Power cable fault early warning method and device and electronic equipment

Abstract

The invention discloses a power cable fault early warning method and device and electronic equipment. Relates to the field of intelligent power grids, and the method comprises the steps: collecting the original operation data of a power cable according to a preset sampling frequency, the original operation data at least comprising voltage waveform data, current waveform data and partial discharge data; preprocessing the original operation data to obtain preprocessed data; extracting target features in the preprocessed data, wherein the target features at least comprise time domain features, frequency domain features and joint distribution features of the original operation data; comparing the target feature with a standard feature stored in a preset feature library, and detecting whether abnormal disturbance exists in the original operation data; and when it is detected that abnormal disturbance exists in the original operation data, based on the target features, a power cable fault early warning model is adopted to obtain a fault early warning result of the power cable. The technical problem that the early failure of the power cable is difficult to warn timely and accurately in the prior art is solved.

Description

Technical Field

[0001] The present invention relates to the field of smart grids, and in particular to a method, device and electronic equipment for early warning of power cable faults. Background Art

[0002] Power cables are widely used in modern power systems and are extremely important for ensuring the reliability and continuity of power systems. However, cable failures can lead to many serious consequences, such as power outages, equipment damage, economic losses, and security threats. Cable failures include many types, among which early failures are highly concealed, manifested as subtle signs such as slight insulation degradation, partial discharge, or slight changes in electrical parameters. If not discovered and handled in time, they can easily develop into serious permanent failures, increasing the difficulty and cost of repair.

[0003] In the early stage of faults, electrical parameters change slightly and are easily affected by system operating conditions, such as load and power supply fluctuations. It is difficult to effectively establish a prediction model for complex fault types such as intermittent partial discharge faults, and the fault warning sensitivity and reliability are low. Although partial discharge monitoring is a key technology for detecting insulation defects, the partial discharge signal is extremely weak and is easily attacked by electromagnetic interference and noise in the actual environment, which greatly reduces the measurement accuracy. It is also difficult to accurately define the intrinsic connection between different types of partial discharges and the development process of early faults, and it is impossible to accurately estimate the warning time. In addition, most related technologies rely on a single feature for analysis and judgment, while early cable faults are actually the result of the interweaving and joint action of multiple factors. Single feature analysis is difficult to fully grasp the overall picture and development trend of the fault, and it is easy to misjudge or miss the judgment.

[0004] With regard to the problem that it is difficult to provide timely and accurate warnings for early power cable failures in the above-mentioned related technologies, no effective solution has been proposed so far. Summary of the invention

[0005] The embodiments of the present invention provide a power cable fault early warning method, device and electronic equipment to at least solve the technical problem in the related art that it is difficult to timely and accurately warn of early power cable faults.

[0006] According to one aspect of the embodiments of the present invention, there is provided a method for early warning of power cable faults, including: collecting the original operation data of the power cable according to a predetermined sampling frequency, where the original operation data at least includes: voltage waveform data, current waveform data, and partial discharge data; preprocessing the original operation data to obtain preprocessed data, where the preprocessing at least includes filtering processing and noise reduction processing; extracting target features from the preprocessed data, where the target features at least include the time-domain features and frequency-domain features of the original operation data, and the joint distribution features of the original operation data in the time domain and frequency domain; comparing the target features with the standard features stored in a preset feature library to detect whether there are abnormal disturbances in the original operation data; in the case where abnormal disturbances are detected in the original operation data, based on the target features, using a power cable fault early warning model to obtain the fault early warning result of the power cable.

[0007] According to another aspect of the embodiments of the present invention, there is also provided a power cable fault early warning device, including: a data collection module for collecting the original operation data of the power cable according to a predetermined sampling frequency, where the original operation data at least includes: voltage waveform data, current waveform data, and partial discharge data; a data preprocessing module for preprocessing the original operation data to obtain preprocessed data, where the preprocessing at least includes filtering processing and noise reduction processing; a feature extraction module for extracting target features from the preprocessed data, where the target features at least include the time-domain features and frequency-domain features of the original operation data, and the joint distribution features of the original operation data in the time domain and frequency domain; a feature detection module for comparing the target features with the standard features stored in a preset feature library to detect whether there are abnormal disturbances in the original operation data; a fault early warning module for, in the case where abnormal disturbances are detected in the original operation data, based on the target features, using a power cable fault early warning model to obtain the fault early warning result of the power cable.

[0008] According to another aspect of the embodiments of the present invention, there is also provided a non-volatile storage medium storing multiple instructions suitable for being loaded and executed by a processor to perform any one of the power cable fault early warning methods.

[0009] According to another aspect of the embodiments of the present invention, there is also provided an electronic device including one or more processors and a memory, where the memory is used to store one or more programs, and when the one or more programs are executed by the one or more processors, the one or more processors are caused to implement any one of the power cable fault early warning methods.

[0010] In an embodiment of the present invention, by collecting the original operation data of a power cable according to a predetermined sampling frequency, where the original operation data at least includes: voltage waveform data, current waveform data, and partial discharge data; preprocessing the original operation data to obtain preprocessed data, where the preprocessing at least includes filtering and noise reduction; extracting target features from the preprocessed data, where the target features at least include the time-domain features, frequency-domain features of the original operation data, and the joint distribution features of the original operation data in the time domain and frequency domain; comparing the target features with the standard features stored in a preset feature library to detect whether there are abnormal disturbances in the original operation data; in the case where abnormal disturbances are detected in the original operation data, based on the target features, using a power cable fault warning model to obtain the fault warning result of the power cable, achieving the purpose of accurately identifying power cable faults by adopting a multi-source data collection, preprocessing, and multi-dimensional feature extraction method that integrates disturbance features, combining a machine learning model with a preset feature library comparison, thereby realizing the technical effect of improving the safety and reliability of the power system and reducing the operation and maintenance costs, and further solving the technical problem that it is difficult to timely and accurately warn of early faults of power cables in the related art. Description of the Drawings

[0011] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The schematic embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation to the present invention. In the drawings:

[0012] Figure 1 is a flowchart of a power cable fault warning method according to an embodiment of the present invention;

[0013] Figure 2 is a schematic diagram of an optional power cable fault warning process framework according to an embodiment of the present invention;

[0014] Figure 3 is a schematic diagram of an optional power cable fault warning system framework according to an embodiment of the present invention;

[0015] Figure 4 is a schematic diagram of a power cable fault warning device according to an embodiment of the present invention. Detailed Embodiments

[0016] To enable those skilled in the art to better understand the solution of the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

[0017] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily need to be used to describe a specific order or sequence. It should be understood that such data can be interchanged under appropriate circumstances so that the embodiments of the present invention described here can be implemented in an order different from those illustrated or described here. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product or device including a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0018] According to an embodiment of the present invention, an embodiment of a method for early warning of power cable faults is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer-executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that here.

[0019] Figure 1 is a flowchart of a method for early warning of power cable faults according to an embodiment of the present invention. As Figure 1 shown, the method includes the following steps:

[0020] Step S102, collect the original operation data of the power cable according to a predetermined sampling frequency, where the original operation data at least includes: voltage waveform data, current waveform data, and partial discharge data.

[0021] Optionally, collecting the voltage waveform data of the power cable means collecting the voltage waveforms under normal operation and fault conditions of the power cable, which may include, but are not limited to, the peak value of the voltage, the shape of the waveform, the voltage fluctuation conditions, etc. These data can reflect the stability of the cable insulation state. Collecting the current waveform data of the power cable means collecting the current waveform data during cable operation, which may include, but are not limited to, the peak value of the current, the shape of the waveform, the current change trend, etc. These data can reveal the abnormality of the internal current distribution of the cable, especially the current change caused by partial discharge. Partial discharge is a tiny discharge phenomenon caused by uneven electric field in the insulation material of power equipment. The partial discharge data of the collected power cable may include, but are not limited to, parameters such as the intensity, frequency, and phase of the detected discharge pulse, as well as the change trend of the discharge pulse over time.

[0022] It should be noted that high-precision voltage and current waveform data can provide real-time monitoring of the operating state of the power cable, facilitating the capture of any subtle abnormal fluctuations, which can be used as an omen of early faults. The collection of partial discharge data can provide early warning of the deterioration of the insulation material, because partial discharge is often a precursor to damage to the cable insulation layer. Timely detection and analysis of partial discharge data can effectively prevent permanent cable faults. By setting a reasonable sampling frequency and collecting the above three types of original operating data, it is possible to ensure the acquisition of information that comprehensively reflects the operating state of the power cable, providing a solid data foundation for subsequent preprocessing, feature extraction, and fault warning.

[0023] In an optional embodiment, the original operating data of the power cable is collected according to a predetermined sampling frequency, including: collecting the voltage waveform data and current waveform data of the power cable according to a first predetermined sampling frequency, and collecting the partial discharge data of the power cable according to a second predetermined sampling frequency, to obtain the original operating data, where the first predetermined sampling frequency is less than the second predetermined sampling frequency.

[0024] Optionally, according to the characteristics of the power cable operation signals, the sampling frequency corresponding to the voltage waveform data and the current waveform data (i.e., the first sampling frequency) can be set to 50 kilohertz (kHz). This frequency can meet the requirement of capturing early fault disturbance signals, ensure that the details of waveform changes are recorded, and provide a basis for subsequent signal analysis. The second predetermined sampling frequency is applied to the partial discharge data. Considering the characteristics of partial discharge signals being weak and changing rapidly, the corresponding sampling frequency (i.e., the second sampling frequency) can be set to a higher 40 megahertz (MHz). Such a high sampling frequency can accurately capture every detail of the partial discharge signal, which is crucial for analyzing partial discharge characteristics and judging its type and severity. The voltage and current waveform data can reflect the macroscopic electrical characteristics of the power cable under normal operating conditions, and their changes are relatively slow. Therefore, a lower sampling frequency (50 kHz) is sufficient to record the characteristics of these signals. The partial discharge data involves more refined changes in electrical parameters, especially the instantaneousness and weakness of the signals, and requires a higher sampling frequency (40 MHz) to ensure the integrity and accuracy of the data. The difference in sampling frequencies can ensure the accurate acquisition of different types of data, capture both the changes in macroscopic electrical characteristics and detailed records of tiny disturbance signals such as partial discharges, and provide comprehensive data support for subsequent fault warning.

[0025] Step S104: Preprocess the original operation data to obtain preprocessed data, where the preprocessing includes at least filtering and noise reduction.

[0026] Optionally, the purpose of filtering is to remove noise and irrelevant interference in the signal and retain the signal components related to faults. In the operation data of power cables, the voltage and current waveforms may be affected by external factors such as power supply fluctuations, load changes, and electromagnetic interference, and background noise may also be mixed into the partial discharge data. Therefore, filtering is necessary to ensure the purity of the data and improve the accuracy of subsequent analysis. Through filtering and wavelet noise reduction preprocessing, the noise and interference in the original operation data are effectively removed, and the useful components of the signal are highlighted, providing a high-quality data basis for subsequent feature extraction and fault warning analysis. The preprocessed data is more pure, which helps to accurately extract features and efficiently train the model, thereby improving the performance and reliability of the entire warning system.

[0027] In an optional embodiment, preprocessing the original operation data to obtain preprocessed data includes: filtering the original operation data using the infinite impulse response filter algorithm to obtain filtered data; and performing noise reduction processing on the filtered data using the wavelet transform algorithm to obtain preprocessed data.

[0028] Optionally, an Infinite Impulse Response Filter (IIR filter) is a digital signal processing technique used to remove noise or interference at specific frequencies from a signal. In the power cable fault warning method, the IIR filter is used for the preliminary filtering of the original operation data to remove high-frequency noise and other irrelevant interference signals. This is because when a power cable is operating, the voltage and current waveforms may be mixed with noise caused by electromagnetic interference, grid fluctuations, etc. These noises will mask or distort the signal characteristics related to faults, affecting the accuracy of subsequent analysis. Wavelet transform is an effective signal analysis tool, especially suitable for the processing of non-stationary signals. During the preprocessing process, wavelet transform is used for further noise reduction to improve the signal quality. Through IIR filtering and wavelet noise reduction, these interferences can be effectively removed, highlighting the signal characteristics related to faults, and providing cleaner and more reliable data for subsequent feature extraction and fault warning model training. By removing noise and interference, the preprocessed data can more accurately reflect the actual operating state of the cable, improve the accuracy of feature extraction and the sensitivity of fault warning, which is of great significance for the early detection and prevention of cable faults.

[0029] In an optional embodiment, the original operation data is preprocessed to obtain preprocessed data, including: using a wavelet transform algorithm to perform noise reduction processing on the filtered data to obtain noise-reduced data; and performing enhancement processing on the data features associated with a predetermined fault type in the noise-reduced data to obtain preprocessed data.

[0030] Optionally, wavelet transform is a powerful signal analysis tool that can decompose a signal into components of different frequencies and time scales, thereby accurately locating and removing noise in the time-frequency domain. In the power cable fault warning method, wavelet transform is used to process the data preliminarily filtered by the IIR filter to further reduce the noise interference in the signal. Based on the noise-reduced data, through feature enhancement processing, the key features related to the predetermined fault type can be further highlighted, improving the accuracy of subsequent fault detection and warning. The data after wavelet transform noise reduction and feature enhancement processing is the preprocessed data. These data not only remove noise interference, but also strengthen the feature signals related to the fault type, providing better-quality data input for subsequent feature extraction and fault warning models.

[0031] In the above method, performing noise reduction through wavelet transform and enhancing the features associated with the predetermined fault type are key steps in the power cable fault warning method to improve data quality and prediction accuracy. The preprocessed data is cleaner and richer in features, which can significantly improve the performance of the fault warning model, detect and warn of cable faults early, and ensure the safe operation of power cables.

[0032] Optionally, the predetermined fault types may include, but are not limited to, deterioration of cable insulation materials, partial discharge faults, weak electrical parameter changes, and intermittent partial discharges. Among them, the deterioration of cable insulation materials is a common cause of cable faults, which may initially show minor changes such as a decrease in insulation resistance and partial discharges. By enhancing the features related to insulation deterioration, such as the pulse frequency and energy distribution of partial discharges, the change in the insulation state can be detected at an early stage. A partial discharge fault is a phenomenon that occurs in a local area of the cable insulation due to the electric field strength exceeding the breakdown strength of the material. The features related to partial discharges include the morphology, frequency, and energy distribution of discharge pulses. The enhancement of these features helps to identify the early signs of partial discharges and issue timely warnings. Weak electrical parameter changes refer to the fact that during the operation of the cable, minor changes in electrical parameters may be an omen of early faults, such as minor distortions in voltage and current waveforms and spectral changes in signals. Through feature enhancement, these changes can be highlighted, improving the detection ability for early faults. Intermittent partial discharges mean that partial discharges sometimes occur intermittently, which poses challenges to fault detection. Feature enhancement can focus on capturing the weak signals of intermittent partial discharges, analyzing their repeatability and patterns, and providing data support for establishing an effective prediction model. The abnormal operation state of the cable, that is, in addition to specific fault types, abnormal changes in the operation state of the cable may also be an omen of faults. By analyzing the statistical features (such as mean, variance, peak value, etc.) of voltage and current waveforms, it can be determined whether the cable is operating stably and whether there are potential fault risks.

[0033] In an optional embodiment, a wavelet transform algorithm is used to perform noise reduction processing on the filtered data to obtain noise-reduced data, including: determining the wavelet decomposition level corresponding to the filtered data; based on the wavelet decomposition level, performing multi-resolution wavelet decomposition on the filtered data to obtain multiple wavelet coefficients, where the multiple wavelet coefficients include high-frequency wavelet coefficients representing data details and low-frequency wavelet coefficients representing data profiles, the high-frequency wavelet coefficients correspond to frequencies greater than a preset frequency, and the low-frequency wavelet coefficients correspond to frequencies less than or equal to the preset frequency; performing denoising processing on the high-frequency wavelet coefficients based on a preset threshold to obtain processed high-frequency wavelet coefficients; and performing wavelet reconstruction on the processed high-frequency wavelet coefficients and low-frequency wavelet coefficients to obtain noise-reduced data.

[0034] Optionally, the selection of the number of wavelet decomposition levels is based on the spectral characteristics of the signal and the noise frequency to be removed. In power cable fault warning, a sufficient number of decomposition levels are selected to separate high-frequency noise and low-frequency signal features, but at the same time, over-decomposition should be avoided to prevent affecting the signal structure and quality. For example, for the partial discharge monitoring signal of a power cable, due to its weakness and complexity, a relatively high number of decomposition levels may be required to capture detailed features. Wavelet decomposition is based on the principle of multi-resolution analysis, decomposing the signal into wavelet coefficients of different frequency components, including high-frequency wavelet coefficients and low-frequency wavelet coefficients. Among them, the high-frequency wavelet coefficients correspond to higher decomposition levels and represent the details of the signal, such as mutations and spikes. These wavelet coefficients are usually associated with high-frequency disturbances such as noise and partial discharges. The low-frequency wavelet coefficients correspond to lower decomposition levels and reflect the general profile of the signal, such as the baseline and trend. These coefficients retain the basic features of the signal and are very important for monitoring the long-term health of the cable. Threshold denoising sets a threshold and regards those high-frequency wavelet coefficients smaller than the threshold as noise, and sets them to zero or performs soft threshold processing. The selection of the threshold needs to comprehensively consider the signal-to-noise ratio of the signal, the statistical distribution of wavelet coefficients, and the characteristics of the noise. For example, in power cable signal processing, a signal-to-noise ratio-based threshold method can be used for denoising based on the weakness of the partial discharge signal and the statistical characteristics of the noise. Wavelet reconstruction is the process of recombining the denoised high-frequency wavelet coefficients and the unprocessed low-frequency wavelet coefficients to restore the signal. Through wavelet reconstruction, denoised data can be generated, that is, the noise-reduced data. The noise-reduced data not only retains the basic features of the signal but also removes most of the noise interference, improves the signal quality, and makes subsequent feature extraction and fault warning analysis more accurate. Through precise wavelet decomposition and threshold denoising, the quality of the noise-reduced data can be ensured, providing more reliable data support for the intelligent operation and maintenance of power cables.

[0035] Optionally, the wavelet transform algorithm can be used, but not limited to, to denoise the acquired original signal. Select the Daubechies wavelet basis, and determine the number of decomposition levels according to the characteristics of the signal and the noise level. After performing multi-level wavelet decomposition on the signal, remove the noise coefficients through a threshold processing method (such as the soft threshold method), retain the effective signal coefficients, and then perform wavelet reconstruction to obtain the denoised waveform data. The wavelet transform formula is as follows:

[0036]

[0037] Where Wf(a,b) is the wavelet transform coefficient, a is the scale factor, b is the translation factor, f(t) is the original signal, and ψ(t) is the wavelet basis function.

[0038] Further, perform normalization processing on the denoised waveform data to make the data have a unified scale within a specific range, facilitating subsequent analysis and comparison. The normalization formula is:

[0039]

[0040] where x norm is the normalized signal, x is the original signal, x min and x max are the minimum and maximum values of the original signal respectively.

[0041] Step S106, extract the target features from the preprocessed data, where the target features at least include the time-domain features, frequency-domain features of the original operation data, and the joint distribution features of the original operation data in the time domain and frequency domain;

[0042] Optionally, feature extraction is the core link in the power cable fault warning method, aiming to extract key information related to faults from the preprocessed data and provide input for the subsequent fault warning model. Feature extraction mainly involves the analysis of time-domain features, frequency-domain features, and time-frequency domain features. Among them, time-domain features can directly reflect the changes of signals on the time axis. For power cable fault warning, time-domain features can capture the instantaneous characteristics of fault disturbances. Through frequency-domain features, the frequency changes of signals can be captured, revealing the non-linear changes or structural damage information inside the cable. The joint distribution features (i.e., time-frequency domain features) combine the analysis of the time domain and frequency domain. Through time-frequency analysis methods such as wavelet transform and short-time Fourier transform (STFT), the joint distribution features of signals in time and frequency, i.e., time-frequency diagrams, are obtained. Time-frequency domain features can capture the transient signal features generated by faults such as partial discharge. By setting the extraction of target features to comprehensively analyze the time domain, frequency domain, and time-frequency domain, these features complement each other and can comprehensively reflect the operating state of the cable from multiple dimensions. Inputting these features into the fault warning model, the model can learn the relationship between these features and cable faults, and can identify whether there are early fault disturbances in the cable, thus realizing high-precision and high-sensitivity fault prediction and warning.

[0043] In an optional embodiment, extracting the target features from the preprocessed data includes: extracting the time-domain features from the preprocessed data, where the time-domain features at least include the mean, variance, and peak value of the preprocessed data; extracting the transient features from the preprocessed data, where the transient features include the rising moment, falling moment, and pulse width in the waveform of the preprocessed data; performing a fast Fourier transform on the preprocessed data to obtain the frequency-domain features; using the wavelet transform time-frequency analysis method to obtain the joint distribution features of the original operation data in the time domain and frequency; where the target features include time-domain features, transient features, frequency-domain features, and joint distribution features.

[0044] Optionally, the time-domain features directly reflect the changes of the signal on the time axis and are the basis for analyzing the operating state of the cable. The time-domain features mentioned in the claims include the mean, variance, and peak value of the data. These statistical characteristics can provide information about the overall level and fluctuation of the signal. For example, the mean can reflect the average level of the signal and can be used to detect changes in the signal baseline; the variance is used to measure the magnitude of the signal fluctuation and helps to identify the stability of the signal. An abnormally high variance may indicate the presence of a fault. The peak value is the maximum value of the signal, and the change in the peak value can indicate whether there are abnormal spike phenomena in the signal, which is a common symptom of early faults. The transient features focus on the rapid changes in the signal, such as the rising time, falling time, and pulse width of the waveform. These features are closely related to small and instantaneous disturbances such as partial discharge. The extraction of transient features helps to capture the transient behavior of early faults and analyze the characteristics and development process of the faults. Through the Fast Fourier Transform (FFT), the preprocessed time-domain data can be transformed into the frequency domain to analyze the frequency composition of the signal. The frequency-domain features include the main frequency component, harmonic content, etc. The changes in these features may be related to faults such as insulation aging and poor contact of the cable. The extraction of frequency-domain features can reveal the distribution of the signal at different frequencies and provide important clues for fault type identification. Wavelet transform can not only analyze the frequency components of the signal but also consider the time dimension simultaneously, providing the time-frequency distribution characteristics of the signal. Compared with the Fast Fourier Transform, wavelet transform is more effective in analyzing non-stationary and instantaneous events such as partial discharge and short circuit. Through wavelet transform, the transient features related to faults in the signal can be highlighted, and the change trends in the time domain and frequency domain can be analyzed to help identify weak signals of early faults. The set target features include not only time-domain features, transient features, and frequency-domain features but also specifically mention joint distribution features, that is, the time-frequency features obtained through wavelet transform. Comprehensive analysis of these multi-source features can more comprehensively evaluate the operating state of the cable, overcome the limitations of single-feature analysis, and improve the accuracy and reliability of fault warning.

[0045] Optionally, extract the target features from the preprocessed data, specifically including:

[0046] Step S1061, calculate the basic statistical features such as the mean, variance, and peak-to-peak value of the waveform data for the preprocessed data. These features can reflect the overall amplitude and fluctuation of the signal. The mean calculation formula is:

[0047]

[0048] where μ is the mean, N is the number of data points, and x(i) is the i-th data point. The variance calculation formula is:

[0049]

[0050] The peak-to-peak value is the difference between the maximum and minimum values of the signal.

[0051] Step S1062: Extract features such as the rise time, fall time, and pulse width of the waveform. These features are closely related to the occurrence and development process of faults. The rise time is defined as the time required for the signal to rise from a certain amplitude, the fall time is the time required for the signal to fall from a certain amplitude, and the pulse width is the duration during which the signal amplitude is greater than a certain value. By monitoring and analyzing these features, it is possible to preliminarily determine whether there are early fault disturbances in the cable.

[0052] Step S1063: Perform a fast Fourier transform (FFT) on the waveform data to convert the time-domain signal into a frequency-domain signal and obtain a spectrogram. Analyze features such as the frequency components, harmonic content, and frequency distribution in the spectrogram. Changes in the main frequency components may be related to faults such as cable insulation aging and partial discharge, and an increase in the harmonic content may indicate the presence of non-linear faults in the cable. The FFT calculation formula is:

[0053]

[0054] where X(k) is the frequency-domain signal, k is the frequency index, x(n) is the time-domain signal, and N is the number of data points.

[0055] Step S1064: Calculate the power spectral density (PSD) to further analyze the energy distribution of the signal at different frequencies. The PSD can be obtained through the Fourier transform of the autocorrelation function, and the formula is:

[0056]

[0057] where P xx (f) is the power spectral density, R xx (m) is the autocorrelation function, and f is the frequency.

[0058] Step S1065: Time-frequency domain feature extraction: Adopt time-frequency analysis methods such as wavelet transform to obtain the joint distribution features of the signal in time and frequency, that is, the time-frequency diagram. The time-frequency diagram can intuitively display the change of the signal frequency over time, which helps to capture the transient signal features generated by faults such as partial discharge. The STFT calculation formula is:

[0059]

[0060] where STFT(t,f) is the result of the short-time Fourier transform, and w(τ - t) is the window function.

[0061] Step S1066: Extract the energy distribution features from the time-frequency diagram, calculate the energy integrals within different time periods and frequency bands, analyze the variation trend of the energy concentration region, and the variation law of the energy of different frequency components over time. These features can provide important bases for the identification of fault types and the judgment of the degree of fault development.

[0062] Step S1067: Deeply study the various waveform disturbance features that may occur before the permanent fault of the cable equipment, and establish a waveform disturbance feature library. Compare and match the extracted time-domain, frequency-domain, and time-frequency-domain features with the standard features in the feature library to analyze whether there are abnormal disturbances in the current operating state of the cable. By comparing the features of the partial discharge signals, judge whether there is a specific type of partial discharge phenomenon, as well as its severity and development trend.

[0063] Step S108: Compare the target features with the standard features stored in the preset feature library to detect whether there are abnormal disturbances in the original operating data;

[0064] Optionally, the preset feature library is established based on a large amount of historical cable operating data and known fault cases, and stores the set of characteristic parameters of various types of power cables in the normal operating state, early fault state, and fault state. These characteristic parameters include, but are not limited to, time-domain features (such as the mean, variance, peak-to-peak value, rise time, fall time, pulse width, etc. of the waveform), frequency-domain features (such as the main frequency component, harmonic content, frequency distribution, etc.), and time-frequency-domain features (such as the variation law of the energy of different frequency components over time). Comparing the target features extracted from the preprocessed data with the standard features in the feature library can quickly identify possible early fault signs and achieve the timely warning of power cable faults.

[0065] Step S110: In the case of detecting abnormal disturbances in the original operating data, based on the target features, use the power cable fault warning model to obtain the fault warning result of the power cable.

[0066] Optionally, in the data after feature extraction and preprocessing, by comparing the extracted target features with the standard features in the preset feature library, it can be identified whether there are abnormal disturbances related to early faults. If the value of a certain feature or certain features exceeds the normal operating range, or has significant similarity with the early fault features in the feature library, this indicates that abnormal disturbances have occurred in the cable operation data. At this time, the system needs to further analyze whether these abnormal disturbances indicate the risk of early faults in the cable. Once abnormal disturbances are detected, the next step is to use the power cable fault warning model to deeply analyze these abnormalities and predict the fault development trend of the cable. The warning model can be an algorithm based on machine learning, such as a combined model of support vector machine (SVM) and long short-term memory network (LSTM). These models are trained to identify and predict the possible fault modes of power cables under different operating conditions.

[0067] In an optional embodiment, before obtaining the fault warning result of the power cable by using the power cable fault warning model based on the target features in the case where abnormal disturbances are detected in the original operation data, the method further includes: obtaining multiple groups of historical operation data of the power cable and the state labels corresponding to the multiple groups of historical operation data respectively, where each group of historical operation data includes voltage waveform data, current waveform data, and partial discharge data of the power cable in the corresponding historical period, and the state labels are normal operating state, early fault state, and fault state; preprocessing the multiple groups of historical operation data to obtain multiple groups of preprocessed data, where the preprocessing at least includes filtering and noise reduction; extracting features from the multiple groups of preprocessed data to obtain multiple groups of features, where each group of features includes time-domain features, frequency-domain features of the corresponding historical operation data, and joint distribution features of the corresponding historical operation data in the time domain and frequency domain; training the initial model with the multiple groups of features as the input and the state labels corresponding to the multiple groups of historical operation data respectively as the output to obtain the power cable fault warning model.

[0068] Optionally, obtain multiple sets of historical operation data of the power cable, which cover the operation conditions of the cable under different states, including normal operation state, early fault state, and fault state. The historical data can include, but is not limited to, voltage waveform data, current waveform data, and partial discharge data, which are all important parameters reflecting the operation state of the cable. Preprocess the historical operation data, and this process includes filtering and noise reduction. Filtering is used to remove interference signals, and noise reduction further reduces the noise in the signal through techniques such as wavelet transform to ensure the purity of the data. The preprocessed data, that is, the preprocessing data, is used as the input for subsequent feature extraction and model training. Extract features from multiple sets of preprocessing data, and the extracted features include time-domain features, frequency-domain features, and time-frequency domain joint distribution features. Time-domain features involve the statistical characteristics of the signal, frequency-domain features reflect the frequency composition of the signal, and time-frequency domain joint distribution features are obtained through methods such as wavelet transform and can reveal the complex changes of the signal in time and frequency. These features constitute a multi-dimensional feature vector describing the operation state of the cable. Use the extracted multiple sets of features as the input and the state labels corresponding to the historical operation data as the output to train the initial model. The state labels clearly indicate the operation state (normal, early fault, fault) of the cable corresponding to the data points, helping the model learn the feature patterns under different states. The training process can include, but is not limited to, support vector machine (SVM) and long short-term memory network (LSTM) algorithms, aiming to build a model that can accurately predict the cable state.

[0069] In the above way, the power cable fault warning model can learn the feature patterns of normal operation, early fault, and fault states from historical data and has the ability to identify and warn of potential cable faults. After the model training is completed, it can be applied to real-time monitoring data. Through preprocessing, feature extraction, and classification prediction of real-time data, timely warning of early faults of power cables can be achieved, thereby improving the safety and reliability of the power system.

[0070] Optionally, build a fault prediction model that combines support vector machine (SVM) and long short-term memory network (LSTM). According to the model training and evaluation results, combined with the actual operation requirements and safety standards of the cable, set a reasonable fault warning threshold. When the model prediction value exceeds the set threshold, it is determined that the cable may have an early fault risk and trigger the warning mechanism. Specifically include:

[0071] Step S1101, build a fault prediction model that combines support vector machine (SVM) and long short-term memory network (LSTM), which can accurately classify and predict the operation state of the cable according to the extracted disturbance features.

[0072] Step S1102: According to the cable operation historical data, including normal operation status, early fault statuses of different degrees, and actual data after faults occur. Divide these data into a training set and a test set according to a certain ratio. The training set is used to train the model, and the test set is used to evaluate the performance of the model.

[0073] Step S1103: Perform feature engineering processing on the training set data. Use the extracted time-domain, frequency-domain, and time-frequency-domain features as the input vectors of the model, and mark the corresponding output labels according to the actual operation status of the cable (normal, early fault, fault). Then, use the training set data to train the selected machine learning algorithm so that the model can accurately learn the feature patterns in different operation states and minimize the prediction error.

[0074] Step S1104: According to the model training and evaluation results, combined with the actual operation requirements and safety standards of the cable, set a reasonable fault warning threshold. When the model prediction value exceeds the set threshold, it is determined that the cable may have an early fault risk, and the warning mechanism is triggered. The setting of the warning threshold is adjusted in combination with expert experience and actual operation conditions to ensure the timeliness and accuracy of the warning and avoid false alarms and missed alarms.

[0075] After the above training and learning, an early cable fault prediction and warning model integrating disturbance features is formed to achieve the early prediction and effective warning of permanent cable faults.

[0076] Through the above steps S102 to S110, it is possible to achieve the purpose of accurately identifying power cable faults by using the multi-source data acquisition, preprocessing, and multi-dimensional feature extraction methods integrating disturbance features, combined with the comparison between the machine learning model and the preset feature library. Thus, the technical effects of improving the safety and reliability of the power system and reducing the operation and maintenance costs are achieved, and further, the technical problem of the difficulty in timely and accurately warning of early power cable faults in the related technologies is solved.

[0077] Based on the above embodiments and alternative embodiments, the present invention proposes an alternative implementation manner. Figure 2 It is a schematic diagram of an alternative power cable fault warning process framework according to an embodiment of the present invention. Figure 3 It is a schematic diagram of an alternative power cable fault warning system framework according to an embodiment of the present invention. This method can be applied to a system framework as shown in Figure 3 and a process framework as shown in Figure 2 This method includes:

[0078] S1, Data acquisition and preprocessing: According to the characteristics of cable fault signals, set an appropriate sampling frequency, filter the collected original signals through a filter, and use the Daubechies wavelet basis to perform threshold processing on the coefficients after wavelet decomposition to remove noise coefficients, and then perform wavelet reconstruction to achieve signal denoising. Finally, use an adaptive signal enhancement algorithm to highlight weak fault feature signals. Specifically, it includes:

[0079] S11, For voltage and current waveform data, the sampling frequency is set to 50 kHz to meet the requirement of capturing early fault disturbance signals. Such a sampling frequency can accurately record the detailed changes of the waveform and provide sufficient data support for subsequent analysis. For partial discharge pulse signals (i.e., partial discharge data), since their signals are weak and change rapidly, the sampling frequency is as high as 40 MHz to accurately obtain the detailed characteristics of weak partial discharge signals.

[0080] S12, Filter the collected original signals sequentially through a filter. The filtering algorithm uses the infinite impulse response (IIR) filter algorithm to effectively remove interference signals and improve signal quality.

[0081] S13, Use the wavelet transform algorithm to perform noise reduction processing on the collected original signals. Select the Daubechies wavelet basis and determine the decomposition level according to the characteristics of the signal and the noise level.

[0082] S14, Normalize the denoised waveform data so that the data has a unified scale within a specific range, facilitating subsequent analysis and comparison.

[0083] S2, Extraction of early cable fault characteristics by fusing weak disturbance and partial discharge characteristics: Extract time-domain, frequency-domain characteristics of weak fault feature signals and partial discharge pulse signals, compare and match the extracted time-domain, frequency-domain, and time-frequency domain characteristics with the standard characteristics in the established waveform disturbance feature library, and analyze whether there are abnormal disturbances in the current operating state of the cable. Judge whether there are specific types of partial discharge phenomena, as well as their severity and development trend. Specifically, it includes:

[0084] S21, Time-domain feature extraction: Calculate basic statistical features such as the mean, variance, and peak value of specific waveform data (weak fault feature signals and partial discharge pulse signals) in the preprocessed signals, which can reflect the overall amplitude and fluctuation of the signals.

[0085] S22, Extract features such as the rise time, fall time, and pulse width of specific waveform data in the preprocessed signals. These features are closely related to the occurrence and development process of faults. By monitoring and analyzing these features, it can be preliminarily judged whether there are early fault disturbances in the cable.

[0086] S23, Frequency-domain feature extraction: Perform a fast Fourier transform (FFT) on the specific waveform data in the preprocessed data to convert the time-domain signal into a frequency-domain signal and obtain a spectrogram.

[0087] S24, Calculate the power spectral density (PSD) to further analyze the energy distribution of the original operating data at different frequencies.

[0088] S25, Time-frequency domain feature extraction: Adopt the wavelet transform time-frequency analysis method to obtain the joint distribution characteristics of the original operating data in time and frequency, namely the time-frequency diagram.

[0089] S26, Extract the energy distribution characteristics from the time-frequency diagram.

[0090] S27, Deeply study the diverse waveform disturbance characteristics that may occur before the permanent fault of cable equipment, and establish a waveform disturbance feature library. Compare and match the extracted time-domain, frequency-domain, and time-frequency domain features with the standard features in the feature library to analyze whether there are abnormal disturbances in the current operating state of the cable.

[0091] S28, Judge whether there is a specific type of partial discharge phenomenon, as well as its severity and development trend by comparing the characteristics of partial discharge signals.

[0092] S3, Construction of a cable early fault prediction and warning model based on machine learning: Construct a fault prediction model that combines a support vector machine (SVM) and a long short-term memory network (LSTM). According to the model training and evaluation results, combined with the actual operating requirements and safety standards of the cable, set a reasonable fault warning threshold. When the model prediction value exceeds the set threshold, it is determined that the cable may have an early fault risk and trigger the warning mechanism. Specifically include:

[0093] S31, Construct a fault prediction model that combines a support vector machine (SVM) and a long short-term memory network (LSTM).

[0094] S32, According to the cable operation historical data, including normal operating state, different degrees of early fault states, and actual operating signals after the fault occurs. Divide these data into a training set and a test set according to a certain proportion. The training set is used to train the model and evaluate the performance of the model.

[0095] S33, According to the model training and evaluation results, combined with the actual operating requirements and safety standards of the cable, set a reasonable fault warning threshold. When the model prediction value exceeds the set threshold, it is determined that the cable may have an early fault risk and trigger the warning mechanism.

[0096] It should be noted that this embodiment is not limited to single feature analysis. Instead, it comprehensively considers various disturbance features. After fusing multi-source features, it can comprehensively evaluate the operating state of the cable from multiple perspectives. The features complement and verify each other, effectively reducing misjudgment cases caused by abnormal single features and also reducing the risk of missed judgment due to insignificant changes in some features. At the same time, for medium-voltage power cables in the distribution network, this embodiment establishes a full-life-cycle early warning model for cable faults that dynamically develops from partial discharge, early faults to permanent faults.

[0097] In this embodiment, a power cable fault early warning device is also provided. This device is used to implement the above-mentioned embodiments and preferred implementation manners, and those that have been described will not be repeated. As used hereinafter, the terms "module" and "device" can be a combination of software and / or hardware that can achieve a predetermined function. Although the devices described in the following embodiments are preferably implemented in software, implementation in hardware, or a combination of software and hardware is also possible and contemplated.

[0098] According to an embodiment of the present invention, an apparatus embodiment for implementing the above-mentioned power cable fault early warning method is also provided. Figure 4 It is a schematic structural diagram of a power cable fault early warning device according to an embodiment of the present invention. As Figure 4 shown, the above-mentioned power cable fault early warning device includes: a data acquisition module 400, a data preprocessing module 402, a feature extraction module 404, a feature detection module 406, and a fault early warning module 408, where:

[0099] The data acquisition module 400 is used to acquire the original operating data of the power cable according to a predetermined sampling frequency. Among them, the original operating data at least includes: voltage waveform data, current waveform data, and partial discharge data;

[0100] The data preprocessing module 402 is used to preprocess the original operating data to obtain preprocessed data. Among them, the preprocessing at least includes filtering processing and noise reduction processing;

[0101] The feature extraction module 404 is used to extract target features from the preprocessed data. Among them, the target features at least include the time-domain features, frequency-domain features of the original operating data, and the joint distribution features of the original operating data in the time domain and frequency domain;

[0102] The feature detection module 406 is used to compare the target features with the standard features stored in the preset feature library to detect whether there are abnormal disturbances in the original operating data;

[0103] The fault early warning module 408 is used to, when it is detected that there are abnormal disturbances in the original operating data, based on the target features, adopt the power cable fault early warning model to obtain the fault early warning result of the power cable.

[0104] It should be noted that the above-mentioned various modules can be implemented by software or hardware. For example, for the latter, it can be implemented in the following ways: the above-mentioned various modules can be located in the same processor; or, the above-mentioned various modules are located in different processors in any combination.

[0105] It should be noted here that the above data acquisition module 400, data preprocessing module 402, feature extraction module 404, feature detection module 406, and fault warning module 408 correspond to steps S102 to S110 in the embodiment. The examples and application scenarios implemented by the above modules and the corresponding steps are the same, but are not limited to the content disclosed in the above embodiment. It should be noted that the above modules can run in a computer terminal as part of the device.

[0106] It should be noted that the optional or preferred implementation manners of this embodiment can be referred to the relevant descriptions in the embodiment, and will not be elaborated here.

[0107] The above power cable fault warning device may further include a processor and a memory. The above data acquisition module 400, data preprocessing module 402, feature extraction module 404, feature detection module 406, fault warning module 408, etc. are all stored in the memory as program modules, and the corresponding functions are implemented by the processor executing the above program modules stored in the memory.

[0108] The processor contains a kernel, and the kernel retrieves the corresponding program module from the memory. One or more kernels can be set. The memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of, for example, read-only memory (ROM) or flash memory (flash RAM), and the memory includes at least one storage chip.

[0109] According to the embodiment of the present application, an embodiment of a non-volatile storage medium is also provided. Optionally, in this embodiment, the above non-volatile storage medium includes a stored program, wherein when the above program runs, it controls the device where the non-volatile storage medium is located to execute any one of the above power cable fault warning methods.

[0110] Optionally, in this embodiment, the above non-volatile storage medium can be located in any one of the computer terminals in the computer terminal group in the computer network, or in any one of the mobile terminals in the mobile terminal group. The above non-volatile storage medium includes a stored program.

[0111] Optionally, when the program is running, control the device where the non-volatile storage medium is located to perform the following functions: collect the original operation data of the power cable at a predetermined sampling frequency, where the original operation data at least includes: voltage waveform data, current waveform data, and partial discharge data; preprocess the original operation data to obtain preprocessed data, where the preprocessing at least includes filtering and noise reduction; extract the target features from the preprocessed data, where the target features at least include the time-domain features, frequency-domain features of the original operation data, and the joint distribution features of the original operation data in the time domain and frequency domain; compare the target features with the standard features stored in the preset feature library to detect whether there are abnormal disturbances in the original operation data; in the case of detecting abnormal disturbances in the original operation data, based on the target features, use the power cable fault warning model to obtain the power cable fault warning result.

[0112] According to an embodiment of the present application, an embodiment of a processor is further provided. Optionally, in this embodiment, the above-mentioned processor is used to run a program, where the above-mentioned program, when running, executes any one of the above-mentioned power cable fault warning methods.

[0113] According to an embodiment of the present application, an embodiment of a computer program product is further provided. When executed on a data processing device, it is adapted to execute a program initialized with the steps of any one of the above-mentioned power cable fault warning methods.

[0114] Optionally, when the above-mentioned computer program product is executed on a data processing device, it is adapted to execute a program initialized with the following method steps: collect the original operation data of the power cable at a predetermined sampling frequency, where the original operation data at least includes: voltage waveform data, current waveform data, and partial discharge data; preprocess the original operation data to obtain preprocessed data, where the preprocessing at least includes filtering and noise reduction; extract the target features from the preprocessed data, where the target features at least include the time-domain features, frequency-domain features of the original operation data, and the joint distribution features of the original operation data in the time domain and frequency domain; compare the target features with the standard features stored in the preset feature library to detect whether there are abnormal disturbances in the original operation data; in the case of detecting abnormal disturbances in the original operation data, based on the target features, use the power cable fault warning model to obtain the power cable fault warning result.

[0115] An embodiment of the present invention provides an electronic device, which includes a processor, a memory, and a program stored in the memory and executable on the processor. When the processor executes the program, the following steps are implemented: acquiring original operation data of a power cable at a predetermined sampling frequency, where the original operation data at least includes: voltage waveform data, current waveform data, and partial discharge data; preprocessing the original operation data to obtain preprocessed data, where the preprocessing at least includes filtering processing and noise reduction processing; extracting target features from the preprocessed data, where the target features at least include time-domain features, frequency-domain features of the original operation data, and joint distribution features of the original operation data in the time domain and frequency domain; comparing the target features with standard features stored in a preset feature library to detect whether there are abnormal disturbances in the original operation data; and in the case where abnormal disturbances are detected in the original operation data, based on the target features, using a power cable fault warning model to obtain a fault warning result of the power cable.

[0116] The order of the above embodiments of the present invention is only for description and does not represent the advantages or disadvantages of the embodiments.

[0117] In the above embodiments of the present invention, the descriptions of the respective embodiments have their own emphases. For parts not detailed in a certain embodiment, reference may be made to the relevant descriptions of other embodiments.

[0118] In several embodiments provided by the present application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only illustrative. For example, the above module division can be a logical function division. In actual implementation, there can be other division methods. For example, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed mutual coupling or direct coupling or communication connection can be through some interfaces. The indirect coupling or communication connection of modules or modules can be in an electrical or other form.

[0119] The modules described as separate components above may or may not be physically separated. The components displayed as modules may or may not be physical modules, that is, they can be located in one place or distributed to multiple modules. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0120] In addition, in each embodiment of the present invention, the functional modules can be integrated in a processing module, or each module can exist physically alone, or two or more modules can be integrated in one module. The above integrated modules can be implemented in the form of hardware or in the form of software functional modules.

[0121] If the above integrated module is implemented in the form of a software functional module and sold or used as an independent product, it can be stored in a computer-readable non-volatile storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a non-volatile storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods of various embodiments of the present invention. The aforementioned non-volatile storage medium includes: various media such as USB flash drives, read-only memories (ROMs), random access memories (RAMs), external hard drives, magnetic disks, or optical discs that can store program codes.

[0122] The above are only the preferred embodiments of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the principle of the present invention, several improvements and refinements can be made, and these improvements and refinements should also be regarded as the protection scope of the present invention.

Claims

1. A power cable fault early warning method, characterized in that: include: Collecting original operation data of the power cable according to a predetermined sampling frequency, wherein the original operation data at least includes: voltage waveform data, current waveform data and partial discharge data; Preprocessing the original operation data to obtain preprocessed data, wherein the preprocessing at least includes filtering processing and noise reduction processing; Extracting target features from the preprocessed data, wherein the target features include at least time domain features and frequency domain features of the original operation data, and joint distribution features of the original operation data in the time domain and the frequency domain; Comparing the target feature with the standard feature stored in a preset feature library to detect whether there is abnormal disturbance in the original operation data; When abnormal disturbance is detected in the original operation data, a power cable fault warning model is adopted based on the target feature to obtain a fault warning result of the power cable.

2. The method according to claim 1, characterized in that The collecting of the original operation data of the power cable according to the predetermined sampling frequency includes: The voltage waveform data and the current waveform data of the power cable are collected at a first predetermined sampling frequency, and the partial discharge data of the power cable is collected at a second predetermined sampling frequency to obtain the original operation data, wherein the first predetermined sampling frequency is less than the second predetermined sampling frequency.

3. The method according to claim 1, characterized in that The preprocessing of the original operation data to obtain preprocessed data includes: Using an infinite impulse response filter algorithm to filter the original operating data to obtain filtered data; The wavelet transform algorithm is used to perform noise reduction processing on the filtered data to obtain the preprocessed data.

4. The method according to claim 3, characterized in that The preprocessing of the original operation data to obtain preprocessed data includes: Using the wavelet transform algorithm to perform noise reduction processing on the filtered data to obtain noise-reduced data; The data features associated with the predetermined fault type in the noise reduction data are enhanced to obtain the preprocessed data.

5. The method according to claim 4, characterized in that The adopting of a wavelet transform algorithm to perform noise reduction processing on the filtered data to obtain noise-reduced data comprises: Determine the number of wavelet decomposition layers corresponding to the filtered data; Based on the wavelet decomposition layer number, the filtered data is subjected to multi-resolution wavelet decomposition to obtain a plurality of wavelet coefficients, wherein the plurality of wavelet coefficients include high-frequency wavelet coefficients representing data details and low-frequency wavelet coefficients representing data overviews, the high-frequency wavelet coefficients corresponding to a frequency greater than a preset frequency, and the low-frequency wavelet coefficients corresponding to a frequency less than or equal to the preset frequency; Performing denoising processing on the high-frequency wavelet coefficients based on a preset threshold to obtain processed high-frequency wavelet coefficients; The processed high-frequency wavelet coefficients and the low-frequency wavelet coefficients are subjected to wavelet reconstruction to obtain the noise reduction data.

6. The method according to claim 1, characterized in that The extracting target features from the preprocessed data comprises: Extracting the time domain features from the preprocessed data, wherein the time domain features at least include a mean, a variance, and a peak value of the preprocessed data; Extracting transient features from the preprocessed data, wherein the transient features include a rising time, a falling time, and a pulse width in a waveform of the preprocessed data; Performing fast Fourier transform on the preprocessed data to obtain the frequency domain features; Using wavelet transform time-frequency analysis method to obtain the joint distribution characteristics of the original operation data in time domain and frequency; Among them, the target features include the time domain features, the transient features, the frequency domain features and the joint distribution features.

7. The method according to any one of claims 1 to 6, characterized in that In the case where the abnormal disturbance is detected in the original operation data, before obtaining the fault warning result of the power cable by using the power cable fault warning model based on the target feature, the method further includes: Acquire multiple groups of historical operation data of the power cable and status labels corresponding to the multiple groups of historical operation data, wherein each group of historical operation data includes voltage waveform data, current waveform data and partial discharge data of the power cable in a corresponding historical period; Preprocessing the multiple sets of historical operation data to obtain multiple sets of preprocessed data, wherein the preprocessing at least includes filtering processing and noise reduction processing; Performing feature extraction on the multiple groups of preprocessed data to obtain multiple groups of features, wherein each group of features includes time domain features and frequency domain features of the corresponding historical operation data, and joint distribution features of the corresponding historical operation data in the time domain and the frequency domain; The multiple groups of features are used as input, and the state labels corresponding to the multiple groups of historical operation data are used as output, and the initial model is trained to obtain the power cable fault warning model.

8. A power cable fault warning device, characterized in that: include: A data acquisition module, used to collect original operation data of the power cable according to a predetermined sampling frequency, wherein the original operation data at least includes: voltage waveform data, current waveform data and partial discharge data; A data preprocessing module, used to preprocess the original operation data to obtain preprocessed data, wherein the preprocessing at least includes filtering processing and noise reduction processing; A feature extraction module, used to extract target features from the preprocessed data, wherein the target features include at least time domain features and frequency domain features of the original operation data, and joint distribution features of the original operation data in the time domain and the frequency domain; A feature detection module, used to compare the target feature with the standard feature stored in a preset feature library to detect whether there is abnormal disturbance in the original operation data; The fault warning module is used to obtain a fault warning result of the power cable by adopting a power cable fault warning model based on the target feature when abnormal disturbance is detected in the original operation data.

9. A non-volatile storage medium, characterized in that: The non-volatile storage medium stores a plurality of instructions, and the instructions are suitable for being loaded by a processor and executing the power cable fault early warning method according to any one of claims 1 to 7.

10. An electronic device, characterized in that: It includes one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the power cable fault warning method described in any one of claims 1 to 7.

Images