Traveling wave time delay estimation method and related device

Abstract

The invention belongs to an underground cable fault positioning method, and provides a traveling wave time delay estimation method and a related device in order to solve the problem that a traveling wave signal processing algorithm cannot completely meet high-precision and high-adaptability traveling wave time delay detection requirements in a complex electric power scene. Carrying out adaptive frequency band wavelet deconstruction and detail extraction on the first sensing signal and the second sensing signal respectively, and then carrying out dynamic parameter modal splitting based on variational problem solution on each reconstructed signal in the obtained multi-layer first reconstructed signal group and the second reconstructed signal group corresponding to the scale; the method comprises the following steps: obtaining a plurality of intrinsic frequency band modal components with specific sparsity, taking the first intrinsic frequency band modal component as a main mode, comprehensively considering the energy ratio and cross-correlation sharpness of two main modes in each layer, and determining a final traveling wave time delay estimation result according to the cross correlation coefficient of the two main modes under the optimal common decomposition layer. According to the method, the calculation redundancy is reduced while the precision is ensured, and the real-time performance is improved.

Description

A method and related apparatus for estimating traveling wave time delay Technical Field

[0001] This application pertains to a method for locating faults in underground cables, specifically involving a traveling wave time delay estimation method and related apparatus. Background Technology

[0002] Underground cable connections offer significant advantages in resisting harsh weather conditions such as freezing, rain, snow, and strong winds. However, underground cables consist of multiple cable segments connected together, resulting in numerous cable joints. These joints, buried underground for extended periods, are susceptible to corrosion from water immersion and saline-alkali environments, leading to decreased cable insulation performance and potentially causing short-circuit faults. The traveling wave method, a crucial technique for cable fault location, detects transient traveling waves generated when a fault occurs and propagating towards both ends of the cable. Fault location is then determined based on the time difference between the arrival time of these transient traveling waves at the measurement point. Theoretically, this method is less affected by fault type and transition resistance, possessing high accuracy potential. In recent years, with the development of signal processing and high-precision time synchronization technologies, the traveling wave method has been widely researched and applied in cable fault location. Currently, algorithms for processing traveling wave signals have some value in their respective application scenarios, but all suffer from significant limitations and cannot fully meet the high-precision, highly adaptable traveling wave time delay detection requirements in complex power scenarios. Summary of the Invention

[0003] This application addresses the technical problem that current processing algorithms for traveling wave signals cannot fully meet the requirements for high-precision and highly adaptable traveling wave delay detection in complex power scenarios, and provides a traveling wave delay estimation method and related apparatus.

[0004] To achieve the above objectives, this application adopts the following technical solution: Firstly, this application proposes a traveling wave delay estimation method, comprising: acquiring two current sensing signals located on both sides of a fault point, denoted as the first sensing signal and the second sensing signal, respectively; performing adaptive frequency band wavelet decomposition and detail extraction on the first sensing signal and the second sensing signal, respectively, to obtain a multi-layer first reconstructed signal group and a second reconstructed signal group corresponding to the scale; for each layer, performing dynamic parameter mode decomposition based on variational problem solving on each reconstructed signal in the first reconstructed signal group and the second reconstructed signal group, respectively, to obtain multiple eigenband mode components with specific sparsity; taking the first of the multiple eigenband mode components with specific sparsity as the dominant mode of that layer, and calculating the energy proportion of the dominant mode in the reconstructed signal of that layer; comprehensively considering the energy proportion of the two dominant modes in each layer and the cross-correlation sharpness of the two dominant modes, selecting the optimal common decomposition layer; and determining the final traveling wave delay estimation result based on the cross-correlation coefficient of the two dominant modes under the optimal common decomposition layer.

[0005] Furthermore, after acquiring the two current sensing signals located on both sides of the fault point, the method further includes: performing denoising processing on the first sensing signal and the second sensing signal respectively to obtain a first denoised signal and a second denoised signal; and performing amplitude normalization processing on the first denoised signal and the second denoised signal.

[0006] Furthermore, the method for adaptive frequency band wavelet deconstruction and detail extraction of the first and second sensing signals respectively includes: presetting a maximum decomposition level; cyclically processing the first and second sensing signals layer by layer to separate multiple layers of high-frequency detail components and low-frequency approximation components; for each layer, from the first layer to the maximum decomposition level, sequentially setting the low-frequency approximation components to zero, retaining the high-frequency detail components for inverse wavelet transform, and reconstructing reconstructed signal groups of detail features at different frequency scales, with the first and second sensing signals respectively denoted as the first reconstructed signal group and the second reconstructed signal group of the corresponding layer.

[0007] Furthermore, when performing dynamic parameter mode decomposition based on variational problem solving, the number of modes and penalty parameters are preset, and the initial center frequency is set in combination with the spectral characteristics of the reconstructed signal itself.

[0008] Furthermore, the method for determining the final traveling wave time delay estimation result includes: calculating the cross-correlation coefficient of the two principal modes under the optimal common decomposition layer, determining the time delay corresponding to the maximum value of the cross-correlation coefficient, and using it as the final traveling wave time delay estimation result.

[0009] Furthermore, when comprehensively considering the energy proportion and cross-correlation sharpness of the two main modes in each layer, the optimal common decomposition layer is selected through dynamic weighting.

[0010] Secondly, this application proposes a traveling wave delay estimation system, comprising: a data module for acquiring two current sensing signals located on both sides of a fault point, denoted as the first sensing signal and the second sensing signal, respectively; an extraction module for performing adaptive frequency band wavelet deconstruction and detail extraction on the first sensing signal and the second sensing signal, respectively, to obtain a multi-layer first reconstructed signal group and a second reconstructed signal group corresponding to the scale; a decomposition module for performing dynamic parameter mode decomposition based on variational problem solving on each reconstructed signal in the first reconstructed signal group and the second reconstructed signal group for each layer, to obtain multiple eigenband mode components with specific sparsity; a dominant mode module for taking the first of the multiple eigenband mode components with specific sparsity as the dominant mode of the layer, and calculating the energy proportion of the dominant mode in the reconstructed signal of the layer; a screening module for comprehensively considering the energy proportion of the two dominant modes in each layer and the cross-correlation sharpness of the two dominant modes to screen out the optimal common decomposition layer; and a calculation module for determining the final traveling wave delay estimation result based on the cross-correlation coefficient of the two dominant modes under the optimal common decomposition layer.

[0011] Thirdly, this application proposes an electronic device, including: a memory and one or more processors; the memory is coupled to the processors; wherein the memory stores computer program code, the computer program code including computer instructions, and when the computer instructions are executed by the processor, the electronic device performs the steps of the above-described ripple delay estimation method.

[0012] Fourthly, this application proposes a computer-readable storage medium, characterized in that the computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of the above-described traveling wave delay estimation method.

[0013] Compared with the prior art, this application has the following beneficial effects: This application proposes a traveling wave delay estimation method, which performs adaptive frequency band wavelet deconstruction and detail extraction on the first and second sensing signals respectively, and then performs dynamic parameter mode decomposition based on variational problem solving on each of the obtained first and second reconstructed signal groups to obtain multiple intrinsic frequency band mode components with specific sparsity. The first one is taken as the main mode, and then the energy ratio and cross-correlation sharpness of the two main modes are comprehensively considered. Based on the cross-correlation coefficient of the main modes selected under the optimal common decomposition layer, the final traveling wave delay estimation result is determined. This application achieves microsecond-level traveling wave delay measurement by dynamically selecting the optimal common decomposition layer and determining the final traveling wave delay estimation result based on the cross-correlation coefficient of the principal modes. It concentrates the core transient features of the traveling wave, effectively eliminating interference from secondary components. Time-domain cross-correlation analysis is performed based on the principal modes of the optimal common decomposition layer. Due to the high purity and prominent features of the principal modes, the cross-correlation system has sharper peaks and lower sidelobes, avoiding the peak blurring problem caused by feature scale mismatch or signal aliasing in traditional algorithms. Layer-by-layer wavelet decomposition and reconstruction can adapt to traveling wave features at different frequency scales. Each reconstructed signal can cover the corresponding high-frequency details. Combined with principal mode selection, it reduces computational redundancy while ensuring accuracy and improving real-time performance.

[0014] This application also proposes a traveling wave delay estimation system, an electronic device, and a computer-readable storage medium, which possess all the advantages of the aforementioned traveling wave delay estimation methods. Attached Figure Description

[0015] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0016] Figure 1 is a schematic diagram of the working principle of existing two-end traveling wave ranging; Figure 2 is a flowchart of one traveling wave time delay estimation method of this application; Figure 3 is another flowchart of the traveling wave time delay estimation method of this application; Figure 4 is a schematic diagram of two-end current traveling wave acquisition in an embodiment of this application; Figure 5 is a schematic diagram of two-end fault waveform in simulation modeling verification in an embodiment of this application; Figure 6 is a schematic diagram of the process of selecting the optimal common decomposition layer in an embodiment of this application; Figure 7 is a schematic diagram of the maximum cross-correlation coefficient in an embodiment of this application; Figure 8 is a schematic diagram of one traveling wave time delay estimation system of this application. Detailed Implementation

[0017] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. The components of the embodiments of this application described and shown in the accompanying drawings can generally be arranged and designed in various different configurations.

[0018] Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0019] It should be noted that similar labels and letters in the following figures indicate similar items. Therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures.

[0020] In the description of the embodiments of this application, it should be noted that if terms such as "upper," "lower," "horizontal," or "inner" indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings, or the orientation or positional relationship commonly used when the product of the invention is in use, they are only for the convenience of describing this application and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation, and therefore should not be construed as a limitation on this application. In addition, terms such as "first" and "second" are only used to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0021] Furthermore, the use of the term "horizontal" does not imply that the component must be absolutely horizontal, but rather that it can be slightly tilted. For example, "horizontal" simply means that its direction is more horizontal than "vertical," and does not mean that the structure must be completely horizontal, but can be slightly tilted.

[0022] In the description of the embodiments of this application, it should also be noted that, unless otherwise explicitly specified and limited, the terms "set," "install," "connect," and "link" should be interpreted broadly. For example, they can refer to a fixed connection, a detachable connection, or an integral connection; they can refer to a mechanical connection or an electrical connection; they can refer to a direct connection or an indirect connection through an intermediate medium; and they can refer to the internal connection of two components. Those skilled in the art can understand the specific meaning of the above terms in this application according to the specific circumstances.

[0023] In centralized photovoltaic and wind farms, 35kV collector lines commonly use underground cable connections. Compared to overhead lines, cable connections offer significant advantages in resisting severe weather conditions such as freezing, rain, snow, and strong winds. However, underground cables consist of multiple cable segments connected together, resulting in numerous cable joints. These joints, buried underground for extended periods, are susceptible to corrosion from water immersion and saline-alkali environments, leading to decreased insulation performance and increasing the risk of short-circuit faults. Manually inspecting faulty cable ends is costly and inefficient, becoming a weak link affecting the reliable operation of new energy power plants. The traveling wave method, an important technique for cable fault location, detects transient traveling waves generated when a fault occurs and propagating towards both ends of the cable, using the time difference between the arrival times of the traveling waves at the measurement point to pinpoint the fault. Theoretically, this method is less affected by the fault type and transition resistance, and has the potential for high accuracy. In recent years, with the development of signal processing technology and high-precision time synchronization technology, the traveling wave method has been widely researched and applied in the field of cable fault location.

[0024] Although traveling wave ranging technology is widely used in cable fault location, the complexity of actual power scenarios presents numerous challenges to traveling wave signal analysis and processing. On the one hand, cable fault types are diverse, including single-phase short circuits, three-phase short circuits, and weak faults at cable joints. The amplitude and frequency characteristics of traveling wave signals generated by different fault types vary significantly. For example, weak faults at joints produce small-amplitude traveling wave signals, whose characteristics are easily masked by interference. On the other hand, the cable operating environment is subject to significant electromagnetic interference, such as high-frequency noise from photovoltaic inverters and interference from wind farm motors, resulting in low signal-to-noise ratios for the acquired traveling wave signals. Furthermore, the propagation frequency range of traveling waves varies considerably between different cable specifications; short-distance cable faults result in higher traveling wave frequencies, while long-distance cable faults result in lower traveling wave frequencies. These factors lead to traveling wave signals exhibiting characteristics of non-stationarity, strong interference, and a large frequency span, posing significant difficulties for denoising, feature extraction, and accurate wavefront timing calibration. Failure to effectively address these issues will directly increase the traveling wave delay detection error, thereby affecting the accuracy of fault location and failing to meet the actual needs of power operation and maintenance for rapid and accurate emergency repairs.

[0025] To address the core challenges in traveling wave signal analysis and processing, various algorithms have been proposed to denoise, extract features, and calibrate wavefront timing. The first is wavelet transform, with Discrete Wavelet Transform (DWT) being a common method in traditional multi-scale wavelet decomposition. This algorithm calibrates the traveling wavefront by finding the timing corresponding to the modulus maxima. The second is Empirical Mode Decomposition (EMD), which decomposes complex non-stationary traveling wave signals into multiple intrinsic frequency band modal components to extract traveling wave features. The third is traditional Variational Mode Decomposition (VMD), which uses a preset number of modes to partition the signal's frequency domain, avoiding the mode aliasing problem of EMP. The fourth is an algorithm combining the Teager energy operator with discrete wavelet transform. This algorithm uses the Teager energy operator to enhance the high-frequency coefficients after discrete wavelet transform, thereby highlighting the traveling wavefront features. In addition, some technologies attempt to combine wavelet decomposition and variational decomposition in a simple way, hoping to combine the advantages of the two algorithms to improve the processing effect.

[0026] However, the existing algorithms and combinations mentioned above all have significant limitations and cannot fully meet the needs of complex power scenarios.

[0027] Figure 1 illustrates the working principle of a conventional double-ended traveling wave distance measurement system. In Figure 1, A, B, and C represent taps on three underground cables. The passive traveling wave measurement device is installed in the branch boxes at both ends of the cable. When a short circuit fault occurs at a tap in the underground cable between two branches, forming a short circuit point (point A in Figure 1), a strong current traveling wave is emitted. This current traveling wave propagates along the cable transmission line to both sides (directions A1 and A2 in Figure 1). When it reaches the passive traveling wave measurement device installed in the cable branch box, it triggers the device to record the waveform. The device then processes the waveform using an internal noise reduction algorithm to extract the arrival time of the traveling wave front, which serves as the basis for fault diagnosis. The distance from the fault location to the passive traveling wave measurement device, i.e., the distance between the fault location and the measurement point, is then calculated using the formula below. :

[0028] in, The distance to the point where the fault occurred. The wave speed of the traveling wave. The time to reach the passive traveling wave measurement device at one end. The time to reach the passive traveling wave measurement device at the other end. This is the total length of the cable line.

[0029] Currently, cable fault traveling wave location is mainly used to detect the following key parameters: 1) Two-end current traveling wave monitoring: The transient current traveling wave signal generated when the fault occurs is captured by a high-frequency current sensor, providing key data for subsequent calculation of the distance to the fault location.

[0030] 2) Two-end traveling wave wavefront time difference calibration: Perform high-accuracy two-end time difference calculation for the detected traveling wave wavefront, typically at the microsecond level.

[0031] In the field of traveling wave ranging in power systems, traveling wave signal analysis and processing technology is the core link in achieving accurate fault location. While mainstream processing algorithms for traveling wave signals have certain value in their respective application scenarios, they all have significant limitations and cannot fully meet the high-precision, highly adaptable traveling wave delay detection requirements in complex power scenarios. The following are some common algorithms, explained in detail: 1) Wavelet Transform Algorithm: Traditional multi-scale wavelet decomposition methods (such as discrete wavelet transform methods) are used to calibrate the traveling wave front. Typically, the moment corresponding to the modulus maxima is taken as the traveling wave front moment. However, discrete wavelet transform suffers from an over-reliance on experience regarding the number of decomposition levels, making it unsuitable for all waveforms. The number of decomposition levels in discrete wavelet transform is directly related to the frequency band and time resolution. A higher decomposition level corresponds to a lower frequency band and a more ambiguous time resolution. Therefore, an excessively high decomposition level will increase the error due to the ambiguity of the time resolution, while an excessively low decomposition level will result in an excessively high frequency band, with too many maxima of high-frequency coefficients, making it difficult to effectively distinguish feature information.

[0032] 2) Empirical Mode Decomposition (EMD) Algorithm: Empirical Mode Decomposition (EMD) extracts traveling wave features by adaptively decomposing a signal into multiple intrinsic frequency band modal components (IMFs). However, this algorithm suffers from a significant "mode aliasing" problem. When a traveling wave signal contains transient components of different frequency ranges (such as fault traveling waves and environmental interference signals), EMD easily mixes components of different frequencies into the same IMF, causing the traveling wave features to be diluted by the interference signal, making it impossible to accurately separate the pure traveling wave principal component. Furthermore, the EMD decomposition process relies on the judgment of local extrema of the signal. When processing traveling wave signals with low signal-to-noise ratios (such as small-amplitude traveling waves generated by weak faults in cable joints), it is prone to generating false IMF components due to misjudgment of local extrema, further interfering with wavefront feature extraction. Moreover, EMD lacks a clear criterion for decomposition termination, usually relying on a manually preset number of iterations. Excessive iteration increases computational redundancy, while insufficient iteration leads to inadequate extraction of traveling wave features, making it difficult to adapt to traveling wave signal processing scenarios with different fault types and cable parameters.

[0033] 3) Traditional Variational Mode Decomposition (VMD) Methods: While this method can achieve frequency domain division of the signal by pre-setting the number of modes and avoid mode aliasing in EMD, it has two major drawbacks: First, it requires manual experience to set the number of modes and penalty parameters. If the preset number of modes is too small, key transient features and interference signals in the traveling wave signal will be forcibly classified into the same mode, making effective separation impossible. If the number of modes is too large, the main traveling wave signal will be over-decomposed into multiple low-energy sub-modes, resulting in fragmentation of the traveling wave features and increasing the difficulty of subsequent wavefront identification. Second, the initial center frequency selection lacks adaptability. It usually adopts a fixed frequency interval or random initialization method. When the frequency range of the traveling wave signal changes due to cable length and fault type (e.g., the traveling wave frequency is higher for short-distance cable faults and lower for long-distance cable faults), a fixed initial center frequency can easily lead to a mismatch between the decomposed modes and the actual frequency band of the traveling wave, thus affecting the accuracy of wavefront timing calibration and failing to meet the fault location requirements of different specifications of collector lines.

[0034] 4) The Teager energy operator uses a wavelet transform algorithm: This algorithm enhances the high-frequency coefficients after the Discrete Wavelet Transform (DWT) to highlight the characteristics of the traveling wave front, but it has significant limitations in its applicable scenarios. First, the Teager energy operator is extremely sensitive to noise. When there is strong electromagnetic interference in the cable operating environment (such as high-frequency noise generated by photovoltaic inverters or motor interference from wind farms), the operator will simultaneously enhance both the high-frequency interference signal and the traveling wave characteristic signal, resulting in an increased proportion of interference components in the enhanced high-frequency coefficients. This, in turn, masks the true extreme point of the traveling wave front and increases the probability of wave front misjudgment. Second, the enhancement effect of this algorithm depends on the number of decomposition levels of the DWT. If the number of decomposition levels is not chosen properly (e.g., too few levels result in the high-frequency coefficients containing too much low-frequency interference, while too many levels result in a decrease in time resolution), the Teager energy operator cannot effectively compensate for the inherent defects of the DWT, and the problem of large wave front timing calibration errors will still occur. Furthermore, the algorithm lacks an adaptive adjustment mechanism. For traveling wave signals of different amplitudes and frequencies, the parameters of the operator (such as the size of the time window) need to be manually adjusted to optimize the enhancement effect. It cannot achieve fully automated traveling wave delay detection, which reduces its practicality and efficiency in fault location of large-scale power line collection lines.

[0035] Therefore, although existing technologies include individual applications or simple combinations of wavelet decomposition and variational decomposition, they have not solved key problems such as adaptive matching of the number of decomposition layers, intelligent setting of the initial center frequency, and selection of the dual-channel common optimal layer, and require manual intervention in parameter adjustment.

[0036] Based on the above, this application proposes a traveling wave time delay estimation method and related apparatus. Through dynamic weighted scoring and sharpness quantification, the entire process is automated. The following detailed description, in conjunction with embodiments and accompanying drawings, further illustrates this application.

[0037] Figure 2 shows a flowchart of the traveling wave time delay estimation method of this application, which may include: S101, acquiring two current sensing signals located on both sides of the fault point, which are respectively denoted as the first sensing signal and the second sensing signal.

[0038] It should be noted that when a cable fault occurs, a transient current traveling wave is generated. This transient current traveling wave will propagate along the cable to both sides. The signal acquisition channels on both sides of the fault point can capture this transient current traveling wave signal respectively, providing raw data support for subsequent time delay calculation.

[0039] S102, perform adaptive frequency band wavelet deconstruction and detail extraction on the first and second sensing signals respectively to obtain multi-layer first reconstructed signal groups and second reconstructed signal groups corresponding to the scale.

[0040] It should be noted that the key features of traveling wave signals are rapidly changing high-frequency components, while environmental interference and measurement noise are mostly low-frequency or irregular interference. Multi-scale wavelet decomposition can break down the signal according to frequency scales, adaptively adjusting parameters during the decomposition process to retain high-frequency detail components that reflect the characteristics of the traveling wave, while discarding low-frequency approximations that do not contain key information. The signal is then reconstructed through inverse wavelet transform, highlighting the characteristics of the traveling wave. By effectively separating the high-frequency feature components and low-frequency interference components in the traveling wave signal, the key features of the traveling wave are highlighted. At the same time, by adaptively decomposing to adapt to traveling wave signals at different frequency scales, the problem of traditional wavelet decomposition relying too much on empirically setting the number of decomposition levels can be avoided, providing a higher quality signal for subsequent mode decomposition.

[0041] S103, for each layer, perform dynamic parameter mode decomposition based on variational problem solving on each reconstructed signal in the first reconstructed signal group and the second reconstructed signal group to obtain multiple intrinsic frequency band mode components with specific sparsity.

[0042] It should be noted that each reconstructed signal in the reconstructed signal group may still contain mixed components of different frequencies. Dynamic parameter mode decomposition based on variational problem solving constructs a variational model that minimizes the sum of the estimated bandwidths of all modes. It adaptively adjusts the decomposition parameters based on the signal's own spectral characteristics, decomposing each reconstructed signal into multiple sparse intrinsic frequency band modal components with non-overlapping frequencies. This achieves accurate frequency domain partitioning of the signal and avoids mode aliasing. Consequently, it enables accurate frequency domain separation of the reconstructed signal, resulting in pure and highly sparsity intrinsic frequency band modal components. This effectively avoids the mode aliasing and spurious component problems of traditional empirical mode decomposition, while also addressing the deficiency of traditional variational mode decomposition parameters relying on manual setting, providing high-quality sub-signals for subsequent master mode selection.

[0043] S104: The first of the multiple intrinsic frequency band mode components with specific sparsity is taken as the dominant mode of the layer, and the energy proportion of the dominant mode in the reconstructed signal of the layer is calculated.

[0044] It should be noted that after the reconstructed signal is decomposed, the multiple eigenband mode components with specific sparsity exhibit a natural ordered arrangement. The first eigenband mode component output is the component with the largest energy proportion, the most complete carrying of the traveling wave core characteristics, and the frequency that best matches the traveling wave characteristic frequency band in the reconstructed signal of that layer. There is no need to perform energy traversal comparisons on all eigenband mode components; directly selecting the first one is sufficient to determine the core feature carrier of that layer. The energy proportion calculation of this main mode is then performed simultaneously, allowing for the retention of quantified feature data for use in the subsequent joint layer selection process of the two signals for performance evaluation.

[0045] S105. Taking into account the energy ratio of the two main modes in each layer and the sharpness of the cross-correlation between the two main modes, the optimal common decomposition layer is selected.

[0046] It should be noted that the two reconstructed signal groups correspond to multiple decomposition layers, each with its own dominant mode. Relying solely on energy percentage may result in defects in the synchronization of the selected decomposition layer between the two signals, while relying solely on synchronization may overlook ineffective layers with insufficient energy. Therefore, two indicators must be considered: energy percentage reflects the feature purity of the dominant mode, and cross-correlation sharpness reflects the time synchronization accuracy of the dominant modes between the two ends. The optimal decomposition layer is selected through dynamic weight scoring. This avoids the problem of insufficient decomposition layer adaptability caused by single-indicator selection. The selected optimal common decomposition layer simultaneously meets the requirements of high feature purity and good time synchronization, providing the optimal analytical basis for subsequent calculation of the time delay of the two signals and solving the pain point of feature scale mismatch between the two signals in traditional algorithms.

[0047] S106. Based on the cross-correlation coefficients of the two principal modes under the optimal common decomposition layer, determine the final traveling wave time delay estimation result.

[0048] It should be noted that the two principal modes under the optimal common decomposition layer concentrate their respective core characteristics of the traveling wave and have good time synchronization. By calculating the cross-correlation function of the two principal modes, the time difference corresponding to the maximum value of the cross-correlation coefficient is the time difference of the traveling wave propagating from the fault point to the two acquisition points, and this time difference is the traveling wave delay. This application performs cross-correlation analysis based on high-purity, highly synchronized principal modes, resulting in a sharper peak and lower sidelobes in the cross-correlation function. This avoids time delay misjudgment caused by peak ambiguity, significantly improves the accuracy of traveling wave delay estimation, and ensures the accuracy of subsequent fault distance calculation.

[0049] Figure 3 shows another flowchart of the traveling wave delay estimation method of this application, which may include the following steps: Figure 4 shows a schematic diagram of two-end current traveling wave acquisition. A 35kV cable is led out from the overall armored sheath in a branch box, forming three branch lines. A passive traveling wave measurement device is installed on the cable grounding side, and the current traveling wave is detected by installing traveling wave induction coils in phases A, B, and C of the cable. When a short circuit fault occurs at a junction of the underground cable between two branch lines, a strong current traveling wave will be emitted. The current traveling wave is caused by the capacitance effect of the cable to ground. When a sudden short circuit occurs in the cable, the capacitor will discharge charge to ground, thereby causing a high-rate-of-change current traveling wave. The current traveling wave will propagate along the cable transmission line to both sides. When it reaches the passive traveling wave measurement device installed in the branch box, the waveform of the passive traveling wave measurement device is recorded. After processing by a noise reduction algorithm, the arrival time of the traveling wave front is extracted as the basis for fault judgment. Then, the distance to the fault location is determined according to the calculation method given by the formula of the two-end traveling wave method. Specifically: S201, preprocessing.

[0050] Preprocessing of the two sensor signals: It should be noted that the two signals here specifically refer to two sources: one is the current traveling wave of the traveling wave inductors corresponding to phases A, B, and C in the branch box on one side of the fault point, and the other is the current traveling wave of the traveling wave inductors corresponding to phases A, B, and C in the branch box on the other side of the fault point. The two sensor signals are denoted as the first sensor signal and the second sensor signal, respectively.

[0051] First, environmental noise and measurement interference in the first and second sensing signals are removed by wavelet transform technology to obtain relatively clean signals, which are denoted as the first denoised signal and the second denoised signal respectively.

[0052] Then, amplitude normalization processing is performed on the first and second denoised signals to obtain the first and second preprocessed signals, respectively. This is used to scale the energy of the first and second denoised signals to the same order of magnitude, so as to eliminate the influence of amplitude differences on subsequent analysis and ensure the fairness of the comparison.

[0053] S202, multi-scale wavelet hierarchical reconstruction.

[0054] The preprocessed dual signals, namely the first preprocessed signal and the second preprocessed signal, are subjected to adaptive frequency band wavelet deconstruction and detail extraction. A maximum deconstruction level is preset. Through a layer-by-layer iteration, from the first level to the maximum deconstruction level, high-frequency detail components representing the rapidly changing characteristics of the signal are separated at each level, while low-frequency approximation components reflecting the overall signal profile are discarded. Only the retained high-frequency detail components are used for inverse wavelet transform to reconstruct a new signal highlighting the detailed features at a specific frequency scale, denoted as the reconstructed signal. This reconstructed signal highlights the detailed features of the original signal at that specific frequency scale. After the iteration is completed, each channel will obtain a number of reconstructed signals from different frequency scales, equal to the maximum deconstruction level, denoted as the first reconstructed signal group and the second reconstructed signal group, respectively.

[0055] S203, Variational Optimization Modal Splitting and Feature Extraction.

[0056] Each reconstructed signal in the first and second reconstructed signal groups obtained in the previous step is subjected to dynamic parametric mode decomposition based on variational problem solving. Dynamic parametric mode decomposition intelligently sets the initial center frequency by pre-setting the number of modes and penalty parameters, combined with the spectral characteristics of the reconstructed signal itself. Without manual intervention, it adaptively decomposes the reconstructed signal into a series of intrinsic mode functions (IMFs) with specific sparsity, so as to achieve accurate division of the frequency domain of the traveling wave signal.

[0057] In addition, to improve decomposition efficiency and specificity, after identifying the dominant mode in the intrinsic frequency band modal components, the proportion of the dominant mode to the total energy of the reconstructed signal is calculated. This energy proportion can reflect the changing trend of the dominant mode energy intensity at each decomposition scale.

[0058] S204, dynamic weighted joint layer selection and precise time delay calculation.

[0059] After completing wavelet hierarchical reconstruction and variational optimization mode decomposition at all frequency scales, the cross-correlation coefficients of the two main modes under the same decomposition level are first calculated. Based on the peak half-width at half-maximum of the cross-correlation coefficients, the time synchronization sharpness index is defined, which can fully quantify the time synchronization accuracy of the two main modes.

[0060] Then, a crucial multi-weighted layer selection stage begins. Instead of selecting the optimal layer individually for each path, a comprehensive approach is adopted, considering both the energy proportion of all common decomposition layers and the sharpness of the cross-correlation. Sharpness is defined as the reciprocal of the full width at half maximum (FWHM) of the cross-correlation peak, measuring the temporal synchronicity and similarity of two signals. A higher sharpness value (narrower FWHM) indicates a high degree of correlation between the two signals at a precise point in time, with higher consistency in waveform and arrival time. A dynamic weighted scoring system is used, consisting of the geometric mean of the energy proportions of the two signals and the normalized cross-correlation sharpness. If the former is greater than 0.3 (i.e., very strong energy), energy is considered more important, and a weight of 60% for energy and 40% for sharpness is assigned. If the energy is weaker, synchronicity (sharpness) is considered more important, and a weight of 30% for energy and 70% for sharpness is assigned. This design balances the two indicators of "signal characteristic strength" and "synchronicity between the two paths," avoiding misjudgments that might arise from a single indicator, and thus adaptively selecting a common decomposition layer that is optimal for both paths. This optimal common decomposition layer ensures that subsequent delay analysis is performed on the most representative and synchronous features.

[0061] Finally, using the two main modes extracted from the optimal common decomposition layer, the time delay corresponding to the maximum value of the cross-correlation coefficient is found by calculating their cross-correlation coefficients, which is the final traveling wave time delay estimation result.

[0062] This application achieves accurate multi-scale feature extraction of traveling wave signals by combining layer-by-layer wavelet reconstruction with adaptive frequency band division based on variational optimization, avoiding the mode mixing problems of traditional empirical mode decomposition and the parameter dependence problems of traditional variational decomposition methods. Furthermore, a dynamic weighted scoring system is designed based on two indices: energy proportion and cross-correlation sharpness. High energy proportions emphasize energy, while low energy proportions emphasize synchronization, selecting the optimal decomposition layer for both paths to ensure the matching of feature scales in time delay estimation. A sharpness index based on the peak half-width of cross-correlation is defined to quantify the time synchronization accuracy of the two signals, providing an objective basis for joint layer selection and improving the accuracy of time delay estimation. Moreover, from wavelet denoising and variational optimization decomposition of the initial center frequency to joint layer selection, automated processing without manual intervention is achieved, adapting to different line specifications and fault types.

[0063] To illustrate the application of this application, simulation modeling verification of the method was performed. Specifically, a phase-to-phase short-circuit fault model of a three-core cable with parameters identical to those of an actual 35kV medium-voltage three-core cable in a wind farm was built using PSCAD / EMTDC simulation software. The specific verification method is as follows: 1) Fault model construction: The total cable length is 500m, and the fault point is set 200m from the transformer power source and 300m from the end-matching load. The simulation step size is set to 0.02us, corresponding to a sampling frequency of 50MHz. A phase-to-phase short-circuit fault between phases A and B of the cable was simulated, and the double-ended fault waveform was obtained. Figure 5 shows a schematic diagram of the double-ended fault waveform in this simulation modeling verification.

[0064] 2) Signal Processing Flow: The dual-ended fault waveform is processed using the method described in this application. First, wavelet denoising is performed to obtain two clean signals. Then, the dual-ended signals are processed layer by layer using multi-scale wavelet hierarchical reconstruction and variational optimization mode decomposition. The optimal common decomposition layer is selected by dynamic weighting based on energy proportion and cross-correlation sharpness; in this embodiment, it is the second layer. Figure 6 shows a schematic diagram of the process of selecting the optimal common decomposition layer in this embodiment. Finally, cross-correlation analysis is performed on the second-layer principal mode, yielding a maximum cross-correlation coefficient of 0.6763. The absolute value of the time corresponding to this maximum cross-correlation coefficient is the time delay difference between the two signals, which is 0.48 μs. Figure 7 shows a schematic diagram of the maximum cross-correlation coefficient in this embodiment.

[0065] 3) Verification of positioning results: According to the double-ended traveling wave ranging formula, the wave velocity is calculated from the cable electrical parameters and is approximately 2 × 10⁻⁶. 8 The calculated fault distance was 298m, which is only 2m different from the actual fault distance of 300m. The accuracy rate reached 99.33% and the error was only 0.67%, which fully verified the high accuracy and effectiveness of the method in dual-end traveling wave time delay estimation and fault location.

[0066] This application addresses the pain points of traditional methods, such as parameter dependence and feature aliasing. Specifically: First, this application employs a two-layer anti-interference architecture of wavelet denoising and adaptive decomposition, ensuring the stability of traveling wave feature extraction from the source. Wavelet adaptive threshold denoising accurately filters out environmental noise and measurement interference, and amplitude normalization eliminates amplitude differences between the first and last traveling wave signals, providing a clean and comparable signal foundation for subsequent analysis. The preprocessed signal undergoes layer-by-layer wavelet reconstruction to extract high-frequency detail components. Then, dynamic parameter mode decomposition based on variational problem solving achieves precise frequency domain partitioning of the reconstructed signal. Its initial center frequency is intelligently determined by the signal spectrum peak value, avoiding mode aliasing in traditional EMD and the fixed parameter dependence problem of traditional variational optimization decomposition methods. Even in complex interference scenarios, it can still stably extract core features. Furthermore, experimental verification shows that in a strong interference environment with a signal-to-noise ratio as low as 10dB, the relative error of wavefront timing calibration in this application can be controlled within 5%, while the relative errors of the traditional DWT algorithm and the Teager energy operator algorithm reach 15% and 20% or more, respectively, fully demonstrating the strong anti-interference capability and stability of the method in this application. Secondly, this application achieves microsecond-level traveling wave delay measurement through a combination strategy of dynamic joint layer selection and main mode cross-correlation. On the one hand, the energy ratio of each layer's main mode and the reconstructed signal is first calculated, and then the sharpness is calculated by the peak half-width of the cross-correlation of the two main modes. Combined with dynamic weight scoring, a common optimal layer with both high energy ratio and high synchronization is selected. The main mode of this common optimal layer concentrates the core transient characteristics of the traveling wave, effectively eliminating interference from secondary components. On the other hand, this application performs time-domain cross-correlation analysis based on the main mode of the common optimal layer. Due to the high purity and prominent features of the main modes, the peak value of the cross-correlation coefficient is sharper and the sidelobes are lower, avoiding the peak blurring problem caused by feature scale mismatch or signal aliasing in traditional algorithms. In actual testing, for simulated fault scenarios of 35kV collector lines, the time delay detection error of this application can be controlled within 2μs, corresponding to a fault location error of less than 50m, which is far superior to the variational optimization decomposition method based on fixed parameters (approximately 150m) and the empirical mode decomposition algorithm (approximately 200m), meeting stringent accuracy requirements. Furthermore, the core advantage of this application lies in its layer-by-layer decomposition and dynamic joint layer selection mechanism, achieving fully automated traveling wave processing without manual intervention. Multi-dimensional judgment ensures optimal decomposition and layer selection: layer-by-layer wavelet decomposition and reconstruction adapt to the characteristics of traveling waves at different frequency scales, with each reconstructed signal covering the corresponding high-frequency details; the initial center frequency of the mode decomposition based on variational problem solving adapts to the signal spectrum, eliminating the need for manual setting; during joint layer selection, invalid layers with low energy proportions are first filtered out, and then the weights are dynamically adjusted according to the geometric mean of energy to select the two commonly optimal layers.This design allows the application to automatically adapt to traveling wave signals of different specifications of collector lines and different fault types, eliminating the need to adjust parameters such as the number of wavelet decomposition layers and the number of modes in variational optimization decomposition for specific scenarios. This solves the pain point of repeated parameter tuning in traditional algorithms and significantly improves the versatility and ease of use for fault location of large-scale, multi-specification collector lines. Finally, the application reduces computational redundancy and improves real-time performance while ensuring accuracy through a layer-by-layer decomposition and effective layer selection strategy. Traditional algorithms require completing all decomposition steps, continuing to calculate even after obtaining effective features, leading to redundancy. In contrast, the application only needs to complete the layer-by-layer decomposition of the preset maximum number of layers, and then quickly filters out invalid layers through an energy proportion threshold, reducing the computational load of joint layer selection and eliminating meaningless over-decomposition calculations. Taking a 35kV collector line traveling wave signal as an example, the average calculation time of the application is approximately 8ms. Compared with traditional fixed-layer wavelet decomposition algorithms and empirical mode decomposition algorithms, the computational efficiency is improved by 46.7% and 68%, respectively, meeting the real-time requirement of the traveling wave location system to complete delay detection within 20ms, providing technical support for rapid fault determination and alarm push.

[0067] Figure 8 shows a schematic diagram of the traveling wave delay estimation system of this application, which may include: a data module for acquiring two sensing signals located on both sides of the fault point, denoted as the first sensing signal and the second sensing signal, respectively; an extraction module for performing adaptive frequency band wavelet deconstruction and detail extraction on the first sensing signal and the second sensing signal respectively, to obtain a multi-layer first reconstructed signal group and a second reconstructed signal group corresponding to the scale; a decomposition module for performing dynamic parameter mode decomposition based on variational problem solving on each reconstructed signal in the first reconstructed signal group and the second reconstructed signal group for each layer, to obtain multiple eigenband mode components with specific sparsity; a dominant mode module for taking the first of the multiple eigenband mode components with specific sparsity as the dominant mode of the layer, and calculating the energy proportion of the dominant mode in the reconstructed signal of the layer; a screening module for comprehensively considering the energy proportion of the two dominant modes in each layer and the cross-correlation sharpness of the two dominant modes to screen out the optimal common decomposition layer; and a calculation module for determining the final traveling wave delay estimation result based on the cross-correlation coefficient of the two dominant modes under the optimal common decomposition layer.

[0068] It should be noted that, in the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the system embodiments described above are merely illustrative; for instance, the division of each block is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple blocks may be combined or integrated into another device, or some features may be ignored or not executed. The modules described as separate components may or may not be physically separated. The components shown as modules may be one or more physical units, that is, they may be located in one place or distributed in multiple different places. Some or all of the modules can be selected to achieve the purpose of the solution in this embodiment according to actual needs.

[0069] Furthermore, in the various embodiments of the present invention, the modules can be integrated into one processing unit, or each module can exist physically separately, or two or more modules can be integrated into one unit. The integrated unit described above can be implemented in hardware or as a software functional unit.

[0070] This application also provides an electronic device, which may include one or more processors, memory and communication interfaces.

[0071] The memory, communication interface, and processor are coupled together. For example, the memory, communication interface, and processor can be coupled together via a bus.

[0072] The communication interface is used for data transmission with other devices. The memory stores computer program code. This computer program code includes computer instructions, which, when executed by the processor, cause the electronic device to perform the steps of the aforementioned ripple delay estimation method.

[0073] The processor can be a processor or controller, such as a Central Processing Unit (CPU), a general-purpose processor, a Digital Signal Processor (DSP), an Application-Specific Integrated Circuit (ASIC), a Field-Programmable Gate Array (FPGA), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It can implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with this disclosure. The processor can also be a combination that implements computing functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc. The processor can be used to support an electronic device in performing the method steps provided in the above embodiments.

[0074] The bus can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. These buses can be categorized as address buses, data buses, control buses, etc.

[0075] This application provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the steps of the above-described traveling wave delay estimation method.

[0076] The computer-readable storage media involved in this application include random access memory (RAM), memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disks, removable disks, CD-ROMs, or any other form of storage media known in the art.

[0077] The above are merely preferred embodiments of this application and are not intended to limit this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A method for estimating the time delay of a traveling wave, characterized in that, include: Two current sensing signals located on either side of the fault point are acquired and denoted as the first sensing signal and the second sensing signal, respectively. Adaptive frequency band wavelet deconstruction and detail extraction are performed on the first and second sensing signals to obtain multi-layered first and second reconstructed signal groups corresponding to the scale. For each layer, dynamic parameter mode decomposition based on variational problem solving is performed on each reconstructed signal in the first and second reconstructed signal groups to obtain multiple eigenband mode components with specific sparsity. The first of these eigenband mode components with specific sparsity is taken as the dominant mode of that layer, and the energy proportion of the dominant mode in the reconstructed signal of that layer is calculated. Taking into account the energy ratio of the two principal modes in each layer and the cross-correlation sharpness of the two principal modes, the optimal common decomposition layer is selected; based on the cross-correlation coefficient of the two principal modes under the optimal common decomposition layer, the final traveling wave time delay estimation result is determined.

2. The traveling wave time delay estimation method according to claim 1, characterized in that, After acquiring the two current sensing signals located on both sides of the fault point, the method further includes: performing denoising processing on the first sensing signal and the second sensing signal respectively to obtain a first denoised signal and a second denoised signal; and performing amplitude normalization processing on the first denoised signal and the second denoised signal.

3. The traveling wave time delay estimation method according to claim 1, characterized in that, The method for adaptive frequency band wavelet deconstruction and detail extraction of the first and second sensing signals includes: presetting a maximum decomposition level; sequentially cyclically processing the first and second sensing signals to separate multiple layers of high-frequency detail components and low-frequency approximation components; for each layer, from the first layer to the maximum decomposition level, sequentially setting the low-frequency approximation components to zero, retaining the high-frequency detail components, performing inverse wavelet transform, and reconstructing reconstructed signal groups with detail features at different frequency scales, with the first and second sensing signals respectively denoted as the first reconstructed signal group and the second reconstructed signal group of the corresponding layer.

4. The traveling wave time delay estimation method according to claim 1, characterized in that, When performing dynamic parameter mode decomposition based on variational problem solving, the number of modes and penalty parameters are preset, and the initial center frequency is set in combination with the spectral characteristics of the reconstructed signal itself.

5. The traveling wave time delay estimation method according to claim 1, characterized in that, The method for determining the final traveling wave time delay estimation result includes: calculating the cross-correlation coefficient of the two principal modes under the optimal common decomposition layer, determining the time delay corresponding to the maximum value of the cross-correlation coefficient, and using it as the final traveling wave time delay estimation result.

6. The traveling wave time delay estimation method according to claim 1, characterized in that, When comprehensively considering the energy ratio and cross-correlation sharpness of the two main modes in each layer, the optimal common decomposition layer is selected through dynamic weighting.

7. A traveling wave time delay estimation system, characterized in that, include: The data module is used to acquire two current sensing signals located on both sides of the fault point, which are denoted as the first sensing signal and the second sensing signal, respectively. The extraction module is used to perform adaptive frequency band wavelet deconstruction and detail extraction on the first sensing signal and the second sensing signal, respectively, to obtain a multi-layer first reconstructed signal group and a second reconstructed signal group corresponding to the scale. The decomposition module is used to perform dynamic parameter mode decomposition based on variational problem solving on each reconstructed signal in the first and second reconstructed signal groups for each layer, to obtain multiple intrinsic frequency band mode components with specific sparsity. The master mode module is used to select the first of multiple intrinsic frequency band mode components with specific sparsity as the master mode of the layer, and to calculate the energy proportion of the master mode in the reconstructed signal of the layer. The filtering module is used to comprehensively consider the energy ratio of the two main modes in each layer and the cross-correlation sharpness of the two main modes to select the optimal common decomposition layer. The calculation module is used to determine the final traveling wave time delay estimation result based on the cross-correlation coefficient of the two principal modes under the optimal common decomposition layer.

8. The traveling wave time delay estimation system according to claim 7, characterized in that, After acquiring the two current sensing signals located on both sides of the fault point, the method further includes: performing denoising processing on the first sensing signal and the second sensing signal respectively to obtain a first denoised signal and a second denoised signal; and performing amplitude normalization processing on the first denoised signal and the second denoised signal.

9. An electronic device, characterized in that, include: A memory, one or more processors; the memory is coupled to the processors; wherein the memory stores computer program code, the computer program code including computer instructions, and when the computer instructions are executed by the processor, the electronic device performs the steps of the traveling wave delay estimation method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program that, when executed by a processor, implements the steps of the traveling wave delay estimation method as described in any one of claims 1-7.

Images