A cable fault intelligent positioning diagnosis method and system based on double-end traveling wave

Abstract

This invention discloses a method and system for intelligent cable fault location and diagnosis based on double-ended traveling waves, belonging to the field of cable fault location and diagnosis technology. It includes simultaneously setting up monitoring systems at both ends of the cable line, designing an adaptive signal adjustment mechanism, processing double-ended data of transmitted and received signals based on double-ended traveling waves, constructing a topology model of the cable network using the characteristics of the double-ended traveling wave data, and training a fault training model using the double-ended data to locate the fault point, thus completing the intelligent location diagnosis of the cable fault. This invention solves the problems of signal attenuation and noise interference, enhances the ability to identify small partial discharge phenomena, achieves accurate location of fault points in complex multi-branch cable networks, can update the cable network status in real time, adapts to environmental changes, and ensures the timeliness of diagnostic results.

Description

Technical Field

[0001] This invention relates to the field of cable fault location and diagnosis technology, and in particular to a cable fault intelligent location and diagnosis method and system based on double-ended traveling waves. Background Technology

[0002] With the continuous development of the global economy and the acceleration of urbanization, modern society has placed higher demands on the stability and reliability of power supply. Especially in the context of smart cities and industry, the development of various high-tech industries and the improvement of residents' living standards have made it necessary for power networks to not only cover a wider area but also be able to carry a greater load. In this context, the safety and stability of cables, as an important medium for power transmission, have become particularly important. However, due to natural environmental factors (such as extreme weather), human damage, and other unforeseen problems, cable faults occur frequently, posing a serious threat to the normal operation of the power system. Therefore, there is an urgent need for an efficient and accurate cable fault location and diagnosis method to ensure that the power system can be restored to normal operation in the shortest possible time to reduce economic losses and social impact.

[0003] Current intelligent fault location and diagnosis technologies for cables mainly rely on traditional methods such as the single-end traveling wave method or the bridge method. These methods have revealed some limitations in practical applications. When dealing with cable lines in complex terrain conditions, the traditional single-end traveling wave method is difficult to accurately identify small partial discharge phenomena located at the far end of the cable. This is mainly due to the influence of signal attenuation and noise interference. When it comes to multi-branch cable networks, existing technologies often cannot effectively distinguish whether the fault occurs in the main cable or the branch cable, resulting in a decrease in location accuracy. The insufficient ability to identify small partial discharge phenomena and the inaccurate fault location in multi-branch cable networks have become a challenge. Summary of the Invention

[0004] In view of the problems existing in the current intelligent fault location and diagnosis methods and systems for cables based on double-ended traveling waves, this invention is proposed.

[0005] Therefore, this invention addresses the problems of signal attenuation, noise interference, and insufficient identification of small partial discharges in the single-ended traveling wave method under complex terrain and multi-branch cable networks. This invention improves positioning accuracy and reliability by using double-ended traveling wave data processing and convolutional neural network feature extraction.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0007] In a first aspect, embodiments of the present invention provide a cable fault intelligent location and diagnosis method based on double-ended traveling waves, which includes: synchronously setting up monitoring systems at both ends of the cable line; the monitoring system includes a signal transmitting unit and a signal receiving unit; designing an adaptive signal adjustment mechanism to adjust the environmental condition parameters of the signal transmitting unit and the signal receiving unit; and synchronizing the signal transmission and signal reception times.

[0008] Based on the processing of two-end traveling wave data of transmitted and received signals, a convolutional neural network is used to extract the features of the two-end traveling wave data during the processing. The topology model of the cable network is constructed using the features of the two-end traveling wave data. The topology model includes a recurrent neural network model and a two-end data fault training model.

[0009] By using a topology model of the cable network, a database is established based on the original cable length and propagation time. The location of the fault point is calculated by using the time difference between the transmitted and received signals tracked by the database and a recurrent neural network model. A dual-end data fault training model is then used to train the fault point to complete the intelligent location diagnosis of cable faults.

[0010] As a preferred embodiment of the intelligent location and diagnosis method for cable faults based on double-ended traveling waves described in this invention, the method of synchronously setting up a monitoring system at both ends of the cable line includes using time synchronization technology of a satellite navigation system to obtain the parameter configuration of the signal transmitting unit and the signal receiving unit, and combining it with an atomic clock calibration module to dynamically correct the electromagnetic interference problem of the time synchronization technology caused by temperature changes. The atomic clock calibration module includes obtaining the parameter configuration of the transmitting unit and the signal receiving unit for nanosecond-level time synchronization application scenarios by utilizing the quantum properties of atoms under electromagnetic radiation.

[0011] As a preferred embodiment of the intelligent location and diagnosis method for cable faults based on double-ended traveling waves described in this invention, the adaptive signal adjustment mechanism includes establishing a prediction model by combining acquired parameter configurations with historical data. The historical data includes a class of environmental condition parameters collected during previous signal transmission processes. The signal transmission process is adjusted based on the prediction model.

[0012] The establishment of the prediction model includes fusing multi-source heterogeneous data of a class of environmental condition parameters, such as temperature, humidity, and electromagnetic interference intensity, with signal quality indicators such as signal-to-noise ratio, bit error rate, and cable operating status. The cable operating status includes the cable's load current and voltage fluctuations.

[0013] The multi-source heterogeneous data fusion includes using variational autoencoders to enhance the features of a class of environmental condition parameters, extracting deep-level patterns hidden in the load current and voltage fluctuation data of cables. These deep-level patterns include the nonlinear variation patterns of signal-to-noise ratio and bit error rate in the time series. The time series includes aligning time series of different lengths using dynamic time warping based on the time series features in historical data, constructing a hybrid model of long short-term memory network and attention mechanism, designing an intelligent agent based on reinforcement learning, and dynamically adjusting the frequency, power, and modulation mode of the signal transmission unit according to the obtained parameter configuration and signal quality indicators of the signal transmission and reception units to synchronize the time of signal transmission and reception.

[0014] As a preferred embodiment of the intelligent location and diagnosis method for cable faults based on double-ended traveling waves described in this invention, the processing of the double-ended data of the transmitted and received signals includes introducing variational mode decomposition to decompose the received signal during time synchronization and extracting the partial discharge characteristics of different frequency components of the synchronized signal.

[0015] The decomposition of the signal received during time synchronization includes, based on the variational mode decomposition method, an adaptive multi-scale variational mode decomposition method is proposed to accurately decompose the time-synchronized received signal. The adaptive multi-scale variational mode decomposition method includes dynamically adjusting the decomposition scale and center frequency range of the synchronized signal.

[0016] The decomposition scale includes performing spectral analysis on the synchronized signal using Fast Fourier Transform to extract features from the spectral distribution. These features include the main bandwidth, energy distribution ratio, high-frequency noise, and the proportion of low-frequency drift. Based on the extracted features, a dynamic K-value calculation formula based on energy segmentation is proposed.

[0017] ;

[0018] in, Indicates the decomposition scale. This indicates the total number of sub-bands in the spectrum allocation. Indicates the first The energy percentage of each sub-band This indicates the preset energy threshold. Indicates an indicator function;

[0019] The center frequency range includes frequency peaks whose features are extracted from the spectral distribution using a peak detection algorithm. The calculation formula for the peak detection algorithm is as follows:

[0020] ;

[0021] in, Indicates frequency At, spectrum The first derivative, Indicates frequency At, spectrum The second derivative, Indicates frequency, This represents a user-defined scaling factor. Indicates the threshold parameter. Represents the mean of the spectrum. The standard deviation of the spectrum This represents the set of peak points.

[0022] As a preferred embodiment of the cable fault intelligent location and diagnosis method based on double-ended traveling waves described in this invention, the center frequency range further includes introducing signal-to-noise ratio as a constraint condition to optimize the center frequency, wherein the optimization of the center frequency includes performing local signal-to-noise ratio calculation on the frequency bands near each center frequency.

[0023] When the signal-to-noise ratio is lower than the set threshold When a low-frequency band is set, that band consists of noise, and the center frequency corresponding to the noise is excluded.

[0024] When the signal-to-noise ratio is higher than or equal to the set threshold When a high-frequency band is set, that band contains effective signal components, and the corresponding noise center frequency is retained;

[0025] The threshold This includes setting thresholds based on the statistical characteristics of the signal-to-noise ratio distribution. ;

[0026] The formula for calculating the nearby frequency band range is:

[0027] ;

[0028] in, Indicates the bandwidth. Indicates the center frequency;

[0029] The formula for calculating the local signal-to-noise ratio of the frequency bands near each center frequency is as follows:

[0030] ;

[0031] in, Indicates signal power. Indicates noise power. The local signal-to-noise ratio represents the center frequency.

[0032] As a preferred embodiment of the intelligent location and diagnosis method for cable faults based on double-ended traveling waves described in this invention, the method of using a convolutional neural network to extract double-ended traveling wave data features during the processing includes extracting double-ended traveling wave data features of peak point positions by capturing the temporal distribution characteristics of the signal through a convolutional layer. The convolutional neural network includes a one-dimensional convolutional neural network and a two-dimensional convolutional neural network.

[0033] The transmitted and received signals are converted into CNN input form, which is represented as a one-dimensional time series or a two-dimensional time-frequency spectrum.

[0034] If it is time series data, it is input into a one-dimensional convolutional neural network;

[0035] If it is a time-frequency spectrum, it is input into a two-dimensional convolutional neural network;

[0036] When the time distribution characteristics are captured in a one-dimensional convolutional neural network, the one-dimensional convolutional neural network signal is scanned to extract the double-ended traveling wave data features.

[0037] When the temporal distribution characteristics are captured in a two-dimensional convolutional neural network, the double-ended traveling wave data features are extracted by sliding the convolution kernel on the time axis and frequency axis.

[0038] The extracted features of the double-ended traveling wave data are represented in vector form, as shown in the formula:

[0039] ;

[0040] in, Represents the feature vector of a two-ended traveling wave data. Representing feature dimension, This represents the feature dimension of the first double-ended traveling wave data feature vector. This represents the feature dimension of the second double-ended traveling wave data feature vector. Indicates the first The feature dimension of a double-ended traveling wave data feature vector;

[0041] The construction of the cable network topology model includes combining the characteristics of the double-ended traveling wave data with historical data, and training the model through double-ended data faults to construct the topology.

[0042] The formula for combining the characteristics of the double-ended traveling wave data with the historical data is as follows:

[0043] ;

[0044] in, Indicates the weighting coefficient. Represents the feature vector of a two-ended traveling wave data. Represents historical data. This represents the combined feature vector;

[0045] The topology training using a dual-end data fault training model includes combining the feature vectors of nodes in the cable network. This means that the nodes in the cable network are adjusted using the output of the dual-end data fault training model;

[0046] When a fault is detected at a node in the cable network, the weighting factor is reduced.

[0047] As a preferred embodiment of the intelligent cable fault location and diagnosis method based on double-ended traveling waves described in this invention, the following steps are included: establishing a database, which involves creating a cable network topology database based on the N node locations of the cable, and training the cable network topology database using a double-ended data fault training model. The training of the cable network topology database includes capturing a loss function in the database and incorporating double-ended traveling wave data features into the loss function in real time, so that the model can adapt to changes in the cable network and complete the intelligent location and diagnosis of cable faults.

[0048] Secondly, embodiments of the present invention provide a cable fault intelligent location and diagnosis system based on double-ended traveling waves, comprising: an adjustment mechanism module, which synchronously sets up monitoring systems at both ends of the cable line and designs an adaptive signal adjustment mechanism; a processing module, which processes the double-ended data of the transmitted and received signals based on the double-ended traveling waves, extracts the double-ended traveling wave data features during the processing using a convolutional neural network, and constructs a topology model of the cable network using the double-ended traveling wave data features; and a diagnosis module, which uses the topology model of the cable network and a double-ended data fault training model to train the fault location, thereby completing the intelligent location and diagnosis of cable faults.

[0049] Thirdly, embodiments of the present invention provide a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement any of the steps of the above-described intelligent location and diagnosis method for cable faults based on double-ended traveling waves.

[0050] Fourthly, embodiments of the present invention provide a computer-readable storage medium having a computer program stored thereon, wherein: when the computer program is executed by a processor, it implements any step of the above-described intelligent location and diagnosis method for cable faults based on double-ended traveling waves.

[0051] The beneficial effects of this invention are as follows: By combining double-ended traveling wave data processing with intelligent algorithms, this invention improves the accuracy and efficiency of cable fault location. Utilizing a double-ended synchronous monitoring system and an adaptive signal adjustment mechanism, it solves the problems of signal attenuation and noise interference, enhancing the ability to identify small partial discharge phenomena. By introducing a convolutional neural network to extract features and construct a cable network topology model, it achieves accurate fault location in complex multi-branch cable networks. The signal processing process is optimized through dynamic time warping and variational mode decomposition, improving the accuracy and reliability of data processing. The database established based on a recurrent neural network and a double-ended data fault training model can update the cable network status in real time, adapting to environmental changes and ensuring the timeliness of diagnostic results. Attached Figure Description

[0052] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort. Wherein:

[0053] Figure 1 The flowchart illustrates a cable fault intelligent location and diagnosis method and system based on double-ended traveling waves, as provided in one embodiment of the present invention.

[0054] Figure 2 This is a schematic diagram of intelligent fault location for a cable fault location and diagnosis method and system based on double-ended traveling waves, provided as an embodiment of the present invention. Detailed Implementation

[0055] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0056] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0057] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0058] This invention is described in detail with reference to the schematic diagrams. When detailing the embodiments of this invention, for ease of explanation, the cross-sectional views illustrating the device structure may be partially enlarged, not adhering to the usual scale. Furthermore, the schematic diagrams are merely examples and should not be construed as limiting the scope of protection of this invention. In actual fabrication, the three-dimensional spatial dimensions of length, width, and depth should be included.

[0059] Furthermore, in the description of this invention, it should be noted that the terms "upper," "lower," "inner," and "outer," etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. These terms are used solely for the convenience of describing the invention and for simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention. In addition, the terms "first," "second," or "third" are used for descriptive purposes only and should not be construed as indicating or implying relative importance.

[0060] Unless otherwise explicitly specified and limited, the terms "installation," "connection," and "joining" in this invention should be interpreted broadly. For example, they can refer to fixed connections, detachable connections, or integral connections; similarly, they can refer to mechanical connections, electrical connections, or direct connections, or indirect connections through an intermediate medium, or internal connections between two components. Those skilled in the art can understand the specific meaning of the above terms in this invention based on the specific circumstances.

[0061] Example 1

[0062] Reference Figure 1 and Figure 2 This is the first embodiment of the present invention, which provides a cable fault intelligent location and diagnosis method based on double-ended traveling waves, including:

[0063] S1: A monitoring system is installed synchronously at both ends of the cable line. The monitoring system includes a signal transmitting unit and a signal receiving unit. An adaptive signal adjustment mechanism is designed to adjust the environmental condition parameters of the signal transmitting unit and the signal receiving unit to synchronize the signal transmission and signal reception times.

[0064] Among them, the monitoring system installed at both ends of the cable line includes using the time synchronization technology of the satellite navigation system to obtain the parameter configuration of the signal transmitting unit and the signal receiving unit. Combined with the atomic clock calibration module, it dynamically corrects the electromagnetic interference problem of the time synchronization technology caused by temperature changes. The atomic clock calibration module includes obtaining the parameter configuration of the transmitting unit and the signal receiving unit for nanosecond-level time synchronization application scenarios by utilizing the quantum properties of atoms under electromagnetic radiation.

[0065] S1.1: The design of an adaptive signal adjustment mechanism includes establishing a prediction model by combining the acquired parameter configuration with historical data. The historical data includes a type of environmental condition parameters collected in previous signal transmission processes. The signal transmission process is adjusted based on the prediction model.

[0066] Establishing a predictive model involves fusing a class of environmental condition parameters, such as temperature, humidity, and electromagnetic interference intensity, with signal quality indicators, such as signal-to-noise ratio, bit error rate, and cable operating status, using multi-source heterogeneous data. Cable operating status includes the cable's load current and voltage fluctuations.

[0067] Multi-source heterogeneous data fusion includes using variational autoencoders to enhance the features of a class of environmental condition parameters, extracting deep patterns hidden in the load current and voltage fluctuation data of cables, including the nonlinear variation of signal-to-noise ratio and bit error rate in the time series; the time series includes using dynamic time warping to align time series of different lengths based on the time series features in historical data, constructing a hybrid model of long short-term memory network and attention mechanism, designing a reinforcement learning-based intelligent agent, and dynamically adjusting the frequency, power and modulation mode of the signal transmission unit according to the acquired parameter configuration and signal quality index of the signal transmission and reception units to synchronize the time of signal transmission and reception.

[0068] S2: Based on the two-end traveling wave, the two-end data of the transmitted and received signals are processed. The features of the two-end traveling wave data in the processing process are extracted using a convolutional neural network. The topology model of the cable network is constructed using the features of the two-end traveling wave data. The topology model includes a recurrent neural network model and a two-end data fault training model.

[0069] The processing of the two-end data of the transmitted and received signals includes introducing variational mode decomposition to decompose the received signal during time synchronization and extracting the partial discharge characteristics of different frequency components of the synchronized signal.

[0070] The decomposition of the signal received during time synchronization includes the proposal of an adaptive multi-scale variational mode decomposition method based on the variational mode decomposition method. This method is used to accurately decompose the time-synchronized received signal. The adaptive multi-scale variational mode decomposition method includes dynamically adjusting the decomposition scale and center frequency range of the synchronized signal.

[0071] The decomposition scale involves performing spectral analysis on the synchronized signal using Fast Fourier Transform (FFT) to extract features from the spectral distribution. These features include the dominant bandwidth, energy distribution ratio, high-frequency noise, and the proportion of low-frequency drift. Based on the extracted features, a dynamic K-value calculation formula based on energy segmentation is proposed.

[0072] ;

[0073] in, Indicates the decomposition scale. This indicates the total number of sub-bands in the spectrum allocation. Indicates the first The energy percentage of each sub-band This indicates the preset energy threshold. Indicates an indicator function;

[0074] The center frequency range includes the frequency peaks from which features are extracted using a peak detection algorithm. The calculation formula for the peak detection algorithm is as follows:

[0075] ;

[0076] in, Indicates frequency At, spectrum The first derivative, Indicates frequency At, spectrum The second derivative, Indicates frequency, This represents a user-defined scaling factor. Indicates the threshold parameter. Represents the mean of the spectrum. The standard deviation of the spectrum This represents the set of peak points.

[0077] Furthermore, when processing the transmitted and received signals based on the two-end traveling wave method, variational mode decomposition (VMD) is introduced to decompose the time-synchronized signal, extracting the partial discharge characteristics of different frequency components. The spectrum is divided into 10 sub-bands. Analysis reveals that the main frequency bandwidth is concentrated in the range of 30 kHz to 70 kHz, with the fourth sub-band having the highest energy proportion, reaching 40%. High-frequency noise accounts for 15% of the total energy, and low-frequency drift accounts for 10%. The optimal decomposition scale is calculated to be 6 segments. The center frequency range is optimized using a peak detection algorithm. Assuming a frequency of 50 kHz, the point is calculated to meet the peak condition and included in the peak point set to further improve the signal decomposition accuracy. The processed and extracted data from the two-end data are shown in Table 1 below.

[0078] Table 1. Data Extraction Table for Processing Dual-End Data

[0079] Parameter name Numerical values ​​and units Remark Number of sub-bands 10 Total number of sub-bands in the spectrum allocation Lower limit of main frequency bandwidth 30 kHz Lower limit of the main frequency bandwidth range Main frequency bandwidth limit 70 kHz Upper limit of main frequency bandwidth Energy percentage of the 4th sub-band 40% The sub-band with the highest energy proportion in the main frequency band High-frequency noise energy ratio 15% The proportion of high-frequency noise in the total energy Low-frequency drift energy ratio 10% The proportion of low-frequency drift in total energy Optimal Decomposition Scale 6 paragraphs The calculated optimal decomposition scale Peak detection frequency 50 kHz Frequency points in peak detection algorithms

[0080] Table 1 summarizes the key data in the cable fault location method based on double-ended traveling waves. Table 2 shows that the frequency spectrum is divided into 10 sub-bands, with a main frequency bandwidth ranging from 30 to 70 kHz. The fourth sub-band has the highest energy share, reaching 40%, while high-frequency noise and low-frequency drift account for 15% and 10% of the total energy, respectively. The optimal decomposition scale is calculated to be 6 segments, and the peak point at 50 kHz is determined in the peak detection algorithm. These data provide important basis for signal decomposition and feature extraction, thus effectively supporting the construction of cable network topology models and fault diagnosis.

[0081] S2.1: The center frequency range also includes introducing the signal-to-noise ratio as a constraint to optimize the center frequency. Optimizing the center frequency includes calculating the local signal-to-noise ratio of the frequency bands near each center frequency.

[0082] When the signal-to-noise ratio is lower than the set threshold When a low-frequency band is set, that band consists of noise, and the center frequency corresponding to the noise is excluded.

[0083] When the signal-to-noise ratio is higher than or equal to the set threshold When a high-frequency band is set, that band contains effective signal components, and the corresponding noise center frequency is retained;

[0084] threshold This includes setting thresholds based on the statistical characteristics of the signal-to-noise ratio distribution. ;

[0085] The formula for calculating the nearby frequency band range is:

[0086] ;

[0087] in, Indicates the bandwidth. Indicates the center frequency;

[0088] The formula for calculating the local signal-to-noise ratio in the frequency bands near each center frequency is as follows:

[0089] ;

[0090] in, Indicates signal power. Indicates noise power. The local signal-to-noise ratio represents the center frequency.

[0091] S2.2: Using convolutional neural networks to extract double-ended traveling wave data features during the processing includes extracting double-ended traveling wave data features of peak point positions by capturing the temporal distribution characteristics of the signal through convolutional layers. Convolutional neural networks include one-dimensional convolutional neural networks and two-dimensional convolutional neural networks.

[0092] The transmitted and received signals are converted into CNN input form, which is represented as a one-dimensional time series or a two-dimensional time-frequency spectrum.

[0093] If it is time series data, it is input into a one-dimensional convolutional neural network;

[0094] If it is a time-frequency spectrum, it is input into a two-dimensional convolutional neural network;

[0095] When the time distribution characteristics are captured in a one-dimensional convolutional neural network, the one-dimensional convolutional neural network signal is scanned to extract the double-ended traveling wave data features.

[0096] When the temporal distribution characteristics are captured in a two-dimensional convolutional neural network, the double-ended traveling wave data features are extracted by sliding the convolution kernel on the time axis and frequency axis.

[0097] The extracted features of the double-ended traveling wave data are represented in vector form, as shown in the formula:

[0098] ;

[0099] in, Represents the feature vector of a two-ended traveling wave data. Representing feature dimension, This represents the feature dimension of the first double-ended traveling wave data feature vector. This represents the feature dimension of the second double-ended traveling wave data feature vector. Indicates the first The feature dimension of a double-ended traveling wave data feature vector;

[0100] Constructing a topology model for a cable network involves combining the characteristics of double-ended traveling wave data with historical data, and training the model for topology through double-ended data fault training.

[0101] The formula that combines the characteristics of double-ended traveling wave data with historical data is:

[0102] ;

[0103] in, Indicates the weighting coefficient. Represents the feature vector of a two-ended traveling wave data. Represents historical data. This represents the combined feature vector;

[0104] Topology training through dual-end data fault training models includes combining the feature vectors of nodes in the cable network. This means that the nodes in the cable network are adjusted using the output of the dual-end data fault training model;

[0105] When a fault is detected at a node in the cable network, the weighting factor is reduced.

[0106] S3: Utilize the topology model of the cable network, combine the original cable length and propagation time to establish a database, use the database and the time difference between the transmitted and received signals tracked by the recurrent neural network model to calculate the location of the fault point, and use a dual-end data fault training model to train the fault point to complete the intelligent location diagnosis of cable faults.

[0107] The process of establishing the database includes creating a cable network topology database based on the N node locations of the cable, training the cable network topology database using a two-end data fault training model, and capturing the loss function in the database by using two-end traveling wave data features in real time to incorporate the loss function into the database. This process enables the model to adapt to changes in the cable network and complete the diagnosis of intelligent cable fault location.

[0108] In a preferred embodiment, a cable fault intelligent location and diagnosis system based on double-ended traveling waves includes an adjustment mechanism module, which synchronously sets up monitoring systems at both ends of the cable line and designs an adaptive signal adjustment mechanism; a processing module, which processes the double-ended data of the transmitted and received signals based on the double-ended traveling waves, uses a convolutional neural network to extract the double-ended traveling wave data features during the processing, and uses the double-ended traveling wave data features to construct a topology model of the cable network; and a diagnosis module, which uses the topology model of the cable network and a double-ended data fault training model to train the fault location, thereby completing the intelligent location and diagnosis of cable faults.

[0109] The above-mentioned unit modules can be embedded in the processor of the computer device in hardware form or independent of it, or they can be stored in the memory of the computer device in software form, so that the processor can call and execute the corresponding operations of the above modules.

[0110] The computer device can be a terminal, comprising a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad on the computer device's casing, or an external keyboard, touchpad, or mouse.

[0111] In summary, this invention improves the accuracy and efficiency of cable fault location by combining double-ended traveling wave data processing with intelligent algorithms. It solves the problems of signal attenuation and noise interference by utilizing a double-ended synchronous monitoring system and an adaptive signal adjustment mechanism, enhancing the ability to identify small partial discharge phenomena. By introducing a convolutional neural network to extract features and construct a cable network topology model, it achieves precise fault location in complex multi-branch cable networks. The signal processing process is optimized using dynamic time warping and variational mode decomposition methods, improving the accuracy and reliability of data processing. The database established based on a recurrent neural network and a double-ended data fault training model can update the cable network status in real time, adapting to environmental changes and ensuring the timeliness of diagnostic results.

[0112] Example 2

[0113] Reference Figure 1 and Figure 2 This is the second embodiment of the present invention, which provides a cable fault intelligent location and diagnosis method based on double-ended traveling waves. In order to verify the beneficial effects of the present invention, a simulation experiment is conducted for scientific demonstration.

[0114] During the optimization of the center frequency, the statistical analysis of the signal-to-noise ratio (SNR) distribution showed an average SNR of 15 dB and a standard deviation of 3 dB. With a threshold set at 12 dB, the formula for calculating the frequency range near the center frequency assumes a center frequency of 50 kHz and a bandwidth of 10 kHz, resulting in a range of 45 kHz to 55 kHz. When calculating the local SNR for each frequency band near the center frequency, assuming a signal power of 0.8 and a noise power of 0.2, the formula is used to calculate... The local signal-to-noise ratio (SNR) at the center frequency is 4 dB. Since this value is lower than the set threshold of 12 dB, this frequency band is classified as a low-frequency band, and the corresponding center frequency of 50 kHz is excluded. For another center frequency of 70 kHz, the surrounding frequency range is 65 kHz to 75 kHz. Calculated using the same method, the local SNR is 16 dB, which is higher than the set threshold of 12 dB. Therefore, this frequency band is classified as a high-frequency band, and the corresponding center frequency of 70 kHz is retained. The effective optimization of the center frequency is shown in Table 2 below.

[0115] Table 2 Center Frequency Optimization Data Table

[0116] Parameter name Numerical values ​​and units Remark Average signal-to-noise ratio 15 dB Statistical analysis results of signal-to-noise ratio distribution Signal-to-noise ratio standard deviation 3 dB Statistical analysis results of signal-to-noise ratio distribution threshold 12 dB Threshold set according to signal-to-noise ratio distribution Center frequency 1 50 kHz Initial center frequency Bandwidth 10 kHz Calculation parameters for the frequency band range near the center frequency Frequency band range (50 kHz) 45-55 kHz Frequency range around 50 kHz center frequency Signal power (50 kHz) 0.8 Signal power at a center frequency of 50 kHz Noise power (50 kHz) 0.2 Noise power at a center frequency of 50 kHz Local signal-to-noise ratio (50 kHz) 4 dB Local signal-to-noise ratio at a center frequency of 50 kHz Center frequency 2 70 kHz Another initial center frequency Frequency band range (70 kHz) 65-75 kHz Frequency range around 70 kHz center frequency Local signal-to-noise ratio (70 kHz) 16 dB Local signal-to-noise ratio at center frequency 70 kHz

[0117] Table 2 shows key data in the center frequency optimization process, including the average and standard deviation of the signal-to-noise ratio (SNR), the set threshold, and the frequency band range, signal power, noise power, and local SNR near the two center frequencies. By calculating and comparing the local SNR with the threshold, the low-frequency center frequency of 50 kHz was excluded, while the high-frequency center frequency of 70 kHz was retained, achieving effective optimization of the center frequency. A comparison of this invention with existing technologies is shown in Table 3 below:

[0118] Table 3 Comparison of the present invention with the prior art

[0119] Comparison Projects Existing technology Technical solution of the present invention Signal processing methods Single-ended traveling wave method or bridge method is difficult to handle signal attenuation problems in complex terrain and multi-branch cable networks. Two-end traveling wave data processing combined with adaptive multi-scale variational mode decomposition effectively reduces signal attenuation and noise interference. Partial discharge feature extraction capability Insufficient ability to identify small partial discharge phenomena Introducing a convolutional neural network (CNN) to extract features from double-ended traveling wave data significantly improves the accuracy of identifying small partial discharges. Frequency optimization mechanism Lack of signal-to-noise ratio optimization methods targeting the center frequency Based on the signal-to-noise ratio (SNR) distribution, a threshold is set, the center frequency is dynamically optimized, and low SNR frequency bands are excluded. Topology modeling It is difficult to accurately distinguish between the fault points of the main cable and the branch cable. Cable network topology is constructed using a recurrent neural network model and a two-end data fault training model to improve positioning accuracy. Environmental adaptability Unable to adapt to signal quality fluctuations caused by changes in environmental conditions such as temperature and humidity Design an adaptive signal adjustment mechanism to dynamically correct the signal transmission process by combining historical data and predictive models. Database update capability Database updates are lagging, making it difficult to reflect the status of the cable network in real time. Real-time capture of loss functions from the database and incorporation of two-ended traveling wave data features ensure that the model adapts to dynamic changes in the network.

[0120] Table 3 clearly demonstrates the optimizations and improvements of the technical solution of this invention in terms of signal processing, feature extraction, frequency optimization, topology modeling, environmental adaptability, and database updates.

[0121] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A method for intelligent location and diagnosis of cable faults based on double-ended traveling waves, characterized in that: include, A monitoring system is installed synchronously at both ends of the cable line. The monitoring system includes a signal transmitting unit and a signal receiving unit. An adaptive signal adjustment mechanism is designed to adjust the environmental condition parameters of the signal transmitting unit and the signal receiving unit to synchronize the signal transmission and signal reception time. Based on the processing of two-end traveling wave data of transmitted and received signals, a convolutional neural network is used to extract the features of the two-end traveling wave data during the processing. The topology model of the cable network is constructed using the features of the two-end traveling wave data. The topology model includes a recurrent neural network model and a two-end data fault training model. By using the topology model of the cable network, a database is established in combination with the length and propagation time of the original cable. The location of the fault point is calculated by using the time difference between the transmitted and received signals tracked by the database and the recurrent neural network model. A dual-end data fault training model is used to train the fault point to complete the diagnosis of intelligent location of cable faults. Processing the two-end data of the transmitted and received signals includes introducing variational mode decomposition to decompose the received signal during time synchronization and extracting the partial discharge characteristics of different frequency components of the synchronized signal. The decomposition of the signal received during time synchronization includes the proposal of an adaptive multi-scale variational mode decomposition method based on the variational mode decomposition method. This method is used to accurately decompose the time-synchronized received signal. The adaptive multi-scale variational mode decomposition method includes dynamically adjusting the decomposition scale and center frequency range of the synchronized signal. The decomposition scale involves performing spectral analysis on the synchronized signal using Fast Fourier Transform (FFT) to extract features from the spectral distribution. These features include the dominant bandwidth, energy distribution ratio, high-frequency noise, and the proportion of low-frequency drift. Based on the extracted features, a dynamic K-value calculation formula based on energy segmentation is proposed. ； in, Indicates the decomposition scale. This indicates the total number of sub-bands in the spectrum allocation. Indicates the first The energy percentage of each sub-band This indicates the preset energy threshold. Indicates an indicator function; The center frequency range includes the frequency peaks from which features are extracted using a peak detection algorithm. The calculation formula for the peak detection algorithm is as follows: ； in, Indicates frequency At, spectrum The first derivative, Indicates frequency At, spectrum The second derivative, Indicates frequency, This represents a user-defined scaling factor. Indicates the threshold parameter. Represents the mean of the spectrum. The standard deviation of the spectrum This represents the set of peak points.

2. The intelligent cable fault location and diagnosis method based on double-ended traveling waves as described in claim 1, characterized in that: The monitoring system installed at both ends of the cable line includes using time synchronization technology of satellite navigation system to obtain the parameter configuration of signal transmitting unit and signal receiving unit, combined with atomic clock calibration module to dynamically correct electromagnetic interference problems of time synchronization technology caused by temperature changes. The atomic clock calibration module includes obtaining the parameter configuration of transmitting unit and signal receiving unit for nanosecond-level time synchronization application scenarios by utilizing the quantum properties of atoms under electromagnetic radiation.

3. The intelligent cable fault location and diagnosis method based on double-ended traveling waves as described in claim 2, characterized in that: The adaptive signal adjustment mechanism includes establishing a prediction model by acquiring parameter configurations and combining them with historical data. The historical data includes a type of environmental condition parameters collected during previous signal transmission processes. The signal transmission process is adjusted based on the prediction model. The establishment of the prediction model includes fusing multi-source heterogeneous data of a class of environmental condition parameters, such as temperature, humidity, and electromagnetic interference intensity, with signal quality indicators such as signal-to-noise ratio, bit error rate, and cable operating status. The cable operating status includes the cable's load current and voltage fluctuations. The multi-source heterogeneous data fusion includes using variational autoencoders to enhance the features of a class of environmental condition parameters, extracting deep-level patterns hidden in the load current and voltage fluctuation data of cables. These deep-level patterns include the nonlinear variation patterns of signal-to-noise ratio and bit error rate in the time series. The time series includes aligning time series of different lengths using dynamic time warping based on the time series features in historical data, constructing a hybrid model of long short-term memory network and attention mechanism, designing an intelligent agent based on reinforcement learning, and dynamically adjusting the frequency, power, and modulation mode of the signal transmission unit according to the obtained parameter configuration and signal quality indicators of the signal transmission and reception units to synchronize the time of signal transmission and reception.

4. The intelligent cable fault location and diagnosis method based on double-ended traveling waves as described in claim 3, characterized in that: The center frequency range also includes introducing signal-to-noise ratio as a constraint to optimize the center frequency. The optimization of the center frequency includes performing local signal-to-noise ratio calculations on the frequency bands near each center frequency. When the signal-to-noise ratio is lower than the set threshold When a low-frequency band is set, that band consists of noise, and the center frequency corresponding to the noise is excluded. When the signal-to-noise ratio is higher than or equal to the set threshold When a high-frequency band is set, that band contains effective signal components, and the corresponding noise center frequency is retained; The threshold This includes setting thresholds based on the statistical characteristics of the signal-to-noise ratio distribution. ; The formula for calculating the nearby frequency band range is: ； in, Indicates the bandwidth. Indicates the center frequency; The formula for calculating the local signal-to-noise ratio of the frequency bands near each center frequency is as follows: ； in, Indicates signal power. Indicates noise power. The local signal-to-noise ratio represents the center frequency.

5. The intelligent cable fault location and diagnosis method based on double-ended traveling waves as described in claim 4, characterized in that: The extraction of double-ended traveling wave data features during the processing using a convolutional neural network includes extracting the peak point location of the double-ended traveling wave data features by capturing the temporal distribution characteristics of the signal through convolutional layers. The convolutional neural network includes a one-dimensional convolutional neural network and a two-dimensional convolutional neural network. The transmitted and received signals are converted into CNN input form, which is represented as a one-dimensional time series or a two-dimensional time-frequency spectrum. If it is time series data, it is input into a one-dimensional convolutional neural network; If it is a time-frequency spectrum, it is input into a two-dimensional convolutional neural network; When the time distribution characteristics are captured in a one-dimensional convolutional neural network, the one-dimensional convolutional neural network signal is scanned to extract the double-ended traveling wave data features. When the temporal distribution characteristics are captured in a two-dimensional convolutional neural network, the double-ended traveling wave data features are extracted by sliding the convolution kernel on the time axis and frequency axis. The extracted features of the double-ended traveling wave data are represented in vector form, as shown in the formula: ； in, Represents the feature vector of a two-ended traveling wave data. Representing feature dimension, This represents the feature dimension of the first double-ended traveling wave data feature vector. This represents the feature dimension of the second double-ended traveling wave data feature vector. Indicates the first The feature dimension of a double-ended traveling wave data feature vector; The construction of the cable network topology model includes combining the characteristics of the double-ended traveling wave data with historical data, and training the model through double-ended data faults to construct the topology. The formula for combining the characteristics of the double-ended traveling wave data with the historical data is as follows: ； in, Indicates the weighting coefficient. Represents the feature vector of a two-ended traveling wave data. Represents historical data. This represents the combined feature vector; The topology training using a dual-end data fault training model includes combining the feature vectors of nodes in the cable network. This means that the nodes in the cable network are adjusted using the output of the dual-end data fault training model; When a fault is detected at a node in the cable network, the weighting factor is reduced.

6. The intelligent cable fault location and diagnosis method based on double-ended traveling waves as described in claim 5, characterized in that: The establishment of the database includes creating a cable network topology database based on the N node locations of the cable, and training the cable network topology database using a two-end data fault training model. The training of the cable network topology database includes capturing the loss function in the database and incorporating the two-end traveling wave data features into the loss function in the database in real time, so that the model can adapt to changes in the cable network and complete the diagnosis of intelligent location of cable faults.

7. A cable fault intelligent location and diagnosis system based on double-ended traveling waves, based on the cable fault intelligent location and diagnosis method based on double-ended traveling waves according to any one of claims 1 to 6, characterized in that: include, The adjustment mechanism module has a monitoring system installed simultaneously at both ends of the cable line, and an adaptive signal adjustment mechanism is designed. The processing module processes the two-end data of the transmitted and received signals based on the two-end traveling wave, uses a convolutional neural network to extract the features of the two-end traveling wave data during the processing, and uses the features of the two-end traveling wave data to construct a topology model of the cable network. The diagnostic module utilizes the topology model of the cable network and employs a two-end data fault training model to train the fault location, thereby completing the intelligent location diagnosis of cable faults.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, it implements the steps of the intelligent cable fault location and diagnosis method based on double-ended traveling waves as described in any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, it implements the steps of the intelligent location and diagnosis method for cable faults based on double-ended traveling waves as described in any one of claims 1 to 6.

Images