A traveling wave ranging method and system based on dynamic analysis of variable-scale waveforms

Abstract

The present invention provides a traveling wave ranging method and system based on dynamic analysis of variable-scale waveforms, including: screening candidate data segments containing fault points according to a pre-trained anomaly detection neural network; roughly locating the section where the fault point is located based on a waveform morphology movement algorithm for the candidate data segments containing the fault point; and accurately locating the fault point based on the determined section where the fault point is located using a morphology analysis algorithm. The present invention can screen out various data segments near the fault point through an algorithm, and can be used to limit the range of the fault point according to the dynamic analysis of the waveform morphology. The introduction of a traveling wave ranging method based on dynamic analysis of variable-scale waveforms can maximize the analysis of various complex waveforms and improve the ranging accuracy.

Description

Technical Field

[0001] The present invention belongs to the technical field of traveling wave double-terminal fault distance measurement for power transmission lines, and in particular relates to a traveling wave distance measurement method and system based on dynamic analysis of variable-scale waveforms. Background Art

[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.

[0003] The inspection and restoration of fault points in transmission lines play a decisive role in the safe and stable operation of the power system. However, the location of the fault point cannot be manually inspected like looking for a needle in a haystack. The location of the fault point needs to be given by the impedance method or the traveling wave method, and then manually searched and restored. This patent uses the traveling wave ranging method. High-precision ranging can increase the speed of manual search for fault points. The current methods used for traveling wave ranging are wavelet transform and neural network algorithm. However, both methods have shortcomings from the statistical results. When using the wavelet method, the fault location is generally selected as the modulus maximum after the wavelet transform. When using the neural network, the waveform is generally divided into several categories and then the optimized least squares method is used to train the waveform containing the fault point and other waveforms to find the fault point. However, the disadvantage of both is that there are many types of fault waveforms, which makes it impossible to improve the ranging accuracy of complex waveforms in practice. Summary of the invention

[0004] In order to solve the above problems, the present invention proposes a traveling wave ranging method and system based on dynamic analysis of variable-scale waveforms. The present invention forms a data segment with a data point and a small section of other data points around it. Obviously, the data segment near the fault point is different from the data segment at other locations. Therefore, the various data segments near the fault point can be screened out by an algorithm, and the range of the fault point can be limited according to the dynamic analysis of the waveform morphology. The introduction of the traveling wave ranging method based on dynamic analysis of variable-scale waveforms can analyze various complex waveforms to the maximum extent and improve the ranging accuracy.

[0005] According to some embodiments, a first solution of the present invention provides a traveling wave ranging method based on dynamic analysis of variable-scale waveforms, which adopts the following technical solutions:

[0006] A traveling wave ranging method based on variable scale waveform dynamic analysis, comprising:

[0007] Filter candidate data segments containing fault points based on a pre-trained anomaly detection neural network;

[0008] For the candidate data segment containing the fault point, the section where the fault point is located is roughly located based on the waveform morphology movement algorithm;

[0009] Based on the determined section where the fault point is located, the morphological analysis algorithm is used to accurately locate the fault point.

[0010] Furthermore, the step of screening candidate data segments containing fault points according to the pre-anomaly detection neural network includes:

[0011] Construct a training sample set based on the relationship between transmission line data and typical fault points;

[0012] Using the training sample set to train the anomaly detection neural network;

[0013] Acquire transmission line data in real time and screen candidate data segments containing fault points based on the trained anomaly detection neural network.

[0014] Furthermore, for the candidate data segment containing the fault point, the section where the fault point is located is roughly located based on the waveform morphology movement algorithm, specifically:

[0015] Decompose the candidate data segment containing the fault point into a initial high-frequency detail data with different resolutions and 1 approximate data using wavelet algorithm;

[0016] Take the absolute value of the initial high-frequency detail data and normalize them;

[0017] The section where the fault point is located is roughly located according to the amplitude of the initial high-frequency detail data.

[0018] Furthermore, the section where the fault point is located is roughly located according to the amplitude of the initial high-frequency detail data, specifically:

[0019] Add weights to a initial high-frequency detail data arranged in order from high to low resolution and accumulate them to obtain weighted cumulative high-frequency detail data;

[0020] Filter according to the amplitude of weighted accumulated high-frequency detail data.

[0021] Furthermore, the screening according to the amplitude of the weighted cumulative high-frequency detail data includes:

[0022] If there is no non-maximum data location, record the location queue index s is empty, and then the position of the maximum amplitude in the weighted cumulative high-frequency detail data is taken as the section where the fault point is located;

[0023] Otherwise, the weighted accumulated high-frequency detail data is in index s The value at the position is set to zero, and the low-resolution data in the initial high-frequency detail data is set at index s The value at the position is set to zero, and the position of the maximum amplitude in the weighted cumulative high-frequency detail data is recorded to roughly locate the section where the fault point is located.

[0024] Furthermore, the screening according to the frequency of the weighted cumulative high-frequency detail data also includes:

[0025] Find the maximum position of the high-frequency detail data amplitude in front of the maximum position in the initial high-frequency detail data of low resolution, and record it as the first maximum position in front of the current maximum value;

[0026] Reduce the value of the low-resolution initial high-frequency detail data near the first maximum value before the current maximum value to 70% of the original value, find the maximum value before the current first maximum value again and update the maximum value position to the latest first maximum value position, until there is no maximum value before the first maximum value;

[0027] If the position distance between the first maximum value and the maximum value is within the set range, the position queue index s Clear and add the 20 positions before the current maximum value and all positions after it;

[0028] Otherwise, the current maximum position is updated and used as the final maximum position to find the data sample segment, and the final maximum position is moved to the vicinity of the midpoint of the data segment, that is, the final maximum position is in the interval near the midpoint of the data segment.

[0029] Furthermore, based on the determined section where the fault point is located, the fault point is accurately located using a morphological analysis algorithm, specifically:

[0030] The interval near the midpoint of the final maximum value position data segment obtained by screening is used as the morphological analysis interval;

[0031] The data standard deviation of the data interval before the midpoint is obtained by using the highest resolution high-frequency wavelet data of the segment near the midpoint of the final maximum value position data segment obtained by screening, and this is used as the disturbance benchmark;

[0032] Determine the magnification n by making the data maximum values ​​near the midpoint of the initial high-frequency detail data with the highest resolution all smaller than the magnification multiplied by the standard deviation; find the first maximum value near the midpoint that is larger than the magnification multiplied by the standard deviation based on the magnification;

[0033] If the data disturbance of the small interval before the data midpoint of the high-frequency wavelet is small, the first maximum value position before it is taken as the fault point position;

[0034] Otherwise, first remove high-frequency disturbances through morphological filtering, calculate the variance of the data near the midpoint of the linear fitting data, extend the fitted linear function to the interval after the midpoint, subtract the high-frequency data from the fitting data, and set the difference to 0 where it is less than 3 times the variance; smooth the deviation data with EMA (exponential moving average), and then analyze the specific fault point based on the morphology of the smoothed data.

[0035] According to some embodiments, a second solution of the present invention provides a traveling wave ranging system based on dynamic analysis of variable-scale waveforms, which adopts the following technical solutions:

[0036] A traveling wave ranging system based on variable scale waveform dynamic analysis, comprising:

[0037] a data screening module configured to screen candidate data segments containing fault points based on a pre-trained anomaly detection neural network;

[0038] A rough positioning module is configured to roughly locate the section where the fault point is located based on a waveform morphology movement algorithm for the candidate data segment containing the fault point;

[0039] The precise positioning module is configured to precisely locate the fault point by using a morphological analysis algorithm based on the determined section where the fault point is located.

[0040] According to some embodiments, a third aspect of the present invention provides a computer-readable storage medium.

[0041] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps in a traveling wave ranging method based on dynamic analysis of variable-scale waveforms as described in the first aspect above.

[0042] According to some embodiments, a fourth aspect of the present invention provides a computer device.

[0043] A computer device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps in the traveling wave ranging method based on dynamic analysis of variable-scale waveforms as described in the first aspect above are implemented.

[0044] Compared with the prior art, the present invention has the following beneficial effects:

[0045] The present invention only needs to take the data segment composed of the actual fault point and its surrounding data as the normal class sample and the other data points and their surrounding data segments as the abnormal class sample, so there is no need to pay attention to the diversity of sample types caused by the complexity of the waveform. Similarly, the waveform shape movement only needs to make the modulus maximum position fall into a small segment, and there is no need to give the specific location of the fault point. Therefore, it can handle complex waveforms and has good robustness to noise. Finally, we only need to analyze the shape of the wave in a small interval and give the final fault point. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] The accompanying drawings in the specification, which constitute a part of the present invention, are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute improper limitations on the present invention.

[0047] Figure 1 It is a flow chart of a traveling wave ranging method based on dynamic analysis of variable-scale waveforms according to an embodiment of the present invention;

[0048] Figure 2 is a program flow chart of the traveling wave ranging method based on variable scale waveform dynamic analysis according to an embodiment of the present invention;

[0049] Figure 3 It is an algorithm flow chart of an anomaly detection neural network in a traveling wave ranging method based on variable scale waveform dynamic analysis described in an embodiment of the present invention. DETAILED DESCRIPTION

[0050] The present invention will be further described below in conjunction with the accompanying drawings and embodiments.

[0051] It should be noted that the following detailed descriptions are all illustrative and intended to provide further explanation of the present invention. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present invention belongs.

[0052] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit exemplary embodiments according to the present invention. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.

[0053] In the absence of conflict, the embodiments of the present invention and the features of the embodiments may be combined with each other.

[0054] Embodiment 1

[0055] like Figure 1-Figure 3As shown, this embodiment provides a traveling wave ranging method based on dynamic analysis of variable-scale waveforms. This embodiment uses the method applied to a server as an example for illustration. It can be understood that the method can also be applied to a terminal, and can also be applied to a terminal, a server, and a system, and is implemented through the interaction between the terminal and the server. The server can be an independent physical server, or a server cluster or a distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network servers, cloud communications, middleware services, domain name services, security services CDN, and big data and artificial intelligence platforms. The terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart speaker, a smart watch, etc., but is not limited thereto. The terminal and the server can be directly or indirectly connected via wired or wireless communication, and the present invention is not limited here. In this embodiment, the method includes the following steps:

[0056] Step S1: Screening candidate data segments containing fault points based on a pre-trained anomaly detection neural network;

[0057] Step S2: for the candidate data segment containing the fault point, roughly locate the section where the fault point is located based on the waveform morphology movement algorithm;

[0058] Step S3: Based on the determined section where the fault point is located, the fault point is accurately located using a morphological analysis algorithm.

[0059] Specifically, in step S1, the candidate data segments containing the fault points are screened according to the pre-anomaly detection neural network, including:

[0060] Step S1.1: construct a training sample set based on the relationship between transmission line data and typical fault points;

[0061] Step S1.2: training anomaly detection neural network using the training sample set;

[0062] Step S1.3: Acquire the transmission line data in real time, and screen the candidate data segments containing the fault points based on the trained anomaly detection neural network.

[0063] The anomaly detection neural network algorithm is specifically:

[0064] Given a training sample set X = {Xn|n = 1, 2, ..., N} and a Hilbert space kernel generator (hereinafter referred to as RKHS) function Φ, RKHS maps the input data space to the feature space F, and introduces a linear or hyperplane transformation function f(X n )=ω T Φ(X n)-r, in order to separate the features of positive sample data and negative sample data. In order to obtain the coefficient matrix ω and the offset radius r, the following function needs to be optimized globally:

[0065]

[0066] Here ν∈(0,1) is a hyperparameter used to control the distance of the hyperplane from the origin and the amount of deviation of the data sample from the hyperplane, and <·,·> represents the dot product. In order to apply the neural network, this embodiment is modified as follows:

[0067] Add the sigmoid activation function g(·) to transform the original linear transformation into a nonlinear transformation

[0068] Add a weight matrix V to represent the linear mapping from the input space to the latent space

[0069] Replace the RKHS function with g(V·) to make it a learnable parameter

[0070] The new optimized function is:

[0071]

[0072] The strategy for optimizing the objective function is to first fix the parameter r and only adjust the parameters ω and V, then fix ω and V and adjust r. It can be proved that the optimal r is the ν quantile of the sequence <ω, g(V Xn)>.

[0073] The proof is as follows:

[0074] make:

[0075]

[0076]

[0077] Because ω and V are fixed, they can be regarded as constants and do not affect the optimization of the objective function. Therefore, constants can be added or deleted arbitrarily to the objective function containing the variable r. The new objective function can be written as:

[0078]

[0079] Taking the derivative of ||r|| we get:

[0080]

[0081] Let F′(||r||)=0 and we get

[0082]

[0083] here Indicates that N samples satisfy The number of samples, similarly N Indicates that N samples satisfy The number of samples. It can be seen that r is a sequence The ν quantile of , the proof is complete.

[0084] The neural network designed for V and ω learning and the entire anomaly detection algorithm process are as follows Figure 3 shown.

[0085] After the anomaly detection, we can find the data segment containing the fault point. According to the calibrated positive (including the fault point) and negative samples (excluding the fault point), we can view the score of each sample. The lower the score, the more likely it is to contain a fault point. It can be seen that the negative samples and positive samples are clustered in different scoring segments, but there is still some overlap between the two, which means that for most samples, it is possible to distinguish whether they contain a fault point.

[0086] Secondly, we can obtain 32 samples with low scores and 32 samples with high scores from the positive and negative samples respectively, and observe the effect of anomaly detection intuitively. It can be clearly seen that this algorithm can effectively filter out data segments containing fault points and eliminate data segments that obviously do not contain fault points.

[0087] Specifically, in step S2, for the candidate data segment containing the fault point, the section where the fault point is located is roughly located based on the waveform morphology movement algorithm, specifically:

[0088] Step S2.1: Decompose the candidate data segment containing the fault point into a initial high-frequency detail data with different resolutions and 1 approximate data using a wavelet algorithm;

[0089] Step S2.2: taking the absolute values ​​of the initial high-frequency detail data and normalizing them;

[0090] Step S2.3: Roughly locate the section where the fault point is located according to the amplitude of the initial high-frequency detail data.

[0091] Among them, in step S2.3, the section where the fault point is located is roughly located according to the amplitude of the initial high-frequency detail data, specifically:

[0092] Step S2.3.1: adding weights to a pieces of initial high-frequency detail data arranged in descending order of resolution and accumulating the weighted accumulated high-frequency detail data;

[0093] Step S2.3.2: Filter according to the amplitude of the weighted accumulated high-frequency detail data.

[0094] In step S2.3.2, the screening according to the amplitude of the weighted cumulative high-frequency detail data includes:

[0095] If there is no non-maximum data location, record the location queue index s is empty, and then the position of the maximum amplitude in the high-frequency detail data after cumulative calculation is taken as the section where the fault point is located;

[0096] Otherwise, the weighted accumulated high-frequency detail data is in index s The value at the position is set to zero, and the low-resolution data in the initial high-frequency detail data is set at index s The value at the position is set to zero, and the position of the maximum amplitude in the weighted cumulative high-frequency detail data is recorded to roughly locate the section where the fault point is located;

[0097] Find the maximum position of the high-frequency detail data amplitude in front of the maximum position in the initial high-frequency detail data of low resolution, and record it as the first maximum position in front of the current maximum value;

[0098] Reduce the value of the low-resolution initial high-frequency detail data near the first maximum value before the current maximum value to 70% of the original value, find the maximum value before the current first maximum value again and update the maximum value position to the latest first maximum value position, until there is no maximum value before the first maximum value;

[0099] If the position distance between the first maximum value and the maximum value is within the set range, the position queue index s Clear and add the 20 positions before the current maximum value and all positions after it;

[0100] Otherwise, the current maximum position is updated and used as the final maximum position to find the data sample segment, and the final maximum position is moved to the vicinity of the midpoint of the data segment, that is, the final maximum position is in the interval near the midpoint of the data segment.

[0101] Specifically, in order to limit the location of the fault point, the wavelet algorithm can be used to find the fault point at multiple scales and filter the samples of the fault point within the specified interval. The algorithm flow is as shown in Algorithm 1:

[0102] Algorithm 1 Waveform Movement

[0103] Input: Decompose the data segment into 8 high-frequency detail data (recorded as yd1, yd2, yd3, yd4, yd5, yd6, yd7, yd8 in order from high resolution to low resolution) and 1 approximate data ya by using wavelet algorithm, take the absolute value of the high-frequency data of wavelet and normalize them; add weights (e.g., equal weights) to the first 7 high-resolution high-frequency data and accumulate them, recorded as y;

[0104]

[0105]

[0106] Specifically, in step S3, based on the determined section where the fault point is located, the fault point is accurately located using a morphological analysis algorithm, specifically:

[0107] The interval near the midpoint of the final maximum value position data segment obtained by screening is used as the morphological analysis interval;

[0108] The data standard deviation of the data interval before the midpoint is obtained by using the highest resolution high-frequency wavelet data of the segment near the midpoint of the final maximum value position data segment obtained by screening, and this is used as the disturbance benchmark;

[0109] The magnification factor n is determined by making the data maximum values ​​near the midpoint of the initial high-frequency detail data with the highest resolution all smaller than the magnification factor multiplied by the standard deviation;

[0110] Based on the magnification, find the first maximum value near the midpoint that is greater than the magnification multiplied by the standard deviation;

[0111] If the data disturbance of the small interval before the data midpoint of the high-frequency wavelet is small, the first maximum value position before it is taken as the fault point position;

[0112] Otherwise, first remove high-frequency disturbances through morphological filtering, calculate the variance of the data near the midpoint of the linear fitting data, extend the fitted linear function to the interval after the midpoint, make a difference between the high-frequency data and the fitting data, and set the difference to 0 where it is less than 3 times the variance; smooth the deviation data with EMA (exponential moving average), and then analyze the specific fault point according to the shape of the smoothed data (slowly increasing data, rapidly increasing data, convex data, and concave data).

[0113] According to the morphology of the smoothed data, the specific fault point is obtained, which is:

[0114] 1. Data slowly increasing form: Set a threshold and compare the average amplitude of the 20 points before and after the current position. If the ratio of the amplitude before and after does not meet the threshold, continue to move the position in the direction of decreasing amplitude. If it meets the threshold, stop moving the position and return to the fault point.

[0115] 2. Rapidly increasing data form: Set the threshold, analyze the average amplitude ratio of the 20 points before and after the current point, and the average amplitude ratio of the 40 points and 20 points before the current point. If both meet the threshold, move the current point in the direction of amplitude attenuation, otherwise return to the position 40 points lower than the current point.

[0116] 3. Protruding shape: Analyze the symmetry and normality of the protruding shape and set thresholds respectively. If both thresholds are met, skip the protruding shape and analyze the shape of the previous data to see whether it is gradually increasing, rapidly increasing, convex or concave. Otherwise, return to the position;

[0117] 4. Data concave shape: directly return to the current position.

[0118] Specifically, the algorithm in step S2 is used to filter samples and retain the best 32 candidate samples. Among these 32 samples, the samples are arranged from low to high according to the time sequence of each sample in the traveling wave and the median is taken as the sample for morphological analysis. The algorithm for morphological analysis is as shown in Algorithm 2.

[0119] Algorithm 2 Morphological analysis fixed point

[0120] Input: the absolute value yd1 of the sample data segment and its wavelet transformed data; the midpoint of the data segment is denoted as mid, and the analysis interval is [mid-120, mid+900];

[0121]

[0122]

[0123]

[0124] After the morphological analysis algorithm, the fault point determined by the morphological analysis is obtained.

[0125] Embodiment 2

[0126] This embodiment provides a traveling wave ranging system based on variable-scale waveform dynamic analysis, including:

[0127] a data screening module configured to screen candidate data segments containing fault points based on a pre-trained anomaly detection neural network;

[0128] A rough positioning module is configured to roughly locate the section where the fault point is located based on a waveform morphology movement algorithm for the candidate data segment containing the fault point;

[0129] The precise positioning module is configured to precisely locate the fault point by using a morphological analysis algorithm based on the determined section where the fault point is located.

[0130] The examples and application scenarios implemented by the above modules and corresponding steps are the same, but are not limited to the contents disclosed in the above embodiment 1. It should be noted that the above modules as part of the system can be executed in a computer system such as a set of computer executable instructions.

[0131] The description of each embodiment in the above embodiments has different emphases. For parts not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0132] The proposed system can be implemented in other ways. For example, the system embodiment described above is only illustrative, and the division of the modules is only a logical function division. In actual implementation, there may be other division methods, such as multiple modules can be combined or integrated into another system, or some features can be ignored or not executed.

[0133] Embodiment 3

[0134] This embodiment provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the steps in the traveling wave ranging method based on dynamic analysis of variable-scale waveforms as described in the first embodiment above are implemented.

[0135] Embodiment 4

[0136] This embodiment provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the steps in the traveling wave ranging method based on dynamic analysis of variable-scale waveforms as described in the first embodiment are implemented.

[0137] Those skilled in the art will appreciate that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of hardware embodiments, software embodiments, or embodiments combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage and optical storage, etc.) containing computer-usable program code.

[0138] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0139] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0140] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process in the computer or other programmable device. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0141] A person skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes of the embodiments of the above-mentioned methods. The storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.

[0142] Although the above describes the specific implementation mode of the present invention in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present invention. Those skilled in the art should understand that various modifications or variations that can be made by those skilled in the art on the basis of the technical solution of the present invention without creative work are still within the scope of protection of the present invention.

Claims

1. A traveling wave ranging method based on dynamic analysis of variable-scale waveforms, characterized in that: include: Filter candidate data segments containing fault points based on a pre-trained anomaly detection neural network; For the candidate data segment containing the fault point, the section where the fault point is located is roughly located based on the waveform morphology movement algorithm; specifically: Decompose the candidate data segment containing the fault point into a initial high-frequency detail data with different resolutions and 1 approximate data using wavelet algorithm; Take the absolute value of the initial high-frequency detail data and normalize them; The fault point section is roughly located according to the amplitude of the initial high-frequency detail data; specifically: Add weights to a initial high-frequency detail data arranged in order from high to low resolution and accumulate them to obtain weighted cumulative high-frequency detail data; Filter based on the amplitude of weighted accumulated high-frequency detail data, including: If there is no non-maximum data location, record the location queue index s is empty, and then the position of the maximum amplitude in the weighted cumulative high-frequency detail data is taken as the section where the fault point is located; Otherwise, the weighted accumulated high-frequency detail data is in index s The value at the position is set to zero, and the low-resolution data in the initial high-frequency detail data is set at index s The value at the position is set to zero, and the position of the maximum amplitude in the weighted cumulative high-frequency detail data is recorded to roughly locate the section where the fault point is located; Also includes: Find the maximum position of the high-frequency detail data amplitude in front of the maximum position in the initial high-frequency detail data of low resolution, and record it as the first maximum position in front of the current maximum value; Reduce the value of the low-resolution initial high-frequency detail data near the first maximum value before the current maximum value to 70% of the original value, find the maximum value before the current first maximum value again and update the maximum value position to the latest first maximum value position, until there is no maximum value before the first maximum value; If the position distance between the first maximum value and the maximum value is within the set range, the position queue index s Clear and add the 20 positions before the current maximum value and all positions after it; Otherwise, the current maximum value position is updated and used as the final maximum value position to find the data sample segment, and the final maximum value position is moved to the vicinity of the midpoint of the data segment, that is, the final maximum value position is in the interval near the midpoint of the data segment; Based on the determined section where the fault point is located, the morphological analysis algorithm is used to accurately locate the fault point; Specifically: The interval near the midpoint of the final maximum value position data segment obtained by screening is used as the morphological analysis interval; The data standard deviation of the data interval before the midpoint is obtained by using the highest resolution high-frequency wavelet data of the segment near the midpoint of the final maximum value position data segment obtained by screening, and this is used as the disturbance benchmark; Determine the magnification n by making the data maximum values ​​near the midpoint of the initial high-frequency detail data with the highest resolution all smaller than the magnification multiplied by the standard deviation; find the first maximum value near the midpoint that is larger than the magnification multiplied by the standard deviation based on the magnification; If the data disturbance of the small interval before the data midpoint of the high-frequency wavelet is small, the first maximum value position before it is taken as the fault point position; Otherwise, first use the morphological filtering method to remove high-frequency disturbances, calculate the variance of the data near the midpoint of the linear fitting data, extend the fitted linear function to the interval after the midpoint, subtract the high-frequency data from the fitting data, and set the difference to 0 where it is less than 3 times the variance; perform EMA smoothing on the deviation data, and then analyze the specific fault point based on the morphology of the smoothed data.

2. A traveling wave ranging method based on variable scale waveform dynamic analysis as claimed in claim 1, characterized in that: The step of screening candidate data segments containing fault points according to the pre-anomaly detection neural network includes: Construct a training sample set based on the relationship between transmission line data and typical fault points; Using the training sample set to train the anomaly detection neural network; Acquire transmission line data in real time and screen candidate data segments containing fault points based on the trained anomaly detection neural network.

3. A traveling wave ranging system based on dynamic analysis of variable-scale waveforms, using a traveling wave ranging method based on dynamic analysis of variable-scale waveforms as claimed in claim 1, characterized in that: include: a data screening module configured to screen candidate data segments containing fault points based on a pre-trained anomaly detection neural network; A rough positioning module is configured to roughly locate the section where the fault point is located based on a waveform morphology movement algorithm for the candidate data segment containing the fault point; The precise positioning module is configured to precisely locate the fault point by using a morphological analysis algorithm based on the determined section where the fault point is located.

4. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps in a traveling wave ranging method based on dynamic analysis of variable-scale waveforms as described in any one of claims 1-2 are implemented.

5. A computer device comprising a memory, a processor and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps in the traveling wave ranging method based on variable-scale waveform dynamic analysis as described in any one of claims 1-2 are implemented.

Images