CN114609475A - Alternating current fault arc detection method and system - Google Patents

Alternating current fault arc detection method and system Download PDF

Info

Publication number
CN114609475A
CN114609475A CN202210277340.3A CN202210277340A CN114609475A CN 114609475 A CN114609475 A CN 114609475A CN 202210277340 A CN202210277340 A CN 202210277340A CN 114609475 A CN114609475 A CN 114609475A
Authority
CN
China
Prior art keywords
current
harmonic
calculating
neural network
fault arc
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202210277340.3A
Other languages
Chinese (zh)
Inventor
刘晓琳
荆涛
黄德顺
米哲
陆飞宇
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Civil Aviation University of China
Original Assignee
Civil Aviation University of China
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Civil Aviation University of China filed Critical Civil Aviation University of China
Priority to CN202210277340.3A priority Critical patent/CN114609475A/en
Publication of CN114609475A publication Critical patent/CN114609475A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
    • G01R31/08—Locating faults in cables, transmission lines, or networks
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R19/00—Arrangements for measuring currents or voltages or for indicating presence or sign thereof
    • G01R19/25—Arrangements for measuring currents or voltages or for indicating presence or sign thereof using digital measurement techniques
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Evolutionary Biology (AREA)
  • Data Mining & Analysis (AREA)
  • Medical Informatics (AREA)
  • Software Systems (AREA)
  • Computer Hardware Design (AREA)
  • Geometry (AREA)
  • Testing Of Short-Circuits, Discontinuities, Leakage, Or Incorrect Line Connections (AREA)

Abstract

The invention discloses an alternating current fault arc detection method and a system, comprising the following steps: acquiring discrete current signals in the aviation cable, decomposing the discrete current signals into various order modal components by adopting an empirical mode decomposition algorithm, reconstructing the current signals, dividing the reconstructed current signals according to periods, acquiring Hurst index, inter-harmonic variance and wavelet energy entropy three-dimensional characteristic data of each period, constructing a CNN-LSTM hybrid neural network, and inputting the Hurst index, inter-harmonic variance and wavelet energy entropy three-dimensional characteristics into the CNN-LSTM hybrid neural network to realize alternating current arc fault detection. The method can enhance the alternating current arc fault recognition capability of the aviation cable, improve the arc fault detection accuracy and robustness, and enable the arc fault recognition speed to be higher due to the simpler structure of the CNN-long and short term memory neural network LSTM.

Description

Alternating current fault arc detection method and system
Technical Field
The invention relates to the technical field of intelligent detection of fault arcs of aviation power distribution systems, in particular to an alternating current fault arc detection method and system.
Background
Aviation cables are a major component of electrical conductor interconnect systems, primarily for power transmission and signal transmission. Because the airplane works in complex environments such as vibration, friction, humidity, temperature change and the like for a long time, the cable is easy to crack and damage, the electric connection part of the cable is easy to loosen, and alternating-current fault electric arcs are easy to occur. When alternating current electric arcs occur, the temperature of the electric arc center can reach 5000-15000 ℃, if the electric arc center is not identified and detected in time and protective measures are taken, cable burning and even fire disasters are easily caused by fault electric arcs. Therefore, the method for identifying the alternating-current fault arc with high efficiency, simplicity and accuracy is researched and applied to an aviation power distribution system, the arc fault is identified in time, effective circuit protection is carried out, cable burning and fire hazard are prevented, and life and property safety of people is protected.
The traditional thermal protection circuit breaker and the solid-state power controller detect the arc by setting a threshold, and once the threshold is fixed, the traditional thermal protection circuit breaker and the solid-state power controller can detect and protect the circuit of the fault arc with the voltage and current amplitude exceeding the threshold; however, in a specific environment, the amplitude of voltage and current is lower than a fixed threshold value when a fault arc occurs due to looseness of cable connection or condensation of water drops between lines, so that the arc fault is difficult to effectively identify by simply adjusting the threshold value, and in addition, the load connected with the aviation cable is complex and can be switched at any time, the amplitude of the voltage and current can be changed.
Disclosure of Invention
The embodiment of the invention provides an alternating current fault arc detection method, which comprises the following steps:
acquiring a discrete current signal in an aviation cable;
decomposing the discrete current signal into modal components of each order by adopting an empirical mode decomposition algorithm and reconstructing the current signal;
dividing the reconstructed current signals according to periods, and acquiring three-dimensional characteristic data of the Hurst index, the inter-harmonic variance and the wavelet energy entropy of each period;
constructing a CNN-LSTM hybrid neural network;
and inputting the three-dimensional characteristics of the Hurst index, the inter-harmonic variance and the wavelet energy entropy into the CNN-LSTM hybrid neural network to realize alternating current arc fault detection.
Further, the step of selecting a proper modal component to reconstruct the current signal by comparing the normalized correlation coefficient of each order of modal component and the original discrete current signal with the preset threshold value comprises:
setting r for 1-initialized counter iPreset ofNumber of cycles imax；
Performing correlation operation on each order modal component and the original discrete current signal
Figure BDA0003553550570000021
Wherein IMFi represents each order modal component of the discrete current signal,
Figure BDA0003553550570000022
representing a discrete current signal;
comparison riAnd rPreset ofIf r isi≥rPreset ofThen r is preservediOtherwise, the counter i is incremented by one and r is recalculatediUntil i reaches the cycle number imax；
Adding the stored modal components to obtain a reconstructed current signal;
further, the Hurst index feature calculating step comprises the following steps:
dividing x (i) into M subsequences with length of n, and calculating the mean value of the subsequences
Figure BDA0003553550570000023
Wherein, x (i) is to divide the reconstructed current signal into a plurality of time sequences with the length of N according to periods;
mean value adjustment sequenceColumn Di,m＝x(i,m)-μn,m,i＝1,2,...,n；
Calculating the accumulated deviation of the sub-sequence
Figure BDA0003553550570000024
i＝1,2,...,n；
Calculating subsequence of sub-sequence range
Figure BDA0003553550570000025
Calculating the standard deviation of the subsequences
Figure BDA0003553550570000026
Calculating each rescaling range Rd/SdAnd average of all M subsequences rescaling range
Figure BDA0003553550570000031
Comparing N with N/10, if N>N/32, then by
Figure BDA0003553550570000032
Linear fitting is carried out to estimate the Hurst index, and the operation is finished, wherein R represents the polar difference of the subsequence, S represents the standard deviation of the subsequence, H represents the Hurst index, and C represents a constant;
otherwise, making n equal to n/2, and returning to continue the operation.
Further, the inter-harmonic variance feature calculation step includes:
calculating the discrete Fourier transform of the time series x (i)
Figure BDA0003553550570000033
Selecting X (k) intermediate harmonic components to form a new discrete Fourier transform set
Figure BDA0003553550570000034
Where β is the upper frequency space limit under consideration;
computing time series inter-harmonic variance
Figure BDA0003553550570000035
Wherein
Figure BDA0003553550570000036
Representing the time series inter-harmonic means.
Further, the wavelet energy entropy feature calculation step comprises the following steps:
computing the energy E of the ith sample point in the a-scale in the time series x (i)a(i)；
Calculating the total energy of the time series x (i) in the a-scale
Figure BDA0003553550570000037
Calculating the total energy of all scales of the time series x (i)
Figure BDA0003553550570000038
Calculating pa＝E(a)/E；
Calculating energy entropy of wavelet under scale a
Figure BDA0003553550570000039
Where m denotes the number of size layers under wavelet transform.
Further, the method also comprises the step of testing and training the CNN-LSTM hybrid neural network, and the steps comprise:
dividing a Hurst index, inter-harmonic variance and wavelet energy entropy three-dimensional characteristic data set into a training set and a testing set according to a ratio of 8: 2;
searching parameters such as an optimal initial learning rate, batch size and the like by adopting a grid search method, and determining the iteration times W of the CNN-LSTM hybrid neural network, wherein the initialization training times v is 1;
performing model training through a convolutional neural network CNN, and extracting characteristic information of three-dimensional characteristic data of a Hurst index, an inter-harmonic variance and a wavelet energy entropy;
learning characteristic information by using a long-short term memory neural network (LSTM);
updating parameters of the CNN-LSTM hybrid neural network by using error back propagation;
comparing the training times v with the size of W, and if v is larger than W, finishing the training;
otherwise, continuing the iterative training operation;
and inputting a test set, identifying the fault arc and outputting a detection result.
The invention also provides an alternating current fault arc detection system, comprising:
the signal conditioning module is used for acquiring current signals in the aviation cable through the current transformer, inputting the current signals into the signal conditioning module, acquiring and amplifying the current signals, and performing analog-to-digital conversion through the analog-to-digital converter to obtain discrete current signals;
the interference filtering module is used for decomposing the discrete current signals into various orders of modal components IMF through an empirical mode decomposition algorithm, then selecting modal component reconstruction current according to a normalized correlation coefficient and a preset coefficient threshold value, dividing the obtained reconstruction current signals according to cycles, and sequentially solving the Hurst index, the inter-harmonic variance and the wavelet energy entropy three-dimensional characteristic data of the current signals of each cycle;
and the fault arc identification module is used for inputting the Hurst index, the inter-harmonic variance and the wavelet energy entropy three-dimensional characteristic data into the fault identification module, and training and testing the data through the CNN-LSTM hybrid neural network to realize alternating current arc fault detection.
Further, still include:
an aviation power supply;
an AC fault arc interrupter;
a cable to be tested;
a current transformer;
an aviation load;
the aviation power supply is connected with one end of the alternating-current fault arc circuit breaker, the alternating-current fault arc circuit breaker is connected with the cable to be tested, one line of the cable to be tested penetrates through the current transformer, and the cable to be tested is connected with the aviation load.
Compared with the prior art, the embodiment of the invention provides an alternating-current fault arc detection method which has the following beneficial effects:
the aviation cable alternating current fault arc identification method and the aviation cable alternating current fault arc identification system have the beneficial effects that: the fault information is obtained from three dimensions of H-I-W, the alternating current arc fault identification capability of the aviation cable can be enhanced, the arc fault detection accuracy and robustness are improved, and the CNN-long short-term memory neural network LSTM is simpler in structure, so that the arc fault identification speed is higher, and effective basis is improved for the next circuit protection action.
Drawings
FIG. 1 is a flow chart of intelligent AC fault arc identification based on a CNN-long short-term memory neural network LSTM;
FIG. 2 is a schematic structural diagram of an aviation cable AC fault arc identification system employed in the method of the present invention;
FIG. 3 is a schematic diagram of the internal structure of the current transformer;
FIG. 4 is a flow diagram of interference filtering based on modal decomposition;
FIG. 5 is a flow chart of three-dimensional feature extraction from H-I-W;
FIG. 6 is a raw current waveform of an exemplary test circuit under linear load with an arc fault;
FIGS. 7(a) - (c) are three-dimensional characteristic waveforms of an exemplary test current signal H-I-W;
FIG. 8 is a training absolute error curve of an exemplary test current signal based on a CNN-LSTM network;
FIG. 9 is an AC arc fault detection accuracy curve of the method of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Referring to fig. 1 to 9, an embodiment of the present invention provides an ac fault arc detection method, including:
the aviation cable alternating current fault arc detection method provided by the invention comprises the following steps of:
step 1) constructing an aviation cable alternating current fault arc detection system; as shown in fig. 2, the fault detection system includes a system to be detected 1 and an upper computer 2; wherein: the system to be tested 1 comprises an aviation power supply 3, an alternating-current fault arc breaker 4, a cable to be tested 5, a current transformer 6 and an aviation load 7, wherein the aviation power supply 3 is connected with one end of the alternating-current fault arc breaker 4, the alternating-current fault arc breaker 4 is connected with the cable to be tested 5, one line of the cable to be tested 5 penetrates through the current transformer 6, and the cable to be tested 5 is connected with the aviation load 7; the upper computer 2 comprises a signal conditioning module 8, an interference filtering module 9 and a fault arc identification module 10, the output of the current transformer 6 is connected with the input of the signal conditioning module 8, the output of the signal conditioning module 8 is connected with the input end of the interference filtering module 9, the output of the interference filtering module 9 is connected with the input of the fault arc identification module 10, and the output of the fault arc identification module 10 is connected with one end of the alternating-current fault arc circuit breaker 4.
The aviation power supply 3 is an alternating current power supply of 115V/400Hz, the current transformer is an industrial-grade clamping type transformer, and a cable to be tested penetrates through the center of the current transformer.
And 2) acquiring a current signal in the aviation cable through the current transformer 6, wherein the current transformer 6 has an internal structure as shown in fig. 3, the cable penetrates through the current transformer 6, a sinusoidal alternating current signal flows through the cable to generate a variable magnetic field, when a clamping joint 201 of the current transformer 6 is closed, a coil 202 generates current, the current is input into the signal conditioning module 8 through an output end 203, the acquisition and amplification of the current signal are realized, and then analog-to-digital conversion is performed through an analog-to-digital converter to obtain a discrete current signal.
Step 3) a specific method of the interference filtering module 9 is as shown in fig. 4, the current signal is amplified by the signal conditioning module 8, then is input into the interference filtering module 9 to be decomposed by using an empirical mode decomposition algorithm, and a proper modal component is selected to reconstruct the current signal by comparing a normalized correlation coefficient of each modal component and the original signal with a preset threshold value, so as to realize interference filtering, and the specific process is as follows:
firstly, initializing i to 1, and setting a preset value r according to actual conditionsPreset ofNumber of cycles imax；
Performing empirical mode decomposition on the current signal to obtain a series of modal components;
thirdly, each order of modal component is related to the original signal
Figure BDA0003553550570000061
To obtain ri；
R isiAnd rPreset ofAnd (6) comparing. If ri≥rPreset ofThen r is preservediIn c, otherwise, the timer i is added with one, and the operation is continued until i reaches the cycle number imax；
And fifthly, adding the modal components in the c to obtain a reconstructed signal y.
Step 4) dividing the reconstructed current signal y obtained in the step 3 into a plurality of time sequences x (i) with the length of N according to periods, and sequentially extracting three-dimensional characteristics of the Hurst index, the intermittent wave equation and the wavelet energy entropy from the current signal of each period, wherein the flow is shown in FIG. 5, and the specific process is as follows:
hurst index characterization:
dividing x (i) into M subsequences with length of n, calculating average value of subsequences
Figure BDA0003553550570000071
② calculating a mean adjustment sequence Di,m＝x(i,m)-μn,m,i＝1,2,...,n；
Calculating the accumulated deviation of subsequence
Figure BDA0003553550570000072
i＝1,2,...,n；
Fourthly, calculating the sequence range of the subsequence
Figure BDA0003553550570000073
Calculating the standard deviation of the subsequence
Figure BDA0003553550570000074
Calculating each rescaling range Rd/SdAnd average of all M subsequences rescaling range
Figure BDA0003553550570000075
Comparing N with N/10, if N>N/32, then by
Figure BDA0003553550570000076
Estimating H by linear fitting, and finishing the operation; otherwise, let n equal to n/2, return to ② to continue the operation.
Inter-harmonic variance characteristics:
computing the discrete Fourier transform of the time series x (i)
Figure BDA0003553550570000077
Selecting X (k) intermediate harmonic component to form new discrete Fourier transform set
Figure BDA0003553550570000078
Where β is the upper frequency space limit under consideration;
calculating time series inter-harmonic variance
Figure BDA0003553550570000079
Wavelet energy entropy characteristics:
computing the energy E of the ith sampling point in the a scale in the time sequence x (i)a(i)；
Calculating the total energy of the time series x (i) under a scale
Figure BDA00035535505700000710
Calculating the total energy of all scales of the time series x (i)
Figure BDA00035535505700000711
Fourthly, calculating pa＝E(a)/E；
Fifthly, calculating the energy entropy of the wavelet under the scale a
Figure BDA0003553550570000081
Step 5) inputting the calculation result obtained in the step 4 into the arc fault recognition module 10, and training and testing the arc fault through a CNN-LSTM (convolutional neural network combined with long and short term memory neural network) hybrid neural network, so as to realize alternating current arc fault detection, wherein a detection flow chart is shown in FIG. 1, and the specific detection process is as follows:
dividing a characteristic data set into a training set and a test set according to a ratio of 8:2, and performing labeling processing on the data set;
searching parameters such as optimal initial learning rate, batch size and the like by adopting a grid search method, determining the algorithm iteration times W of the recognition module, and setting the initialization training times v to be 1;
carrying out model training through a convolutional neural network CNN, and extracting the characteristic information of the H-I-W three-dimensional characteristic data;
fourthly, learning characteristic information by using the long-term and short-term memory neural network LSTM;
utilizing error back propagation to update network parameters;
comparing the training times v with the size of W, finishing the training if v is larger than W, and returning to the third step of continuing the training operation if the iteration times is increased by one;
seventhly, inputting a test set, identifying fault electric arcs and outputting detection results;
inputting the detection result of the fault arc identification module 10 into the circuit breaker 4 to perform action control, and finishing the operation.
An example test current signal containing normal and arc faults as shown in fig. 6 is selected, H-I-W three-dimensional features are extracted by a method in step 4 and are shown in fig. 7(a) - (c), the H-I-W three-dimensional features are input into a CNN-LSTM hybrid neural network in step 5 to be trained and tested, the absolute error of network training is shown in fig. 8, and the test result is shown in fig. 9. The experimental result shows that the method can effectively, quickly and accurately identify and detect the alternating-current fault arc of the aviation cable, and the network structure is simpler and the control performance is better.
Although the embodiments of the present invention have been disclosed in the foregoing for illustrative purposes, those skilled in the art will appreciate that various modifications, additions and substitutions are possible, without departing from the scope and spirit of the invention as disclosed in the accompanying drawings.

Claims (8)

1. An ac fault arc detection method, comprising:
acquiring a discrete current signal in an aviation cable;
decomposing the discrete current signal into modal components of each order by adopting an empirical mode decomposition algorithm, and reconstructing the current signal;
dividing the reconstructed current signals according to periods, and acquiring three-dimensional characteristic data of the Hurst index, the inter-harmonic variance and the wavelet energy entropy of each period;
constructing a CNN-LSTM hybrid neural network;
and inputting the three-dimensional characteristics of the Hurst index, the inter-harmonic variance and the wavelet energy entropy into the CNN-LSTM hybrid neural network to realize alternating current arc fault detection.
2. An ac fault arc detection method as claimed in claim 1, wherein said step of reconstructing a current signal comprises:
initialization counter i ═ 1 and preset value rPreset ofNumber of cycles imax；
Performing correlation operation on each order of modal component of the discrete current signal decomposed by the empirical mode decomposition algorithm and the normalized correlation coefficient of the discrete current signal to obtain the normalized correlation coefficient ri：
Figure FDA0003553550560000011
Wherein IMFi represents each order modal component of the discrete current signal,
Figure FDA0003553550560000012
representing a discrete current signal;
comparison riAnd rPreset ofIf r isi≥rPreset ofThen r is preservediOtherwise, the counter i is incremented by one and r is recalculatediUntil i reaches the cycle number imax；
And adding the stored modal components to obtain a reconstructed current signal.
3. An ac fault arc detection method as claimed in claim 2, wherein said Hurst exponential characteristic calculation step comprises:
dividing x (i) into M subsequences with length of n, and calculating the mean value of the subsequences
Figure FDA0003553550560000013
Wherein, x (i) is to divide the reconstructed current signal into a plurality of time sequences with the length of N according to the period;
calculating subsequence mean adjustment sequence Di,m＝x(i,m)-μn,m,i＝1,2,...,n；
Calculating the accumulated deviation of the sub-sequence
Figure FDA0003553550560000021
Calculating subsequence of sub-sequences
Figure FDA0003553550560000022
Calculating the standard deviation of the subsequences
Figure FDA0003553550560000023
Calculating each rescaling range Rd/SdAnd average of all M subsequences rescaling range
Figure FDA0003553550560000024
Comparing N with N/10, if N>N/32, then by
Figure FDA0003553550560000025
Linear fitting is carried out to estimate Hurst exponential characteristics, and operation is finished, wherein R represents a subsequence range, S represents a standard deviation of a subsequence, H represents a Hurst exponent, and C represents a constant;
otherwise, making n equal to n/2, and returning to continue the operation.
4. An ac fault arc detection method as claimed in claim 3, wherein said inter-harmonic variance feature calculation step comprises:
calculating the discrete Fourier transform of the time series x (i)
Figure FDA0003553550560000026
Selecting X (k) intermediate harmonic components to form a new discrete Fourier transform set
Figure FDA0003553550560000027
Where β is the upper frequency space limit under consideration;
computing time series inter-harmonic variance
Figure FDA0003553550560000028
Wherein
Figure FDA0003553550560000029
Representing the time series inter-harmonic means.
5. An AC fault arc detection method as claimed in claim 3, wherein said wavelet energy entropy feature calculation step comprises:
computing the energy E of the ith sample point in the a-scale in the time series x (i)a(i)；
Calculating the total energy of the time series x (i) in the a-scale
Figure FDA00035535505600000210
Calculating the total energy of all scales of the time series x (i)
Figure FDA00035535505600000211
Calculating pa＝E(a)/E；
Calculating energy entropy of wavelet under scale a
Figure FDA0003553550560000031
Where m represents the number of size layers under wavelet transform.
6. The ac fault arc detection method of claim 1, further comprising testing and training a CNN-LSTM hybrid neural network, comprising the steps of:
dividing a Hurst index, inter-harmonic variance and wavelet energy entropy three-dimensional characteristic data set into a training set and a testing set according to a ratio of 8: 2;
searching for optimal initial learning rate and batch size parameters by adopting a grid search method, and determining the iteration times W of the CNN-LSTM hybrid neural network, wherein the initialization training times v is 1;
performing model training through a convolutional neural network CNN, and extracting characteristic information of three-dimensional characteristic data of a Hurst index, an inter-harmonic variance and a wavelet energy entropy;
learning characteristic information by using a long-short term memory neural network (LSTM);
updating parameters of the CNN-LSTM hybrid neural network by using error back propagation;
comparing the training times v with the size of W, and if v is larger than W, finishing the training;
otherwise, continuing the iterative training operation;
and inputting a test set, and identifying the fault arc.
7. An ac fault arc detection system, comprising:
the signal conditioning module is used for acquiring current signals in the aviation cable through the current transformer, inputting the current signals into the signal conditioning module, acquiring and amplifying the current signals, and performing analog-to-digital conversion through the analog-to-digital converter to obtain discrete current signals;
the interference filtering module is used for decomposing the current signals into modal components of each order through an empirical mode decomposition algorithm, then selecting modal component reconstruction currents according to the normalized correlation coefficient and the preset coefficient threshold value, dividing the obtained reconstruction current signals according to periods, and sequentially obtaining the Hurst index, the inter-harmonic variance and the wavelet energy entropy three-dimensional characteristic data of the current signals of each period;
and the fault arc identification module is used for inputting the Hurst index, the inter-harmonic variance and the wavelet energy entropy three-dimensional characteristic data into the fault identification module, and training and testing the data through the CNN-LSTM hybrid neural network to realize alternating current arc fault detection.
8. An aircraft cable ac fault arc detection system as claimed in claim 7, further comprising:
an aviation power supply;
an AC fault arc interrupter;
a cable to be tested;
a current transformer;
an aviation load;
the aviation power supply is connected with one end of the alternating-current fault arc circuit breaker, the alternating-current fault arc circuit breaker is connected with the cable to be tested, one line of the cable to be tested penetrates through the current transformer, and the cable to be tested is connected with the aviation load.
CN202210277340.3A 2022-03-18 2022-03-18 Alternating current fault arc detection method and system Pending CN114609475A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202210277340.3A CN114609475A (en) 2022-03-18 2022-03-18 Alternating current fault arc detection method and system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202210277340.3A CN114609475A (en) 2022-03-18 2022-03-18 Alternating current fault arc detection method and system

Publications (1)

Publication Number Publication Date
CN114609475A true CN114609475A (en) 2022-06-10

Family

ID=81865956

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202210277340.3A Pending CN114609475A (en) 2022-03-18 2022-03-18 Alternating current fault arc detection method and system

Country Status (1)

Country Link
CN (1) CN114609475A (en)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109425808A (en) * 2017-08-25 2019-03-05 福特全球技术公司 The self-adapting detecting of DC arc fault in vehicle high-voltage system based on Hurst index
CN115616364A (en) * 2022-12-16 2023-01-17 中国科学技术大学先进技术研究院 Fault arc detection method, device, equipment and storage medium
CN115905835A (en) * 2022-11-15 2023-04-04 国网四川省电力公司电力科学研究院 Low-voltage alternating current arc fault diagnosis method fusing multidimensional characteristics
CN116203309A (en) * 2022-11-18 2023-06-02 南方电网数字电网研究院有限公司 Fluxgate excitation signal processing method, device, server and storage medium

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109425808A (en) * 2017-08-25 2019-03-05 福特全球技术公司 The self-adapting detecting of DC arc fault in vehicle high-voltage system based on Hurst index
CN115905835A (en) * 2022-11-15 2023-04-04 国网四川省电力公司电力科学研究院 Low-voltage alternating current arc fault diagnosis method fusing multidimensional characteristics
CN115905835B (en) * 2022-11-15 2024-02-23 国网四川省电力公司电力科学研究院 Low-voltage alternating current arc fault diagnosis method integrating multidimensional features
CN116203309A (en) * 2022-11-18 2023-06-02 南方电网数字电网研究院有限公司 Fluxgate excitation signal processing method, device, server and storage medium
CN116203309B (en) * 2022-11-18 2023-12-12 南方电网数字电网研究院有限公司 Fluxgate excitation signal processing method, device, server and storage medium
CN115616364A (en) * 2022-12-16 2023-01-17 中国科学技术大学先进技术研究院 Fault arc detection method, device, equipment and storage medium

Similar Documents

Publication Publication Date Title
CN114609475A (en) Alternating current fault arc detection method and system
Jiang et al. Series arc fault detection based on random forest and deep neural network
CN106845010B (en) Low-frequency oscillation dominant mode identification method based on improved SVD noise reduction and Prony
Roy et al. Detection, classification, and estimation of fault location on an overhead transmission line using S-transform and neural network
CN110068759B (en) Fault type obtaining method and device
Xiao et al. High-impedance fault detection method based on one-dimensional variational prototyping-encoder for distribution networks
Zaro et al. Power quality detection and classification using S-transform and rule-based decision tree
CN111007364A (en) Method for identifying early self-recovery fault of cable
Ngaopitakkul et al. Combination of discrete wavelet transform and probabilistic neural network algorithm for detecting fault location on transmission system
CN114077846A (en) RF-LSTM-based fault current multi-domain identification method and storage medium
Panigrahi et al. A power quality detection and classification algorithm based on FDST and hyper-parameter tuned light-GBM using memetic firefly algorithm
CN113376551B (en) Method, system, device and storage medium for detecting series fault arc
CN117272143A (en) Power transmission line fault identification method and device based on gram angle field and residual error network
Yang et al. Identification and detection of dc arc fault in photovoltaic power generation system
Liu et al. Detection of AC arc faults of aviation cables based on HIW three-dimensional features and CNN-LSTM neural network
Yang et al. Fault Classification in Distribution Systems Using Deep Learning With Data Preprocessing Methods Based on Fast Dynamic Time Warping and Short-Time Fourier Transform
Li et al. Detection and localization of short-duration variations using sliding window SVD and sparse signal decomposition
Feng et al. Identification of faults using sparse optimization
CN113866614A (en) Multi-scenario user side low-voltage direct-current switch arc fault diagnosis method and device
CN113740671A (en) Fault arc identification method based on VMD and ELM
Chiradeja et al. Identification of fault types for single circuit transmission line using discrete wavelet transform and artificial neural networks
Aslan et al. ANN based fault location for medium voltage distribution lines with remote-end source
Alhaj et al. Power system harmonics estimation using sliding window based LMS
Gangwar et al. Detection and classification of faults on transmission line using time-frequency approach of current transients
CN117269660A (en) Fault arc detection method and system based on variation coefficient difference algorithm

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination