature are fitted to construct a relationship model between the attenuation coefficient and the fault distance, including: extracting multiple target local extreme points of the signal of the beat frequency feature, and segmenting the extreme point sequence to fit each segment of the sequence; solving the fitting parameters using the least squares method to obtain the attenuation coefficient; and constructing a relationship model between the attenuation coefficient and the fault distance.
[0010] In a possible implementation, the information entropy change characteristics of the high-frequency fault component are extracted to obtain a target information entropy sequence, including: coarse-graining the high-frequency fault component to obtain signals of multiple time scales; reconstructing the phase space of the signal of each scale based on the target embedding dimension and time delay; calculating the sorting pattern of each reconstructed vector to count the occurrence probability of each sorting pattern; calculating the permutation entropy, and constructing a multi-scale permutation entropy curve to obtain the entropy curve characteristics and the target information entropy sequence.
[0011] In a possible implementation, the line fault is judged based on the relational model and the target information entropy sequence, and the fault location is obtained, including: calculating the information entropy change rate according to the target information entropy sequence, and judging the fault state based on the target threshold; calculating the beat frequency attenuation rate using the relational model; locating the fault point based on the beat frequency attenuation rate and the information entropy change rate of multiple sampling points using triangulation or a pre-trained neural network model.
[0012] In a possible implementation, constructing a relationship model between the attenuation coefficient and the fault distance includes: Among them, k1, k2, k3, k4, k5 represent model parameters, α init represents the initial attenuation coefficient, a mean represents the average attenuation coefficient, and Δα represents the change in the attenuation coefficient.
[0013] In a possible implementation, the high-frequency fault component is coarse-grained to obtain multiple signals of different time scales, including: Where y represents the processed signal, τ represents the length of the time window, α represents the time index, j represents the index of the time window, and Y HF [i] represents the high-frequency fault component with sequence index i, σ represents the standard deviation, N represents the total length of the signal, and γ represents the effective signal length factor.
[0014] The embodiment of the present application also provides a line fault detection system based on the beat frequency attenuation principle and high-frequency signal information entropy, including: a sampling unit, a preprocessing unit, a calculation unit and a decision unit; wherein the sampling unit is used to collect high-frequency transient characteristic signals at a target node on the line based on a preset sampling frequency; the preprocessing unit is used to extract high-frequency fault components in the high-frequency transient characteristic signals using wavelet transform combined with high-pass filtering; the calculation unit is used to perform Hilbert transform on the high-frequency fault components and perform envelope spectrum analysis on the transformed signals to extract beat frequency features; the attenuation characteristics of the beat frequency features are fitted to construct a relationship model between the attenuation coefficient and the fault distance; the information entropy change characteristics of the high-frequency fault components are extracted to obtain a target information entropy sequence; the decision unit is used to judge the line fault based on the relationship model and the target information entropy sequence, and obtain the fault location.
[0015] An embodiment of the present application also provides a line fault detection device based on the beat frequency attenuation principle and high-frequency signal information entropy, characterized in that it includes: a processor, a memory, and a system bus; wherein the processor and the memory are connected via the system bus; the memory is used to store one or more programs, and the one or more programs include instructions, and when the instructions are executed by the processor, the processor executes the method described in the above embodiment.
[0016] In the line fault detection method, system and device based on the beat frequency attenuation principle and high-frequency signal information entropy provided above, the embodiment of the present application can effectively solve the problem of insufficient positioning accuracy of traditional methods in high-resistance fault detection and complex network environments by analyzing the beat frequency attenuation characteristics and information entropy of the high-frequency signal generated by the fault. Not only the sensitivity of fault detection is improved, but also the anti-interference ability of the method is enhanced. In addition, by combining physical principles (beat frequency attenuation) and information theory (signal entropy), the method improves the adaptability and generalization ability of the algorithm while maintaining the theoretical basis, including faster and more accurate fault identification and location, lower false alarm rate, and better adaptability to various complex fault types. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0018] Figure 1 A flow chart of a line fault detection method based on the beat frequency attenuation principle and high-frequency signal information entropy provided in an embodiment of the present application;
[0019] Figure 2 A schematic block diagram of a line fault detection system based on the beat frequency attenuation principle and high-frequency signal information entropy provided in an embodiment of the present application. DETAILED DESCRIPTION
[0020] Various exemplary embodiments of the present disclosure will now be described in detail with reference to the accompanying drawings. It should be noted that the relative arrangement of components and steps, numerical expressions and numerical values set forth in these embodiments do not limit the scope of the present disclosure unless otherwise specifically stated.
[0021] Those skilled in the art can understand that the terms "first", "second" and the like in the embodiments of the present disclosure are only used to distinguish different steps, devices or modules, etc., and neither represent any specific technical meaning nor represent the necessary logical order between them. It should also be understood that in the embodiments of the present disclosure, "multiple" can refer to two or more, and "at least one" can refer to one, two or more. It should also be understood that for any component, data or structure mentioned in the embodiments of the present disclosure, in the absence of explicit limitation or contrary revelation given in the context, it can generally be understood as one or more. In addition, the term "and / or" in the present disclosure is only a description of the association relationship of the associated objects, indicating that there can be three relationships, for example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in the present disclosure generally indicates that the associated objects before and after are an "or" relationship. It should also be understood that the description of each embodiment in the present disclosure emphasizes the differences between the embodiments, and the same or similar parts can refer to each other. For the sake of brevity, they will not be repeated one by one.
[0022] At the same time, it should be understood that, for ease of description, the sizes of the various parts shown in the drawings are not drawn according to the actual proportional relationship. The following description of at least one exemplary embodiment is actually only illustrative and is by no means intended to limit the present disclosure and its application or use. The techniques, methods and devices known to ordinary technicians in the relevant fields may not be discussed in detail, but where appropriate, the techniques, methods and devices should be considered part of the specification. It should be noted that similar numbers and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further discussed in subsequent drawings.
[0023] In order to make the purpose, technical solution and advantages of the embodiments of the present application clearer, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0024] Figure 1 A flow chart of a line fault detection method based on the beat frequency attenuation principle and high-frequency signal information entropy provided in an embodiment of the present application.
[0025] It is important to understand that when a power line fails, a high-frequency transient oscillation signal will be generated at the fault point. When these signals propagate on the line, due to the distributed parameter characteristics of the line, multiple oscillation components with similar frequencies will be formed, thus generating a beat frequency phenomenon. The envelope of the beat frequency signal exhibits an attenuation characteristic, and its attenuation rate is closely related to the distance from the fault point to the observation point.
[0026] The beat frequency signal can be expressed as: s(t) = A1cos(ω1t) + A2cos(ω2t), where A1 and A2 are the amplitudes of two oscillation components with similar frequencies, and ω1 and ω2 are the corresponding angular frequencies. Expanding the above formula, we can get in, is the average amplitude. It can be seen that the beat frequency signal consists of a high-frequency carrier and a low frequency envelope The frequency of the envelope is the beat frequency:
[0027] Furthermore, information entropy is an important indicator for describing signal uncertainty. For high-frequency transient signals generated by faults, their information entropy will change as the signal propagates on the line. By analyzing the signal information entropy at different observation points, the fault location information can be obtained. The information entropy of a discrete signal is defined as: Among them, p i is the probability of the ith value in the signal. For a continuous signal, its information entropy can be calculated by dividing the signal into several intervals: Where p(x) is the probability density function of the signal.
[0028] Based on the above principle, the present application embodiment designs a line fault detection method, which mainly includes the following steps: Figure 1As shown, at step S101, a high-frequency transient characteristic signal at a target node on the line is collected based on a preset sampling frequency. In one implementation scenario, a high-frequency sensor is installed at a key node of the line to collect the high-frequency transient signal generated by the fault. The sampling frequency should be no less than 100kHz to ensure that the high-frequency component generated at the moment of the fault can be captured.
[0029] In step S102, the high-frequency fault component in the high-frequency transient characteristic signal is extracted by wavelet transform combined with high-pass filtering. Specifically, the high-frequency transient characteristic signal is decomposed into detail coefficients and approximate coefficients of different frequency bands by using the target wavelet basis function; based on the preset number of decomposition layers, the target high-frequency detail coefficients after decomposition are retained to reconstruct the high-frequency transient characteristic signal; a high-pass filter with a preset adjustable cutoff frequency is used, and the reconstructed signal is adaptively filtered according to the lower limit of the frequency band decomposed by the target layer to obtain the high-frequency fault component.
[0030] Considering that high-frequency transient signals in power systems are usually distributed in a wide frequency band ranging from kHz to MHz, and are often masked by strong power frequency and its harmonic signals, the embodiment of the present application adopts a composite method of wavelet transform combined with high-pass filtering to extract high-frequency components containing fault characteristics. The specific steps are as follows.
[0031] Perform multi-scale discrete wavelet transform (DWT) decomposition on the sampled signal: First, select a suitable wavelet basis function (such as db4 or sym4, etc.) to decompose the signal X(t) into detail coefficients d j (k) and the approximation coefficient a of the last layer J (k): Among them, j represents the decomposition scale, k is the time shift parameter, ψ j,k (t) and They are wavelet function and scaling function respectively. By adjusting the number of decomposition layers J, the signal frequency band can be flexibly divided.
[0032] Select the appropriate decomposition layer number J0 according to the fault feature frequency band and retain the high-frequency detail coefficients: Based on prior knowledge and actual tests, determine the frequency band range containing the main fault information and select the corresponding wavelet decomposition layer J0. Only retain the high-frequency detail coefficients of layers 1 to J0 to reconstruct the high-frequency signal: This step can effectively remove low-frequency interference and highlight high-frequency fault characteristics.
[0033] Adaptively high-pass filter the reconstructed high-frequency signal to further suppress the residual low-frequency components: select a high-order FIR or IIR high-pass filter h(t) with an adjustable cutoff frequency, whose cutoff frequency is automatically determined according to the lower limit of the frequency band of the J0 layer wavelet decomposition. Filter the reconstructed high-frequency signal: Y HF(t) = h(t)*X HF (t), where * represents the convolution operation. The order and type of the filter need to balance the computational complexity and filtering effect. Kaiser window, Chebyshev and other optimization design methods can be used. For the convenience of subsequent analysis, the amplitude of the filtered high-frequency signal can also be normalized.
[0034] Through the above composite steps, high-frequency transient signals containing key fault characteristics can be effectively extracted, laying the foundation for subsequent analysis.
[0035] In step S103, the high-frequency fault component is subjected to Hilbert transform, and the transformed signal is subjected to envelope spectrum analysis to extract the beat frequency feature. Further, the signal of the high-frequency fault component is subjected to Hilbert transform to construct an analytical signal; the instantaneous amplitude and instantaneous phase are extracted from the analytical signal to obtain the target envelope signal; the mutation point of the target envelope signal is detected to divide the envelope signal into multiple quasi-steady-state segments; the corresponding local envelope spectrum is obtained by estimating the periodogram of each segment of the signal; the peak value of each local envelope spectrum is found within a preset frequency range, and the comprehensive beat frequency component is obtained by fusion calculation to extract the beat frequency feature.
[0036] It should be noted that the beat frequency is a periodic intensity variation phenomenon caused by the superposition of two sinusoidal signals with similar frequencies. In the power system, the superposition of the reflected wave and the incident wave at the fault point will produce obvious beat frequency characteristics, and its frequency is closely related to the fault distance. The embodiment of the present application uses the Hilbert transform combined with the adaptive envelope spectrum analysis method to accurately extract the beat frequency characteristics, and the specific steps are as follows.
[0037] Perform discrete Hilbert transform on normalized high frequency signal: Fast discrete Hilbert transform using FFT: Wherein, DFT and IDFT represent discrete Fourier transform and its inverse transform respectively, and sign(ω) is the sign function.
[0038] Construct an analytical signal and extract its instantaneous amplitude and phase: Among them, the instantaneous amplitude (envelope): Instantaneous phase:
[0039] Adaptive segmentation of the envelope signal: The sliding variance method can be used to detect the mutation point of the envelope signal and divide the signal into multiple quasi-steady-state segments: {A1[n], A2[n], ..., A K [n]}, where K is the number of segments determined adaptively.
[0040] For each signal segment, the improved Welch periodogram is estimated to obtain the local envelope spectrum. For the i-th signal segment, the window is added and framed: A i,m[n] = A i [n]·w[n-mR], m=0, 1, ..., M-1; perform FFT on each frame: Averaging gives the power spectrum estimate: Where w[n] is the Hanning window, R is the frame shift, and N is the number of FFT points.
[0041] Identify the beat frequency component in the local envelope spectrum, and find the peak value within the preset frequency range for the envelope spectrum of each signal segment. Then, combine the results of multiple segments to obtain the final beat frequency estimate. The weighted average method can be used to fuse the results of multiple segments: Among them, the weight w i It can be determined based on the energy or peak significance of each signal segment.
[0042] Through the above-mentioned adaptive and multi-scale analysis steps, the beat frequency characteristics of the fault signal can be accurately extracted, effectively coping with the challenge of signal non-stationarity in actual systems.
[0043] At step S104, the attenuation characteristics of the beat frequency feature are fitted to construct a relationship model between the attenuation coefficient and the fault distance. In one embodiment, multiple target local extreme value points of the signal of the beat frequency feature are extracted, and the extreme value point sequence is processed in sections to fit each section of the sequence; the fitting parameters are solved using the least squares method to obtain the attenuation coefficient; and the relationship model between the attenuation coefficient and the fault distance is constructed.
[0044] It should be understood that the attenuation characteristics of the beat frequency signal are closely related to the distance from the fault point. The embodiment of the present application proposes a beat frequency attenuation characteristic analysis method based on piecewise exponential fitting to adapt to multiple attenuation modes that may exist in an actual system.
[0045] To extract the local maximum point sequence of the beat frequency signal, the sliding window method can be used to find the local maximum point in the envelope signal A[n]: {(n i , A i )|A[n i ]>A[n i -δ]and A[n i ]>A[n i +δ], i=1, 2, ..., N}, where δ is the local maximum value judgment window, and its size is adaptively adjusted according to the beat frequency period. In order to improve the robustness, the curvature analysis can be combined to eliminate the false maximum points.
[0046] The maximum point sequence is segmented. The improved CUSUM algorithm can be used to detect the change points of the decay rate and divide the maximum point sequence into multiple subsequences: Wherein, M is the number of segments determined adaptively.
[0047] Furthermore, we can perform logarithmic linear regression on each sequence segment, and for the kth sequence segment: Use weighted least squares method to solve the parameters: Among them, the weight It can be determined according to the sample reliability, N k is the number of samples in the kth segment.
[0048] Evaluate the goodness of fit and perform model selection. The coefficient of determination R can be calculated for each segment of the fit. 2 The Akaike information criterion (AIC) was used to comprehensively evaluate the fitting effect. If the improvement of the segmented fitting was not significant, it would degenerate into a single exponential model.
[0049] Next, extract the key attenuation characteristic parameters:
[0050] Initial decay rate: α init =α (1) ;
[0051] Average decay rate:
[0052] Decay rate changes:
[0053] Then, a multi-parameter fault distance estimation model is established: Among them, k1, k2, k3, k4, k5 represent model parameters, α init represents the initial attenuation coefficient, α mean represents the average attenuation coefficient, and Δα represents the change in the attenuation coefficient. Among them, k1, k2, k3, k4, and k5 can be obtained through a large amount of simulation and measured data training.
[0054] Through the above steps, the attenuation characteristics of the beat frequency signal can be fully characterized, and an accurate quantitative relationship with the fault distance can be established, providing a reliable basis for fault location. This method can adapt to multiple attenuation modes that may appear in complex systems, improving positioning accuracy and scope of application.
[0055] At step S105, the information entropy change characteristics of the high-frequency fault component are extracted to obtain the target information entropy sequence. In an implementation scenario, the high-frequency fault component is coarse-grained to obtain signals of multiple different time scales; the signal of each scale is reconstructed in phase space based on the target embedding dimension and time delay; the sorting mode of each reconstructed vector is calculated to count the occurrence probability of each sorting mode; the permutation entropy is calculated, and a multi-scale permutation entropy curve is constructed to obtain the entropy curve characteristics and the target information entropy sequence.
[0056] It should be understood that information entropy is an important indicator for describing signal uncertainty. The present application embodiment introduces the concept of multi-scale permutation entropy to improve the sensitivity and reliability of fault detection by analyzing the complexity variation characteristics of high-frequency signals at different time scales.
[0057] Multi-scale decomposition of high-frequency signals: HF [i] Perform coarse-graining processing to obtain signals of different time scales τ:
[0058] Where y represents the processed signal, τ represents the length of the time window, α represents the time index, j represents the index of the time window, and Y HF [i] represents the high-frequency fault component with sequence index i, σ represents the standard deviation, N represents the total length of the signal, and γ represents the effective signal length factor. For a signal with a sampling rate of 20kHz and a total length of 1 second (N=20000), the following parameter settings can be considered: τ=64, α=1.2, β=1.5, σ=τ / 5, γ=τ / 4.
[0059] Reconstruct the phase space of the signal at each scale: select the appropriate embedding dimension m and time delay λ, and construct the reconstruction vector:
[0060] Compute the ordering mode of each reconstructed vector: Arrange the elements in ascending order and get the sort index π i =[r1, r2, ..., r m ]. There are m! possible sorting modes. Statistical probability of occurrence of each sorting mode: Calculate the permutation entropy:
[0061] Secondly, a multi-scale permutation entropy curve is constructed and the entropy curve features are extracted, including but not limited to the entropy mean, entropy standard deviation and entropy curve slope.
[0062] Through the above steps, a characteristic sequence reflecting the statistical complexity changes of high-frequency signals on multiple time scales can be obtained. This multi-scale analysis method can capture subtle changes in signals and is of great significance for early detection of faults. In practical applications, appropriate parameters need to be selected based on system characteristics and computing resource limitations.
[0063] In step S106, the line fault is judged based on the relational model and the target information entropy sequence, and the fault location is obtained. Specifically, the information entropy change rate is calculated based on the target information entropy sequence, and the fault state is judged based on the target threshold; the beat frequency attenuation rate is calculated using the relational model; based on the beat frequency attenuation rate and information entropy change rate of multiple sampling points, the fault point is located using triangulation or a pre-trained neural network model.
[0064] Furthermore, the multi-scale information entropy change rate calculation can calculate the sliding change rate of entropy for each time scale τ: dH / dt(t, t) = [H(τ, t+Δt)-H(τ, t)] / Δt, t = 1, 2, ..., N-Δt, where H(τ, t) is the sample entropy at time t under the τ scale, and Δt is the sliding window length. This step can capture the changes in the dynamic characteristics of the system under different time scales.
[0065] Next, construct an adaptive multi-feature fusion criterion:
[0066]
[0067] AND max(S(f))>λ2(t)·μ(S(f))for anyτ∈[τ min ,τ max ],
[0068] Where λ1(t) and λ2(t) are dynamic threshold coefficients, σ(·) and μ(·) represent standard deviation and mean operations, respectively. This adaptive criterion can automatically adjust the sensitivity according to the system operation status and improve the reliability of detection.
[0069] It is also possible to precisely locate the moment when a fault occurs: where τ opt is the optimal time scale, w1 and w2 are weight coefficients. Through weighted fusion of multiple features, the initial moment of the fault can be determined more accurately.
[0070] Before positioning, it is necessary to analyze the beat frequency attenuation characteristics: f , t f +T w ] Perform wavelet packet transform in the time window to extract the characteristic frequency band energy sequence E(t). Use nonlinear least squares method to fit the attenuation model: E(t) = A·exp(-αt)·cos(2πf0t+φ) to obtain the attenuation coefficient α and characteristic frequency f0. This step can accurately characterize the transient process characteristics after the fault.
[0071] In another implementation scenario, the fault detection threshold can also be set to be that when the beat frequency characteristics and the information entropy change rate simultaneously meet the following conditions, it is determined that a fault has occurred: the beat frequency amplitude exceeds a preset threshold; the information entropy change rate exceeds a preset threshold.
[0072] Regarding fault location, the triangulation method can be used to locate the fault point based on the beat frequency attenuation rate and information entropy change rate of multiple observation points.
[0073] Suppose there are three observation points A, B, and C, and their distances to the fault point are d A ,d B ,d C , then:
[0074]
[0075] Among them, k1, b1, k2, and b2 are unknown coefficients that can be determined through offline experiments. By combining the above equations, the distance from the fault point to each observation point can be solved, thereby determining the fault location.
[0076] The fault distance can also be estimated based on the neural network model with α and f0 as input: L = NN(α, f0), where NN(·) represents the pre-trained neural network model. This model can adaptively learn the complex nonlinear relationship between system parameters and fault distance.
[0077] Through the above algorithm, rapid fault detection, accurate location and distance estimation can be achieved. The algorithm has strong anti-interference ability and adaptability, and is suitable for complex and changeable power system operating environment.
[0078] Figure 2 A schematic block diagram of a line fault detection system based on the beat frequency attenuation principle and high-frequency signal information entropy provided in an embodiment of the present application. It should be understood that the system shown in the figure is exemplary and not restrictive. This means that the system architecture involved is not limited to a specific form or design, but is presented as an example. In other words, the architecture shown in the figure can be regarded as a way of expression to clearly describe related concepts and relationships, and does not exclude other forms of architecture. Therefore, when interpreting the architecture in the picture, it should be understood that the model is flexible and diverse, and its purpose is to provide an exemplary description rather than a restrictive provision on a specific form.
[0079] The embodiment of the present application includes a sampling unit 201, a preprocessing unit 202, a calculation unit 203 and a decision unit 204; wherein the sampling unit 201 is used to collect high-frequency transient characteristic signals at a target node on the line based on a preset sampling frequency; the preprocessing unit 202 is used to extract high-frequency fault components in the high-frequency transient characteristic signals by using wavelet transform combined with high-pass filtering; the calculation unit 203 is used to perform Hilbert transform on the high-frequency fault components and perform envelope spectrum analysis on the transformed signals to extract beat frequency features; fit the attenuation characteristics of the beat frequency features to construct a relationship model between the attenuation coefficient and the fault distance; extract the information entropy change characteristics of the high-frequency fault components to obtain a target information entropy sequence; the decision unit 204 is used to judge the line fault based on the relationship model and the target information entropy sequence, and obtain the fault location.
[0080] Furthermore, an embodiment of the present application also provides a line fault detection device based on the beat frequency attenuation principle and high-frequency signal information entropy, including: a processor, a memory, and a system bus; the processor and the memory are connected through the system bus; the memory is used to store one or more programs, and the one or more programs include instructions, which, when executed by the processor, enable the processor to execute any of the above methods.
[0081] Furthermore, an embodiment of the present application also provides a computer program product, which, when executed on a terminal device, enables the terminal device to execute any of the above methods.
[0082] It can be known from the description of the above implementation mode that those skilled in the art can clearly understand that all or part of the steps in the above-mentioned embodiment method can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product can be stored in a storage medium such as ROM / RAM, a disk, an optical disk, etc., including several instructions for enabling a computer device (which can be a personal computer, a server, or a network communication device such as a media gateway, etc.) to execute the methods described in the various embodiments of the present application or certain parts of the embodiments.
[0083] It should be noted that the various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments, and the same or similar parts between the various embodiments can be referred to each other. For the device disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part description.
[0084] It should also be noted that, in the embodiments of the present application, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that there is any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that the process, method, article or equipment including a series of elements includes not only those elements, but also includes other elements that are not clearly listed, or also includes elements inherent to such process, method, article or equipment. In the absence of more restrictions, the elements limited by the statement "comprise one..." do not exclude the presence of other identical elements in the process, method, article or equipment including the elements.
[0085] The above description of the disclosed embodiments enables professionals and technicians in the field to implement or use the present application. Various modifications to these embodiments will be apparent to professionals and technicians in the field, and the general principles defined in the embodiments of the present application can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to these embodiments shown in the embodiments of the present application, but will conform to the widest range consistent with the principles and novel features disclosed in the embodiments of the present application.
Claims
1. A line fault detection method based on the beat frequency attenuation principle and high-frequency signal information entropy, characterized in that: include: Collect high-frequency transient characteristic signals at target nodes on the line based on a preset sampling frequency; The high-frequency fault component in the high-frequency transient characteristic signal is extracted by wavelet transform combined with high-pass filtering; Perform Hilbert transform on the high-frequency fault component and perform envelope spectrum analysis on the transformed signal to extract the beat frequency characteristics; Fit the attenuation characteristics of the beat frequency feature to build a relationship model between the attenuation coefficient and the fault distance; Extract the information entropy change characteristics of high-frequency fault components to obtain the target information entropy sequence; Based on the relational model and the target information entropy sequence, the line fault is judged and the fault location is obtained.
2. The line fault detection method according to claim 1, characterized in that: in, The high-frequency fault components in the high-frequency transient characteristic signals are extracted by wavelet transform combined with high-pass filtering, including: The target wavelet basis function is used to decompose the high-frequency transient characteristic signal into detail coefficients and approximate coefficients of different frequency bands; Based on the preset number of decomposition layers, the target high-frequency detail coefficients after decomposition are retained to reconstruct the high-frequency transient characteristic signal; The reconstructed signal is adaptively filtered using a high-pass filter with a preset adjustable cutoff frequency according to the lower limit of the frequency band decomposed by the target layer to obtain the high-frequency fault component.
3. The line fault detection method according to claim 1, characterized in that: in, Perform Hilbert transform on the high-frequency fault component and perform envelope spectrum analysis on the transformed signal to extract the beat frequency features, including: Performing Hilbert transform on the signal of high-frequency fault component to construct analytical signal; Extract the instantaneous amplitude and instantaneous phase by analyzing the signal to obtain the target envelope signal; Detecting a mutation point of a target envelope signal to divide the envelope signal into a plurality of quasi-steady-state segments; Obtain the corresponding local envelope spectrum by estimating the periodogram of each signal segment; The peak value of each local envelope spectrum is found within the preset frequency range, and the comprehensive beat frequency component is obtained through fusion calculation to extract the beat frequency characteristics.
4. The line fault detection method according to claim 1, characterized in that: in, The attenuation characteristics of the beat frequency feature are fitted to construct a relationship model between the attenuation coefficient and the fault distance, including: Extract multiple target local extreme value points of the signal with beat frequency characteristics, and process the extreme value point sequence in sections to fit each section of the sequence; The fitting parameters are solved using the least square method to obtain the attenuation coefficient; Construct a relationship model between attenuation coefficient and fault distance.
5. The line fault detection method according to claim 1, characterized in that: in, Extract the information entropy change characteristics of high-frequency fault components to obtain the target information entropy sequence, including: Coarse-grain processing is performed on high-frequency fault components to obtain signals of multiple different time scales; Reconstruct the phase space of the signal at each scale based on the target embedding dimension and time delay; Calculate the sorting mode of each reconstructed vector to count the occurrence probability of each sorting mode; The permutation entropy is calculated and a multi-scale permutation entropy curve is constructed to obtain the entropy curve characteristics and the target information entropy sequence.
6. The line fault detection method according to claim 1, characterized in that: in, Based on the relational model and the target information entropy sequence, the line fault is judged and the fault location is obtained, including: The information entropy change rate is calculated according to the target information entropy sequence, and the fault state is judged based on the target threshold; The beat frequency attenuation rate is calculated using the relational model; Based on the beat frequency attenuation rate and information entropy change rate of multiple sampling points, the fault point is located using triangulation or a pre-trained neural network model.
7. The line fault detection method according to claim 4, characterized in that: in, Construct a relationship model between attenuation coefficient and fault distance, including: Among them, k1, k2, k3, k4, k5 represent model parameters, α init represents the initial attenuation coefficient, α mean represents the average attenuation coefficient, and Δα represents the change in the attenuation coefficient.
8. The line fault detection method according to claim 5, characterized in that: in, The high-frequency fault components are processed in a coarse-grained manner to obtain signals of multiple different time scales, including: Where y represents the processed signal, τ represents the length of the time window, α represents the time index, j represents the index of the time window, and Y HF [i] represents the high-frequency fault component with sequence index i, σ represents the standard deviation, N represents the total length of the signal, and γ represents the effective signal length factor.
9. A line fault detection system based on the beat frequency attenuation principle and high frequency signal information entropy, characterized in that: include: Sampling unit, preprocessing unit, calculation unit and decision unit; among them, The sampling unit is used to collect high-frequency transient characteristic signals at a target node on the line based on a preset sampling frequency; The preprocessing unit is used to extract high-frequency fault components in the high-frequency transient characteristic signal by using wavelet transform combined with high-pass filtering; The calculation unit is used to perform Hilbert transform on the high-frequency fault component and perform envelope spectrum analysis on the transformed signal to extract the beat frequency feature; fit the attenuation characteristics of the beat frequency feature to construct a relationship model between the attenuation coefficient and the fault distance; extract the information entropy change characteristics of the high-frequency fault component to obtain the target information entropy sequence; The decision unit is used to judge the line fault based on the relational model and the target information entropy sequence, and obtain the fault location.
10. A line fault detection device based on the beat frequency attenuation principle and high frequency signal information entropy, characterized in that: include: A processor, a memory, and a system bus; wherein the processor and the memory are connected via the system bus; The memory is used to store one or more programs, wherein the one or more programs include instructions, and when the instructions are executed by the processor, the processor executes the method according to any one of claims 1 to 8.
Citation Information
Cited By
Power transmission and distribution on-line insulation detection method and system
CN119178974A
A method and system for online insulation detection of power transmission and distribution
CN119178974B
Power cable fault discharge sound recognition method and system based on multi-characteristic fusion
CN120544610A
Method for judging abnormity of ultrasonic welding head according to beat frequency
CN121163654A
Photovoltaic inverter fault processing method and device based on acoustic characteristics
CN121171259A
