Method for capturing weak traveling wave based on energy rate of change kurtosis and phase spectrum order

Abstract

The present application relates to power transmission and transformation equipment monitoring technical field, specifically to a kind of weak traveling wave capture method based on energy change rate kurtosis and phase spectrum order degree, comprising: line acquisition traveling wave signal and are processed to obtain preprocessed signal;Energy mutation event is obtained based on sliding window kurtosis screening;Based on cross-scale phase gradient entropy, it is judged whether fault occurs.For the online detection of traveling wave signal in prior art, the problem of relatively difficult extraction, sliding window kurtosis based on energy change rate and cross-scale phase gradient entropy are introduced twice screening process, by detecting the "kurtosis" of energy change rate instead of energy "amplitude", extremely weak signal mutation can be reliably captured from strong noise, solve the false negative problem caused by low signal-to-noise ratio of traditional method.“Cross-scale phase gradient entropy” criterion, from the internal structure order of signal, can effectively distinguish the fault traveling wave with physical law and random noise, greatly reduce false positive rate.

Description

Technical Field

[0001] This invention relates to the field of power transmission and transformation equipment monitoring technology, specifically to a method for capturing weak traveling waves based on energy change rate kurtosis and phase spectrum order. Background Technology

[0002] In modern power systems, the safe operation of transmission and distribution lines is of paramount importance. Common fault types include high-impedance grounding faults (such as those caused by trees snagging on the ground or broken wires falling to the ground) and intermittent discharges during the early stages of insulation degradation in cables, bushings, and other equipment. These faults share the common characteristic of small fault currents, resulting in extremely weak traveling wave signals with amplitudes often only tens to hundreds of millivolts. These signals are easily drowned out by various background noises on the line (such as switching operations, radio interference, and white noise), resulting in a very low signal-to-noise ratio.

[0003] Existing traveling wave fault detection technologies mostly rely on setting a fixed amplitude or energy threshold for the traveling wave to trigger waveform recording and alarms. This method has significant drawbacks: if the threshold is set too high, weak initial fault signals will be ignored, leading to missed detections and allowing the fault to escalate into a serious accident, missing the optimal opportunity for preventative maintenance; if the threshold is set too low, various noise interferences on the line will frequently trigger the system, resulting in massive amounts of invalid data and false alarms, placing a huge burden on maintenance personnel and reducing system reliability.

[0004] Some improved techniques attempt to incorporate wavelet analysis, extracting energy or modulus maxima features from specific frequency bands for identification. However, at extremely low signal-to-noise ratios, the wavelet coefficient features of faulty signals are also contaminated by noise, making effective differentiation difficult. Other studies introduce information entropy or wavelet entropy as criteria to measure signal complexity. However, these methods are typically used for subsequent analysis of already captured signals and do not fundamentally solve the core challenge of reliably "triggering capture" in a strong noisy environment. Summary of the Invention

[0005] To address the aforementioned problems in existing technologies, a method and system for capturing weak traveling waves based on energy change rate kurtosis and phase spectrum order is provided.

[0006] The specific technical solution is as follows: A weak traveling wave acquisition method based on energy change rate kurtosis and phase spectrum order, comprising: Step S1: Acquiring and processing the traveling wave signal of the line to obtain a preprocessed signal; Step S2: Calculating the sliding window kurtosis of the energy change rate of the preprocessed signal, and filtering out energy mutation events based on the sliding window kurtosis; Step S3: Extracting wavelet coefficient phases from the energy mutation events and calculating the cross-scale phase gradient entropy, and determining whether a fault has occurred based on the cross-scale phase gradient entropy.

[0007] On the other hand, in step S1, a current transformer or a capacitive voltage divider is used to collect data from the line; when the data acquisition device is the current transformer, the traveling wave signal is a current traveling wave signal; when the data acquisition device is the capacitive voltage divider, the traveling wave signal is a voltage traveling wave signal.

[0008] On the other hand, step S2 includes: step S21: calculating the sliding window kurtosis of the preprocessed signal; step S22: comparing the sliding window kurtosis with a preset kurtosis threshold to determine whether an energy mutation occurs; if yes, proceed to step S23; if no, return to step S1 to measure the next sliding window kurtosis; step S23: extracting preset lengths of signals from the forward and backward signals according to the time point at which the energy mutation is detected and constructing the energy mutation event.

[0009] On the other hand, step S21 includes: step S211: calculating the short-time energy sequence using the sliding window method on the preprocessed signal; step S212: calculating the energy change rate by performing first-order difference on the short-time energy sequence to obtain the energy change rate sequence; step S213: performing sliding window based on the energy change rate sequence and calculating the kurtosis to obtain the sliding window kurtosis.

[0010] On the other hand, step S3 includes: step S31: performing continuous wavelet transformation on the signal segment corresponding to the energy mutation event to obtain wavelet coefficients; step S32: extracting phase information from the wavelet coefficients and constructing a wavelet phase matrix; step S33: calculating the cross-scale phase gradient entropy based on the wavelet phase matrix; step S34: comparing the cross-scale phase gradient entropy with the gradient entropy threshold to determine whether a fault has occurred.

[0011] On the other hand, after performing step S3, the method further includes: step S4: extracting and outputting the associated signal segment based on the time point of the fault occurrence.

[0012] A weak traveling wave acquisition system based on energy change rate kurtosis and phase spectrum order is provided for implementing the above-mentioned method. The weak traveling wave acquisition system includes: a signal acquisition module, which acquires and processes traveling wave signals from the line to obtain a preprocessed signal; a kurtosis filtering module connected to the signal acquisition module, which calculates the sliding window kurtosis of the energy change rate on the preprocessed signal and filters out energy mutation events based on the sliding window kurtosis; and a gradient entropy filtering module connected to the kurtosis filtering module, which extracts wavelet coefficient phases from the energy mutation events and calculates cross-scale phase gradient entropy, determining whether a fault has occurred based on the cross-scale phase gradient entropy.

[0013] On the other hand, the kurtosis filtering module includes: a kurtosis calculation module, which calculates the sliding window kurtosis of the preprocessed signal; a kurtosis comparison module, which is connected to the kurtosis calculation module; the kurtosis comparison module compares the sliding window kurtosis with a preset kurtosis threshold to determine whether an energy mutation occurs; and an event construction module, which is connected to the kurtosis comparison module; the event construction module extracts a preset length of signal from the forward and backward signals according to the time point at which the energy mutation is detected and constructs the energy mutation event.

[0014] On the other hand, the kurtosis calculation module includes: an energy calculation module, which calculates a short-time energy sequence using a sliding window method on the preprocessed signal; a difference calculation module, which is connected to the energy calculation module; the difference calculation module performs first-order difference calculation on the short-time energy sequence to obtain an energy change rate sequence; and a kurtosis generation module, which is connected to the difference calculation module; the kurtosis generation module performs a sliding window calculation based on the energy change rate sequence to obtain the sliding window kurtosis.

[0015] On the other hand, the gradient entropy screening module includes: a wavelet transform module, which performs continuous wavelet transform on the signal segment corresponding to the energy mutation event to obtain wavelet coefficients; a phase calculation module, which is connected to the wavelet transform module; the phase calculation module extracts phase information from the wavelet coefficients and constructs a wavelet phase matrix; a gradient entropy calculation module, which is connected to the phase calculation module; the gradient entropy calculation module calculates the cross-scale phase gradient entropy based on the wavelet phase matrix; and a comparison module, which compares the cross-scale phase gradient entropy with a gradient entropy threshold to determine whether a fault has occurred.

[0016] The above technical solution has the following advantages or beneficial effects: Addressing the relatively difficult problem of online detection and extraction of traveling wave signals in existing technologies, a two-stage screening process based on the sliding window kurtosis of the energy change rate and cross-scale phase gradient entropy is introduced. By detecting the "kurtosis" of the energy change rate rather than the energy "amplitude," extremely weak signal abrupt changes can be reliably captured from strong noise, solving the problem of missed detection caused by the low signal-to-noise ratio of traditional methods. The "cross-scale phase gradient entropy" criterion, based on the inherent structural order of the signal, can effectively distinguish between physically regular fault traveling waves and random noise, greatly reducing the false alarm rate. Attached Figure Description

[0017] Embodiments of the invention will be described more fully with reference to the accompanying drawings. However, the drawings are for illustration and explanation only and do not constitute a limitation on the scope of the invention.

[0018] Figure 1 This is an overall schematic diagram of an embodiment of the present invention;

[0019] Figure 2 This is a schematic diagram of step S2 in an embodiment of the present invention;

[0020] Figure 3 This is a schematic diagram of step S21 in an embodiment of the present invention;

[0021] Figure 4 This is a schematic diagram of step S3 in an embodiment of the present invention;

[0022] Figure 5 This is a schematic diagram of step S4 in an embodiment of the present invention;

[0023] Figure 6 This is a schematic diagram of the system in an embodiment of the present invention;

[0024] Figure 7 This is a schematic diagram of the kurtosis filtering module in an embodiment of the present invention;

[0025] Figure 8 This is a schematic diagram of the kurtosis calculation module in an embodiment of the present invention;

[0026] Figure 9 This is a schematic diagram of the gradient entropy filtering module in an embodiment of the present invention. Detailed Implementation

[0027] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] It should be noted that, unless otherwise specified, the embodiments and features described in the present invention can be combined with each other.

[0029] The present invention will be further described below with reference to the accompanying drawings and specific embodiments, but this is not intended to limit the scope of the invention.

[0030] This invention includes: a method for capturing weak traveling waves based on the kurtosis of the rate of energy change and the order of the phase spectrum, such as... Figure 1 As shown, the process includes: Step S1: Acquiring and processing traveling wave signals from the line to obtain a preprocessed signal; Step S2: Calculating the sliding window kurtosis of the energy change rate of the preprocessed signal, and filtering out energy mutation events based on the sliding window kurtosis; Step S3: Extracting wavelet coefficient phases from the energy mutation events and calculating the cross-scale phase gradient entropy, and determining whether a fault has occurred based on the cross-scale phase gradient entropy.

[0031] Specifically, addressing the challenge of online detection and extraction of traveling wave signals in existing technologies, this paper introduces a two-stage screening process based on the sliding window kurtosis of the energy change rate and cross-scale phase gradient entropy. By detecting the "kurtosis" of the energy change rate rather than the energy "amplitude," extremely weak signal abrupt changes can be reliably captured from strong noise, solving the problem of missed detection caused by low signal-to-noise ratio in traditional methods. The "cross-scale phase gradient entropy" criterion, based on the inherent structural order of the signal, can effectively distinguish between physically regular fault traveling waves and random noise, greatly reducing the false alarm rate.

[0032] In actual implementation, the above technical solution is integrated into a monitoring system as a software embodiment. The physical hardware of this monitoring system consists of two parts: sensors and computer equipment for processing. However, in actual product deployment, the above structure can be arbitrarily integrated or split as needed. For example, the computer equipment can be split into an edge computing component and a remote control platform, or the above modules can be integrated into a monitoring and early warning device, etc., without affecting the implementation of the solution.

[0033] The line in question is an operating high-voltage transmission line. Depending on monitoring needs, the traveling wave signal can be identified as a voltage traveling wave signal or a current traveling wave signal. Its essence is to characterize the sudden energy changes on the line caused by a fault event; therefore, the specific signal type will not affect the implementation of the scheme. When the acquisition device is a current transformer, the traveling wave signal is a current traveling wave signal; when the acquisition device is a capacitive voltage divider, the traveling wave signal is a voltage traveling wave signal. Therefore, in step S1, a current transformer or a capacitive voltage divider is used to acquire data from the line. This signal may be pure background noise or it may contain actual traveling wave signals.

[0034] After acquiring the raw signal, a digital filtering process is performed on it. Common digital filtering processes include low-pass filtering, dark baseline subtraction of the sampling device, and Hanning window filtering. These filtering operations can remove most of the background noise, but have little impact on traveling wave signals with abrupt energy changes in time-structure.

[0035] Subsequently, for the preprocessed signals, this scheme performs kurtosis-based screening based on the rate of energy change. Since the traveling wave signal is an isolated spike in energy and its data distribution is extremely sharp, energy mutation events are identified through kurtosis screening as relatively suspicious fault occurrence points, and then the corresponding signal segments are extracted for secondary discrimination.

[0036] The secondary discrimination is transformed into extracting the wavelet signal phase based on wavelet coefficient transformation, and constructing a wavelet phase signal matrix. This matrix represents the traveling wave signal at the signal structure level. The inventors believe that fault traveling waves, as a structured signal, exhibit interconnected and ordered phase changes at different scales (frequency). In contrast, the phase of noise is completely random and uncorrelated at all scales. Therefore, cross-scale phase gradient entropy can effectively measure traveling wave signals.

[0037] In one embodiment, such as Figure 2 As shown, step S2 includes: step S21: calculate the sliding window kurtosis of the preprocessed signal; step S22: compare the sliding window kurtosis with a preset kurtosis threshold to determine whether an energy mutation occurs; if yes, proceed to step S23; if no, return to step S1 to measure the next sliding window kurtosis; step S23: extract a preset length of signal from the forward signal and the backward signal respectively according to the time point when the energy mutation is detected, and construct an energy mutation event.

[0038] Among them, such as Figure 3 As shown, step S21 includes: step S211: calculating the short-time energy sequence using the sliding window method on the preprocessed signal; step S212: calculating the energy change rate by performing first-order difference on the short-time energy sequence to obtain the energy change rate sequence; step S213: performing sliding window based on the energy change rate sequence and calculating the kurtosis to obtain the sliding window kurtosis.

[0039] Specifically, to achieve a better filtering process, in this embodiment, the sliding window kurtosis is first calculated on the preprocessed signal of the digital filter. Specifically, the kurtosis calculation process includes:

[0040] The sliding window method is used to calculate the short-time energy sequence of the input preprocessed signal. In actual online detection, this sliding window method typically involves the preceding digital filtering process processing a segment of the filtered signal, which is then continuously fed into a buffer queue. At the tail of the queue is a buffer with a length of... The sliding window extracts the signal at the current bit position and calculates the short-time energy sequence. :

[0041] ;

[0042] In the formula, It is a short-time energy sequence. For the first The energy amplitude of each sampling point For the current calculation time point, The window length is typically set to be greater than the typical width of a traveling wave front.

[0043] Then, for short-time energy sequences Performing a first-order difference yields the energy change rate sequence. To amplify the abrupt changes in the signal:

[0044] .

[0045] In the formula, The current time point Short-time energy sequence, This is a short-time energy sequence from the previous time point.

[0046] Finally, the kurtosis of the energy change rate sequence within the sliding window is calculated:

[0047]

[0048] In the formula, For ravine, It is the length of the sliding window used to calculate kurtosis. It is inside the window The mean, This corresponds to the time value captured by the kurtosis sliding window.

[0049] Generally speaking, fault traveling waves cause It is an isolated spike with an extremely sharp data distribution and a kurtosis value. It will be very high (far greater than 3). And caused by background noise The fluctuations are relatively random and mild, and the data distribution is closer to a Gaussian distribution with a kurtosis value around 3.

[0050] Based on this characteristic, a kurtosis threshold can be constructed for comparison; for example, a low, fixed kurtosis threshold can be set. (For example, When detected When an energy mutation event is considered to have occurred, a data segment of a certain length (e.g., 512 sampling points) before and after that moment is defined as a "candidate signal segment" of the energy mutation event and sent to the second layer of screening.

[0051] In one embodiment, such as Figure 4 As shown, step S3 includes: step S31: performing continuous wavelet transformation on the signal segment corresponding to the energy mutation event to obtain wavelet coefficients; step S32: extracting phase information from the wavelet coefficients and constructing a wavelet phase matrix; step S33: calculating the cross-scale phase gradient entropy based on the wavelet phase matrix; step S34: comparing the cross-scale phase gradient entropy with the gradient entropy threshold to determine whether a fault has occurred.

[0052] Specifically, to effectively screen energy mutation events, in this embodiment, continuous wavelet transform is first performed on the signal segment corresponding to the energy mutation event to obtain wavelet coefficients. :

[0053]

[0054] in, It is a candidate signal segment. It is the mother wavelet (Morlet wavelet is recommended, as it performs excellently in time-frequency localization). It is a scale factor. It is the time shift factor. Wavelet coefficients It is a complex number that contains amplitude and phase information.

[0055] Subsequently, the phase information of the wavelet coefficients is extracted to construct a wavelet phase matrix. .

[0056] Finally, the cross-scale phase gradient entropy of the wavelet phase matrix is ​​calculated, including:

[0057] First, calculate the wavelet phase matrix along the scale axis. The gradient is used to obtain the phase gradient matrix. :

[0058] ;

[0059] Then, on the timeline Above, for the phase gradient matrix The probability distribution of the values ​​in each column (i.e., at each time point) is statistically analyzed. The gradient value range is then determined. Divided into Each small interval is counted, and the statistics fall within each interval. The number of gradient values ​​within a given range, and their probabilities are calculated. .

[0060] Then calculate each time point "Instantaneous phase gradient entropy" :

[0061] ;

[0062] Finally, the average instantaneous entropy of the entire candidate signal segment is calculated to obtain the final cross-scale phase gradient entropy. :

[0063] .

[0064] in, It is the length of the candidate signal segment.

[0065] For fault traveling waves, their cross-scale phase relationships exhibit regularity, leading to phase gradients. The value of entropy will be concentrated in a few intervals, and the probability distribution is uneven, resulting in a calculated entropy value. It will be very low. For noise, its phase is completely random, and the phase gradient... The value of will be uniformly distributed across all intervals, and the probability distribution is close to a uniform distribution. The calculated entropy value It will be very high.

[0066] To address this, a phase entropy threshold is set. This threshold is also insensitive because it distinguishes between the two essential states of "order" and "randomness." If the calculated... If the candidate signal segment exhibits structured morphological characteristics, it is considered a fault traveling wave. Otherwise, it is classified as noise.

[0067] When a faulty traveling wave is confirmed, a fault is considered to have occurred, and a fault warning signal is issued.

[0068] In one embodiment, such as Figure 5 As shown, after executing step S3, the following step is also included: step S4: extract and output the associated signal segment based on the time point of the fault occurrence.

[0069] Specifically, to facilitate fault location, in this embodiment, when a fault is determined to occur, associated signal segments are extracted and output according to the time point of the fault occurrence. Signals of preset lengths are extracted forward and backward according to the time point of the fault occurrence as associated signal segments, and timestamps are added for subsequent fault location, type identification and equipment status assessment.

[0070] A weak traveling wave capture system based on energy change rate kurtosis and phase spectrum order is used to implement the above method; such as Figure 6 As shown, the weak traveling wave acquisition system includes: a signal acquisition module 1, which acquires and processes traveling wave signals from the line to obtain a preprocessed signal; a kurtosis filtering module 2, which is connected to the signal acquisition module 1; the kurtosis filtering module 2 calculates the sliding window kurtosis of the energy change rate of the preprocessed signal and filters out energy mutation events based on the sliding window kurtosis; and a gradient entropy filtering module 3, which is connected to the kurtosis filtering module 2; the gradient entropy filtering module 3 extracts wavelet coefficient phases from the energy mutation events and calculates the cross-scale phase gradient entropy, and determines whether a fault has occurred based on the cross-scale phase gradient entropy.

[0071] Specifically, addressing the challenge of online detection and extraction of traveling wave signals in existing technologies, this paper introduces a two-stage screening process based on the sliding window kurtosis of the energy change rate and cross-scale phase gradient entropy. By detecting the "kurtosis" of the energy change rate rather than the energy "amplitude," extremely weak signal abrupt changes can be reliably captured from strong noise, solving the problem of missed detection caused by low signal-to-noise ratio in traditional methods. The "cross-scale phase gradient entropy" criterion, based on the inherent structural order of the signal, can effectively distinguish between physically regular fault traveling waves and random noise, greatly reducing the false alarm rate.

[0072] In actual implementation, the above technical solution is integrated into a monitoring system as a software embodiment. The physical hardware of this monitoring system consists of two parts: sensors and computer equipment for processing. However, in actual product deployment, the above structure can be arbitrarily integrated or split as needed. For example, the computer equipment can be split into an edge computing component and a remote control platform, or the above modules can be integrated into a monitoring and early warning device, etc., without affecting the implementation of the solution.

[0073] The line in question is an operating high-voltage transmission line. Depending on monitoring needs, the traveling wave signal can be identified as a voltage traveling wave signal or a current traveling wave signal. Its essence is to characterize the sudden energy changes on the line caused by a fault event; therefore, the specific signal type will not affect the implementation of the scheme. When the acquisition device is a current transformer, the traveling wave signal is a current traveling wave signal; when the acquisition device is a capacitive voltage divider, the traveling wave signal is a voltage traveling wave signal. Therefore, in step S1, a current transformer or a capacitive voltage divider is used to acquire data from the line. This signal may be pure background noise or it may contain actual traveling wave signals.

[0074] After acquiring the raw signal, a digital filtering process is performed on it. Common digital filtering processes include low-pass filtering, dark baseline subtraction of the sampling device, and Hanning window filtering. These filtering operations can remove most of the background noise, but have little impact on traveling wave signals with abrupt energy changes in time-structure.

[0075] Subsequently, for the preprocessed signals, this scheme performs kurtosis-based screening based on the rate of energy change. Since the traveling wave signal is an isolated spike in energy and its data distribution is extremely sharp, energy mutation events are identified through kurtosis screening as relatively suspicious fault occurrence points, and then the corresponding signal segments are extracted for secondary discrimination.

[0076] The secondary discrimination is transformed into extracting the wavelet signal phase based on wavelet coefficient transformation, and constructing a wavelet phase signal matrix. This matrix represents the traveling wave signal at the signal structure level. The inventors believe that fault traveling waves, as a structured signal, exhibit interconnected and ordered phase changes at different scales (frequency). In contrast, the phase of noise is completely random and uncorrelated at all scales. Therefore, cross-scale phase gradient entropy can effectively measure traveling wave signals.

[0077] In one embodiment, such as Figure 7 As shown, the kurtosis filtering module 2 includes: a kurtosis calculation module 21, which calculates the sliding window kurtosis of the preprocessed signal; a kurtosis comparison module 22, which is connected to the kurtosis calculation module 21; the kurtosis comparison module 22 compares the sliding window kurtosis with a preset kurtosis threshold to determine whether an energy mutation occurs; and an event construction module 23, which is connected to the kurtosis comparison module 22; the event construction module 23 extracts a preset length of signal from the forward and backward signals according to the time point when the energy mutation is detected and constructs an energy mutation event.

[0078] In one embodiment, such as Figure 8 As shown, the kurtosis calculation module 21 includes: an energy calculation module 211, which calculates a short-time energy sequence using a sliding window method on the preprocessed signal; a difference calculation module 212, which is connected to the energy calculation module 211; the difference calculation module 212 performs first-order difference calculation on the short-time energy sequence to obtain an energy change rate sequence; and a kurtosis generation module 213, which is connected to the difference calculation module 212; the kurtosis generation module 213 performs a sliding window calculation based on the energy change rate sequence to obtain the sliding window kurtosis.

[0079] In one embodiment, such as Figure 9 As shown, the gradient entropy screening module 3 includes: a wavelet transform module 31, which performs continuous wavelet transformation on the signal segment corresponding to the energy mutation event to obtain wavelet coefficients; a phase calculation module 32, which is connected to the wavelet transform module 31; the phase calculation module 32 extracts phase information from the wavelet coefficients and constructs a wavelet phase matrix; a gradient entropy calculation module 33, which is connected to the phase calculation module 32; the gradient entropy calculation module 33 calculates the cross-scale phase gradient entropy based on the wavelet phase matrix; and a comparison module 34, which compares the cross-scale phase gradient entropy with the gradient entropy threshold to determine whether a fault has occurred.

[0080] Compared with the prior art, the present invention has the following significant advantages:

[0081] 1. By detecting the "kurtosis" of the rate of energy change rather than the "amplitude" of energy, extremely weak signal abrupt changes can be reliably captured from strong noise, solving the problem of missed detection caused by low signal-to-noise ratio in traditional methods.

[0082] 2. The original "cross-scale phase gradient entropy" criterion, based on the inherent structural order of the signal, can effectively distinguish between fault traveling waves with physical laws and random noise, greatly reducing the false alarm rate.

[0083] 3. The thresholds corresponding to the core criteria (kurtosis and phase entropy) of this method are based on the statistical and structural characteristics of the signal and are not sensitive to operating conditions and background noise levels, thus overcoming the problem that traditional thresholds require repeated manual tuning.

[0084] 4. The algorithms involved in this invention (differential calculation, kurtosis calculation, CWT calculation, entropy calculation) have moderate computational complexity and can be efficiently implemented on the embedded platform (such as DSP, FPGA) of existing power system protection and monitoring devices without increasing expensive hardware costs, thus having strong practicality.

[0085] Those skilled in the art will understand that various aspects, or possible implementations of various aspects, of the present invention can be embodied as systems, methods, or computer program products. Therefore, various aspects, or possible implementations of various aspects, of the present invention can take the form of entirely hardware embodiments, entirely software embodiments (including firmware, resident software, etc.), or embodiments combining software and hardware aspects, all collectively referred to herein as "circuit," "module," or "system." Furthermore, various aspects, or possible implementations of various aspects, of the present invention can take the form of computer program products, which are computer instructions stored in memory.

[0086] The memory can be a computer-readable signal medium or a computer-readable storage medium. Computer-readable storage media include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or apparatuses, or any suitable combination thereof, such as random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, and portable read-only memory (CD-ROM).

[0087] A processor in a computer reads computer instructions stored in memory, enabling the processor to execute the functional actions specified in each step or combination of steps in a flowchart; and to generate means for implementing the functional actions specified in each block or combination of blocks in a flowchart.

[0088] It should be understood that a processor in a computer can be understood as one or more application-specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), complex programmable logic devices (CPLDs), field-programmable gate arrays (FPGAs), general-purpose processors, controllers, microcontrollers (MCUs), microprocessors, or other electronic components used to execute the aforementioned computer instructions.

[0089] Computer instructions may be executed entirely on the user's local computer, partially on the user's local computer, as a separate software package, partially on the user's local computer and partially on a remote computer, or entirely on a remote computer or server. It should also be noted that in some alternative implementations, the functions indicated by the steps in the flowchart or the blocks in the block diagram may not occur in the order shown in the diagram. For example, depending on the functions involved, two consecutive steps or blocks may actually be executed approximately simultaneously, or these blocks may sometimes be executed in reverse order.

[0090] Of course, in practical applications, the various components of a computer system are coupled together through a bus system. The bus system is used to enable communication and connection between these components. In addition to the data bus, the bus system also includes a power bus, a control bus, and a status signal bus.

[0091] The above are merely preferred embodiments of the present invention and are not intended to limit the implementation methods and protection scope of the present invention. Those skilled in the art should recognize that any equivalent substitutions and obvious changes made based on the description and illustrations of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for capturing weak traveling waves based on the kurtosis of the rate of energy change and the order of the phase spectrum, characterized in that, include: Step S1: Acquire traveling wave signals from the line and process them to obtain preprocessed signals; Step S2: Calculate the sliding window kurtosis of the energy change rate for the preprocessed signal, and filter out energy mutation events based on the sliding window kurtosis; Step S3: Extract the phase of wavelet coefficients for the energy mutation event and calculate the cross-scale phase gradient entropy. Determine whether a fault has occurred based on the cross-scale phase gradient entropy. Step S3 includes: Step S31: Perform continuous wavelet transform on the signal segment corresponding to the energy mutation event to obtain wavelet coefficients; Step S32: Extract phase information from the wavelet coefficients and construct the wavelet phase matrix; Step S33: Calculate the cross-scale phase gradient entropy based on the wavelet phase matrix; Step S34: Compare the cross-scale phase gradient entropy with the gradient entropy threshold to determine whether a fault has occurred; In step S33, the calculation process of the cross-scale phase gradient entropy includes: First, calculate the wavelet phase matrix along the scale axis. The gradient is used to obtain the phase gradient matrix; Then, on the timeline Above, probability distribution statistics are performed on the values ​​of each column of the phase gradient matrix, including the gradient value range. Divided into Each small interval is counted, and the statistics fall within each interval. Calculate the number of gradient values ​​within a given range and their probabilities. Then calculate the instantaneous phase gradient entropy at each time point; Finally, the instantaneous phase gradient entropy of the entire candidate signal segment is calculated.

2. The weak traveling wave acquisition method according to claim 1, characterized in that, In step S1, a current transformer or a capacitor divider is used to collect data from the line. When the acquisition device is the current transformer, the traveling wave signal is a current traveling wave signal; When the acquisition device is the capacitor voltage divider, the traveling wave signal is a voltage traveling wave signal.

3. The weak traveling wave acquisition method according to claim 1, characterized in that, Step S2 includes: Step S21: Calculate the sliding window kurtosis for the preprocessed signal; Step S22: Compare the sliding window kurtosis with a preset kurtosis threshold to determine whether an energy mutation occurs; If so, proceed to step S23; If not, return to step S1 to measure the kurtosis of the next sliding window; Step S23: Based on the time point at which the energy mutation is detected, extract a preset length of signal from the forward and backward signals respectively and construct the energy mutation event.

4. The weak traveling wave acquisition method according to claim 3, characterized in that, Step S21 includes: Step S211: Calculate the short-time energy sequence of the preprocessed signal using the sliding window method; Step S212: Perform first-order difference calculation on the short-time energy sequence to obtain the energy change rate sequence; Step S213: Perform a sliding window based on the energy change rate sequence and calculate the kurtosis to obtain the sliding window kurtosis.

5. The weak traveling wave acquisition method according to claim 1, characterized in that, After performing step S3, the procedure further includes: Step S4: Extract and output the associated signal segment based on the time point of the fault occurrence.

6. A weak traveling wave acquisition system based on energy change rate kurtosis and phase spectrum order, characterized in that, Used to implement the method as described in any one of claims 1-5; The weak traveling wave acquisition system includes: The signal acquisition module acquires traveling wave signals from the line and processes them to obtain a preprocessed signal. A kurtosis filtering module, wherein the kurtosis filtering module is connected to the signal acquisition module; The kurtosis filtering module calculates the sliding window kurtosis of the energy change rate of the preprocessed signal, and obtains energy mutation events based on the sliding window kurtosis. Gradient entropy filtering module, wherein the gradient entropy filtering module is connected to the kurtosis filtering module; The gradient entropy filtering module extracts the wavelet coefficient phase of the energy mutation event and calculates the cross-scale phase gradient entropy, and determines whether a fault has occurred based on the cross-scale phase gradient entropy.

7. The weak traveling wave acquisition system according to claim 6, characterized in that, The kurtosis filtering module includes: A kurtosis calculation module calculates the sliding window kurtosis on the preprocessed signal; A kurtosis comparison module, which is connected to the kurtosis calculation module; The kurtosis comparison module compares the sliding window kurtosis with a preset kurtosis threshold to determine whether an energy mutation occurs. An event construction module, which is connected to the kurtosis comparison module; The event construction module extracts a preset length of signal from the forward and backward signals based on the time point at which the energy mutation is detected, and constructs the energy mutation event.

8. The weak traveling wave acquisition system according to claim 7, characterized in that, The kurtosis calculation module includes: An energy calculation module, wherein the energy calculation module uses a sliding window method to calculate a short-time energy sequence from the preprocessed signal; A differential calculation module, which is connected to the energy calculation module; The difference calculation module performs first-order difference calculation on the short-time energy sequence to obtain the energy change rate sequence; A kurtosis generation module, which is connected to the difference calculation module; The kurtosis generation module performs a sliding window operation based on the energy change rate sequence and calculates the kurtosis to obtain the sliding window kurtosis.

9. The weak traveling wave acquisition system according to claim 6, characterized in that, The gradient entropy filtering module includes: The wavelet transform module performs continuous wavelet transformation on the signal segment corresponding to the energy mutation event to obtain wavelet coefficients. A phase calculation module, which is connected to the wavelet transform module; The phase calculation module extracts phase information from the wavelet coefficients and constructs a wavelet phase matrix. A gradient entropy calculation module, which is connected to the phase calculation module; The gradient entropy calculation module calculates the cross-scale phase gradient entropy based on the wavelet phase matrix. The comparison module compares the cross-scale phase gradient entropy with a gradient entropy threshold to determine whether a fault has occurred.

Images