A fault detection method for a distribution overhead line

Abstract

The application relates to the technical field of electric power, and discloses a fault detection method for a power distribution overhead line, which comprises the following steps: obtaining a target signal set, using wavelet transform to decompose the target signal, obtaining a plurality of high-frequency detail coefficients and low-frequency approximation coefficients, using empirical mode decomposition to decompose the low-frequency approximation coefficients, obtaining a plurality of intrinsic mode functions, then constructing a multi-dimensional feature vector corresponding to the target signal, and finally inputting each multi-dimensional feature vector corresponding to each signal in the target signal set into a preset line fault prediction model to obtain a fault classification detection result; that is, the signal is processed through the fusion of wavelet transform and empirical mode decomposition, the fault features in the non-stationary signal can be more effectively extracted, hidden faults in the power distribution overhead line can be quickly and accurately detected, the risk of serious line faults caused by hidden faults can be effectively reduced, and the stable operation of the power system can be ensured.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electric power, in particular to a fault detection method for distribution overhead line. BACKGROUND

[0002] In the power system, line fault detection is a key link to ensure the safe and stable operation of the entire system. With the continuous expansion of the power grid, its complexity is also constantly improving, especially for long-distance distribution overhead lines, the concealment and complexity of line faults are increasingly prominent. The distribution overhead line is directly exposed to the external environment and is easily affected by weather conditions (such as lightning, freezing, wind deflection, etc.) and external interference (such as tree contact, foreign object intrusion, etc.), and is prone to various hidden faults. This kind of fault usually shows non-typical and short-term signal characteristics, such as local arc discharge and intermittent contact fault.

[0003] The signal of the hidden fault has non-stationary, transient and low energy characteristics, which makes the existing fault analysis methods face many difficulties in processing such signals. At present, most fault analysis methods are based on linear analysis tools (such as Fourier transform) or use simple feature extraction methods. These methods work well when dealing with stationary signals, but have limited processing capacity for non-stationary signals, making it difficult to accurately extract key fault features from the signal.

[0004] More seriously, if the hidden fault cannot be effectively repaired for a long time, it is likely to gradually evolve into a serious line fault, which will pose a major threat to the stability of the power system, and may even cause a large-scale power outage, causing great losses to social production and life.

[0005] Therefore, it is of great importance to develop a technical solution that can quickly process non-stationary signals and accurately extract hidden fault features for the intelligent operation of the power system. SUMMARY

[0006] Therefore, it is necessary to propose a fault detection method for distribution overhead line to effectively extract fault features in non-stationary signals, so as to quickly and accurately detect hidden faults in distribution overhead line, effectively reduce the risk of serious line faults caused by hidden faults, and ensure the stable operation of the power system.

[0007] To achieve the above purpose, the present application provides a fault detection method for distribution overhead line in the first aspect, which comprises:

[0008] Obtaining a target signal set, the target signal set comprising three phase current signals, three phase voltage signals and an electric field intensity signal of the distribution overhead line;

[0009] decomposing the target signal using wavelet transform to obtain a plurality of layers of high-frequency detail coefficients and low-frequency approximation coefficients, wherein the target signal is any one of the target signal set;

[0010] decomposing the low-frequency approximation coefficients using empirical mode decomposition to obtain a plurality of intrinsic mode functions;

[0011] determining mutual information between each intrinsic mode function and each layer of high-frequency detail coefficients, and taking the intrinsic mode function corresponding to mutual information greater than a mutual information threshold as a high-frequency intrinsic mode function;

[0012] constructing a multi-dimensional feature vector corresponding to the target signal according to all high-frequency intrinsic mode functions and all high-frequency detail coefficients;

[0013] inputting each multi-dimensional feature vector corresponding to each signal in the target signal set into a preset line fault prediction model to obtain a fault classification detection result.

[0014] Optionally, the decomposing the target signal using wavelet transform to obtain a plurality of layers of high-frequency detail coefficients and low-frequency approximation coefficients comprises:

[0015] iteratively decomposing the target signal using the wavelet transform until the number of decomposition layers of the current iteration decomposition is equal to a preset number of decomposition layers, to obtain a plurality of layers of high-frequency detail coefficients corresponding to all iteration decompositions, and the low-frequency approximation coefficient corresponding to the current iteration decomposition.

[0016] Optionally, the method further comprises:

[0017] transforming the target signal using short-time Fourier transform to obtain a time-frequency spectrogram;

[0018] extracting a signal dominant frequency from the time-frequency spectrogram;

[0019] determining the preset number of decomposition layers according to the signal dominant frequency.

[0020] Optionally, the determining the preset number of decomposition layers according to the signal dominant frequency comprises:

[0021] determining the preset number of decomposition layers by using the formula

[0022] wherein, the preset number of decomposition layers, is a floor function, is a logarithm with base 2, is a signal sampling frequency of the target signal, is the signal dominant frequency. ​

[0023] Optionally, the using the empirical mode decomposition to decompose the low-frequency approximation coefficient to obtain a plurality of intrinsic mode functions comprises:

[0024] The using the empirical mode decomposition to iteratively decompose the low-frequency approximation coefficient until a residual component obtained by a current iteration decomposition has an energy less than or equal to an energy threshold value to obtain a plurality of initial intrinsic mode functions corresponding to all the iteration decompositions, and the first preset initial intrinsic mode function is taken as an intrinsic mode function.

[0025] Optionally, the method further comprises:

[0026] determining the energy of the target signal;

[0027] taking a preset percentage of the energy of the target signal as the energy threshold value.

[0028] Optionally, before the inputting the plurality of multi-dimensional feature vectors corresponding to each signal in the target signal set into a preset line fault prediction model to obtain a fault classification detection result, the method further comprises:

[0029] using a short-time Fourier transform to transform the target signal to obtain a time-frequency spectrogram;

[0030] sequentially arranging all the high-frequency detail coefficients to obtain a coefficient vector;

[0031] determining the energy of each intrinsic mode function, and sequentially arranging the energy of all the intrinsic mode functions to obtain an energy vector;

[0032] constructing a three-dimensional tensor corresponding to the target signal according to the time-frequency spectrogram, the coefficient vector and the energy vector;

[0033] The inputting the plurality of multi-dimensional feature vectors corresponding to each signal in the target signal set into a preset line fault prediction model to obtain a fault classification detection result comprises:

[0034] inputting the plurality of multi-dimensional feature vectors corresponding to each signal in the target signal set and the plurality of three-dimensional tensors into the preset line fault prediction model to obtain the fault classification detection result.

[0035] Optionally, the method further comprises:

[0036] determining a current amplitude, a voltage drop amplitude, a harmonic content and a fault time according to three phase current signals and three phase voltage signals;

[0037] using a preset fuzzy rule to determine a fault type detection result according to the current amplitude, the voltage drop amplitude, the harmonic content and the fault time.

[0038] According to the fault type detection result and the fault classification detection result, a target fault classification detection result is determined.

[0039] Optionally, the mutual information between each intrinsic mode function and each layer of high-frequency detail coefficient is determined by:

[0040] determining the joint probability distribution between each intrinsic mode function and each layer of high-frequency detail coefficient, determining the marginal probability distribution of each intrinsic mode function, and determining the marginal probability distribution of each layer of high-frequency detail coefficient;

[0041] According to the joint probability distribution between each intrinsic mode function and each layer of high-frequency detail coefficient, the marginal probability distribution of each intrinsic mode function, and the marginal probability distribution of each layer of high-frequency detail coefficient, the mutual information between each intrinsic mode function and each layer of high-frequency detail coefficient is determined.

[0042] Optionally, the target signal set is obtained by:

[0043] Three initial phase current signals, three initial phase voltage signals and an initial electric field intensity signal collected by an electromagnetic mutual inductor are obtained, wherein the electromagnetic mutual inductor has been installed on the overhead distribution line.

[0044] Each of the initial phase current signals, the initial phase voltage signals and the initial electric field intensity signal is subjected to Gaussian kernel denoising processing to obtain three phase current signals, three phase voltage signals and an electric field intensity signal.

[0045] The three phase current signals, the three phase voltage signals and the electric field intensity signal are all taken as elements to form the target signal set.

[0046] To achieve the above-mentioned purpose, the second aspect of the present application provides a fault detection device for an overhead distribution line, which comprises:

[0047] An acquisition module is configured to acquire a target signal set, wherein the target signal set comprises three phase current signals, three phase voltage signals and an electric field intensity signal of the overhead distribution line.

[0048] A first decomposition module is configured to use wavelet transform to decompose a target signal to obtain a plurality of layers of high-frequency detail coefficients and low-frequency approximation coefficients, wherein the target signal is any one of the target signal set.

[0049] A second decomposition module is configured to use empirical mode decomposition to decompose the low-frequency approximation coefficients to obtain a plurality of intrinsic mode functions.

[0050] determine mutual information between each intrinsic modal function and each layer of high frequency detail coefficient, and take the intrinsic modal function corresponding to mutual information greater than a mutual information threshold as a high frequency intrinsic modal function;

[0051] construct a multi-dimensional feature vector corresponding to the target signal according to all high frequency intrinsic modal functions and all high frequency detail coefficients;

[0052] predicting module, for inputting each multi-dimensional feature vector corresponding to each signal in the target signal set into a preset line fault prediction model to obtain a fault classification detection result.

[0053] To achieve the above object, the present application provides a computer readable storage medium storing a computer program in a third aspect, the computer program is executed by a processor, so that the processor executes the method as any one of the first aspect.

[0054] To achieve the above object, the present application provides a computer device in a fourth aspect, comprising a memory and a processor, the memory stores a computer program, the computer program is executed by the processor, so that the processor executes the method as any one of the first aspect.

[0055] The embodiment of the present application has the following beneficial effects: the above method obtains a target signal set, the target signal set includes three phase current signals, three phase voltage signals and electric field intensity signals of the distribution overhead line, uses wavelet transform to decompose the target signal to obtain a plurality of layers of high frequency detail coefficients and low frequency approximation coefficients, wherein the target signal is any one signal in the target signal set, and uses empirical mode decomposition to decompose the low frequency approximation coefficient to obtain a plurality of intrinsic modal functions, then determines mutual information between each intrinsic modal function and each layer of high frequency detail coefficient, and takes the intrinsic modal function corresponding to mutual information greater than a mutual information threshold as a high frequency intrinsic modal function, then constructs a multi-dimensional feature vector corresponding to the target signal according to all high frequency intrinsic modal functions and all high frequency detail coefficients, finally inputs each multi-dimensional feature vector corresponding to each signal in the target signal set into a preset line fault prediction model to obtain a fault classification detection result; that is, by fusing wavelet transform and empirical mode decomposition to process the signal, the fault features in the non-stationary signal can be more effectively extracted, so that the hidden fault in the distribution overhead line can be quickly and accurately detected, reliable support is provided for the intelligent operation of the power system, the risk of serious line fault caused by hidden fault is effectively reduced, and the stable operation of the power system is ensured. BRIEF DESCRIPTION OF DRAWINGS

[0056] In order to make the technical solutions in the embodiments of the present application or the prior art clearer, the accompanying drawings needed in the embodiments or prior art description will be briefly introduced below. Obviously, the accompanying drawings in the following description only represent some embodiments of the present application, and for those skilled in the art, other drawings can also be obtained without creative effort.

[0057] Wherein:

[0058] Figure 1 It is a schematic diagram of a fault detection method of a distribution overhead line in an embodiment of the present application.

[0059] Figure 2 It is a schematic diagram of a fault detection device of a distribution overhead line in an embodiment of the present application.

[0060] Figure 3 It is an internal structure diagram of a computer device in some embodiments. DETAILED DESCRIPTION

[0061] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments only represent some embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative effort fall within the scope of the present application.

[0062] In a power system, line fault detection is a key link to ensure the safe and stable operation of the entire system. With the continuous expansion of the power grid, its complexity is also increasing, especially for long-distance distribution overhead lines, the concealment and complexity of line faults are increasingly significant. The distribution overhead line is directly exposed to the external environment and is easily affected by weather conditions (such as lightning, freezing, wind deflection, etc.) and external interference (such as tree contact, foreign matter intrusion, etc.), and is prone to diversified concealed faults. Such faults usually exhibit non-typical and short-lived signal characteristics, such as local arc discharge and intermittent contact faults.

[0063] The signals of concealed faults have non-stationary, transient and low-energy characteristics, which makes it difficult for existing fault analysis methods to handle such signals. At present, most fault analysis methods are based on linear analysis tools (such as Fourier transform) or use simple feature extraction methods. These methods work well when dealing with stationary signals, but have limited processing capacity for non-stationary signals, making it difficult to accurately extract key fault features from the signals.

[0064] More seriously, if the hidden fault cannot be effectively repaired for a long time, it is extremely likely to gradually evolve into a serious line fault, which further threatens the stability of the power system, and even may cause a large-scale power outage accident, causing great loss to social production and life.

[0065] Therefore, it is of great importance to develop a technical solution capable of quickly processing non-stationary signals and accurately extracting hidden fault features for ensuring the intelligent operation of the power system.

[0066] In view of the above problems, the present application provides a fault detection method for a distribution overhead line, which can more effectively extract fault features in non-stationary signals, thereby quickly and accurately detecting hidden faults in the distribution overhead line, effectively reducing the risk of serious line faults caused by hidden faults, and ensuring the stable operation of the power system. The specific implementation principle will be described in detail in the following embodiments.

[0067] In a first aspect, the present application provides a fault detection method for a distribution overhead line.

[0068] Please refer to Figure 1 , a schematic diagram of a fault detection method for a distribution overhead line in an embodiment of the present application, which comprises:

[0069] Step 110: Obtain a target signal set, which includes three phase current signals, three phase voltage signals and an electric field intensity signal of the distribution overhead line.

[0070] Wherein, the distribution overhead line here refers to the line to be detected for faults.

[0071] As for the signal collection method, in some embodiments, a signal collection device can be installed on the distribution overhead line to collect the three phase current signals, three phase voltage signals and electric field intensity signals of the distribution overhead line through the signal collection device; wherein the signal collection device can be configured with multiple high-speed collection channels, i.e. each signal collection can be configured with a high-speed collection channel.

[0072] It can be understood that by configuring multiple high-speed collection channels, simultaneous high-speed collection of signals between different channels can be achieved, thereby ensuring the time sequence consistency of the collected signals, avoiding time offset errors between signals in different channels due to different time sequences of the collected signals, and thus affecting the subsequent detection results.

[0073] Further, in some embodiments, the signal acquisition device also supports multiple sampling frequencies and multiple resolution configurations, which can be adjusted according to actual needs; for example, after detecting a fault, it can be adjusted to a high-frequency mode for signal acquisition, and after detecting no fault, it can be adjusted to a low-frequency mode or a normal mode for signal acquisition.

[0074] Step 120: using wavelet transform, decomposing the target signal to obtain multiple layers of high-frequency detail coefficients and low-frequency approximation coefficients, wherein the target signal is any one signal in the target signal set.

[0075] Wherein, the wavelet transform is WT, that is, Wavelet Transform.

[0076] For the decomposition mode of the multiple layers of high-frequency detail coefficients, in some embodiments, the target signal can be sequentially decomposed using wavelet transform to obtain the multiple layers of high-frequency detail coefficients obtained by decomposition.

[0077] For the decomposition mode of the low-frequency approximation coefficient, in some embodiments, the target signal can be sequentially decomposed using wavelet transform to obtain multiple layers of low-frequency approximation coefficients, and then in the multiple layers of low-frequency approximation coefficients, any one layer of low-frequency approximation coefficient is taken as the low-frequency approximation coefficient obtained by decomposition.

[0078] Step 130: using empirical mode decomposition, decomposing the low-frequency approximation coefficient to obtain multiple intrinsic mode functions.

[0079] Wherein, the empirical mode decomposition is EMD, that is, Empirical Mode Decomposition.

[0080] It should be noted that the present application decomposes the low-frequency approximation coefficient obtained by wavelet transform decomposition through empirical mode decomposition, which can finely separate and focus on the fault features of hidden faults, so as to improve the subsequent detection results.

[0081] For the decomposition mode of the multiple intrinsic mode functions, in some embodiments, the low-frequency approximation coefficient is sequentially decomposed using empirical mode decomposition to obtain the multiple intrinsic mode functions obtained by decomposition.

[0082] Step 140: determining the mutual information between each intrinsic mode function and each layer of high-frequency detail coefficient, and taking the intrinsic mode function corresponding to the mutual information greater than the mutual information threshold value as the high-frequency intrinsic mode function.

[0083] Wherein, the mutual information threshold value can be obtained and set in advance by the operator according to a large amount of experience, experiment or statistics, of course, it can also be set by the operator according to actual needs.

[0084] It should be noted that there is a mutual information between each layer of high-frequency detail coefficient and the nth eigenmode function, that is, for the nth eigenmode function, there are multiple mutual information, and in multiple mutual information, as long as there is at least one mutual information greater than the mutual information threshold, the nth eigenmode function is taken as a high-frequency eigenmode function.

[0085] Step 150: constructing a multi-dimensional feature vector corresponding to the target signal according to all high-frequency eigenmode functions and all high-frequency detail coefficients.

[0086] For the construction method of the multi-dimensional feature vector, in some embodiments, all high-frequency eigenmode functions and all high-frequency detail coefficients can be directly taken as the multi-dimensional feature vector, that is, all high-frequency eigenmode functions and all high-frequency detail coefficients can be arranged in sequence to form a multi-dimensional feature vector.

[0087] For the construction method of the multi-dimensional feature vector, in some embodiments, all high-frequency eigenmode functions and all high-frequency detail coefficients can be directly taken as the multi-dimensional feature vector, that is, all high-frequency eigenmode functions and all high-frequency detail coefficients can be arranged in sequence to form a multi-dimensional feature vector.

[0088] Further, for the determination method of the weight of each data in the high-frequency feature set, in some embodiments, the weight of each data in the high-frequency feature set can be determined by using the formula ; wherein, is the weight of the nth data in the high-frequency feature set, is the nth data in the high-frequency feature set, is an L2 norm symbol, is the total number of data in the high-frequency feature set.

[0089] Further, for the multi-dimensional feature vector constructed according to the weight, skewness, kurtosis and approximate entropy of all data in the high-frequency feature set, in some embodiments, the expression of the constructed multi-dimensional feature vector is: ; wherein, is the multi-dimensional feature vector, is the weight of the nth data in the high-frequency feature set, is the skewness of the nth data in the high-frequency feature set, is the nth data in the high-frequency feature set, is the kurtosis of the nth data in the high-frequency feature set, is the approximate entropy of the nth data in the high-frequency feature set, is the total number of data in the high-frequency feature set.

[0090] It should be noted that this application constructs a multi-dimensional feature vector corresponding to the target signal by adopting weights, skewness, kurtosis and approximate entropy, which can accurately extract key fault features from complex hidden fault signals to effectively improve the accuracy of subsequent model fault detection.

[0091] Step 160: Input each multidimensional feature vector corresponding to each signal in the target signal set into a preset line fault prediction model to obtain a fault classification detection result.

[0092] The preset line fault prediction model herein refers to a pre-trained model used to predict and output fault classification detection results based on the multi-dimensional feature vectors corresponding to the input signals.

[0093] It should be noted that each multidimensional feature vector corresponding to each signal in the target signal set can be obtained by repeating steps 120 to 150 multiple times, which will not be described in detail here.

[0094] Regarding the method for obtaining the preset line fault prediction model, in some embodiments, the multidimensional feature vectors corresponding to each signal in a large number of historical signal sets and the corresponding fault classification labels can be obtained, and then the multidimensional feature vectors corresponding to each signal in the large number of historical signal sets and the corresponding fault classification labels are sequentially input into the initial line fault prediction model for training. After training to a certain extent, a trained preset line fault prediction model is obtained; wherein, the fault classification label can be used as the true value of the initial line fault prediction model, that is, by comparing the true value with the fault classification detection result output during the training process, it can be determined whether the initial line fault prediction model is well trained and meets the expected requirements.

[0095] Regarding the type of initial line fault prediction model, in some embodiments, a CNN-LSTM-Transformer model can be selected as the initial line fault prediction model; wherein CNN stands for Convolutional Neural Network, and LSTM stands for Long Short-Term Memory.

[0096] In the embodiment of the present application, by fusing wavelet transform and empirical mode decomposition to process the signal, the fault characteristics in the non-stationary signal can be more effectively extracted, thereby quickly and accurately detecting the hidden faults in the distribution overhead lines, providing reliable support for the intelligent operation of the power system, effectively reducing the risk of serious line failures caused by hidden faults, and ensuring the stable operation of the power system.

[0097] In addition to the above-mentioned effects of quickly and accurately detecting hidden faults, reducing the risk of serious line faults, and ensuring the stable operation of the power system, the power distribution overhead line fault detection method has the following advantages: ensuring time sequence consistency: by configuring multiple high-speed acquisition channels in the signal acquisition device, simultaneous high-speed acquisition of signals between different channels is achieved, ensuring the time sequence consistency of the collected signals, avoiding time offset errors caused by different signal times, and thus providing a reliable data foundation for subsequent accurate fault detection; flexible sampling configuration: the signal acquisition device supports configuration of multiple sampling frequencies and multiple resolutions, and can be adjusted according to actual needs, for example, adjusting to high-frequency mode for signal acquisition after detecting a fault, which can capture fault signal characteristics in more detail, and adjusting to low-frequency or normal mode when there is no fault, which meets the collection needs in different scenarios and saves collection resources and storage space to a certain extent; fine separation of fault features: using empirical mode decomposition to decompose the low-frequency approximation coefficients obtained by wavelet transform decomposition can finely separate and focus on the fault features of hidden faults, and through this multi-level decomposition and combined processing method, more subtle and critical fault information can be extracted from complex signals, improving the extraction ability of hidden fault features and helping to more accurately identify fault types and locations; precise selection of high-frequency intrinsic mode functions: by determining the mutual information between each intrinsic mode function and each layer of high-frequency detail coefficients, and selecting the intrinsic mode function corresponding to the mutual information greater than the mutual information threshold as the high-frequency intrinsic mode function, this selection method can effectively find the part with strong correlation with high-frequency detail coefficients from numerous intrinsic mode functions, further highlighting the features related to faults, removing irrelevant or interfering information, and improving the quality and effectiveness of the features; improve fault detection accuracy: use weight, skewness, kurtosis, and approximate entropy to construct a multi-dimensional feature vector, which can describe the characteristics of the data from different angles, accurately extract key fault features from complex hidden fault signals, and thus effectively improve the accuracy of subsequent model fault detection and reduce the possibility of misjudgment and omission. Select an advanced model architecture: select the CNN-LSTM-Transformer model as the initial line fault prediction model, which combines the advantages of convolutional neural network (CNN) in feature extraction, long short-term memory network (LSTM) in time series processing, and Transformer in capturing long-distance dependencies, can more effectively process information in the multi-dimensional feature vector, and further improve the accuracy and reliability of fault classification and detection.

[0098] In a possible implementation, the step 120 in the above embodiment includes: using the wavelet transform to iteratively decompose the target signal until the number of decomposition layers of the current iteration is equal to the preset number of decomposition layers, to obtain the multi-layer high-frequency detail coefficients corresponding to all the iterations and the low-frequency approximation coefficient corresponding to the current iteration.

[0099] The preset number of decomposition layers can be set by the operator according to a large amount of experience, experiments or statistics, or can be set by the operator according to actual needs.

[0100] For the value of the preset number of decomposition layers, in some embodiments, the preset number of decomposition layers can be set to 6.

[0101] For the iterative decomposition manner of the multi-layer high-frequency detail coefficients, in some embodiments, assuming that the number of layers of the high-frequency detail coefficients is six, the wavelet transform can be used to decompose the target signal to obtain the first layer of high-frequency detail coefficients, decompose the first layer of high-frequency detail coefficients to obtain the second layer of high-frequency detail coefficients, decompose the second layer of high-frequency detail coefficients to obtain the third layer of high-frequency detail coefficients, decompose the third layer of high-frequency detail coefficients to obtain the fourth layer of high-frequency detail coefficients, decompose the fourth layer of high-frequency detail coefficients to obtain the fifth layer of high-frequency detail coefficients, decompose the fifth layer of high-frequency detail coefficients to obtain the sixth layer of high-frequency detail coefficients, to obtain the six layers of high-frequency detail coefficients corresponding to all the iterations.

[0102] For the iterative decomposition manner of the low-frequency approximation coefficient, in some embodiments, assuming that the number of layers of the high-frequency detail coefficients is six, the wavelet transform can be used to decompose the target signal to obtain the first layer of low-frequency approximation coefficients, decompose the first layer of low-frequency approximation coefficients to obtain the second layer of low-frequency approximation coefficients, decompose the second layer of low-frequency approximation coefficients to obtain the third layer of low-frequency approximation coefficients, decompose the third layer of low-frequency approximation coefficients to obtain the fourth layer of low-frequency approximation coefficients, decompose the fourth layer of low-frequency approximation coefficients to obtain the fifth layer of low-frequency approximation coefficients, decompose the fifth layer of low-frequency approximation coefficients to obtain the sixth layer of low-frequency approximation coefficients, and take the sixth layer of low-frequency approximation coefficients as the low-frequency approximation coefficient corresponding to the current iteration.

[0103] In the embodiments of the present application, the multi-layer high-frequency detail coefficients and the low-frequency approximation coefficient can be more accurately obtained by iteratively decomposing the target signal, to provide a more reliable basis for subsequent fault feature extraction and detection.

[0104] It can be understood that the accurate acquisition of multi-level coefficients: through iterative decomposition until the preset decomposition layer number is reached, the high-frequency detail coefficients and the final low-frequency approximation coefficients of the target signal at different decomposition levels can be systematically obtained, and such multi-level decomposition method helps to more comprehensively understand the characteristics of the signal and provides rich information for subsequent fault feature extraction; adapt to different faults: different faults may exhibit different characteristics at different decomposition levels, and by presetting the decomposition layer number, the depth of decomposition can be adjusted according to actual needs to better adapt to the feature extraction needs of different types of hidden faults and improve the accuracy of fault detection; improve the quality of feature extraction: the iterative decomposition method can gradually refine the signal, so that the obtained high-frequency detail coefficients and low-frequency approximation coefficients can more accurately reflect the essential characteristics of the signal, which helps to more effectively extract key fault features in the subsequent steps, reduces the interference of irrelevant information or noise, and improves the quality of fault feature extraction; enhance the versatility of the method: the preset decomposition layer number can be set according to actual conditions, so that the method has certain versatility and flexibility, whether it is facing different types, different severity of faults, or different running environment of overhead distribution lines, the preset decomposition layer number can be adjusted to optimize the fault detection effect.

[0105] In a feasible implementation manner, the method in the above embodiment further includes: using a short-time Fourier transform to transform the target signal to obtain a time-frequency spectrogram; performing frequency extraction on the time-frequency spectrogram to obtain a signal dominant frequency; and determining the preset decomposition layer number according to the signal dominant frequency.

[0106] For the determination manner of the preset decomposition layer number, in some embodiments, a signal acquisition frequency of the target signal can be acquired, and the preset decomposition layer number can be determined according to the signal acquisition frequency and the signal dominant frequency.

[0107] In the embodiments of the present application, by dynamically determining the preset decomposition layer number according to the signal dominant frequency, the adaptability and accuracy of fault detection can be improved.

[0108] It can be understood that the adaptability is enhanced: different overhead power distribution line fault signals have different dominant frequencies, the time-frequency spectrum is obtained through short-time Fourier transform and the dominant frequency of the signal is extracted, and then the preset decomposition layer number is determined according to the dominant frequency, so that the method can adaptively adjust according to the characteristics of different fault signals, and the adaptability of the fault detection method to different types of faults is enhanced; the accuracy is improved: the preset decomposition layer number directly affects the effect of fault feature extraction, if the decomposition layer number is set unreasonably, the key fault features may be ignored or too much irrelevant information interference may be caused, the preset decomposition layer number is dynamically determined according to the signal dominant frequency, the high-frequency detail coefficients and low-frequency approximation coefficients related to the fault can be more accurately obtained, and thus the key fault features can be more effectively extracted, and the accuracy of fault detection is improved; the resource utilization is optimized: dynamically determining the preset decomposition layer number can avoid over-decomposition or under-decomposition, over-decomposition will increase the calculation complexity and time cost, and under-decomposition may not be able to sufficiently extract fault features, by reasonably determining the decomposition layer number according to the signal dominant frequency, the utilization of computing resources can be optimized under the premise of ensuring the accuracy of fault detection, and the detection efficiency is improved.

[0109] In a possible implementation, the preset decomposition layer number is determined according to the signal dominant frequency in the above embodiments, including:

[0110] The preset decomposition layer number is determined by using the formula

[0111] Wherein, is the preset decomposition layer number, is a down rounding symbol, is a logarithm with base 2, is a signal sampling frequency of a target signal, is a signal dominant frequency.

[0112] In the embodiments of the present application, the preset decomposition layer number is determined according to the signal dominant frequency by using the formula, the scientific and reasonable determination of the decomposition layer number is realized, and the accuracy and efficiency of fault detection are improved.

[0113] ​It can be understood that the number of decomposition layers is scientifically determined: the preset number of decomposition layers is calculated by a formula, avoiding the subjectivity and randomness of manually setting the number of decomposition layers, making the determination of the number of decomposition layers more scientific and reasonable. The formula comprehensively considers the signal sampling frequency and the signal dominant frequency, both of which are key factors affecting signal decomposition effect. Therefore, the number of decomposition layers calculated by the formula can better adapt to the signal characteristics; improve the accuracy of fault detection: a reasonable number of decomposition layers can ensure that fault features are fully extracted during signal decomposition, while avoiding interference from irrelevant information or noise. The number of decomposition layers determined by the formula can more accurately obtain high-frequency detail coefficients and low-frequency approximation coefficients related to faults, thereby more effectively extracting key fault features and improving the accuracy of fault detection; optimize the utilization of computing resources: dynamically determining the preset number of decomposition layers can avoid over-decomposition or insufficient decomposition, thereby optimizing the utilization of computing resources. Over-decomposition increases computational complexity and time cost, while insufficient decomposition may not fully extract fault features. By reasonably determining the number of decomposition layers through the formula, the detection efficiency can be improved and the computational cost can be reduced while ensuring the accuracy of fault detection.

[0114] In a feasible implementation manner, step 130 in the above embodiment, using empirical mode decomposition, decomposes the low-frequency approximation coefficient to obtain a plurality of intrinsic mode functions, includes: using empirical mode decomposition, iteratively decomposing the low-frequency approximation coefficient until the energy of the residual component obtained by the current iteration decomposition is less than or equal to the energy threshold, obtaining a plurality of initial intrinsic mode functions corresponding to all iteration decompositions, and taking the first preset initial intrinsic mode functions as the intrinsic mode functions.

[0115] Among them, the energy threshold and the specific number of the first preset initial intrinsic mode functions can be obtained by the operator according to a large amount of experience, experiment or statistics and pre-set, of course, can also be set by the operator according to the actual demand.

[0116] For the value of the energy threshold, in some embodiments, the present application can preferably set the energy threshold to 0.7.

[0117] For the specific number of the first preset initial intrinsic mode functions, in some embodiments, the present application can preferably set the first preset initial intrinsic mode functions to the first 3 initial intrinsic mode functions.

[0118] For the iterative decomposition of multiple initial intrinsic modal functions, in some embodiments, the target signal can be decomposed using empirical mode decomposition to obtain a first initial intrinsic modal function and a first residual component, the first residual component is decomposed to obtain a second initial intrinsic modal function and a second residual component, the second residual component is decomposed to obtain a third initial intrinsic modal function and a third residual component, the third residual component is decomposed to obtain a fourth initial intrinsic modal function and a fourth residual component, the fourth residual component is decomposed to obtain a fifth initial intrinsic modal function and a fifth residual component, the fifth residual component is decomposed to obtain a sixth initial intrinsic modal function and a sixth residual component, at this time, the sixth residual component is less than or equal to the energy threshold, and the multiple initial intrinsic modal functions are six initial intrinsic modal functions.

[0119] It should be noted that if the first three initial intrinsic modal functions are set as the intrinsic modal functions, and the energy of the residual component obtained after two times of decomposition is less than or equal to the energy threshold, at this time, the number of intrinsic modal functions is only two, and subsequent decomposition is not performed.

[0120] In the embodiments of the present application, the intrinsic modal functions are obtained by setting the energy threshold for iterative decomposition, which effectively avoids the noise interference caused by over-decomposition, and improves the accuracy and reliability of fault feature extraction.

[0121] It can be understood that over-decomposition is avoided: in the actual signal decomposition process, if the decomposition is excessive, a large amount of noise and irrelevant information will be introduced, which will interfere with the extraction of fault features and reduce the accuracy of fault detection. By setting the energy threshold, the decomposition is stopped when the energy of the residual component obtained by iterative decomposition is less than or equal to the threshold, which can effectively avoid the over-decomposition and reduce the influence of noise and irrelevant information on fault feature extraction; improve the accuracy of feature extraction: since over-decomposition is avoided, the obtained intrinsic modal functions can more accurately reflect the essential features related to faults in the signal. Therefore, in the subsequent construction of a multi-dimensional feature vector and fault detection, more accurate and more relevant feature information can be used to improve the accuracy of fault feature extraction and the reliability of the entire fault detection method; enhance the robustness of the method: in the actual fault detection scene of the distribution overhead line, the signal is often affected by various noise and interference. The method of setting the energy threshold for iterative decomposition can resist these noise and interference to some extent, enhance the adaptability and robustness of the method to different environmental conditions, and ensure that fault features can be effectively extracted and accurate fault detection can be achieved under different conditions.

[0122] In a possible implementation, the method in the above embodiment further includes: determining the energy of the target signal; and setting a preset percentage of the energy of the target signal as the energy threshold.

[0123] The preset percentage can be set by an operator according to a large amount of experience, experiments or statistics in advance, or can be set by the operator according to actual requirements.

[0124] For the value of the preset percentage, in some embodiments, the preset percentage can be set to 5% by the present application.

[0125] In the embodiments of the present application, by dynamically determining the energy threshold, the adaptability and accuracy of fault detection can be improved.

[0126] It can be understood that, by determining the energy of the target signal and setting a preset percentage of the energy of the target signal as the energy threshold, the energy threshold can be dynamically adjusted according to the actual situation of the target signal. This method avoids the inadaptability that can be caused by a fixed energy threshold, because different target signals can have different energy levels. The dynamically determined energy threshold can better adapt to various signal conditions, so that the intrinsic mode function can be more accurately determined in the empirical mode decomposition process, the accuracy and reliability of fault feature extraction are improved, and the performance of the entire fault detection method is improved.

[0127] In a possible implementation, before the step 160 in the above embodiment, the method further includes: transforming the target signal by using short-time Fourier transform to obtain a time-frequency spectrum; arranging all high-frequency detail coefficients in sequence to obtain a coefficient vector; determining the energy of each intrinsic mode function, and arranging the energies of all intrinsic mode functions in sequence to obtain an energy vector; and constructing a three-dimensional tensor corresponding to the target signal according to the time-frequency spectrum, the coefficient vector and the energy vector.

[0128] The step 160 in the above embodiment includes: inputting each multi-dimensional feature vector corresponding to each signal in the target signal set into a preset line fault prediction model to obtain a fault classification detection result, including: inputting each multi-dimensional feature vector corresponding to each signal in the target signal set and each three-dimensional tensor into the preset line fault prediction model to obtain the fault classification detection result.

[0129] For the construction method of the three-dimensional tensor, in some embodiments, the product between the time-frequency spectrum, the coefficient vector and the energy vector can be taken as the three-dimensional tensor.

[0130] In the embodiments of the present application, by constructing a three-dimensional tensor and combining a multi-dimensional feature vector for fault detection, the accuracy and comprehensiveness of fault classification detection are improved.

[0131] It can be understood that more abundant fault information is provided: by using short-time Fourier transform to obtain a time-frequency spectrum, arranging all high-frequency detail coefficients into a coefficient vector, determining the energy of each intrinsic mode function and arranging it into an energy vector, and finally constructing a three-dimensional tensor, the three-dimensional tensor fuses information of the signal in multiple dimensions such as time domain, frequency domain and energy distribution, provides more abundant and comprehensive feature representation for fault detection, and helps to more accurately capture fault features; the recognition ability of the model for faults is enhanced: the multi-dimensional feature vector corresponding to each signal of the target signal is input into the preset line fault prediction model together with the constructed three-dimensional tensor, the multi-dimensional feature vector mainly describes fault features from the perspective of statistical characteristics of the signal, and the three-dimensional tensor supplements information from the perspective of time-frequency energy distribution, the combination of the two enables the model to learn and understand fault features from different angles and different levels, thereby greatly enhancing the recognition ability of the model for various types of faults and improving the accuracy of fault classification detection; the method is suitable for complex and variable fault conditions: in actual operation of the distribution overhead line, fault conditions are complex and diverse, by introducing the three-dimensional tensor as a feature representation method containing multi-dimensional information, the fault detection method can better adapt to these complex and variable fault conditions, and no matter what type of hidden fault is faced, the method can effectively extract key fault features by using input data fusing multi-dimensional information, achieve accurate fault classification detection, and ensure stable operation of the power system.

[0132] In a feasible implementation manner, the method in the above embodiments further includes: determining a current amplitude, a voltage drop amplitude, a harmonic content and a fault time according to the three-phase current signals and the three-phase voltage signals; determining a fault type detection result according to the current amplitude, the voltage drop amplitude, the harmonic content and the fault time by using a preset fuzzy rule; and determining a target fault classification detection result according to the fault type detection result and the fault classification detection result.

[0133] The preset fuzzy rule can be obtained by an operator according to a large amount of experience, experiments or statistics and preset, or can be set by the operator according to actual requirements.

[0134] For the determination manner of the preset fuzzy rule, in some embodiments, the preset fuzzy rule can be determined according to historical fault data and expert experience.

[0135] For the determination of the target fault classification detection result, in some embodiments, it can be judged whether the fault type detection result and the fault classification detection result are consistent, if consistent, the fault classification detection result is taken as the target fault classification detection result, if inconsistent, the method of the present application is re-executed to perform re-detection.

[0136] Further, in some embodiments, if the number of re-detection reaches the preset number and the fault type detection result and the fault classification detection result are still inconsistent, the last obtained fault classification detection result is taken as the target fault classification detection result; of course, in other embodiments, if the number of re-detection reaches the preset number and the fault type detection result and the fault classification detection result are still inconsistent, the first probability of the fault type detection result is determined, and the second probability of the fault classification detection result is determined, and the target fault classification detection result is determined by comparing the first probability and the second probability, that is, in the case that the first probability is greater than the second probability, the fault type detection result is taken as the target fault classification detection result, in the case that the first probability is less than or equal to the second probability, the fault classification detection result is taken as the target fault classification detection result.

[0137] In the embodiments of the present application, by combining the fault type detection result based on the phase current and the phase voltage signal and the fault classification detection result based on the multi-dimensional feature vector, the accuracy and reliability of fault detection are improved.

[0138] It can be understood that complementary fault information is provided: according to three-phase current signals and three-phase voltage signals, current amplitude, voltage drop amplitude, harmonic content and fault time and other characteristics are determined, and a preset fuzzy rule is used to determine the fault type detection result, which provides fault information from the basic electrical characteristics of current and voltage, and the fault classification detection result obtained in the previous step based on the multi-dimensional feature vector is extracted from the non-stationary, transient and other complex characteristics of the signal, the combination of the two provides complementary fault information, which helps to better understand the fault situation; the accuracy of fault detection is enhanced: since the fault type detection result and the fault classification detection result are obtained from different angles, they can be verified with each other, when they are consistent, the type of fault can be more determined, the accuracy of fault detection is improved, when they are inconsistent, through re-detection or comparison of the probability of two detection results, further analysis and judgment can be made, avoiding the misjudgment or omission of a single detection method, thereby improving the reliability of fault detection; adapt to complex and variable fault conditions: in the actual operation of the distribution overhead line, the fault conditions are complex and diverse, by combining the detection results of the two different methods, the fault detection method can better adapt to these complex and variable fault conditions, whether it is a simple fault or a complex hidden fault, the method can more accurately determine the fault type through comprehensive analysis of multiple information, and ensure the stable operation of the power system.

[0139] In a possible implementation, the mutual information between each eigenmode function and each layer of high-frequency detail coefficients in step 140 in the above embodiment is determined by: determining the joint probability distribution between each eigenmode function and each layer of high-frequency detail coefficients, and determining the marginal probability distribution of each eigenmode function and the marginal probability distribution of each layer of high-frequency detail coefficients; and determining the mutual information between each eigenmode function and each layer of high-frequency detail coefficients according to the joint probability distribution between each eigenmode function and each layer of high-frequency detail coefficients, the marginal probability distribution of each eigenmode function, and the marginal probability distribution of each layer of high-frequency detail coefficients.

[0140] In the embodiments of the present application, the mutual information is accurately calculated, the eigenmode functions with strong correlation with the high-frequency detail coefficients are more accurately screened out, and the quality and effectiveness of fault feature extraction are improved.

[0141] It can be understood that the correlation between the intrinsic modal functions and the high-frequency detail coefficients is accurately quantified by determining the joint probability distribution and the marginal probability distribution between each intrinsic modal function and each layer of high-frequency detail coefficients, and calculating the mutual information, which avoids the uncertainty caused by subjective judgment and experience estimation, and makes the correlation evaluation more objective and accurate; improve the quality of feature extraction: mutual information as an index to measure the correlation between two variables, can effectively filter out the part with strong correlation with high-frequency detail coefficients from the many intrinsic modal functions, by taking the intrinsic modal functions greater than the mutual information threshold as the high-frequency intrinsic modal functions, irrelevant or interference information can be removed, and the fault-related features are further highlighted, thereby improving the quality of fault feature extraction; enhance the accuracy of fault detection: since the selected high-frequency intrinsic modal functions have stronger correlation with high-frequency detail coefficients, they can better reflect the essential characteristics of the fault, therefore, in the subsequent construction of multi-dimensional feature vector and fault detection, more accurate and more relevant feature information can be used to improve the accuracy of fault detection and reduce the possibility of misjudgment and omission.

[0142] In a feasible implementation, the step 110 in the above embodiment of acquiring the target signal set comprises: acquiring three initial phase current signals, three initial phase voltage signals and an initial electric field intensity signal collected by an electromagnetic transformer, wherein the electromagnetic transformer has been installed on a distribution overhead line; performing Gaussian kernel denoising processing on each initial phase current signal, each initial phase voltage signal and the initial electric field intensity signal to obtain three phase current signals, three phase voltage signals and an electric field intensity signal; and taking the three phase current signals, the three phase voltage signals and the electric field intensity signal as elements to compose the target signal set.

[0143] In the embodiments of the present application, the quality of signal acquisition and the accuracy of fault detection are improved by optimizing the signal acquisition device and denoising processing.

[0144] It can be understood that the frequency band compatibility and the fast response: adopting the electromagnetic transformer as the signal acquisition device, different turns ratios can be designed for high-frequency and low-frequency signals respectively, frequency band compatibility is realized, at the same time, the sensor core component is constructed by using the nanocrystalline soft magnetic material, the hysteresis effect is reduced, and the response speed to the sudden fault is improved, which helps to more accurately capture the fault signal in the power distribution overhead line, especially the hidden fault signal, and provides a reliable data basis for subsequent fault detection; high sensitivity signal capture: integrated high sensitivity Hall element, auxiliary capture electric field change, the Hall element has high sensitivity to electric field change, which can more accurately reflect the change of electric field intensity in the power distribution overhead line, thereby helping to more accurately detect the fault; effective denoising processing: each initial phase current signal, each initial phase voltage signal and the initial electric field intensity signal are subjected to Gaussian kernel denoising processing, the Gaussian kernel denoising processing can effectively remove the noise interference in the signal, improve the signal-to-noise ratio of the signal, and make the subsequent fault feature extraction and detection more accurate and reliable; improve the accuracy of fault detection: through optimization of the signal acquisition device and denoising processing, three phase current signals, three phase voltage signals and electric field intensity signals with higher quality are obtained, which can more accurately reflect the actual operation state of the power distribution overhead line, and these signals are used as elements to form a target signal set, and subsequent fault detection processing is performed, which helps to improve the accuracy of fault detection and reduce the possibility of misjudgment and omission.

[0145] The application provides a fault detection device for a power distribution overhead line.

[0146] Please refer to Figure 2 , which is a schematic diagram of a fault detection device for a power distribution overhead line in the embodiment of the application. The device 210 comprises:

[0147] The acquisition module 211 is configured to acquire a target signal set, and the target signal set comprises three phase current signals, three phase voltage signals and an electric field intensity signal of the power distribution overhead line.

[0148] The first decomposition module 212 is configured to use wavelet transform to decompose the target signal to obtain a plurality of high-frequency detail coefficients and low-frequency approximation coefficients, wherein the target signal is any one signal in the target signal set.

[0149] The second decomposition module 213 is configured to use empirical mode decomposition to decompose the low-frequency approximation coefficient to obtain a plurality of intrinsic mode functions.

[0150] The determination module 214 is configured to determine mutual information between each intrinsic mode function and each layer of high-frequency detail coefficient, and take the intrinsic mode function corresponding to the mutual information greater than the mutual information threshold value as the high-frequency intrinsic mode function.

[0151] The constructing module 215 is configured to construct a multi-dimensional feature vector corresponding to the target signal according to all high-frequency intrinsic mode functions and all high-frequency detail coefficients.

[0152] The predicting module 216 is configured to input each multi-dimensional feature vector corresponding to each signal in the target signal set into a preset line fault prediction model to obtain a fault classification detection result.

[0153] In the embodiments of the present application, the related content of the above-mentioned acquisition module 211, the first decomposition module 212, the second decomposition module 213, the determining module 214, the constructing module 215 and the predicting module 216 can be referred to the content in the embodiments shown in the above-mentioned embodiments, and thus, the details are not described herein. Figure 1

[0154] It should be noted that the device 210 of the present application further includes other modules, and it can be understood that the method of the present application has a one-to-one correspondence with the device 210, and thus, the other modules of the device 210 of the present application are the corresponding content of the method of the present application in the above-mentioned embodiments.

[0155] In the embodiments of the present application, by fusing wavelet transform and empirical mode decomposition to process the signal, the fault features in the non-stationary signal can be more effectively extracted, so that the hidden faults in the overhead distribution line can be quickly and accurately detected, the intelligent operation of the power system is provided with reliable support, the risk of serious line fault caused by hidden faults is effectively reduced, and the stable operation of the power system is ensured.

[0156] ​In addition to the above-mentioned effects of quickly and accurately detecting hidden faults, reducing the risk of serious line faults, and ensuring the stable operation of the power system, the power distribution overhead line fault detection device provided by the application also has the following advantages: ensuring time sequence consistency: by configuring multiple high-speed acquisition channels in the signal acquisition device, simultaneous high-speed acquisition of signals between different channels is achieved, ensuring the time sequence consistency of the acquired signals, avoiding time offset errors caused by different signal times, and thus providing a reliable data foundation for subsequent accurate fault detection; flexible sampling configuration: the signal acquisition device supports the configuration of multiple sampling frequencies and multiple resolutions, and can be adjusted according to actual needs, for example, adjusting to a high-frequency mode for signal acquisition after detecting a fault, which can capture fault signal characteristics in more detail, and adjusting to a low-frequency or normal mode when there is no fault, which meets the acquisition needs in different scenarios and saves acquisition resources and storage space to a certain extent; fine separation of fault characteristics: using empirical mode decomposition to decompose the low-frequency approximation coefficients obtained by wavelet transform decomposition can finely separate and focus on the fault characteristics of hidden faults, and through this multi-level decomposition and combined processing method, more subtle and critical fault information can be extracted from complex signals, improving the extraction ability of hidden fault characteristics and helping to more accurately identify fault types and locations; precise screening of high-frequency intrinsic mode functions: by determining the mutual information between each intrinsic mode function and each layer of high-frequency detail coefficients, and taking the intrinsic mode function corresponding to the mutual information greater than the mutual information threshold as the high-frequency intrinsic mode function, this screening method can effectively find the part with stronger correlation with high-frequency detail coefficients from numerous intrinsic mode functions, further highlighting the features related to faults, removing irrelevant or interfering information, and improving the quality and effectiveness of the features; improve fault detection accuracy: use weight, skewness, kurtosis, and approximate entropy to construct a multi-dimensional feature vector, which can describe the characteristics of the data from different angles, accurately extract key fault features from complex hidden fault signals, and thus effectively improve the accuracy of subsequent model fault detection and reduce the possibility of misjudgment and omission. Select an advanced model architecture: select the CNN-LSTM-Transformer model as the initial line fault prediction model, which combines the advantages of convolutional neural network (CNN) in feature extraction, long short-term memory network (LSTM) in time series processing, and Transformer in capturing long-distance dependencies, can more effectively process information in the multi-dimensional feature vector, and further improve the accuracy and reliability of fault classification and detection.

[0157] The application also provides a computer readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the power distribution overhead line fault detection method in any of the above method embodiments.

[0158] The application also provides a computer device in a fourth aspect, comprising a memory and a processor, the memory stores a computer program, and the computer program is executed by the processor to make the processor execute the fault detection method of the distribution overhead line in the above-mentioned method embodiment.

[0159] Figure 3 The internal structure diagram of the computer device in some embodiments is shown. The computer device can be a terminal, a server, or a gateway. As shown in the figure, the computer device comprises a processor, a memory and a network interface connected through a system bus. Figure 3

[0160] The memory comprises a non-volatile storage medium and an internal memory. The non-volatile storage medium of the computer device stores an operating system and can also store a computer program, which, when executed by the processor, can make the processor implement each step in the above-mentioned method embodiment. The internal memory can also store a computer program, which, when executed by the processor, can make the processor execute each step in the above-mentioned method embodiment. Those skilled in the art can understand that the computer device shown in the figure is only a block diagram of part of the structure related to the scheme of the application, and does not constitute a limitation on the computer device to which the scheme of the application is applied. The specific computer device can comprise more or fewer components than those shown in the figure, or combine certain components, or have a different component arrangement. Figure 3

[0161] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be completed by a computer program instructing relevant hardware, and the program can be stored in a non-volatile computer readable storage medium. When the program is executed, it can include the processes of the above-mentioned method embodiments.

[0162] ​​Any reference to storage, memory, database or other medium herein can include non-volatile and / or volatile storage. Non-volatile storage can include read-only memory (ROM), programmable ROM (PROM), electronically programmable ROM (EPROM), or electrically erasable ROM (EEPROM). Volatile storage can include random-access memory (RAM). By way of illustration, and not limitation, RAM is available in many forms such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), Synchlink DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct Rambus dynamic RAM (DRDRAM), and Rambus dynamic RAM (RDRAM). The RAM can also include a basic-oxide-of-silicon (BOS) integrated circuit, or any other type of integrated circuit.

[0163] Any of the technical features of the above embodiments can be combined, and for brevity, not all possible combinations of the technical features of the above embodiments are described, however, any combination of the technical features should be considered as within the scope of the present disclosure, as long as the combination does not result in a contradiction.

[0164] The above embodiments merely express several implementation manners of the present application, and the description is relatively specific and detailed, but should not be understood as a limitation on the patent scope of the present application. It should be noted that, for those skilled in the art, several modifications and improvements can be made without departing from the concept of the present application, and these are within the protection scope of the present application. Therefore, the patent protection scope of the present application should be subject to the appended claims.

Claims

1. A method for detecting faults in a power distribution overhead line, characterized in that: The method comprises: Acquire a target signal set, wherein the target signal set includes three phase current signals, three phase voltage signals, and an electric field strength signal of a distribution overhead line; Decomposing the target signal using wavelet transform to obtain multiple layers of high-frequency detail coefficients and low-frequency approximation coefficients, wherein the target signal is any one signal in the target signal set; Decomposing the low-frequency approximation coefficients using empirical mode decomposition to obtain a plurality of intrinsic mode functions; Determine the mutual information between each intrinsic mode function and each layer's high-frequency detail coefficient, and take the intrinsic mode function corresponding to the mutual information greater than the mutual information threshold as the high-frequency intrinsic mode function; Constructing a multidimensional feature vector corresponding to the target signal based on all high-frequency intrinsic mode functions and all high-frequency detail coefficients; Inputting each multidimensional feature vector corresponding to each signal in the target signal set into a preset line fault prediction model to obtain a fault classification detection result; The method further comprises: Determine the current amplitude, voltage drop amplitude, harmonic content and fault time based on the three phase current signals and three phase voltage signals; Determine a fault type detection result based on the current amplitude, the voltage drop amplitude, the harmonic content, and the fault time using a preset fuzzy rule; A target fault classification detection result is determined according to the fault type detection result and the fault classification detection result.

2. The method according to claim 1, characterized in that The target signal is decomposed by using wavelet transform to obtain multiple layers of high-frequency detail coefficients and low-frequency approximation coefficients, including: The target signal is iteratively decomposed using the wavelet transform until the number of decomposition layers of the iterative decomposition is equal to the preset number of decomposition layers, and multiple layers of high-frequency detail coefficients corresponding to all iterative decompositions and the low-frequency approximation coefficients corresponding to the iterative decomposition are obtained.

3. The method according to claim 2, characterized in that The method further comprises: Using short-time Fourier transform, transforming the target signal to obtain a time-frequency spectrum; Performing frequency extraction on the time-frequency spectrum to obtain a signal dominant frequency; The preset number of decomposition layers is determined according to the dominant frequency of the signal.

4. The method according to claim 3, characterized in that The step of determining the preset number of decomposition layers according to the dominant frequency of the signal includes: Using the formula Determining the preset number of decomposition layers; in, is the preset decomposition level, is the floor symbol, is the logarithm to base 2, is the signal sampling frequency of the target signal, is the dominant frequency of the signal.

5. The method according to claim 1, wherein The low-frequency approximation coefficients are decomposed using empirical mode decomposition to obtain a plurality of intrinsic mode functions, including: Using the empirical mode decomposition, the low-frequency approximation coefficients are iteratively decomposed until the energy of the residual component obtained by the iterative decomposition is less than or equal to the energy threshold, and multiple initial intrinsic mode functions corresponding to all iterative decompositions are obtained, and the preset initial intrinsic mode functions are all used as intrinsic mode functions.

6. The method according to claim 5, characterized in that The method further comprises: determining the energy of the target signal; A preset percentage of the energy of the target signal is used as the energy threshold.

7. The method according to claim 1, characterized in that Before inputting each multidimensional feature vector corresponding to each signal in the target signal set into a preset line fault prediction model to obtain a fault classification detection result, the method further includes: Using short-time Fourier transform, transforming the target signal to obtain a time-frequency spectrum; Arrange all high-frequency detail coefficients in sequence to obtain a coefficient vector; Determine the energy of each eigenmode function and arrange the energies of all eigenmode functions in sequence to obtain an energy vector; Constructing a three-dimensional tensor corresponding to the target signal according to the time-frequency spectrum, the coefficient vector, and the energy vector; The step of inputting each multidimensional feature vector corresponding to each signal in the target signal set into a preset line fault prediction model to obtain a fault classification detection result includes: Each multidimensional feature vector and each three-dimensional tensor corresponding to each signal in the target signal set is input into the preset line fault prediction model to obtain the fault classification detection result.

8. The method according to claim 1, characterized in that Determining the mutual information between each intrinsic mode function and each layer of high-frequency detail coefficients includes: Determining a joint probability distribution between each intrinsic mode function and each layer of high-frequency detail coefficients, and determining a marginal probability distribution of each intrinsic mode function, and determining a marginal probability distribution of each layer of high-frequency detail coefficients; The mutual information between each eigenmode function and each layer of high-frequency detail coefficients is determined according to the joint probability distribution between each eigenmode function and each layer of high-frequency detail coefficients, the marginal probability distribution of each eigenmode function, and the marginal probability distribution of each layer of high-frequency detail coefficients.

9. The method according to claim 1, characterized in that The acquiring of the target signal set includes: Acquiring three initial phase current signals, three initial phase voltage signals, and an initial electric field strength signal collected by an electromagnetic mutual inductor, wherein the electromagnetic mutual inductor has been installed on the power distribution overhead line; Performing Gaussian kernel denoising processing on each initial phase current signal, each initial phase voltage signal and the initial electric field strength signal to obtain three phase current signals, three phase voltage signals and an electric field strength signal; The three phase current signals, the three phase voltage signals and the electric field strength signal are all taken as elements to form the target signal set.

Images