A Method and System for Enhancing Line Partial Discharge Signals Based on Wavelet Transform

Abstract

This application relates to the field of signal enhancement technology, specifically to a method and system for enhancing partial discharge signals on transmission lines based on wavelet transform. The method includes: acquiring partial discharge signals on transmission lines; calculating the pulse saliency at each time moment; extracting power frequency phase information from the partial discharge signals, calculating the pulse specificity at each time moment, obtaining pulse evaluation values ​​at each time moment, and filtering out abnormal times; decomposing the partial discharge signals into detail coefficients of multiple decomposition layers using wavelet transform; determining the detail level of each detail coefficient at each decomposition layer to adjust the wavelet threshold of each decomposition layer, processing the detail coefficients of each decomposition layer, and reconstructing the signal using inverse wavelet transform to obtain the enhanced partial discharge signal. This application effectively suppresses noise interference while preserving and enhancing the characteristic information of the real partial discharge to the maximum extent, significantly improving the signal-to-noise ratio of the enhanced partial discharge signal.

Description

Technical Field

[0001] This application relates to the field of signal enhancement technology, specifically to a method and system for enhancing line partial discharge signals based on wavelet transform. Background Technology

[0002] Partial discharge (PD) is a microscopic discharge phenomenon in localized areas of the insulation material of electrical equipment. It is a sign of insulation aging and deterioration. In transmission lines, the detection of partial discharge is crucial for assessing insulation condition and preventing faults. Wavelet transform, due to its excellent time-frequency localization characteristics, is widely used in signal denoising to enhance partial discharge signals.

[0003] The denoising performance of wavelet transform algorithms depends on the choice of threshold. However, partial discharge signals on transmission lines are affected by periodic pulse interference synchronized with the power frequency. The transient characteristics of this type of interference are highly similar to those of real partial discharge pulses, making it difficult for wavelet transform to effectively distinguish real PD pulses from periodic electromagnetic interference pulses with similar transient characteristics at the signal level. As a result, in complex and noisy backgrounds, if the threshold is set too small, periodic interference pulses with high amplitude transient characteristics will be mistakenly retained as signal components, resulting in insufficient improvement in the signal-to-noise ratio of the reconstructed signal. If the threshold is set too large, real partial discharge pulses will be over-filtered out, causing significant signal distortion and loss of effective information, thereby reducing the accuracy and reliability of PD detection. Summary of the Invention

[0004] To address the aforementioned technical problems, a method and system for enhancing line partial discharge signals based on wavelet transform are provided to resolve existing issues.

[0005] The solution to the technical problem in this application is to provide a method and system for enhancing line partial discharge signals based on wavelet transform, including the following steps:

[0006] In a first aspect, embodiments of this application provide a method for enhancing line partial discharge signals based on wavelet transform, the method comprising the following steps:

[0007] Partial discharge signals from transmission lines are collected; the deviation of signal amplitude at different times and the trend of signal amplitude variation in the neighborhood are analyzed, and the pulse saliency at each time is calculated; power frequency phase information is extracted from the partial discharge signals, the periodicity characteristics of pulse saliency at different times of the same phase in the partial discharge signals are analyzed, the pulse specificity at each time is calculated, and the pulse evaluation value at each time is obtained by combining the pulse saliency, and abnormal times are screened out; the partial discharge signals are decomposed into detail coefficients of multiple decomposition layers using wavelet transform; for each detail coefficient, the corresponding mapping point is determined in the adjacent decomposition layer to which it belongs, and the average level of pulse saliency at abnormal times in the corresponding time period after each detail coefficient is mapped back to the original partial discharge signal is analyzed; the detail of each detail coefficient under each decomposition layer is determined by combining the detail coefficient of each detail coefficient and its corresponding mapping point, as well as the number of decomposition layers, so as to adjust the wavelet threshold of each decomposition layer, process the detail coefficients of each decomposition layer, and use inverse wavelet transform to reconstruct the signal to obtain the enhanced partial discharge signal.

[0008] Preferably, the calculation of the impulse saliency at each time point includes:

[0009] Calculate the mean value of the signal amplitude at all times within the partial discharge signal, and record it as the reference amplitude; calculate the difference between the signal amplitude at each time within the partial discharge signal and the reference amplitude, and record it as the relative deviation;

[0010] Calculate the sum of the absolute values ​​of the rate of change of the signal amplitude between any two adjacent time points within the neighborhood of each time point;

[0011] The pulse significance is the product of the relative deviation and the summation.

[0012] Preferably, the extraction of power frequency phase information includes: extracting a power frequency reference signal from the partial discharge signal using a phase-locked loop algorithm, dividing the power frequency reference signal into each power frequency period, and obtaining the power frequency phase of the partial discharge signal at each moment within each power frequency period.

[0013] Preferably, the calculation of the pulse specificity at each time point includes:

[0014] For any moment within each power frequency cycle of the partial discharge signal, select the moment within all other power frequency cycles that has the same power frequency phase as the stated moment and record it as the reference moment;

[0015] The difference in pulse significance between any given time point and each control time point is denoted as the relative difference; the ratio between the relative difference and the pulse significance at each control time point is denoted as the relative ratio.

[0016] The pulse specificity is the sum of the relative comparisons between any given time and all control times.

[0017] Preferably, the pulse evaluation value is the product of pulse specificity and pulse saliency.

[0018] Preferably, the step of filtering out abnormal moments includes: using the average pulse evaluation value of all moments within the partial discharge signal as a segmentation threshold; and selecting moments with pulse evaluation values ​​greater than or equal to the segmentation threshold as abnormal moments.

[0019] Preferably, the process of obtaining the mapping point is as follows: two decomposition layers adjacent to each decomposition layer are denoted as neighboring layers; any detail coefficient under each decomposition layer is time-mapped in each neighboring layer, and the point in each neighboring layer that is closest to the position of the any detail coefficient is selected and denoted as the mapping point.

[0020] Preferably, determining the level of detail of each detail coefficient under each decomposition layer includes:

[0021] The detail coefficients under each decomposition layer are time-mapped in the partial discharge signal to obtain the specific time period corresponding to the partial discharge signal for each detail coefficient under each decomposition layer, and the average value of the pulse evaluation value of all abnormal moments included in the specific time period is calculated.

[0022] The reciprocal of the difference between the number of the decomposition layer to which each detail coefficient belongs and the preset value is used as the weight factor; the detail coefficients of each decomposition layer and the corresponding detail coefficients of all mapping points are weighted and summed using the weight factor as the weight;

[0023] The level of detail is the normalized result of the product between the average value and the weighted summation.

[0024] Preferred, the first Each decomposition layer corresponds to an adjusted wavelet threshold, including: ,and ,in, For the first Each decomposition layer corresponds to an adjusted wavelet threshold. For the first Intermediate variables corresponding to each decomposition layer; For the first Each decomposition layer corresponds to the wavelet threshold before adjustment. For the first The first decomposition layer A detailed coefficient, For the first The first decomposition layer The level of detail corresponding to each detail coefficient For the first The number of all detail coefficients in each decomposition layer.

[0025] Secondly, embodiments of this application also provide a line partial discharge signal enhancement system based on wavelet transform, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of any of the above-described wavelet transform-based line partial discharge signal enhancement methods.

[0026] This application has at least the following beneficial effects:

[0027] This application analyzes the deviation of signal amplitude from the overall level in partial discharge signals and the steepness of signal amplitude changes within the neighborhood, calculating the pulse salience at each moment. Its advantage lies in considering abrupt changes in signal amplitude, enabling sensitive capture of all transient events with pulse morphological characteristics. It also calculates the pulse specificity at each moment, considering the repetition of pulse salience at a fixed phase in each power frequency cycle, reflecting the specificity of pulse phenomena at different times. This assesses the randomness of real PD pulses, identifies periodic pulse interference noise, and improves the detection capability of real partial discharge signals. Finally, it obtains pulse evaluation values ​​at each moment and filters out abnormal moments. Its advantage lies in comprehensively assessing the probability of partial discharge phenomena occurring at different times in the partial discharge signal, filtering out abnormal moments to extract real partial discharge events. Wavelet transform is used to decompose the partial discharge signal into detail coefficients of multiple decomposition layers. The detail level of each detail coefficient in each decomposition layer is determined. The beneficial effect is that it takes into account the abnormal time when each detail coefficient contains PD pulses in the original signal period, as well as the amplitude of each detail coefficient in adjacent decomposition layers. It reflects the feature significance of each detail coefficient at different scales and evaluates the possibility that the detail coefficient represents the detailed information corresponding to the real partial discharge event. The wavelet threshold of each decomposition layer is adjusted to process the detail coefficients of each decomposition layer. The signal is then reconstructed using inverse wavelet transform to obtain the enhanced partial discharge signal. The beneficial effect is that by dynamically adjusting the threshold through the multi-scale features of the signal, noise interference is effectively suppressed while preserving and enhancing the feature information of the real partial discharge to the maximum extent, and the signal-to-noise ratio of the enhanced partial discharge signal is significantly improved. Attached Figure Description

[0028] The wavelet transform-based method for enhancing partial discharge signals of lines according to this application will be further described in detail below with reference to the accompanying drawings.

[0029] Figure 1 A flowchart illustrating the steps of a wavelet transform-based method for enhancing partial discharge signals in a line, as provided in an embodiment of this application.

[0030] Figure 2 This is a flowchart illustrating the steps of a method for obtaining the detail level of each detail coefficient under each decomposition layer, as provided in an embodiment of this application. Detailed Implementation

[0031] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description of the wavelet transform-based method and system for enhancing partial discharge signals in lines, in conjunction with the accompanying drawings and implementation examples, is provided. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0032] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains.

[0033] Please see Figure 1 The diagram illustrates a flowchart of a wavelet transform-based method for enhancing partial discharge signals in a line, according to an embodiment of this application. The method includes the following steps:

[0034] Step 1: Collect partial discharge signals on the transmission line.

[0035] In power systems, transmission lines play a vital role as the transmission medium. Their safe and stable operation is crucial for ensuring the normal operation of the power system. During long-term operation, various reasons can lead to insulation defects, causing partial discharge phenomena on the transmission lines. This accelerates insulation aging and ultimately affects the insulation performance of the transmission lines, thus preventing them from transmitting power normally.

[0036] Secondly, partial discharge refers to the electrical discharge phenomenon that occurs in localized areas of the insulation material on a transmission line. This phenomenon is insufficient to cause immediate breakdown of the insulation, but its long-term presence can lead to deterioration of the insulation material, potentially resulting in complete insulation failure. In other words, partial discharge is an early signal of cable insulation aging and damage. When partial discharge begins to occur in localized areas of the insulation on a transmission line, the insulation layer has not yet broken down; only a localized electrical bridge is formed, allowing the line to continue operating normally. This constitutes the partial discharge phenomenon. Therefore, it is necessary to detect partial discharge phenomena on the line.

[0037] Based on the above analysis, a high-frequency current transformer is used to collect partial discharge signals on the line. The frequency of the signal collection is 100MHz. As another implementation method, the implementer can set it according to the actual situation. Secondly, the maximum and minimum value normalization method is used to normalize the signal amplitude at all times in the partial discharge signal. The maximum and minimum value normalization method is a well-known technology and will not be described in detail here.

[0038] Thus, the partial discharge signal on the transmission line was obtained.

[0039] Step 2: Analyze the deviation of signal amplitude at each moment in the partial discharge signal and the trend of signal amplitude change in the neighborhood, and calculate the pulse saliency at each moment; extract the power frequency phase information from the partial discharge signal, analyze the periodic characteristics of the pulse saliency at different moments of the same phase in the partial discharge signal, calculate the pulse specificity at each moment, and combine the pulse saliency to obtain the pulse evaluation value at each moment, and screen out abnormal moments.

[0040] Since circuits typically operate in complex electromagnetic environments with various interference sources, these noise signals may mix with partial discharge signals, resulting in a large amount of useless information in the detected signal. This affects the accurate identification and analysis of the discharge signal. Secondly, noise signals may be mistaken for discharge signals, leading to inaccurate detection of partial discharge phenomena. Furthermore, the presence of noise may cause errors in signal delay and amplitude measurement, affecting the accurate determination of the discharge source location. By denoising and enhancing the partial discharge signal, the accuracy of partial discharge phenomenon identification can be improved.

[0041] First, when partial discharge occurs on the line, it generates instantaneous pulses. These pulses are characterized by high amplitude and rapid change rate in the partial discharge signal. Therefore, the pulse significance is calculated based on the abrupt changes in the partial discharge signal.

[0042] Calculate the average value of the signal amplitude at all times within the partial discharge signal, and record it as the reference amplitude;

[0043] The difference between the signal amplitude and the reference amplitude at each moment within the partial discharge signal is calculated and denoted as the relative deviation.

[0044] In this embodiment, the absolute value of the difference between the signal amplitude and the reference amplitude at each moment within the partial discharge signal is calculated and denoted as the relative deviation.

[0045] Taking each time point as the center, all times within its neighborhood are recorded as the neighborhood time period;

[0046] In this embodiment, the number of all times within the neighborhood time period is 21. As for other implementation methods, the implementer can set it according to the actual situation.

[0047] Calculate the sum of the absolute values ​​of the rate of change of the signal amplitude between all two adjacent time points within the neighborhood time period;

[0048] It should be noted that the calculation of the rate of change is a well-known technique and will not be elaborated upon here.

[0049] The product of the relative deviation and the summation is used as the pulse significance at each time point;

[0050] It should be noted that the greater the relative deviation, the greater the degree to which the signal amplitude at that moment deviates from the average level of the entire signal; the greater the sum, the more significant the change in signal amplitude within the local area at that moment; and the greater the significance of the obtained pulse, the more the signal amplitude at that moment conforms to the pulse characteristics in the time domain, and the greater the possibility that it belongs to a real partial discharge phenomenon.

[0051] Secondly, when affected by noise sources such as power electronic device switches, thyristor voltage regulators, and frequency converters, both the actual partial discharge phenomenon and the noise interference will generate pulses in the partial discharge signal, causing the actual partial discharge phenomenon and noise to have similar pulse characteristics in the time domain. Therefore, the noise-generated pulses will interfere with the actual partial discharge phenomenon. Direct noise removal may mistakenly remove the actual discharge signal as noise, or misjudge the noise as a discharge signal.

[0052] However, the operating principles of noise sources such as power electronic device switches, thyristor voltage regulators, and frequency converters determine that they will repeat within the power frequency cycle, exhibiting fixed phases and patterns. Therefore, these noise signals show obvious periodicity across different power frequency cycles. Real partial discharge signals are typically sparse; during most pulse-free periods, the signal mainly consists of strong power frequency components and their harmonics. Therefore, partial discharge events are temporally random and do not repeat within every power frequency cycle. Thus, partial discharge signals exhibit randomness and sparsity across different power frequency cycles. Based on this analysis, by dividing the partial discharge signal into power frequency cycles, the differences in pulse characteristics of the same phase under different power frequency cycles are analyzed, and pulse specificity is calculated, specifically:

[0053] The power frequency reference signal is extracted from the partial discharge signal using a phase-locked loop algorithm. The power frequency reference signal is divided into each power frequency period, and the power frequency phase of the partial discharge signal at each moment in each power frequency period is obtained.

[0054] In this embodiment, the zero-crossing detection method is used to extract the power frequency phase. Both the phase-locked loop algorithm and the zero-crossing detection method are well-known technologies and will not be elaborated upon here. Secondly, the power frequency reference signal and the partial discharge signal are in one-to-one correspondence and completely synchronized in time. Therefore, the power frequency phase of the power frequency reference signal at time t is the same as the power frequency phase of the partial discharge signal at time t. The specific process of the zero-crossing detection method is as follows: zero-crossing detection is performed on the power frequency reference signal. The zero-crossing point where the signal amplitude changes from negative to positive is selected and marked as the power frequency phase 0°. The interval between two adjacent zero-crossing points from negative to positive is one power frequency cycle. Since the sampling frequency of the known data... and power frequency Then the number of sampling points for one power frequency cycle can be obtained, that is... This allows us to calculate the power frequency phase at each moment within each power frequency cycle.

[0055] For any moment within each power frequency cycle of the partial discharge signal, select the moment within all other power frequency cycles that has the same power frequency phase as the stated moment and record it as the reference moment;

[0056] The difference in pulse significance between any given time point and each control time point is denoted as the relative difference; the ratio between the relative difference and the pulse significance at each control time point is denoted as the relative ratio.

[0057] The sum of the relative comparisons between any given time and all control times is calculated as the pulse specificity at any given time.

[0058] It should be noted that the greater the relative difference, the greater the relative comparison, and the greater the pulse specificity. This indicates that the pulse characteristics on the same power frequency phase under different power frequency cycles are significantly different, reflecting that the pulse phenomenon at that moment is highly specific. Conversely, the smaller the difference, the smaller the pulse characteristics on the same power frequency phase under different power frequency cycles, reflecting that the pulse phenomenon at that moment is stable and repeatable, and the higher the probability that it is caused by noise sources such as power electronic device switches, thyristor voltage regulators, and frequency converters.

[0059] Furthermore, based on pulse specificity and pulse saliency, a pulse evaluation value is determined to evaluate the partial discharge signal, specifically as follows:

[0060] The product of pulse specificity and pulse saliency is used as the pulse evaluation value at each time step;

[0061] It should be noted that the larger the pulse evaluation value, the more likely the partial discharge signal at that moment is a real partial discharge phenomenon.

[0062] The average value of the pulse evaluation at all times within the partial discharge signal is used as the segmentation threshold.

[0063] The moment when the pulse evaluation value is greater than or equal to the segmentation threshold is selected as the abnormal moment;

[0064] It should be noted that by filtering out abnormal moments, real partial discharge events are extracted from the partial discharge signals.

[0065] This concludes the discovery of the abnormal moment when the partial discharge signal was obtained.

[0066] Step 3: Use wavelet transform to decompose the partial discharge signal into detail coefficients of multiple decomposition layers; for each detail coefficient, determine the corresponding mapping point in the adjacent decomposition layer to which it belongs; analyze the average level of pulse significance at abnormal moments in the time period corresponding to each detail coefficient after mapping back to the original partial discharge signal; combine the detail coefficients of each detail coefficient and its corresponding mapping point, as well as the number of decomposition layers to which it belongs, to determine the detail level of each detail coefficient in each decomposition layer.

[0067] In the process of denoising partial discharge information to enhance the signal, wavelet transform, as a signal processing method, is suitable for non-stationary and nonlinear signals, and features multi-resolution analysis and good time-frequency localization analysis. Based on the difference in energy distribution between partial discharge signals and noise signals in the wavelet domain, an appropriate threshold can be selected to distinguish them and remove useless signals.

[0068] The core idea of ​​wavelet transform lies in using a set of wavelet bases to perform multi-resolution analysis of the original signal through scaling and translation operations. Specifically, it involves performing wavelet decomposition on the signal layer by layer to obtain wavelet coefficients at different decomposition levels, namely detail coefficients and approximation coefficients. The detail coefficients are then compared with the thresholds of each scale level, processed by a threshold function, to obtain estimated detail coefficients. Finally, the estimated detail coefficients and approximation coefficients are used to reconstruct the signal for denoising and enhancement.

[0069] If the wavelet threshold is set too small, a large number of wavelet coefficients representing noise will be retained, and the reconstructed signal will still contain a lot of background noise, causing random noise pulses to be misjudged as real partial discharges. If the threshold is set too large, the wavelet coefficients corresponding to the pulses of real partial discharges will be treated as noise and filtered out, resulting in the loss of a lot of valuable details.

[0070] The partial discharge signal is decomposed by wavelet and then downsampled to obtain the detail coefficients corresponding to different decomposition layers.

[0071] In this embodiment, the wavelet decomposition algorithm is a well-known technology and will not be described in detail here. The wavelet basis function of the wavelet decomposition algorithm is sym8, the number of decomposition layers is 6, and the downsampling factor is 2. That is, after each decomposition layer, the length of the signal will be halved. As for other implementation methods, the implementer can set them according to the actual situation.

[0072] A real, energy-concentrated transient event, namely a partial discharge event, has its energy distributed across a series of continuous frequency bands. In wavelet transform, this manifests as high-amplitude coefficients appearing near the same time region at different decomposition levels. These high-coefficient points exhibit scale correlation on the time-scale plane, forming a continuous ridge.

[0073] Furthermore, the flowchart of the method for obtaining the detail level of each detail coefficient under each decomposition layer provided in the embodiments of this application is as follows: Figure 2 As shown, it specifically includes:

[0074] The detail coefficients under each decomposition layer are time-mapped in the partial discharge signal to obtain the specific time period corresponding to the partial discharge signal for each detail coefficient under each decomposition layer, and the average value of the pulse evaluation value of all abnormal moments included in the specific time period is calculated.

[0075] It should be noted that, for ease of understanding, since wavelet decomposition is used to decompose the signal into 6 levels, assuming the length of the original partial discharge signal is 1000, after the first level of decomposition, the length of the detail coefficients is 500, after the second level, the length of the detail coefficients is 250, after the third level, the length of the detail coefficients is 125, and so on. Therefore, for each detail coefficient after the first level of decomposition, there are two time points in the original partial discharge signal; for each detail coefficient after the second level of decomposition, there are four time points in the original partial discharge signal; and for each detail coefficient after the third level of decomposition, there are six time points in the original partial discharge signal. Furthermore, there are also correspondences between each decomposition level and adjacent decomposition levels. That is, for each detail coefficient after the second level of decomposition, there are two detail coefficients in the first level of decomposition, and for each detail coefficient after the third level of decomposition, there are two detail coefficients in the second level of decomposition. Thus, each detail coefficient corresponding to each decomposition level is mapped back to the time point of the original signal.

[0076] The two decomposition layers adjacent to each decomposition layer are denoted as neighboring layers;

[0077] For each decomposition layer, any detail coefficient is temporally mapped to each neighboring layer. The point in each neighboring layer that is closest to the position of the given detail coefficient is selected and recorded as the mapping point.

[0078] It should be noted that, for ease of understanding, the first... Decomposition layer Detail coefficients Corresponding in the original signal If there are 1 point, then the time position range in the original signal is 1. arrive -1, taking the time of the midpoint within the time range as... At the center position corresponding to the original signal, select the first... The center position within each decomposition layer and The closest detail coefficient is selected as the mapping point. The center position within each decomposition layer and The closest detail coefficient is used as the mapping point.

[0079] The reciprocal of the difference between the number of the decomposition layer to which each detail coefficient belongs and the preset value is used as the weighting factor;

[0080] In this embodiment, the reciprocal of the absolute value of the difference between the number of the decomposition layer to which each detail coefficient belongs and the preset value is used as the weighting factor;

[0081] In this embodiment, the preset value is 3, representing the number of intermediate layers in the decomposition, i.e., the number of decomposition layers. As another implementation method, the implementer can set it according to the actual situation.

[0082] It should be noted that the difference between the number of decomposition layers to which each detail coefficient belongs and the preset value may be equal to 0. When calculating the reciprocal, in order to avoid the denominator being equal to 0, a preset value greater than 0 is added to the denominator. In this embodiment, the preset value greater than 0 is 1. As for other implementation methods, the implementer can set it according to the actual situation.

[0083] Using the weighting factor as the weight, a weighted sum is performed on each detail coefficient under each decomposition layer and the corresponding detail coefficients of all mapping points.

[0084] The normalized result of the product between the average value and the weighted sum is used as the detail level of each detail coefficient under each decomposition layer;

[0085] In this embodiment, the maximum-minimum normalization method is used for normalization. The maximum-minimum normalization method is a well-known technique and will not be described in detail here.

[0086] It should be noted that since the energy of partial discharge signals is mainly distributed at certain intermediate scales, i.e., at the intermediate decomposition layers, the larger the weighting factor, the closer the decomposition layer is to the intermediate layer, and the more signal feature information it contains. Therefore, it is given a higher weight in the weighted summation. The larger the weighted summation result, the more significant the detail coefficients of the decomposition layer are at multiple scales, meaning that a high-energy event has left a strong trace across multiple consecutive frequency bands. This reflects that the detail coefficient is more likely to be a transient event with concentrated energy, i.e., a partial discharge event. The larger the average value, the stronger the pulse characteristics within the original signal that generated this detail coefficient, reflecting a more likely obvious partial discharge event. The greater the detail, the more likely the detail coefficient at the decomposition layer is to characterize the detailed information of the real partial discharge phenomenon.

[0087] At this point, the level of detail corresponding to each detail coefficient under each decomposition layer is obtained.

[0088] Step 4: Based on the level of detail, adjust the wavelet threshold of each decomposition layer, process the detail coefficients of each decomposition layer, and use inverse wavelet transform to reconstruct the signal to obtain the enhanced partial discharge signal.

[0089] Furthermore, threshold selection is a crucial step in the wavelet thresholding denoising algorithm, directly impacting the denoising effect. Higher detail indicates that the detail coefficients possess significant features across multiple scales, rather than being random noise. By increasing the wavelet threshold, noise is filtered out, preserving the true partial discharge pulses. Based on the detail level, the wavelet thresholds at different decomposition levels of the wavelet transform are adjusted, specifically as follows:

[0090]

[0091]

[0092] in, For the first Each decomposition layer corresponds to an adjusted wavelet threshold. For the first Intermediate variables corresponding to each decomposition layer; For the first Each decomposition layer corresponds to the wavelet threshold before adjustment. For the first The first decomposition layer A detailed coefficient, For the first The first decomposition layer The level of detail corresponding to each detail coefficient For the first The number of all detail coefficients in each decomposition layer;

[0093] In this embodiment, the wavelet threshold before adjustment is selected as a hard threshold. The calculation of the hard threshold is a well-known technique, and the calculation process of the hard threshold is as follows: ,in, For the first The noise standard deviation of each decomposition layer For the first The length of the decomposition layer, i.e., the length of the first decomposition layer. The number of all detail coefficients in each decomposition layer.

[0094] Based on the adjusted wavelet threshold, the detail coefficients of different decomposition layers are processed, and the signal is reconstructed by using inverse wavelet transform to obtain the enhanced partial discharge signal.

[0095] It should be noted that wavelet transform and inverse wavelet transform are well-known techniques and will not be elaborated here. By adjusting the wavelet threshold, the noise represented by the detail coefficients whose absolute values ​​are less than the adjusted wavelet threshold is removed, thereby obtaining an enhanced partial discharge signal with a significantly improved signal-to-noise ratio.

[0096] Based on the same inventive concept as the above methods, this application also provides a line partial discharge signal enhancement system based on wavelet transform, including a memory, a processor, and a computer program stored in the memory and running on the processor. When the processor executes the computer program, it implements the steps of any one of the above-described line partial discharge signal enhancement methods based on wavelet transform.

[0097] It should be understood that, although Figure 1 The steps in the flowchart are shown sequentially as indicated by the arrows, but these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order in which these steps are executed, and they can be performed in other orders. Figure 1 At least some of the steps in the process may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least some of the sub-steps or stages of other steps.

[0098] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0099] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application. Therefore, any simple modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of this application, without departing from the content of the technical solution of this application, shall fall within the protection scope of the technical solution of this application.

Claims

1. A method for enhancing a line partial discharge signal based on wavelet transform, characterized in that, The method includes the following steps: Partial discharge signals from transmission lines are collected; the deviation of signal amplitude at different times in the partial discharge signal and the trend of signal amplitude variation in the neighborhood are analyzed, and the pulse saliency at each time is calculated; power frequency phase information is extracted from the partial discharge signal, the periodicity characteristics of pulse saliency at different times of the same phase in the partial discharge signal are analyzed, the pulse specificity at each time is calculated, and the pulse evaluation value at each time is obtained by combining the pulse saliency, and abnormal times are screened out; the partial discharge signal is decomposed into detail coefficients of multiple decomposition layers using wavelet transform; for each detail coefficient, the corresponding mapping point is determined in the adjacent decomposition layer to which it belongs, and the average level of pulse saliency at abnormal times in the corresponding time period after each detail coefficient is mapped back to the original partial discharge signal is analyzed; the detail of each detail coefficient under each decomposition layer is determined by combining the detail coefficient of each detail coefficient and its corresponding mapping point, as well as the number of decomposition layers, so as to adjust the wavelet threshold of each decomposition layer, process the detail coefficients of each decomposition layer, and use inverse wavelet transform to reconstruct the signal to obtain the enhanced partial discharge signal; The calculation of the pulse saliency at each time point includes: Calculate the mean value of the signal amplitude at all times within the partial discharge signal, and record it as the reference amplitude; calculate the difference between the signal amplitude at each time within the partial discharge signal and the reference amplitude, and record it as the relative deviation; Calculate the sum of the absolute values ​​of the rate of change of the signal amplitude between any two adjacent time points within the neighborhood of each time point; The pulse significance is the product of the relative deviation and the summation; The extraction of power frequency phase information includes: extracting a power frequency reference signal from the partial discharge signal using a phase-locked loop algorithm, dividing the power frequency reference signal into each power frequency period, and obtaining the power frequency phase of the partial discharge signal at each moment within each power frequency period; The calculation of the pulse specificity at each time point includes: For any moment within each power frequency cycle of the partial discharge signal, select the moment within all other power frequency cycles that has the same power frequency phase as the stated moment and record it as the reference moment; The difference in pulse significance between any given time point and each control time point is denoted as the relative difference; the ratio between the relative difference and the pulse significance at each control time point is denoted as the relative ratio. The pulse specificity is the sum of the relative comparisons between any given time and all control times; The pulse evaluation value is the product of pulse specificity and pulse significance.

2. The wavelet transform based line partial discharge signal enhancement method as claimed in claim 1, wherein, The process of filtering out abnormal moments includes: using the average pulse evaluation value of all moments within the partial discharge signal as a segmentation threshold; and selecting moments where the pulse evaluation value is greater than or equal to the segmentation threshold as abnormal moments.

3. The wavelet transform based line partial discharge signal enhancement method as claimed in claim 1, wherein, The process of obtaining the mapping point is as follows: the two decomposition layers adjacent to each decomposition layer are denoted as neighboring layers; any detail coefficient under each decomposition layer is time-mapped in each neighboring layer, and the point in each neighboring layer that is closest to the position of the any detail coefficient is selected and denoted as the mapping point.

4. The wavelet transform based line partial discharge signal enhancement method as claimed in claim 1, wherein, The determination of the level of detail of each detail coefficient at each decomposition layer includes: time-mapping each detail coefficient under each decomposition layer to the partial discharge signal to obtain a specific time period corresponding to each detail coefficient under each decomposition layer in the partial discharge signal, and calculating an average value of pulse evaluation values of all abnormal time points contained in the specific time period; taking an inverse of a difference between a layer number of a decomposition layer to which each detail coefficient belongs and a preset value as a weight factor, and performing weighted summation on each detail coefficient under each decomposition layer and the corresponding mapping point corresponding detail coefficient by taking the weight factor as a weight value; the detail degree is a normalized result of a product between the average value and a result of the weighted summation.

5. The wavelet transform based line partial discharge signal enhancement method as claimed in claim 1, wherein, The first decomposition layer corresponds to an adjusted wavelet threshold value, comprising: , and wherein, the first decomposition layer corresponds to an adjusted wavelet threshold value, the first decomposition layer corresponds to an intermediate variable; the first decomposition layer corresponds to a wavelet threshold value before adjustment, the first decomposition layer corresponds to the first detail coefficient, the first decomposition layer corresponds to the first detail coefficient corresponding to a detail degree, the first decomposition layer corresponds to the number of all detail coefficients.

6. A wavelet transform based line partial discharge signal enhancement system comprising a memory, a processor and a computer program stored in the memory and running on the processor, characterized in that, The processor implements the steps of the wavelet transform-based line partial discharge signal enhancement method according to any one of claims 1-5 when executing the computer program.

Images