Partial discharge signal matching method and device, storage medium and electronic equipment

Abstract

The application discloses a partial discharge signal matching method and device, a storage medium and an electronic device. The method comprises the following steps: collecting a partial discharge signal of a power device, wherein the partial discharge signal comprises an incident pulse signal and a reflected pulse signal; constructing a feature matrix according to local extreme points included in the partial discharge signal; dividing the feature matrix to obtain a minimum value feature matrix and a maximum value feature matrix; determining a matching path between the minimum value feature matrix and the maximum value feature matrix by using a dynamic time warping algorithm; and obtaining a matching result of the incident pulse signal and the reflected pulse signal based on the matching path. The application solves the technical problems of low matching accuracy of incident waves and reflected waves and weak anti-noise interference capability in the related art.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of power operation and maintenance, in particular, to a partial discharge signal matching method and device, a storage medium and an electronic device. BACKGROUND

[0002] With the expansion of the scale of the power system and the complexity of power equipment, equipment fault diagnosis and early warning become increasingly important. Partial discharge (PD) as one of the precursors of power equipment failure refers to the partial discharge phenomenon caused by electrical stress under the working state of the equipment. These discharges usually manifest in the form of high-frequency pulse signals. By monitoring and analyzing these signals, the health status of the equipment can be effectively evaluated, potential faults can be discovered in a timely manner, and the stable operation of the power system can be ensured.

[0003] Existing partial discharge signal matching technologies mostly rely on traditional correlation analysis methods, which usually assume a fixed linear relationship between signals. However, partial discharge signals in the power system are often affected by multiple factors such as equipment aging, electrical stress, and external environmental changes, resulting in nonlinear changes in the signals. Traditional methods show low matching accuracy when dealing with nonlinear characteristics such as time delay, amplitude variation, and frequency drift in the signals, making it difficult to meet the demand for high-precision monitoring and fault warning. Existing matching algorithms are highly sensitive to noise, which can easily lead to false matches or missed detections, affecting the reliability of fault detection. At the same time, the matching methods of existing technologies are usually based on static data analysis, lacking the ability to adapt to dynamic running states of equipment and changes in the environment. In the power system, equipment states and environmental conditions can change at any time, and traditional methods fail to effectively respond to these changes, resulting in inaccurate capture of faults or delayed response, affecting the timeliness of fault warning. Therefore, the existing technology has poor dynamic adaptability, and the calculation is complex, real-time performance is poor, which cannot meet the demand of the power system for fast response and real-time monitoring.

[0004] To address the above problems, no effective solutions have been proposed so far. SUMMARY

[0005] The embodiments of the present application provide a partial discharge signal matching method, device, storage medium and electronic device to at least solve the technical problems of low matching accuracy of incident and reflected waves and weak anti-noise interference ability in the related art.

[0006] According to an aspect of the embodiments of the present application, a partial discharge signal matching method is provided, comprising: collecting a partial discharge signal of a power device, wherein the partial discharge signal comprises an incident pulse signal and a reflected pulse signal; constructing a feature matrix according to local extreme points included in the partial discharge signal; dividing based on the feature matrix to obtain a minimum value feature matrix and a maximum value feature matrix; determining a matching path between the minimum value feature matrix and the maximum value feature matrix using a dynamic time warping algorithm; and obtaining a matching result of the incident pulse signal and the reflected pulse signal based on the matching path.

[0007] Optionally, the constructing of the feature matrix according to the local extreme points included in the partial discharge signal comprises: determining extreme point sizes and extreme point positions of the local extreme points; determining extreme point left and right slopes of left and right fitting line segments of the local extreme points based on the extreme point sizes and the extreme point positions; and obtaining the feature matrix according to the extreme point sizes, the extreme point positions, and the extreme point left and right slopes.

[0008] Optionally, the partial discharge signal is a time sequence, and the obtaining of the feature matrix according to the extreme point sizes, the extreme point positions, and the extreme point left and right slopes comprises: converting the extreme point left and right slopes into circumferential angles; converting the extreme point positions into ratios between the extreme point positions and a length of the time sequence; and obtaining a standardized feature matrix based on the extreme point sizes, the circumferential angles, and the ratios.

[0009] Optionally, the collecting of the partial discharge signal of the power device comprises: collecting incident pulse signals and reflected pulse signals of the partial discharge through a sensor of the power device; and obtaining the partial discharge signal based on the incident pulse signals and the reflected pulse signals, wherein the preprocessing comprises at least one of the following: denoising, filtering, smoothing, and normalization.

[0010] Optionally, the local extreme points are multiple, and the determining of the matching path between the minimum value feature matrix and the maximum value feature matrix using the dynamic time warping algorithm comprises: obtaining a starting trend feature based on a slope of a fitting line segment between a starting point of the partial discharge signal and a first local extreme point of the multiple local extreme points; obtaining an ending trend feature based on a slope of a fitting line segment between an ending point of the partial discharge signal and a last local extreme point of the multiple local extreme points; and processing based on the starting trend feature, the ending trend feature, the minimum value feature matrix, and the maximum value feature matrix using the dynamic time warping algorithm to obtain the matching path.

[0011] Optionally, the matching path between the minimum feature matrix and the maximum feature matrix is determined by using a dynamic time warping algorithm, including: determining a local distance matrix according to the minimum feature matrix and the maximum feature matrix, wherein each element in the local distance matrix represents a distance between a minimum value included in the minimum feature matrix and a maximum value included in the maximum feature matrix; determining a cumulative cost matrix based on the local distance matrix by using the dynamic time warping algorithm; and determining the matching path based on a diagonal line of the cumulative cost matrix.

[0012] Optionally, the matching result of the incident pulse signal and the reflected pulse signal is obtained based on the matching path, including: in a case where the matching path is greater than a predetermined distance threshold, determining that the matching result is that a difference between the incident pulse signal and the reflected pulse signal is greater than a predetermined difference threshold; and the method further includes: in a case where the matching result is that the difference between the incident pulse signal and the reflected pulse signal is greater than the predetermined difference threshold, determining that the power equipment has an abnormality.

[0013] According to another aspect of the embodiments of the present application, a partial discharge signal matching device is provided, including: a collection module configured to collect a partial discharge signal of a power equipment, wherein the partial discharge signal includes an incident pulse signal and a reflected pulse signal; a feature construction module configured to construct a feature matrix according to local extreme points included in the partial discharge signal; a division module configured to divide the feature matrix to obtain a minimum feature matrix and a maximum feature matrix; a path generation module configured to determine a matching path between the minimum feature matrix and the maximum feature matrix by using a dynamic time warping algorithm; and a matching module configured to obtain a matching result of the incident pulse signal and the reflected pulse signal based on the matching path.

[0014] According to another aspect of the embodiments of the present application, a nonvolatile storage medium is provided, which stores a plurality of instructions, and the instructions are adapted to be loaded and executed by a processor to implement any one of the partial discharge signal matching methods.

[0015] According to another aspect of the embodiments of the present application, an electronic device is provided, including: one or more processors and a memory, the memory being configured to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors are caused to implement any one of the partial discharge signal matching methods.

[0016] In the embodiment of the present application, a nonlinear alignment method is adopted to collect the partial discharge signal of the power equipment, wherein the partial discharge signal includes incident pulse signals and reflected pulse signals; a feature matrix is constructed according to the local extreme points included in the partial discharge signal; the feature matrix is divided to obtain a minimum value feature matrix and a maximum value feature matrix; a dynamic time warping algorithm is adopted to determine the matching path between the minimum value feature matrix and the maximum value feature matrix; and the matching result of the incident pulse signals and the reflected pulse signals is obtained based on the matching path. The purpose of matching the incident wave and the reflected wave is achieved, the technical effects of improving the anti-noise interference ability and the dynamic adaptability are realized, and the technical problems of low matching precision of the incident wave and the reflected wave and weak anti-noise interference ability in the related art are solved. BRIEF DESCRIPTION OF DRAWINGS

[0017] The drawings described herein are used to provide further understanding of the present application, and form a part of the present application. The illustrative embodiments of the present application and their descriptions serve to explain the present application, and do not constitute improper limitations on the present application. In the drawings:

[0018] Figure 1 is a flowchart of an optional partial discharge signal matching method provided by the embodiment of the present application;

[0019] Figure 2 is a first schematic diagram of an optional partial discharge signal matching method provided by the embodiment of the present application;

[0020] Figure 3 is a second schematic diagram of an optional partial discharge signal matching method provided by the embodiment of the present application;

[0021] Figure 4 is a schematic diagram of an optional partial discharge signal matching device provided by the embodiment of the present application. DETAILED DESCRIPTION

[0022] In order to enable the persons skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, but not all the embodiments. Based on the embodiments in the present application, all the other embodiments obtained by the persons skilled in the art without creative labor should be within the protection scope of the present application.

[0023] It is to be noted that the terms "first", "second", and the like in the description and in the claims of the present application and in the above-described drawings are intended to distinguish similar objects and not necessarily to describe a particular sequential or chronological order. It is to be understood that the use of the terms so-termed "first", "second", etc. can be interchanged, where appropriate, to refer to the same element in order to describe the embodiments of the present application described herein in other than the order illustrated or described herein. Furthermore, the terms "comprising" and "having" and any variations thereof are intended to cover a non-exclusive inclusion, for example, a process, method, system, product, or apparatus that comprises a list of steps or units not necessarily limited to those explicitly listed, but can include other steps or units not expressly listed or inherent to such process, method, product, or apparatus.

[0024] In the signal detection process of partial discharge, the accurate matching of incident wave and reflected wave is crucial for fault diagnosis. Due to the influence of factors such as equipment failure or aging, the incident wave and reflected wave will change to varying degrees, therefore, matching these waveforms and extracting the characteristic information therein is the core of improving the accuracy of fault recognition. However, the existing waveform matching methods often face great technical challenges when dealing with complex partial discharge signals. The existing partial discharge waveform matching technology usually relies on linear correlation algorithm or similarity measurement method, which assumes that the time and amplitude relationship between signals is fixed and linear. However, in practical applications, partial discharge signals are often affected by various factors, including equipment aging, electrical stress, environmental changes, etc., which make the signals exhibit nonlinear characteristics. The traditional linear matching method cannot effectively deal with the problems of time delay, amplitude variation and frequency drift in the signal, resulting in reduced waveform matching accuracy, and further affecting the accuracy and reliability of fault diagnosis.

[0025] Partial discharge signals in power systems are also often disturbed by noise. These noise sources include electromagnetic interference, equipment switching operation and external environmental changes, etc. Traditional matching algorithms have weak processing ability for noise, which easily leads to signal mismatching or missed detection, affecting the effect of fault detection. At the same time, existing technologies are mostly based on static data analysis, lacking dynamic adaptability to the running state of power equipment and environmental changes. In actual power systems, the state of equipment and environmental conditions change at any time, and traditional methods often cannot respond to these changes in time, resulting in failure to accurately capture faults or delayed response. Therefore, the existing technology has significant deficiencies in dynamic adaptability, computational complexity and real-time performance, and it is difficult to meet the needs of modern power systems for fast response and real-time monitoring. How to improve the waveform matching accuracy, enhance the anti-noise interference ability, and improve the dynamic adaptability has become the main challenge faced by partial discharge detection technology.

[0026] To solve the above problems, the embodiment of the present application provides a method for matching partial discharge signals. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order from that shown here.

[0027] Figure 1 An optional method for matching partial discharge signals is provided according to the embodiment of the present application, as shown in the flowchart of Figure 1 The method comprises the following steps:

[0028] Step S102, collecting partial discharge signals of the power equipment, wherein the partial discharge signals include incident pulse signals and reflected pulse signals;

[0029] It can be understood that the partial discharge signals are collected from the power equipment by the sensor, and these signals contain two parts of incident pulse signals and reflected pulse signals. The collected incident pulse signals and reflected pulse signals are the basis for subsequent analysis, providing raw data for subsequent signal processing and matching, and ensuring the accuracy and reliability of the signal source.

[0030] In an optional embodiment, collecting partial discharge signals of the power equipment comprises: collecting incident pulse signals and reflected pulse signals of partial discharge by a sensor of the power equipment; and pre-processing the incident pulse signals and the reflected pulse signals to obtain the partial discharge signals, wherein the pre-processing comprises at least one of the following: denoising, filtering, smoothing, and normalization.

[0031] It can be understood that the incident pulse signals and the reflected pulse signals generated by partial discharge are collected in real time by the sensor installed on the power equipment. These signals are the direct manifestation of the partial discharge phenomenon, and contain key information of the health status of the equipment. The collected raw signals are usually affected by noise interference, equipment vibration, electromagnetic interference and other factors, resulting in a decrease in signal quality. In order to improve the accuracy and reliability of subsequent processing, the collected incident pulse signals and reflected pulse signals are pre-processed.

[0032] The specific operation of pre-processing includes at least one of the following: denoising, filtering, smoothing, and normalization. Denoising is used to remove random noise in the signal and improve the signal-to-noise ratio. Filtering is used to remove high-frequency or low-frequency interference components in the signal through a filter, and to retain useful signal characteristics. Smoothing is used to smooth the signal to reduce sharp fluctuations in the signal and make it smoother. Normalization is used to adjust the amplitude of the signal to a uniform range (such as 0 to 1), eliminating the influence of amplitude difference between different signals.

[0033] Step S104, constructing a feature matrix according to the local extreme points included in the partial discharge signal;

[0034] It can be understood that the collected partial discharge signal is analyzed, the local extreme points (including maximum points and minimum points) are extracted, and the feature matrix is constructed based on the extreme points. By extracting the local extreme points and constructing the feature matrix, the dimension of the signal can be significantly reduced, while the key feature information of the signal is retained, the complexity of subsequent calculation is reduced, and the efficiency of signal processing is improved.

[0035] The matching method of the incident wave and the reflected wave of the partial discharge can be represented by the local extreme approximation. Specifically, in the change process of the time series, the subsequence segment between adjacent extreme points changes monotonously, so these subsequences can be fitted by connecting the line segments of the extreme points. For a single extreme point, the slope change of the fitting line segments on the left and right thereof reflects the morphological characteristics of the original time series. Therefore, the slope of the fitting line segment is crucial for representing the change trend of the signal.

[0036] For the incident wave and the reflected wave time series signal , is the value of the Xth k point, where For the following four cases:

[0037]

[0038] If satisfies case 1) or 2), the kth data point is defined as a local minimum value point of the time series X, and if satisfies case 3) or 4), the kth data point is defined as a local maximum value point of the time series X, and the local minimum value point and the local maximum value point are collectively referred to as a local extreme value point.

[0039] In an optional embodiment, constructing a feature matrix according to the local extreme points included in the partial discharge signal comprises: determining the extreme point size and the extreme point position of the local extreme points; determining the extreme point left and right slopes of the left and right fitting line segments of the local extreme points based on the extreme point size and the extreme point position; and obtaining the feature matrix according to the extreme point size, the extreme point position, and the extreme point left and right slopes.

[0040] It can be understood that the preprocessed partial discharge signal is analyzed, and the local extreme points (including maximum points and minimum points) are extracted. For each extreme point, the specific value (i.e., the extreme point size) and the position in the time series (i.e., the extreme point position) are recorded. These information are important components of the signal morphological characteristics.

[0041] For each local extremum point, the slopes of the fitting line segments on its left and right sides are calculated. The left slope is a straight line fitted through the signal points on the left side of the extremum point, and the slope of the line is calculated to reflect the trend of the signal on the left side of the extremum point. The right slope is a straight line fitted through the signal points on the right side of the extremum point, and the slope of the line is calculated to reflect the trend of the signal on the right side of the extremum point. These slope information can further describe the morphological changes of the signal near the extremum point.

[0042] The extremum point size, extremum point position and left and right slopes of each extremum point are integrated into a feature vector, and the feature vectors of all extremum points are combined into a feature matrix. The feature matrix not only retains the key morphological features of the signal, but also enhances the description ability of the signal trend through the slope information.

[0043] By extracting local extremum points and their related features, the time series of the original signal is greatly simplified into a feature matrix, significantly reducing the data dimension and reducing the complexity of subsequent calculations. The feature matrix not only contains the size and position of the extremum point, but also introduces the left and right slope information, which can more comprehensively describe the morphological changes of the signal and improve the representation ability of the local discharge signal features. In the subsequent dynamic time warping (DTW, Dynamic Time Warping) matching process, the multi-dimensional information (size, position, slope) in the feature matrix can more accurately reflect the dynamic characteristics of the signal, thereby improving the matching accuracy between the incident pulse signal and the reflected pulse signal. The extremum point and its slope information have a certain robustness to noise, and even under noise interference, these key features can still accurately reflect the essential characteristics of the signal, thereby improving the anti-interference ability of the system. The construction process of the feature matrix converts the complex signal processing problem into the analysis of key features, reduces redundant information, and further optimizes the calculation efficiency of the entire matching algorithm, making it more suitable for real-time monitoring and fast response.

[0044] Optionally, for the extracted extremum points, an extremum sequence X is obtained The extremum sequence does not include the start point and the end point of X.

[0045] Wherein: , ; The extremum point position corresponds to is the extremum point type, is the extremum point size.

[0046] The extremum sequence X is converted into an extremum feature matrix Y:

[0047]

[0048] Wherein: ​

[0049] , , corresponding extreme point size, corresponding extreme point position, and represent the slope of the fitting line segment left and right of the extreme point, and the calculation formula is as follows:

[0050]

[0051] In an optional embodiment, the partial discharge signal is a time series, and a feature matrix is obtained according to the extreme point size, the extreme point position, and the slope left and right of the extreme point, including: converting the slope left and right of the extreme point into a circular angle; converting the extreme point position into a ratio between the extreme point position and the length of the time series; and obtaining a normalized feature matrix based on the extreme point size, the circular angle, and the ratio.

[0052] It can be understood that the slope left and right of the extreme point directly reflects the change trend of the signal near the extreme point, but the numerical range of the slope may be large and the unit is not unified. In order to eliminate this difference, the slope is converted into a circular angle (i.e. an angle value). Specifically, the slope value is mapped to an angle value (usually between 0 and 360 degrees) through an arctangent function (arctan), so as to realize the standardization of the slope. This conversion not only unifies the dimension, but also makes the change of the slope more intuitive. The position of the extreme point is absolute in the time series, but the length of the time series of different signals may be different. In order to eliminate this difference, the position of the extreme point is divided by the total length of the time series to obtain a normalized ratio. This ratio represents the relative position of the extreme point in the time series, and can better reflect the distribution of the extreme point in the entire signal. Based on the extreme point size, the converted circular angle (left and right slope), and the normalized extreme point position ratio, a feature matrix is constructed. Through this standardization processing, each feature vector in the feature matrix has a unified dimension and numerical range, thereby improving the comparability and consistency of the feature matrix.

[0053] By converting the slope into a circular angle and converting the extreme point position into a relative ratio, the dimensional differences between different features are eliminated, so that each feature in the feature matrix can be directly compared and calculated, improving the uniformity and consistency of the feature matrix. The standardized feature matrix can more accurately reflect the morphological characteristics of the signal, so that the features between different signals can be directly compared and matched, improving the accuracy and reliability of the matching algorithm. The feature matrix after standardization has stronger robustness to noise and signal changes. Even when the signal is disturbed or changed to a certain extent, the standardized features remain relatively stable, thereby improving the anti-interference ability of the matching algorithm. By converting the extreme point position into a relative ratio, the standardized feature matrix can adapt to time series signals of different lengths, so that the matching algorithm is no longer dependent on the specific length of the signal, thereby improving the universality and flexibility of the algorithm.

[0054] Optionally, when measuring the similarity between two time series, in order to eliminate the dimensional differences between different features, the following standardization processing is required: The standardization methods for left and right slopes , and extreme point position are as follows:

[0055]

[0056] and Convert the slope into a circular angle, Convert the position into the ratio of the extreme point position to the length of the time series. The extreme feature matrix after standardization processing is as follows:

[0057]

[0058] The extreme feature matrix not only retains the morphological change characteristics of the time series, but also retains the correlation between different feature quantities.

[0059] Step S106, based on the feature matrix, the minimum value feature matrix and the maximum value feature matrix are obtained by division;

[0060] It can be understood that the constructed feature matrix is further divided into minimum value feature matrix and maximum value feature matrix, which correspond to the minimum value point and the maximum value point in the signal respectively. This division method can finely process different features of the signal, reduce redundant information, further reduce the calculation complexity, and at the same time improve the feature matrix's representation ability of the signal form, providing a more accurate feature basis for subsequent matching.

[0061] Optionally, in order to reduce the dynamic time warping time complexity and further improve the calculation efficiency, the following steps are taken: Further divided into minimum value feature matrix and maximum value feature matrix When corresponding to the minimum point, When corresponding to the maximum point, , and is and the number of columns of the minimum value feature matrix and the maximum value feature matrix, corresponding to the number of minimum points and maximum points.

[0062] In step S108, a dynamic time warping algorithm is used to determine the matching path between the minimum value feature matrix and the maximum value feature matrix.

[0063] It can be understood that the dynamic time warping (DTW) algorithm is used to calculate the similarity between the minimum value feature matrix and the maximum value feature matrix, and to determine the best matching path therebetween. The DTW algorithm can effectively handle non-linear changes in signals, such as time delay, amplitude change, and frequency drift, and find the best matching path through dynamic programming, significantly improving the accuracy and robustness of signal matching, and is particularly suitable for complex signal environments.

[0064] The dynamic time warping algorithm is an algorithm for calculating the similarity of two time series, which can handle non-linear time delay, amplitude change, and frequency drift in signals. DTW aligns the time axes of two signals non-linearly, minimizes the cumulative distance between the signals, and finds the best matching path.

[0065] In an alternative embodiment, there are multiple local extreme points, and the dynamic time warping algorithm is used to determine the matching path between the minimum value feature matrix and the maximum value feature matrix, including: obtaining a starting trend feature based on the slope of the fitting line segment between the starting point of the partial discharge signal and the first local extreme point of the multiple local extreme points; obtaining an ending trend feature based on the slope of the fitting line segment between the termination point of the partial discharge signal and the last local extreme point of the multiple local extreme points; and using the dynamic time warping algorithm to process the starting trend feature, the ending trend feature, the minimum value feature matrix, and the maximum value feature matrix to obtain the matching path.

[0066] It can be understood that for the time series of partial discharge signals, the first local extreme point (maximum point or minimum point) is found. The slope of the fitting line segment between the starting point of the signal and the first local extreme point is calculated. This slope reflects the trend of the signal from the starting point to the first extreme point, which is called the starting trend feature. It can capture the initial change characteristics of the signal and provide important starting information for matching. In the time series of partial discharge signals, the last local extreme point (maximum point or minimum point) is found. The slope of the fitting line segment between the termination point of the signal and the last local extreme point is calculated. This slope reflects the trend of the signal from the last extreme point to the termination point, which is called the ending trend feature. It can capture the ending change characteristics of the signal and provide important ending information. The starting trend feature and the ending trend feature are added to the minimum value feature matrix and the maximum value feature matrix respectively as an extended part of the feature matrix. The extended minimum value feature matrix and the extended maximum value feature matrix are processed using the dynamic time warping (DTW) algorithm. The DTW algorithm calculates the best matching path between two feature matrices through dynamic programming method. In this process, the starting trend feature and the ending trend feature provide global information for matching, helping the algorithm to align the feature points in the two feature matrices more accurately, so as to obtain a high-precision matching path. Based on the above processing, the DTW algorithm outputs the matching path between the minimum value feature matrix and the maximum value feature matrix. This path reflects the best corresponding relationship between the incident pulse signal and the reflected pulse signal at the feature level.

[0067] Optionally, the starting point and the termination point of the original time series reflect important information of the sequence, and in the transformation process of the extreme value feature matrix, and In addition to the minimum value points and the maximum value points, the first column and the last column are increased by the feature vectors of the starting point and the termination point, respectively. The feature vectors of the starting point and the termination point are represented as follows: the left slope of the starting point is 0, and the right slope is the slope of the fitting line segment between the starting point and the first extreme point. The right slope of the termination point is 0, and the left slope is the slope of the fitting line segment between the termination point and the last extreme point. The position feature and the value range feature are represented in the same way as the extreme points. and The first column and the last column of the feature vectors are represented as follows:

[0068]

[0069] In an alternative embodiment, the matching path between the minimum feature matrix and the maximum feature matrix is determined by using a dynamic time warping algorithm, including: determining a local distance matrix based on the minimum feature matrix and the maximum feature matrix, wherein each element in the local distance matrix represents the distance between a minimum value included in the minimum feature matrix and a maximum value included in the maximum feature matrix; determining a cumulative cost matrix based on the local distance matrix by using the dynamic time warping algorithm; and determining the matching path based on the diagonal line of the cumulative cost matrix.

[0070] It can be understood that the local distance matrix is calculated based on the minimum feature matrix and the maximum feature matrix. The local distance matrix is a two-dimensional matrix, whose rows correspond to the minimum value points in the minimum feature matrix and whose columns correspond to the maximum value points in the maximum feature matrix. Each element in the local distance matrix represents the distance between a minimum value point in the minimum feature matrix and a maximum value point in the maximum feature matrix. This distance can be calculated by Euclidean distance, Manhattan distance or other suitable distance measurement methods. The cumulative cost matrix is calculated based on the local distance matrix by using the dynamic time warping (DTW) algorithm. The cumulative cost matrix is a dynamic programming matrix used to record the minimum cumulative cost from the starting point to the current point. The calculation process of the cumulative cost matrix follows the rules of dynamic programming: the cumulative cost of each element is equal to its local distance plus the minimum cumulative cost path from the starting point to that point. Based on the cumulative cost matrix, the matching path is determined by backtracking from the right bottom corner (i.e. the last element) of the matrix to the left top corner (the starting point) along the minimum cost path. The matching path is usually searched along the diagonal direction of the cumulative cost matrix, because the diagonal direction represents the optimal alignment relationship between two feature points. The matching path reflects the best matching relationship between the minimum feature matrix and the maximum feature matrix.

[0071] By the dynamic programming method, the DTW algorithm can find the globally optimal matching path, not just a local optimal solution. This global optimality ensures the accuracy and reliability of the matching result. The calculation of the local distance matrix can accurately measure the similarity or difference between the minimum value points and the maximum value points, providing a reliable basis for subsequent matching. The DTW algorithm dynamically aligns the feature points in the two feature matrices through the cumulative cost matrix, which can effectively handle nonlinear changes in the signal (such as time delay, amplitude change and frequency drift). This dynamic alignment makes the algorithm adaptable to complex signal feature changes. The diagonal line search based on the cumulative cost matrix can quickly determine the optimal matching path, reducing the possibility of mismatch. At the same time, the dynamic programming property of the cumulative cost matrix makes the algorithm more robust to noise and signal changes.

[0072] In step S110, the matching result of the incident pulse signal and the reflected pulse signal is obtained based on the matching path.

[0073] It can be understood that the matching path determined according to the DTW algorithm finally obtains the matching result of the incident pulse signal and the reflected pulse signal. In this way, high-precision matching of the incident wave and the reflected wave is realized, the corresponding relationship between the signals can be accurately identified, the reliability of fault diagnosis is improved, the anti-noise interference ability and dynamic adaptability of the system are enhanced, and the problems of low matching precision and weak anti-interference ability of the traditional method are solved.

[0074] In an optional embodiment, based on the matching path, the matching result of the incident pulse signal and the reflected pulse signal is obtained, including: in the case that the matching path is greater than a predetermined distance threshold, determining that the matching result is that the difference between the incident pulse signal and the reflected pulse signal is greater than a predetermined difference threshold; the method further includes: in the case that the matching result is that the difference between the incident pulse signal and the reflected pulse signal is greater than the predetermined difference threshold, determining that the power equipment has an abnormality.

[0075] It can be understood that after the matching path between the minimum feature matrix and the maximum feature matrix is calculated, the cumulative distance (or cost) of the matching path is evaluated. This cumulative distance reflects the overall similarity or difference between the incident pulse signal and the reflected pulse signal. According to actual application requirements or empirical data, a predetermined distance threshold (or difference threshold) is set. This threshold is used to determine whether the difference between the signals is within a normal range. If the cumulative distance of the matching path is greater than the predetermined distance threshold, it is determined that the difference between the incident pulse signal and the reflected pulse signal is greater than the predetermined difference threshold. This indicates that there is a significant mismatch between the two signals at the feature level. If the cumulative distance of the matching path is less than or equal to the predetermined distance threshold, it is considered that the difference between the incident pulse signal and the reflected pulse signal is within a normal range. In the case that the matching result indicates that the difference between the incident pulse signal and the reflected pulse signal is greater than the predetermined difference threshold, it is further determined that the power equipment has an abnormality. This abnormality may be related to a partial discharge phenomenon, indicating that the equipment may have a fault or aging problem. If the difference is within the normal range, it is considered that the power equipment is in good operating condition and no significant abnormality is detected.

[0076] The cumulative distance of the matching path quantifies the difference between the incident pulse signal and the reflected pulse signal into a specific value. This quantification provides a clear basis for subsequent fault judgment. When the matching result shows that the signal difference is greater than the predetermined threshold, it can be determined in time that the power equipment has an abnormality, potential faults are discovered in advance, equipment damage is avoided, and the safe operation of the power system is ensured. Through clear threshold judgment and quantitative analysis, the error of human judgment is reduced, and the reliability and stability of the entire fault diagnosis system are enhanced.

[0077] Optionally, after the original time series is transformed into the maximum value feature matrix and the minimum value feature matrix, the sequence dimension is greatly reduced, the time complexity is reduced, and the calculation efficiency is improved. The number of rows of the extreme value feature matrix is the same, representing four-dimensional features, and the variable dimensions are one-to-one corresponding. However, the number of maximum value points and minimum value points of the time series may be different, so the number of columns of the two feature matrices of the same time series will be different. For two time series of different lengths, four extreme value feature matrices are obtained by extreme value feature extraction. The length of the time series, the proportion of the maximum value points, and the proportion of the minimum value points all affect the number of extreme value feature matrices, resulting in different lengths of the columns of the feature matrices although the number of rows is the same. The DTW algorithm can effectively solve the matching problem of unequal lengths, so the DTW algorithm is used to measure the similarity distance between the extreme value feature matrices. Two time series X and Y of lengths n and m are extracted to obtain four extreme value feature matrices: the minimum value feature matrix and (the number of columns is and ), the maximum value feature matrix and (the number of columns is and ). The feature matrices of the same extreme value points are similar, and the dynamic time warping distance is calculated twice. The FLEDTW distance between X and Y is defined as follows:

[0078]

[0079] where is the bending path of the minimum value feature matrix and corresponding to the distance matrix , is the bending path of the maximum value feature matrix and corresponding to the distance matrix , which satisfy the three conditions of boundary, continuity and monotonicity.

[0080] The cumulative cost matrix and is constructed by dynamic programming method, and the above problem is transformed into the following formula:

[0081]

[0082]

[0083] where , , and their definitions are as follows:

[0084]

[0085] wherein: x represents an incident wave signal parameter (such as signal identification), y represents a reflected wave signal parameter (such as signal identification).

[0086] The embodiment increases the multi-dimensional extreme value feature, considers the original sequence shape feature, reasonably matches the extreme value points in the calculation of dynamic time warping, and improves the algorithm measurement accuracy.

[0087] Through the above step S102, the partial discharge signal of the power equipment is collected, wherein the partial discharge signal includes an incident pulse signal and a reflected pulse signal; step S104, according to the local extreme point included in the partial discharge signal, a feature matrix is constructed; step S106, based on the feature matrix, division is performed to obtain a minimum value feature matrix and a maximum value feature matrix; step S108, a dynamic time warping algorithm is used to determine the matching path between the minimum value feature matrix and the maximum value feature matrix; step S110, based on the matching path, a matching result of the incident pulse signal and the reflected pulse signal is obtained. The purpose of matching the incident wave and the reflected wave can be achieved, and the technical effects of improving the anti-noise interference ability and the dynamic adaptability are achieved, thereby solving the technical problems of low matching precision of the incident wave and the reflected wave and weak anti-noise interference ability in the related art.

[0088] Based on the above embodiment and optional embodiment, the application provides an optional implementation, which is applied to the application scene of high-frequency partial discharge pulse incident wave and reflected wave matching based on dynamic time warping. By using the dynamic time warping (DTW) algorithm to effectively perform nonlinear alignment on the two time sequences of incident pulses and reflected pulses in the partial discharge signal, the problems of time delay, amplitude change, noise interference and the like in the signal are overcome, and the precision and robustness of the partial discharge signal matching are significantly improved. The specific implementation process of the method is as follows:

[0089] 1. Signal acquisition and preprocessing: the incident pulses and reflected pulses of the partial discharge signal are collected in real time by the sensor of the power equipment. The collected incident pulse and reflected pulse signals need to undergo a series of preprocessing steps, including denoising, filtering, smoothing, normalization and the like, to remove external noise interference and improve the quality of the signals. Once the signal preprocessing is completed, more consistent and reliable data can be provided for subsequent matching analysis. Figure 2 is a first schematic diagram of an optional partial discharge signal matching method provided by the embodiment of the application, as shown in Figure 2As shown, the graph shows the time-varying trend of the incident pulse and the reflected pulse, with the horizontal axis (X-axis) representing time, from 1 to 24, representing the number of hours in a day (e.g., from 1 am to 12 pm). The vertical axis (Y-axis) represents the numerical value of the pulse, ranging from 0 to 40. The numerical values of the incident pulse and the reflected pulse fluctuate throughout the day. The numerical value of the incident pulse is generally higher than that of the reflected pulse, and reaches a peak at certain time periods (e.g., the 5th hour and the 23rd hour). The numerical value of the reflected pulse is relatively stable, but also fluctuates slightly, rising at the 11th hour and the 23rd hour. The incident pulse and the reflected pulse are analyzed and compared to monitor and evaluate the performance of a certain phenomenon or device.

[0090] 2. Local extreme value feature extraction: By extracting local extreme points from the incident pulse and the reflected pulse signals, the dimensionality of the time series can be significantly reduced, simplifying the signal structure. Specifically, local extreme points include maximum points and minimum points, which can reflect the morphological characteristics of the signal. By performing a difference operation on the time series, local maximum points and local minimum points can be determined. These extreme points usually appear at the positions of signal peaks or troughs, indicating the turning points of the signal. In the process of time series change, the local extreme points of the incident pulse and the reflected pulse represent the change trend of the signal by fitting the slope of the line segment. By connecting adjacent extreme points to form sub-sequence segments, linear fitting is performed on these sub-sequences. The slope of the fitted line segment reflects the change trend of the local time series. Figure 3 is a second schematic diagram of an optional partial discharge signal matching method provided by an embodiment of the present application, as shown in Figure 3 The graph uses different circles to mark the extreme points, including local maximum points (peak points) and local minimum points (trough points). At certain time points (e.g., the 8th hour and the 16th hour), the extreme point positions of the incident pulse and the reflected pulse are similar, indicating that their change trends are similar at these times. At other time points, the extreme point positions and values of the two differ, which may reflect different signal characteristics or system states. By comparing the change trends and extreme points of the two, the health status of the system can be evaluated, and potential faults or abnormalities can be detected in a timely manner.

[0091] 3. Determine the maximum point and the minimum point: For each time series (e.g., incident wave and reflected wave signals), determine whether each point is a local maximum point or a local minimum point. The specific determination method is as follows:

[0092] Local minimum point: when is the current point, and satisfies one of the following conditions:

[0093] , .

[0094] Local maximum point: when is the current point, and one of the following conditions is met:

[0095] , .

[0096] 4. Limit matrix calculation: For the extracted extreme points, a feature matrix is constructed, which includes the size, position and slope of the left and right fitting line segments of each extreme point. The feature matrix calculation process is as follows:

[0097] For the extreme value sequence , the start point and end point of are not included. Convert the extreme value sequence into an extreme value feature matrix , and calculate the slope of the left and right fitting line segments of the extreme points, i.e. the slope of the left fitting line segment of the extreme point:

[0098]

[0099] The slope of the right fitting line segment of the extreme point:

[0100]

[0101] 5. Standardization processing: Each extreme point has a slope of the left and right fitting line segment, which represents the change rate of the left and right line segments of the extreme point. These slopes are important information reflecting the trend of signal change, but since their units and numerical ranges may differ, direct comparison will affect accuracy. In order to eliminate the dimensional problem of the slope, each slope can be converted into the form of a circular angle, which can standardize the measurement of different slopes and convert them into a unified angle value.

[0102] 6、Feature calculation, in the feature vector calculation stage, in order to further improve the calculation efficiency and matching accuracy of dynamic time warping (DTW) algorithm, the extreme value feature matrix is divided into minimum value feature matrix and maximum value feature matrix. This division strategy reduces the redundant information in the calculation process through the refinement of local extreme value, so that each feature matrix only contains the part which has greater influence on the matching result, thereby significantly reducing the complexity of calculation. In order to further improve the accuracy of matching, considering the importance of the starting point and the ending point of the signal in the time sequence feature, the feature vectors of the starting point and the ending point are added in the feature matrix. The feature vector of the starting point represents the starting trend of the signal by calculating the slope of the fitting line segment between the starting point and the first local extreme point; the feature vector of the ending point represents the ending trend of the signal by calculating the slope of the fitting line segment between the ending point and the last local extreme point. The introduction of these feature vectors effectively retains the global information of the time sequence, further enhancing the representation ability of the extreme value feature matrix to the signal form.

[0103] 7、Similarity measurement, after the extreme value feature matrix is standardized and calculated, the dynamic time warping (DTW) algorithm is used to measure the similarity of the extreme value feature matrix. The DTW algorithm is a dynamic programming-based algorithm that can effectively handle the alignment problem between unequal length time sequences. The present application not only considers the position information, extreme point type and fitting slope in the extreme value feature matrix, but also improves the calculation accuracy and robustness of DTW through multi-level feature matching strategy. Specifically, by calculating the matching path between the minimum value feature matrix and the maximum value feature matrix respectively, the calculation complexity is effectively reduced while ensuring the matching accuracy. For each feature matrix, its cumulative cost matrix is calculated, and the best matching path is constructed based on this. This extended DTW algorithm not only improves the accuracy in the matching process, but also effectively adapts to local discharge signals of different lengths and different forms, providing more accurate measurement results in the matching process of incident waves and reflected waves, ensuring the matching accuracy and the efficiency of the algorithm.

[0104] The above-mentioned optional implementation mode at least realizes the following effects:

[0105] (1) The dynamic time warping algorithm adopted by the present application can effectively solve the precision problem of traditional methods in processing local discharge signals. Unlike traditional linear correlation methods, the DTW algorithm can adapt to nonlinear changes in the waveform, overcoming the influence of time delay and amplitude change on waveform matching. In complex power environment, the local discharge signals of cables and equipment often change due to equipment state and environmental factors, and DTW can accurately correspond incident waves and reflected waves through optimized time sequence matching, thereby greatly improving the accuracy and reliability of waveform matching, ensuring accurate identification and early warning of early faults.

[0106] (2) Partial discharge signals are usually accompanied by noise interference in power systems, especially in distribution networks, which may be affected by electromagnetic interference, switching operations of electrical equipment and other factors, resulting in reduced signal quality. Traditional waveform matching methods are usually sensitive to noise, and false matching or missed detection may occur. The present application can maintain high matching degree even in the presence of noise, greatly improving the reliability of the partial discharge fault detection system and ensuring accurate identification and positioning of fault signals.

[0107] (3) The present application optimizes the calculation process of the DTW algorithm, making the waveform matching not only more accurate, but also significantly improving the real-time performance. During the operation of power equipment, the characteristics of partial discharge signals may vary significantly with changes in the environment and fluctuations in the state of the equipment. Traditional methods have poor adaptability to these dynamic changes and are difficult to respond in a timely manner. The DTW algorithm can adjust the matching process in real time according to the changes in the actual signal, adapting to the changing signal characteristics in the operation of the power system. In this way, not only the timeliness of fault detection is improved, but also the demand for real-time monitoring and fault warning of the power system is met, ensuring the reliability and timeliness of the system under different operating conditions.

[0108] It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described herein can be executed in an order different from that shown herein.

[0109] In this embodiment, a partial discharge signal matching device is also provided, which is used to implement the above-mentioned embodiments and preferred embodiments, and will not be described again. As used below, the term "module" "device" can be a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, hardware or a combination of software and hardware implementation is also possible and contemplated.

[0110] According to the embodiments of the present application, a device embodiment for implementing the partial discharge signal matching method is also provided, Figure 4 is a schematic diagram of an optional partial discharge signal matching device according to the embodiments of the present application, as Figure 4 shown, the above-mentioned partial discharge signal matching device comprises: a collection module 402, a feature construction module 404, a division module 406, a path generation module 408, a matching module 410, and the device will be described below.

[0111] The collection module 402 is used to collect the partial discharge signal of the power equipment, wherein the partial discharge signal comprises an incident pulse signal and a reflected pulse signal;

[0112] The feature construction module 404 is connected with the acquisition module 402, and is configured to construct a feature matrix according to local extreme points included in the partial discharge signal.

[0113] The division module 406 is connected with the feature construction module 404, and is configured to divide, based on the feature matrix, to obtain a minimum value feature matrix and a maximum value feature matrix.

[0114] The path generation module 408 is connected with the division module 406, and is configured to determine a matching path between the minimum value feature matrix and the maximum value feature matrix by using a dynamic time warping algorithm.

[0115] The matching module 410 is connected with the path generation module 408, and is configured to obtain a matching result of the incident pulse signal and the reflected pulse signal based on the matching path.

[0116] The partial discharge signal matching device provided in the embodiment has the acquisition module 402, the feature construction module 404, the division module 406, the path generation module 408, and the matching module 410, so as to achieve the purpose of matching the incident wave and the reflected wave, improve the anti-noise interference ability and the dynamic adaptability, and solve the technical problems of low matching precision of the incident wave and the reflected wave and weak anti-noise interference ability in the related art.

[0117] It should be noted that the above modules can be implemented by software or hardware. For the latter, the above modules can be located in the same processor, or be located in different processors in any combination.

[0118] It should be noted that the acquisition module 402, the feature construction module 404, the division module 406, the path generation module 408, and the matching module 410 correspond to steps S102 to S110 in the embodiment, and have the same instances and application scenarios as the corresponding steps, but are not limited to the disclosure in the above embodiment. It should be noted that the above modules can run in a computer terminal as a part of the device.

[0119] It should be noted that the optional or preferred implementation of the embodiment can refer to the related description in the embodiment, which will not be repeated here.

[0120] The partial discharge signal matching device can further include a processor and a memory. The acquisition module 402, the feature construction module 404, the division module 406, the path generation module 408, and the matching module 410 are stored in the memory as program units, and the processor executes the above program units stored in the memory to realize the corresponding functions.

[0121] The processor includes a core, and the core retrieves corresponding program units in the memory. The core can be set as one or more. The memory can include a non-permanent memory in a computer readable medium, a random access memory (RAM), and / or a non-volatile memory such as a read-only memory (ROM) or a flash memory (flash RAM), and the memory includes at least one memory chip.

[0122] Embodiments of the present application provide a non-volatile storage medium, which stores a program. The program is executed by a processor to implement a partial discharge signal matching method.

[0123] Embodiments of the present application provide an electronic device, which includes a processor, a memory, and a program stored in the memory and executable in the processor. The processor executes the program to implement the following steps: collecting a partial discharge signal of a power equipment, wherein the partial discharge signal includes an incident pulse signal and a reflected pulse signal; constructing a feature matrix according to local extreme points included in the partial discharge signal; dividing based on the feature matrix to obtain a minimum value feature matrix and a maximum value feature matrix; determining a matching path between the minimum value feature matrix and the maximum value feature matrix by using a dynamic time warping algorithm; and obtaining a matching result of the incident pulse signal and the reflected pulse signal based on the matching path. The device in this document can be a server, a PC, etc.

[0124] The present application also provides a computer program product, which, when executed in a data processing device, is adapted to execute a program that is initialized with the following method steps: collecting a partial discharge signal of a power equipment, wherein the partial discharge signal includes an incident pulse signal and a reflected pulse signal; constructing a feature matrix according to local extreme points included in the partial discharge signal; dividing based on the feature matrix to obtain a minimum value feature matrix and a maximum value feature matrix; determining a matching path between the minimum value feature matrix and the maximum value feature matrix by using a dynamic time warping algorithm; and obtaining a matching result of the incident pulse signal and the reflected pulse signal based on the matching path.

[0125] Those skilled in the art should understand that embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage devices, etc.) containing computer-usable program code.

[0126] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flow or blocks Figure 1 means for functionally implementing the steps in one or more flow or blocks

[0127] These computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture including instructions which implement the function specified in the flowchart block or blocks. Figure 1 one or more flow or blocks Figure 1 means for functionally implementing the steps in one or more flow or blocks

[0128] The computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart block or blocks. Figure 1 one or more flow or blocks Figure 1 means for functionally implementing the steps in one or more flow or blocks

[0129] In one typical configuration, the computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0130] The memory can include non-persistent memory and / or volatile memory, such as random access memory (RAM) and / or cache memory, for storing instructions and data used and / or generated by the computing device. The memory can also include non-volatile memory, such as read-only memory (ROM), electrically programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), flash memory, or nonvolatile random access memory (RAM) (e.g., fault tolerant RAM), for storing instructions and data used and / or generated by the computing device. The memory is an example of computer readable media.

[0131] Computer-readable media includes permanent and non-permanent, movable and non-movable media that can be implemented by any method or technology to store information. The information can be computer-readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassette, magnetic tape disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information accessible to a computing device. According to the definition herein, computer-readable media does not include transitory media such as modulated data signals and carriers.

[0132] It should also be noted that the terms "comprising", "containing", or any other variant thereof are intended to cover non-exclusive inclusions, such that a process, method, article or apparatus that comprises a list of elements does not only include those elements, but can also include other elements not expressly listed or inherent to such process, method, article or apparatus. Without more limitations, the element defined by the phrase "comprising a" does not exclude the presence of additional identical elements in the process, method, article or apparatus that includes the element.

[0133] Those skilled in the art will appreciate that embodiments of the present application can be provided as a method, a system or a computer program product. Therefore, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer usable program code.

[0134] The above merely provides embodiments of the present application and is not intended to limit the present application. Various modifications and changes can be made to the present application by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principles of the present application shall be included in the scope of the claims of the present application.

Claims

1. A partial discharge signal matching method, characterized in that: include: Collecting partial discharge signals of power equipment, wherein the partial discharge signals include incident pulse signals and reflected pulse signals; constructing a characteristic matrix according to the local extreme points included in the partial discharge signal; Performing division based on the characteristic matrix to obtain a minimum characteristic matrix and a maximum characteristic matrix; Using a dynamic time warping algorithm, determining a matching path between the minimum characteristic matrix and the maximum characteristic matrix; Obtaining a matching result between the incident pulse signal and the reflected pulse signal based on the matching path; The constructing of a characteristic matrix based on the local extreme point included in the partial discharge signal includes: determining the extreme point size and the extreme point position of the local extreme point; determining the left and right slopes of the extreme point of the fitting line segments on the left and right sides of the local extreme point based on the extreme point size and the extreme point position; and obtaining the characteristic matrix based on the extreme point size, the extreme point position, and the left and right slopes of the extreme point; The partial discharge signal is a time series, and the characteristic matrix is ​​obtained based on the extreme point size, the extreme point position, and the left and right slopes of the extreme point, including: converting the left and right slopes of the extreme point into a circular angle; converting the extreme point position into a ratio between the extreme point position and the length of the time series; and obtaining the standardized characteristic matrix based on the extreme point size, the circular angle, and the ratio.

2. The method according to claim 1, characterized in that The collecting of partial discharge signals of power equipment includes: collecting the incident pulse signal and the reflected pulse signal of the partial discharge through a sensor of the power equipment; Preprocessing is performed based on the incident pulse signal and the reflected pulse signal to obtain the partial discharge signal, wherein the preprocessing includes at least one of the following: denoising, filtering, smoothing, and normalization.

3. The method according to claim 1, characterized in that There are multiple local extreme points, and the dynamic time warping algorithm is used to determine the matching path between the minimum feature matrix and the maximum feature matrix, including: Obtaining a starting trend feature based on a slope of a fitting line segment between a starting point of the partial discharge signal and a first local extreme point among a plurality of local extreme points; Obtaining an end trend feature based on a slope of a fitting line segment between a termination point of the partial discharge signal and a last local extreme point among the multiple local extreme points; Based on the starting trend feature, the ending trend feature, the minimum feature matrix and the maximum feature matrix, the dynamic time warping algorithm is used for processing to obtain the matching path.

4. The method according to claim 1, wherein The adopting of the dynamic time warping algorithm to determine the matching path between the minimum feature matrix and the maximum feature matrix includes: Determine a local distance matrix based on the minimum feature matrix and the maximum feature matrix, wherein each element in the local distance matrix represents the distance between the minimum value included in the minimum feature matrix and the maximum value included in the maximum feature matrix; Determining a cumulative cost matrix based on the local distance matrix using the dynamic time warping algorithm; The matching path is determined based on the diagonal of the cumulative cost matrix.

5. The method according to any one of claims 1 to 4, characterized in that Obtaining a matching result between the incident pulse signal and the reflected pulse signal based on the matching path includes: In a case where the matching path is greater than a predetermined distance threshold, determining that the matching result is that a difference between the incident pulse signal and the reflected pulse signal is greater than a predetermined difference threshold; The method further includes: determining that an abnormality exists in the power device when the matching result is that the difference between the incident pulse signal and the reflected pulse signal is greater than the predetermined difference threshold.

6. A partial discharge signal matching device, characterized in that: include: An acquisition module, configured to acquire a partial discharge signal of the power equipment, wherein the partial discharge signal includes an incident pulse signal and a reflected pulse signal; A feature construction module, configured to construct a feature matrix based on local extreme points included in the partial discharge signal; A partitioning module, configured to perform partitioning based on the characteristic matrix to obtain a minimum characteristic matrix and a maximum characteristic matrix; A path generation module, configured to determine a matching path between the minimum feature matrix and the maximum feature matrix using a dynamic time warping algorithm; A matching module, configured to obtain a matching result between the incident pulse signal and the reflected pulse signal based on the matching path; The constructing of a characteristic matrix based on the local extreme point included in the partial discharge signal includes: determining the extreme point size and the extreme point position of the local extreme point; determining the left and right slopes of the extreme point of the fitting line segments on the left and right sides of the local extreme point based on the extreme point size and the extreme point position; and obtaining the characteristic matrix based on the extreme point size, the extreme point position, and the left and right slopes of the extreme point; The partial discharge signal is a time series, and the characteristic matrix is ​​obtained based on the extreme point size, the extreme point position, and the left and right slopes of the extreme point, including: converting the left and right slopes of the extreme point into a circular angle; converting the extreme point position into a ratio between the extreme point position and the length of the time series; and obtaining the standardized characteristic matrix based on the extreme point size, the circular angle, and the ratio.

7. A non-volatile storage medium, characterized in that: The non-volatile storage medium stores a plurality of instructions, and the instructions are suitable for being loaded by a processor and executed by the partial discharge signal matching method according to any one of claims 1 to 5.

8. An electronic device, characterized in that: include: One or more processors and a memory, the memory being used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the partial discharge signal matching method according to any one of claims 1 to 5.

Images