CN116908618A - Low-voltage distribution network alternating current series arc fault diagnosis method - Google Patents
Low-voltage distribution network alternating current series arc fault diagnosis method Download PDFInfo
- Publication number
- CN116908618A CN116908618A CN202310904639.1A CN202310904639A CN116908618A CN 116908618 A CN116908618 A CN 116908618A CN 202310904639 A CN202310904639 A CN 202310904639A CN 116908618 A CN116908618 A CN 116908618A
- Authority
- CN
- China
- Prior art keywords
- waveform
- load
- data
- value
- arc fault
- 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
Links
- 238000000034 method Methods 0.000 title claims abstract description 44
- 238000003745 diagnosis Methods 0.000 title claims abstract description 32
- 238000009826 distribution Methods 0.000 title claims abstract description 27
- 230000007246 mechanism Effects 0.000 claims abstract description 21
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 16
- 238000005070 sampling Methods 0.000 claims abstract description 14
- 230000004927 fusion Effects 0.000 claims abstract description 10
- 238000010801 machine learning Methods 0.000 claims abstract description 6
- 238000001228 spectrum Methods 0.000 claims abstract description 6
- 239000011159 matrix material Substances 0.000 claims description 39
- 230000006870 function Effects 0.000 claims description 19
- 238000012549 training Methods 0.000 claims description 12
- 230000004913 activation Effects 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 7
- 238000012545 processing Methods 0.000 claims description 7
- 238000011176 pooling Methods 0.000 claims description 6
- 230000008569 process Effects 0.000 claims description 5
- 238000010276 construction Methods 0.000 claims description 3
- 238000013075 data extraction Methods 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 230000000737 periodic effect Effects 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 2
- 238000004422 calculation algorithm Methods 0.000 abstract description 9
- 238000001514 detection method Methods 0.000 abstract description 7
- 238000005516 engineering process Methods 0.000 abstract description 4
- 238000010586 diagram Methods 0.000 description 11
- 238000004590 computer program Methods 0.000 description 7
- 239000000428 dust Substances 0.000 description 7
- 238000013528 artificial neural network Methods 0.000 description 5
- 230000004048 modification Effects 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 238000003860 storage Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 238000012795 verification Methods 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 230000004075 alteration Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000007635 classification algorithm Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000003062 neural network model Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000000191 radiation effect Effects 0.000 description 1
- 238000007637 random forest analysis Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 239000000126 substance Substances 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
- 230000001131 transforming effect Effects 0.000 description 1
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
- G01R31/088—Aspects of digital computing
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R23/00—Arrangements for measuring frequencies; Arrangements for analysing frequency spectra
- G01R23/16—Spectrum analysis; Fourier analysis
- G01R23/163—Spectrum analysis; Fourier analysis adapted for measuring in circuits having distributed constants
-
- 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
- G01R31/081—Locating faults in cables, transmission lines, or networks according to type of conductors
- G01R31/086—Locating faults in cables, transmission lines, or networks according to type of conductors in power transmission or distribution networks, i.e. with interconnected conductors
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- Mathematical Physics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Software Systems (AREA)
- Testing Electric Properties And Detecting Electric Faults (AREA)
Abstract
The application aims to provide a low-voltage distribution network alternating current series arc fault diagnosis method, which is used for collecting alternating current signals of two periods at an outlet of an ammeter at a certain sampling rate; the method comprises the steps of dividing a load waveform into a resistive load waveform, a dimmer load waveform and a complex type load waveform by using the spectrum amplitude and extremum weighted fuzzy entropy value difference of the normalized current waveform through a load waveform classification technology; for resistive load waveforms and dimmer load waveforms, five time-frequency features are extracted and arc fault diagnosis is performed using a conventional machine learning classifier. Aiming at complex load waveforms, a one-dimensional new data set is constructed by using data random fusion operation, and a convolutional neural network called a multichannel attention mechanism is trained to finish the identification of arc faults. Compared with the existing detection algorithm, the alternating current series arc fault diagnosis method provided by the application has the advantages of good adaptability and high accuracy.
Description
Technical Field
The application belongs to the technical field of arc fault diagnosis of low-voltage distribution networks, and particularly relates to a low-voltage distribution network alternating current series arc fault diagnosis method.
Background
With the economic development, the scale of the low-voltage distribution network is larger and larger, the possibility of electrical accidents caused by the rapid increase of distribution lines is also rapidly increased, and the stability of the distribution network is extremely challenged. The distribution network is an important component of the power network, and is used as a terminal for supplying power to users by the power system, and the running stability of the distribution network directly influences the reliability of power supply to the users. When the low-voltage distribution line is damaged or aged due to overlong service time or external force, the series-parallel arc is likely to be caused, and the fault arc is extremely easy to cause electric fire. It follows that the hazard caused by the fault arc is not visible. Moreover, when a fault arc occurs, the self impedance and radiation effect of the fault arc bring about harmonic interference and influence the quality of electric energy, damage electric equipment and influence the power supply economy.
At present, the existing low-voltage distribution network alternating current series arc fault diagnosis algorithm is various and can be mainly summarized into a time-frequency domain detection method and a neural network diagnosis method. The time-frequency domain detection is used for completing identification of alternating current series arc faults by extracting characteristics of current signals in time-frequency domain and utilizing a classifier or a threshold method. The neural network diagnosis method uses one-dimensional waveform of data or converts the data into images to be input into various networks for training and testing. Under the influence of different loads, fault arc current waveforms are various, and similarity exists between the normal and fault waveforms of various loads, so that the identification accuracy of the current algorithm is low, and the robustness is poor. In addition, because of uncertainty of neural network feature learning, one-dimensional current data is input into the network for training, redundant information is learned or normal and fault features are submerged, so that the diagnosis effect of the neural network detection algorithm is poor.
Disclosure of Invention
Therefore, aiming at the defects and shortcomings in the prior art, the application provides a novel method for diagnosing the AC series arc faults of the low-voltage distribution network, in particular to a series arc fault detection algorithm with two-stage classification. The algorithm has the advantages that twelve load waveforms are classified through a classification strategy to weaken the similarity degree of waveforms among loads, so that the number of misjudgment samples caused by waveform similarity is reduced, and the arc fault diagnosis accuracy is improved. Meanwhile, a data random fusion mechanism is provided to enhance the original waveform characteristics and a neural network with an attention mechanism is constructed for diagnosis.
The method comprises the steps of collecting alternating current signals of two periods at an outlet of an ammeter at a certain sampling rate; the method comprises the steps of dividing a load waveform into a resistive load waveform, a dimmer load waveform and a complex type load waveform by using the spectrum amplitude and extremum weighted fuzzy entropy value difference of the normalized current waveform through a load waveform classification technology; for resistive and dimmer load waveforms, five time-frequency features are extracted and arc fault diagnosis is performed using a conventional machine-learned classifier. Aiming at complex load waveforms, a one-dimensional new data set is constructed by using data random fusion operation, and a convolutional neural network called a multichannel attention mechanism is trained to finish the identification of arc faults. Compared with the existing detection algorithm, the alternating current series arc fault diagnosis method provided by the application has the advantages of good adaptability and high accuracy.
The technical scheme adopted for solving the technical problems is as follows:
the low-voltage distribution network alternating current series arc fault diagnosis method is characterized in that the load waveform is divided into a resistive load waveform, a dimmer load waveform and a complex type load waveform by utilizing the spectrum amplitude and extremum weighted fuzzy entropy value difference of the normalized current waveform; aiming at resistive and dimmer load waveforms, extracting various time-frequency characteristics and performing arc fault diagnosis by using a machine learning classifier; aiming at complex load waveforms, a one-dimensional new data set is constructed by using data random fusion operation, and a multichannel attention mechanism convolutional neural network is trained to finish the identification of arc faults.
Further, the method for distinguishing the load waveform into the resistive load waveform, the dimmer load waveform and the complex class load waveform by utilizing the difference of the frequency spectrum amplitude and the extremum weighted fuzzy entropy value of the normalized current waveform specifically comprises the following steps:
collecting alternating current signals X (i) of two periods at the outlet of the ammeter at a certain sampling rate, and carrying out normalization processing according to a formula (1);
wherein X is L For normalized waveform sequence, n is the number of sampling points in two periods, X max 、X min Respectively a maximum value and a minimum value in the two periodic current signals;
acquiring normalized waveform, calculating fundamental wave amplitude fc by Fourier transform, and judging if fc is larger than beta 0 Then the waveform is a resistive load; wherein beta is 0 Is the threshold value of the classified resistive load waveform, beta 0 ＝0.95；
If the waveform is not the resistive load waveform, the extremum weighted fuzzy entropy F is calculated ew The value, judge whether this load waveform is dimmer load waveform, the concrete calculation steps are as follows:
1) Calculating a difference sequence of two-period current waveform samples according to the formula (2);
X(n)＝abs(X L (i+1)-X L (i)),i＝1,2,3···n-1 (2)
wherein X (n) is the obtained difference sequence, and abs ()'s is the absolute value operation;
2) Taking four largest peaks in the difference sequence X (n) and calculating P according to the formula (3) av ；
Wherein M (1), M (2), M (3), M (4) are respectively four maximum peaks;
3) Calculating a fuzzy entropy value FE;
4) To calculate P av Obtaining extremum weighted fuzzy entropy F by dividing FE value ew A value represented by formula (4);
if F is calculated ew A value less than beta 1 The waveform is the current waveform of the dimmer load, beta 1 To be able to separate the classification threshold of the dimmer load waveform, and beta 1 ＝0.38；
If fc is less than beta 0 And F ew Greater than beta 1 The load waveform is classified into a complex load waveform pool.
1. The method for diagnosing an arc fault in ac series connection in a low voltage distribution network according to claim 2, wherein the method comprises the steps of:
the specific steps for calculating the fuzzy entropy value are as follows:
the current waveform { X (i): 1< i < N } with length N is constructed into m-dimensional vector according to the formula (5), namely
Y i m ＝{X(i),X(i+1),···,X(i+m-1)}-X 0 (i) (5)
Wherein N is the number of samples, wherein i is more than or equal to 1 and less than or equal to N-m+1;
obtaining sample Y according to (6) i m And sample Y j m Similarity is D ij m
Wherein d ij m The maximum absolute difference value is 1-j, j-N-m and i is not equal to j;
defining a function according to (7)
In the formula, the similarity tolerance r and the membership function
Construction of an m+1-dimensional vector according to (8)
Obtaining fuzzy entropy with sampling point N according to (9)
FE(m,n,r,N)＝lnφ m (n,r)-lnφ m+1 (n,r) (9)。
Further, five time-frequency domain features, respectively pulse factor C, are extracted for resistive and dimmer load waveforms if Margin factor C mf Kurtosis C k Relative content of fundamental wave F 1 Third harmonic relative content ratio F 3 The formula is as follows:
wherein X is L For current sample data, n is the signal length, fc is the amplitude of the fundamental component after fourier transform, 3rdhc is the amplitude of the third harmonic component after fourier transform, 5rdhc is the amplitude of the fifth harmonic component after fourier transform,
to make arc fault diagnosis using a machine learning classifier.
Further, aiming at complex load waveforms, constructing a one-dimensional new data set by using data random fusion operation and training a multichannel attention mechanism convolutional neural network to finish the identification of arc faults;
the specific steps of constructing a one-dimensional new data set are as follows:
1) To preserve the utilization of the overall information, two data point samples of the cycle are selected as an overall waveform;
2) Dividing local waveform information, dividing an integral sample into four parts, wherein each part comprises half-period waveform data, and each part of local waveform comprises a wave crest/wave trough or zero rest part;
3) And constructing one-dimensional new data, randomly selecting two pieces of local information, splicing the two pieces of local information on the whole waveform, and constructing the one-dimensional new data.
Further, the multi-channel attention mechanism convolutional neural network is a multi-channel convolutional neural network which integrates a multi-head attention mechanism and a GRU unit:
three data extraction channels are constructed by the network, and different structural parameters are respectively set for extracting features of data with different lengths so as to ensure that each data segment can extract an optimal feature group; the first channel is used for extracting features of the whole waveform of two periods, a multi-head attention module is embedded in the beginning part of the network, noise and redundant data in the data are eliminated by using an attention mechanism, and the features are focused on, so that the robustness of the model is improved.
Further, the calculation process of the multi-head attention mechanism is as follows:
1) Dividing an input current waveform into a plurality of sections in parallel;
2) Generating three corresponding matrixes for each section of input data, wherein the three corresponding matrixes are respectively a query matrix, a key matrix and a value matrix;
3) Obtaining the weight fraction of the data by utilizing the dot product of the query matrix and the key matrix;
4) The weight fraction is normalized by softmax and multiplied by a value matrix to obtain the final data with weight attention;
Q＝AW Q (16)
K＝AW K (17)
V＝AW V (18)
wherein Q, K, V represent a query matrix, a key matrix and a value matrix, d, respectively k The dimensions of Q, K, V, A being the input matrix, W Q 、W K 、W V Is a coefficient matrix obtained by performing linear transformation on an input matrix;
5) Combining the learned multiple sections of attention to obtain global attention;
after passing through the attention layer, extracting features through the convolution layer, the pooling layer, the BN layer and the output layer, and connecting a GRU unit at the same time to prevent fitting in training; finally, a characteristic vector is output through a BN layer and a Droutput layer and through a relu activation function;
the second and third channels are used for extracting features from the half-period local waveform segments; directly passing through a convolution layer, a pooling layer, a BN layer and a Droutput layer, connecting with a GRU unit, and finally outputting a feature vector through a relu activation function; the Droutput layer is used to help the network prevent overfitting, and the activation function of the convolution layer uses the leakyrelu.
Further, when the sampling rate is changed, the number of the integral waveform sample points and the number of the local waveform sample points are changed, the number and the size of the convolution layers in the three channels are compared and adjusted according to the number of the input data points, and the parameters of the network are obtained through training.
Compared with the prior art, the application and the preferred scheme thereof have at least the following outstanding advantages:
(1) A rapid and efficient load waveform classification algorithm is innovatively designed, and the primary screening of load types of the waveforms with the series arc faults is realized. The technology effectively reduces the calculation stress of the follow-up neural network model, reduces the number of misjudgment samples caused by waveform similarity among loads, and is beneficial to improving the overall diagnosis accuracy.
(2) A data random fusion mechanism is innovatively proposed for arc fault diagnosis. And extracting the local waveform on the basis of using the one-dimensional original data integral waveform, and fusing the local waveform with integral information to construct a one-dimensional new data set. The mechanism enhances the utilization of waveform information, effectively amplifies weak characteristics of arc faults, and improves the recognition accuracy of a back-end detection algorithm.
(3) The Convolutional Neural Network (CNN) is innovatively improved, a multi-head attention module and GRU units are fused into the CNN, and a plurality of feature extraction channels are constructed for the feature extraction of the self-adaption of the expansion of the whole waveforms and the local waveforms with different lengths respectively, so that an MC-MGCNN network is provided. The network has the ability to learn multiple load waveforms efficiently and identify whether an arc fault has occurred.
Drawings
The application is described in further detail below with reference to the attached drawings and detailed description:
FIG. 1 is an overall flow chart of an algorithm for diagnosing an AC series arc fault of a low voltage distribution network according to an embodiment of the present application;
FIG. 2 is a flow chart of a first level classification technique according to an embodiment of the application;
FIG. 3 is a graph of a data random fusion operation in accordance with an embodiment of the present application;
FIG. 4 is a block diagram of an MC-MGCNN network in accordance with an embodiment of the present application;
FIG. 5 is a flow chart of MC-MGNN network parameter training in accordance with an embodiment of the present application;
FIG. 6 is a flow chart of a second level classification technique S3 according to an embodiment of the application;
FIG. 7 is a resistive load waveform classification threshold β according to an embodiment of the present application 0 A schematic diagram;
FIG. 8 is a dimmer load waveform classification threshold β according to an embodiment of the present application 1 A schematic diagram;
FIG. 9 is a schematic diagram of an confusion matrix according to an embodiment of the application.
Detailed Description
In order to make the features and advantages of the present patent more comprehensible, embodiments accompanied with figures are described in detail below:
it should be noted that the following detailed description is illustrative and is intended to provide further explanation of the application. Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs.
It is noted that the terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of exemplary embodiments according to the present application. As used herein, the singular is also intended to include the plural unless the context clearly indicates otherwise, and furthermore, it is to be understood that the terms "comprises" and/or "comprising" when used in this specification are taken to specify the presence of stated features, steps, operations, devices, components, and/or combinations thereof.
The following description is made of the design and specific use steps of the proposed low-voltage distribution network ac series arc fault diagnosis method by a specific embodiment:
the load types to which the proposed technology is applicable include resistive loads, motor-type loads, power-type loads, gas discharge lamp-type loads, dimmer loads, and hybrid-type loads as described above. The overall flow of the construction and implementation of the proposed technique is shown in fig. 1, firstly, serial arc fault data are acquired according to step S1, then, the serial arc fault data are classified into a first-stage load waveform by step S2, the first-stage load waveform is divided into a simple load waveform and a complex load waveform, and then, a second-stage fault diagnosis classification is performed by step S3.
(1) S1: collecting alternating current signals X (i) of two periods at the outlet of the ammeter at a certain sampling rate, and carrying out normalization processing according to a formula (1);
wherein X is L For normalized waveform sequence, n is the number of sampling points in two periods, X max 、X min Respectively, the maximum and minimum of the two periodic current signals.
(2) First-stage classification S2:
step S21: acquiring the normalized waveform in S1, calculating fundamental wave amplitude fc by using Fourier transform, and judging if fc is larger than beta 0 Then it is a waveform of a resistive load. Wherein beta is 0 Is the threshold value of the classified resistive load waveform, beta 0 ＝0.95；
Step S22: if the waveform is not the resistive load waveform, the extremum weighted fuzzy entropy F is calculated ew The value, judge whether this load waveform is dimmer load waveform, the concrete calculation steps are as follows:
1) Calculating a difference sequence of two-period current waveform samples according to the formula (2);
X(n)＝abs(X L (i+1)-X L (i)),i＝1,2,3···n-1 (2)
wherein X (n) is the obtained difference sequence, and abs ()'s is the absolute value operation;
2) Taking four largest peaks in the difference sequence X (n) and calculating P according to the formula (3) av ；
Wherein M (1), M (2), M (3), M (4) are respectively four maximum peaks;
3) Calculating Fuzzy Entropy (FE) values, specifically comprising the following steps:
(1) the current waveform { X (i): 1< i < N } with length N is constructed into m-dimensional vector according to the formula (4), namely
Y i m ＝{X(i),X(i+1),···,X(i+m-1)}-X 0 (i) (4)
Wherein N is the number of samples, wherein i is more than or equal to 1 and less than or equal to N-m+1;
(2) obtaining sample Y according to (5) i m And sample Y j m Similarity is D ij m
Wherein d ij m The maximum absolute difference value is 1-j, j-N-m and i is not equal to j;
(3) defining a function according to (6)
In the formula, the similarity tolerance r and the membership function
(4) Repeating the above steps to construct m+1-dimensional vector according to formula (7)
(5) Obtaining fuzzy entropy with sampling point N according to (8)
FE(m,n,r,N)＝lnφ m (n,r)-lnφ m+1 (n,r) (8)
4) To calculate P av Obtaining extremum weighted fuzzy entropy F by dividing FE value ew A value as shown in formula (9);
if F is calculated ew A value less than beta 1 The waveform is a dimmerCurrent waveform, beta, of load 1 To be able to separate the classification threshold of the dimmer load waveform, and beta 1 ＝0.38。
Step S23: if fc is less than beta 0 And F ew Greater than beta 1 The load waveform is then classified into a pool of complex load waveforms, the resistive and dimmer load waveforms being referred to as simple load waveforms.
(3) Second stage classification S3:
step S31: acquiring a load waveform classification result of S2;
step S32: five time-frequency domain features, namely pulse factor C, are extracted for a simple load waveform if Margin factor C mf Kurtosis C k Relative content of fundamental wave F 1 Third harmonic relative content ratio F 3 The formula is as follows:
wherein X is L For current sample data, n is the signal length, fc is the amplitude of the fundamental component after fourier transform, 3rdhc is the amplitude of the third harmonic component after fourier transform, 5rdhc is the amplitude of the fifth harmonic component after fourier transform,
step S33: arc fault diagnosis is carried out for simple load waveforms by using conventional machine learning classifiers including but not limited to random forests, SVMs, XGBoost and the like;
step S34: the data random fusion operation is carried out on complex load waveforms, a one-dimensional new data set is constructed, and the specific steps are as follows:
1) Reserving the utilization of integral information, and selecting two data point samples with the cycle number as an integral waveform;
2) Dividing local waveform information, dividing an integral sample into four parts, wherein each part contains half-period waveform data, and each part of local waveform contains a wave crest/wave trough or zero rest position;
3) And constructing one-dimensional new data, randomly selecting two pieces of local information, splicing the two pieces of local information on the whole waveform, and constructing the one-dimensional new data.
The overall operation diagram is shown in fig. 3.
Step S35: aiming at the load waveform of a complex load pool, a multichannel convolutional neural network (MC-MGCNN network) which is called as a multi-head attention mechanism and GRU unit is designed based on the Convolutional Neural Network (CNN), and the specific improvement content of the network is as follows:
the network constructs three data extraction channels, and different structural parameters are respectively set for extracting features of data with different lengths, so that each data segment can be ensured to extract an optimal feature group. The first channel is used for extracting features of the whole waveform of two periods, a multi-head attention module is embedded in the beginning part of the network, noise and redundant data in data are eliminated by using an attention mechanism, the features are focused, and the robustness of the model is improved. The calculation process of the multi-head attention mechanism is as follows:
1) Dividing an input current waveform into a plurality of sections in parallel;
2) Generating three corresponding matrixes for each piece of input data, namely a Query matrix (Query), a Key matrix (Key) and a Value matrix (Value);
3) Obtaining the weight fraction of the data by utilizing the dot product of the query matrix and the key matrix;
4) The weight fraction is normalized by softmax and multiplied by a value matrix to obtain the final data with weight attention;
Q＝AW Q (16)
K＝AW K (17)
V＝AW V (18)
wherein Q, K, V represent a query matrix, a key matrix and a value matrix, d, respectively k The dimensions of Q, K, V, A being the input matrix, W Q 、W K 、W V Is a coefficient matrix obtained by linearly transforming an input matrix.
5) The learned multiple segments of attention are combined to obtain a global attention.
After passing through the attention layer, the characteristics are extracted again through the convolution layer, the pooling layer, the BN layer and the output layer, and meanwhile, a GRU unit is connected, so that the fitting in training is prevented. And finally, outputting the feature vector through a BN layer and a Droutput layer and through a relu activation function.
The second and third channels are used to extract features from the half-period local waveform segments. The characteristic vector is directly output through a convolution layer, a pooling layer, a BN layer and a Droutput layer and is connected with a GRU unit, and finally through a relu activation function. The Droutput layer can help the network prevent overfitting, and the activation function of the convolution layer adopts the leakyrelu. The overall structure of the MC-MGCNN network is shown in FIG. 4, and the specific network structure is shown in Table 1.
When the sampling rate is changed, the number of the integral waveform sample points and the number of the local waveform sample points are changed, the number and the size of the convolution layers in the three channels can be properly compared and adjusted according to the number of the input data points, parameters of the MC-MGCNN network are obtained through training, and the training process is shown in fig. 5:
TABLE 1MC-MGCNN network architecture
Step S36: and diagnosing the waveform of the complex load pool by using the MC-MGCNN network.
An example analysis is provided below to further describe the inventive arrangements:
in this example, the obtained data samples are all from step S1. The case analysis selects 11 loads including 9 single loads and 2 mixed types of loads, respectively. The specific parameters are shown in table 2, where the dimmer load is divided into a small angle (dimming angle 60 °) and a large angle (dimming angle 300 °). Each load collects 900 groups under normal and fault conditions, and a total of 21600 groups of data samples are collected.
Table 2 description of experimental load
The current transformer CPL8100A clamps on a distribution line at the outlet of the ammeter and is matched with a DSOX4024A oscilloscope to collect current signals. First, a threshold discussion of the first stage load waveform classification is made, and the resistive load waveform classification threshold is selected as shown in FIG. 7, when β 0 When the current waveform is set to 0.95, the current waveforms of the resistive load and other loads can be separated to the greatest extent. After the resistive load waveforms are separated, continue to calculate F ew The values classify the dimmer load waveform from the remaining few single load waveforms. As shown in fig. 8, when beta 1 When set to 0.38, the dimmer load waveform is most clearly bounded by other load waveforms.
The test is carried out by adopting 1500 groups of samples of the thermos and the incandescent lamp, the final accuracy is 100%, and the current waveforms of the thermos and the incandescent lamp are correctly classified as resistive load waveforms. And selecting 1500 groups of waveform samples of two angles of the dimmer load for testing, and finally correctly classifying 1495 groups, wherein the accuracy is 99.7%, and successfully judging the load waveform as the dimmer load waveform. In order to continue to verify the effectiveness of the load waveform classification, two combined load types (a switching power supply parallel dust collector and an incandescent lamp connected in parallel with a fluorescent lamp) are extracted for verification, and finally, the two load waveforms are successfully classified into a complex load waveform pool, so that the accuracy is 100%.
Next, the effect of the second level diagnostic categorization is validated. Because of different load types, the diagnosis and verification are carried out on the resistive load waveform and the dimmer load waveform, and an XGBoost classifier is selected for verification in the scheme, wherein the diagnosis accuracy of the resistive load waveform is 100%, and the diagnosis accuracy of the dimmer load waveform is 99.9%. Meanwhile, the diagnosis time of both is about 33.4 ms. The results show that the classified load waveforms can diagnose whether arc faults occur or not under the condition of high speed and high accuracy.
Next, the MC-mgcn network for complex load waveform pool diagnosis is validated to obtain the confusion matrix as shown in fig. 9. The final multi-classification accuracy is 99.02% as can be obtained from the confusion matrix, but if only the two states of normal classification and failure are considered, the final accuracy can reach 99.9%. The model can recognize the normal condition of various load waveforms in percentage, but under the fault condition, 8 samples of the combined load waveform of the parallel switch power supply of the dust collector are misjudged as the load waveform of the dust collector, and 3 samples of the load waveform of the dust collector are misjudged as the combined load waveform of the parallel switch power supply of the dust collector. The arc current of the combined load is formed by adding the branch currents of the two loads, is influenced by the load arc characteristic of the dust collector, the arc current of the combined load of the parallel switch power supply of the dust collector shows a waveform similar to a triangular wave, and has larger burrs and distortion at zero rest positions, so that the waveforms of the two loads have certain similarity, and the model cannot accurately identify the waveform type of the load. Finally, the diagnostic time of this model is around 50.1 ms.
It will be appreciated by those skilled in the art that embodiments of the present application may be provided as a method, system, or computer program product. Accordingly, the present application may take the form of an entirely hardware embodiment, an entirely software embodiment or an embodiment combining software and hardware aspects. Furthermore, the present application may take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, and the like) having computer-usable program code embodied therein.
The present application is described with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the application. It will be understood that each flow and/or block of the flowchart illustrations and/or block diagrams, and combinations of flows and/or blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instruction means which implement the function specified in the flowchart flow or flows and/or block diagram block or blocks.
These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the functions specified in the flowchart flow or flows and/or block diagram block or blocks.
The above description is only a preferred embodiment of the present application, and is not intended to limit the application in any way, and any person skilled in the art may make modifications or alterations to the disclosed technical content to the equivalent embodiments. However, any simple modification, equivalent variation and variation of the above embodiments according to the technical substance of the present application still fall within the protection scope of the technical solution of the present application.
The present application is not limited to the above-mentioned best mode, any person can obtain other various forms of diagnosis method for arc fault in ac series connection of low-voltage distribution network under the teaching of the present application, and all equivalent changes and modifications made according to the scope of the present application should be covered by the present application.
Claims (8)
1. The low-voltage distribution network alternating current series arc fault diagnosis method is characterized in that the load waveform is divided into a resistive load waveform, a dimmer load waveform and a complex type load waveform by utilizing the spectrum amplitude and extremum weighted fuzzy entropy value difference of the normalized current waveform; aiming at the resistive load waveform and the dimmer load waveform, extracting various time-frequency characteristics and performing arc fault diagnosis by using a machine learning classifier; aiming at complex load waveforms, a one-dimensional new data set is constructed by using data random fusion operation, and a multichannel attention mechanism convolutional neural network is trained to finish the identification of arc faults.
2. The method for diagnosing an arc fault in ac series connection of a low voltage distribution network according to claim 1, wherein the method comprises the steps of:
the load waveform is divided into a resistive load waveform, a dimmer load waveform and a complex type load waveform by utilizing the difference of the frequency spectrum amplitude and the extremum weighted fuzzy entropy value of the normalized current waveform, and the method specifically comprises the following steps:
collecting alternating current signals X (i) of two periods at the outlet of the ammeter at a certain sampling rate, and carrying out normalization processing according to a formula (1);
wherein X is L Is normalized toWaveform sequence, n is the number of sampling points of two periods, X max 、X min Respectively a maximum value and a minimum value in the two periodic current signals;
acquiring normalized waveform, calculating fundamental wave amplitude fc by Fourier transform, and judging if fc is larger than beta 0 Then the waveform is a resistive load; wherein beta is 0 Is the threshold value of the classified resistive load waveform, beta 0 ＝0.95；
If the waveform is not the resistive load waveform, the extremum weighted fuzzy entropy F is calculated ew The value, judge whether this load waveform is dimmer load waveform, the concrete calculation steps are as follows:
1) Calculating a difference sequence of two-period current waveform samples according to the formula (2);
X(n)＝abs(X L (i+1)-X L (i)),i＝1,2,3···n-1 (2)
wherein X (n) is the obtained difference sequence, and abs ()'s is the absolute value operation;
2) Taking four largest peaks in the difference sequence X (n) and calculating P according to the formula (3) av ；
Wherein M (1), M (2), M (3), M (4) are respectively four maximum peaks;
3) Calculating a fuzzy entropy value FE;
4) To calculate P av Obtaining extremum weighted fuzzy entropy F by dividing FE value ew A value represented by formula (4);
if F is calculated ew A value less than beta 1 The waveform is the current waveform of the dimmer load, beta 1 To be able to separate the classification threshold of the dimmer load waveform, and beta 1 ＝0.38；
If fc is less than beta 0 And F ew Greater than beta 1 The load waveform is classified into a complex load waveform pool.
3. The method for diagnosing an arc fault in ac series connection in a low voltage distribution network according to claim 2, wherein the method comprises the steps of:
the specific steps for calculating the fuzzy entropy value are as follows:
the current waveform { X (i): 1< i < N } with length N is constructed into m-dimensional vector according to the formula (5), namely
Y i m ＝{X(i),X(i+1),···,X(i+m-1)}-X 0 (i) (5)
Wherein N is the number of samples, wherein i is more than or equal to 1 and less than or equal to N-m+1;
obtaining sample Y according to (6) i m And sample Y j m Similarity is D ij m
Wherein d ij m The maximum absolute difference value is 1-j, j-N-m and i is not equal to j;
defining a function according to (7)
In the formula, the similarity tolerance r and the membership function
Construction of an m+1-dimensional vector according to (8)
Obtaining fuzzy entropy with sampling point N according to (9)
FE(m,n,r,N)＝lnφ m (n,r)-lnφ m+1 (n,r) (9)。
4. A method for diagnosing an ac series arc fault in a low voltage distribution network according to claim 3, wherein:
five time-frequency domain features, respectively pulse factor C, are extracted for resistive and dimmer load waveforms if Margin factor C mf Kurtosis C k Relative content of fundamental wave F 1 Third harmonic relative content ratio F 3 The formula is as follows:
wherein X is L For current sample data, n is the signal length, fc is the amplitude of the fundamental component after fourier transform, 3rdhc is the amplitude of the third harmonic component after fourier transform, 5rdhc is the amplitude of the fifth harmonic component after fourier transform,
to make arc fault diagnosis using a machine learning classifier.
5. The method for diagnosing an arc fault in ac series connection in a low voltage distribution network according to claim 4, wherein the method comprises the steps of:
aiming at complex load waveforms, constructing a one-dimensional new data set by using data random fusion operation and training a multichannel attention mechanism convolutional neural network to finish the identification of arc faults;
the specific steps of constructing a one-dimensional new data set are as follows:
1) To preserve the utilization of the overall information, two data point samples of the cycle are selected as an overall waveform;
2) Dividing local waveform information, dividing an integral sample into four parts, wherein each part comprises half-period waveform data, and each part of local waveform comprises a wave crest/wave trough or zero rest part;
3) And constructing one-dimensional new data, randomly selecting two pieces of local information, splicing the two pieces of local information on the whole waveform, and constructing the one-dimensional new data.
6. The method for diagnosing an arc fault in ac series connection of a low voltage distribution network according to claim 1, wherein the method comprises the steps of:
the multichannel attention mechanism convolutional neural network is a multichannel convolutional neural network integrating a multi-head attention mechanism and GRU units:
three data extraction channels are constructed by the network, and different structural parameters are respectively set for extracting features of data with different lengths so as to ensure that each data segment can extract an optimal feature group; the first channel is used for extracting features of the whole waveform of two periods, a multi-head attention module is embedded in the beginning part of the network, noise and redundant data in the data are eliminated by using an attention mechanism, and the features are focused on, so that the robustness of the model is improved.
7. The method for diagnosing an arc fault in ac series connection in a low voltage distribution network according to claim 6, wherein the method comprises the steps of:
the calculation process of the multi-head attention mechanism is as follows:
1) Dividing an input current waveform into a plurality of sections in parallel;
2) Generating three corresponding matrixes for each section of input data, wherein the three corresponding matrixes are respectively a query matrix, a key matrix and a value matrix;
3) Obtaining the weight fraction of the data by utilizing the dot product of the query matrix and the key matrix;
4) The weight fraction is normalized by softmax and multiplied by a value matrix to obtain the final data with weight attention;
Q＝AW Q (16)
K＝AW K (17)
V＝AW V (18)
wherein Q, K, V represent a query matrix, a key matrix and a value matrix, d, respectively k The dimensions of Q, K, V, A being the input matrix, W Q 、W K 、W V Is a coefficient matrix obtained by performing linear transformation on an input matrix;
5) Combining the learned multiple sections of attention to obtain global attention;
after passing through the attention layer, extracting features through the convolution layer, the pooling layer, the BN layer and the output layer, and connecting a GRU unit at the same time to prevent fitting in training; finally, a characteristic vector is output through a BN layer and a Droutput layer and through a relu activation function;
the second and third channels are used for extracting features from the half-period local waveform segments; directly passing through a convolution layer, a pooling layer, a BN layer and a Droutput layer, connecting with a GRU unit, and finally outputting a feature vector through a relu activation function; the Droutput layer is used to help the network prevent overfitting, and the activation function of the convolution layer uses the leakyrelu.
8. The method for diagnosing an arc fault in ac series connection in a low voltage distribution network according to claim 7, wherein the method comprises the steps of: when the sampling rate is changed, the number of the integral waveform sample points and the number of the local waveform sample points are changed, the number and the size of the convolution layers in the three channels are compared and adjusted according to the number of the input data points, and the parameters of the network are obtained through training.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310904639.1A CN116908618A (en) | 2023-07-24 | 2023-07-24 | Low-voltage distribution network alternating current series arc fault diagnosis method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310904639.1A CN116908618A (en) | 2023-07-24 | 2023-07-24 | Low-voltage distribution network alternating current series arc fault diagnosis method |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116908618A true CN116908618A (en) | 2023-10-20 |
Family
ID=88357866
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310904639.1A Pending CN116908618A (en) | 2023-07-24 | 2023-07-24 | Low-voltage distribution network alternating current series arc fault diagnosis method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116908618A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118376882A (en) * | 2024-06-20 | 2024-07-23 | 华中科技大学 | Power distribution network secondary fault identification method based on SVM-ECOC identification sub-low frequency characteristic energy |
-
2023
- 2023-07-24 CN CN202310904639.1A patent/CN116908618A/en active Pending
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118376882A (en) * | 2024-06-20 | 2024-07-23 | 华中科技大学 | Power distribution network secondary fault identification method based on SVM-ECOC identification sub-low frequency characteristic energy |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109660206B (en) | Wasserstein GAN-based photovoltaic array fault diagnosis method | |
| CN106017879A (en) | Universal circuit breaker mechanical fault diagnosis method based on feature fusion of vibration and sound signals | |
| CN110726898B (en) | Power distribution network fault type identification method | |
| CN110161388A (en) | A kind of the fault type recognition method and its system of high-tension apparatus | |
| CN110702966B (en) | Fault arc detection method, device and system based on probabilistic neural network | |
| CN116908618A (en) | Low-voltage distribution network alternating current series arc fault diagnosis method | |
| CN116738204A (en) | Power load data anomaly detection method based on generation of countermeasure transformers | |
| CN113076986B (en) | Photovoltaic fault arc characteristic selection method combining filtering type and packaging type evaluation strategies | |
| CN116399579A (en) | Transformer mechanical fault diagnosis method, device, computer equipment and storage medium | |
| CN108491503B (en) | Method and system for determining fault type of transformer based on data analysis | |
| CN110068776A (en) | Three-level inverter open-circuit fault diagnostic method based on Support Vector Machines Optimized | |
| CN116432112A (en) | Arc fault detection method based on wavelet packet transformation and residual convolution neural network | |
| CN115712871A (en) | Power electronic system fault diagnosis method combining resampling and integrated learning | |
| Rodrigues et al. | Deep learning for power quality event detection and classification based on measured grid data | |
| CN114021424B (en) | PCA-CNN-LVQ-based voltage sag source identification method | |
| Gao et al. | A two-level classification diagnosis method for AC arc faults based on data random fusion and MC-MGCNN network | |
| CN106407536A (en) | Compound fault diagnosis method for clamping diodes and supporting capacitors of inverter | |
| CN105425172A (en) | Self-adaption fault diagnosis method of micro-grid inverter based on multi-frequency band skewness analysis | |
| CN111898446A (en) | Single-phase earth fault studying and judging method based on multi-algorithm normalization analysis | |
| CN113610119A (en) | Method for identifying power transmission line developmental fault based on convolutional neural network | |
| CN116432027A (en) | Transformer running state detection method and system based on vibration signals | |
| Bin et al. | Study on the fault diagnosis of transformer based on the grey relational analysis | |
| Wang et al. | Real-time Diagnosis based on Signal ConvolutionPooling Processing and Shared Filter Learning for Transistor Open-Circuit Faults in a T-type Inverter | |
| CN116628620A (en) | Non-invasive load identification calculation method | |
| CN116298725A (en) | Fault arc detection method, system and storage medium |
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 |