CN113970680A - Arc detection method and device - Google Patents
Arc detection method and device Download PDFInfo
- Publication number
- CN113970680A CN113970680A CN202111237108.9A CN202111237108A CN113970680A CN 113970680 A CN113970680 A CN 113970680A CN 202111237108 A CN202111237108 A CN 202111237108A CN 113970680 A CN113970680 A CN 113970680A
- Authority
- CN
- China
- Prior art keywords
- neural network
- network model
- convolutional neural
- layer
- frame
- 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
- 238000001514 detection method Methods 0.000 title claims abstract description 27
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 101
- 238000000034 method Methods 0.000 claims abstract description 27
- 239000011159 matrix material Substances 0.000 claims abstract description 22
- 238000010891 electric arc Methods 0.000 claims abstract description 8
- 238000003062 neural network model Methods 0.000 claims abstract description 5
- 238000011176 pooling Methods 0.000 claims description 31
- 238000000605 extraction Methods 0.000 claims description 11
- 238000010276 construction Methods 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 2
- 238000005516 engineering process Methods 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 12
- 238000003860 storage Methods 0.000 description 11
- 230000006870 function Effects 0.000 description 10
- 238000004590 computer program Methods 0.000 description 7
- 238000005070 sampling Methods 0.000 description 7
- 238000001228 spectrum Methods 0.000 description 7
- 230000008569 process Effects 0.000 description 5
- 238000009826 distribution Methods 0.000 description 3
- 230000003287 optical effect Effects 0.000 description 3
- 206010000369 Accident Diseases 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 230000000694 effects Effects 0.000 description 2
- 230000004044 response Effects 0.000 description 2
- 230000004913 activation Effects 0.000 description 1
- 230000004931 aggregating effect Effects 0.000 description 1
- 238000003491 array Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 230000001413 cellular effect Effects 0.000 description 1
- 238000013145 classification model Methods 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 235000019800 disodium phosphate Nutrition 0.000 description 1
- 238000009432 framing Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000009413 insulation Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 239000007787 solid Substances 0.000 description 1
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
Abstract
The present disclosure relates to arc detection technologies, and in particular, to an arc detection method and apparatus. The method comprises the following steps: collecting current signals, and performing frame windowing processing on the collected current signals to obtain a plurality of frame current signals; calculating the Mel frequency cepstrum coefficient of each frame of current signal, and constructing a Mel frequency cepstrum coefficient matrix according to the Mel frequency cepstrum coefficient of each frame of current signal; constructing a convolutional neural network model, and training the convolutional neural network model to obtain a trained convolutional neural network model; and adopting a convolution neural network model which is trained in advance to recognize and classify the Mel frequency cepstrum coefficient matrix, and judging whether the electric arc occurs in the circuit according to the recognition and classification result.
Description
Technical Field
The present disclosure relates to arc detection technologies, and in particular, to an arc detection method and apparatus.
Background
Electrical fires represent a high percentage of fire accidents in today's society, and arc faults are one of the important causes of electrical fires. Fault arcs are generally caused by degradation and breakage of insulation of lines and equipment, or poor electrical connection, and when a fault arc occurs, the occurrence of the arc releases high temperature, which is extremely likely to cause a fire. The pressure, radiation and arc root effects generated by the electric arc not only easily cause damage to electrical equipment, but also even cause personal casualty accidents and significant economic and electric arc faults are the most main causes of electric fire accidents. Therefore, the online accurate and quick identification of the arc fault has important guiding significance for reducing electrical fire.
Disclosure of Invention
The application provides an arc detection method and device, which are used for solving the problem that arc faults cannot be accurately and quickly identified on line in the prior art.
In a first aspect, the present application provides a method of arc detection, the method comprising:
collecting current signals, and performing frame windowing processing on the collected current signals to obtain a plurality of frame current signals;
calculating the Mel frequency cepstrum coefficient of each frame of current signal, and constructing a Mel frequency cepstrum coefficient matrix according to the Mel frequency cepstrum coefficient of each frame of current signal;
constructing a convolutional neural network model, and training the convolutional neural network model to obtain a trained convolutional neural network model;
and adopting a convolution neural network model which is trained in advance to recognize and classify the Mel frequency cepstrum coefficient matrix, and judging whether the electric arc occurs in the circuit according to the recognition and classification result.
Optionally, before performing frame-wise windowing on the acquired current signal, the method further includes:
carrying out pre-emphasis processing on the acquired current signal by adopting a high-pass filter to obtain a current signal after the pre-emphasis processing;
the step of performing frame windowing on the collected current signals to obtain a plurality of frames of current signals comprises the following steps:
and performing frame division and windowing processing on the current signals subjected to the pre-emphasis processing to obtain a plurality of frame current signals.
Optionally, the step of calculating mel-frequency cepstrum coefficients of the current signal of each frame comprises:
and respectively carrying out fast Fourier transform on each frame of current signal, carrying out feature extraction on a transform result by adopting a preset triangular band-pass filter, and carrying out discrete cosine transform on the feature extracted by each filter to obtain a corresponding Mel frequency cepstrum coefficient.
Optionally, the step of constructing a two-dimensional convolutional neural network model, and training the two-dimensional convolutional neural network model to obtain a trained two-dimensional convolutional neural network model includes:
constructing a convolutional neural network model and initializing weight parameters;
inputting sample data into a convolutional neural network model, and carrying out forward propagation on the sample data through a convolutional layer, a pooling layer and a full-link layer to obtain an output value;
calculating an error between the output value and a target value;
if the error is larger than the expected value, the error is transmitted back to the convolutional neural network model, the errors of the fully-connected layer, the pooling layer and the convolutional layer are sequentially obtained, the weight of each layer of the convolutional neural network model is adjusted according to the errors of each layer, sample data is input into the convolutional neural network model, and the step that the sample data is transmitted in the forward direction through the convolutional layer, the pooling layer and the fully-connected layer to obtain an output value is executed;
and if the error is less than or equal to the expected value, determining that the training of the convolutional neural network model is finished, and obtaining the trained convolutional neural network model.
Alternatively,
the general structure of the convolutional neural network comprises: one input layer, five convolution layers and pooling layers, and one full-connection layer.
In a second aspect, the present application provides an arc detection apparatus, the apparatus comprising:
the acquisition unit is used for acquiring current signals and performing frame windowing processing on the acquired current signals to obtain a plurality of frame current signals;
the calculation unit is used for calculating the Mel frequency cepstrum coefficient of each frame of current signal and constructing a Mel frequency cepstrum coefficient matrix according to the Mel frequency cepstrum coefficient of each frame of current signal;
the building unit is used for building a convolutional neural network model and training the convolutional neural network model to obtain a trained convolutional neural network model;
and the judging unit is used for identifying and classifying the Mel frequency cepstrum coefficient matrix by adopting a convolutional neural network model which is trained in advance, and judging whether electric arcs occur in the circuit according to the identification and classification result.
Optionally, before performing frame-wise windowing on the acquired current signal, the apparatus further includes:
the preprocessing unit is used for performing pre-emphasis processing on the acquired current signals by adopting a high-pass filter to obtain current signals after the pre-emphasis processing;
the acquisition unit is specifically used for performing frame windowing on the acquired current signals to obtain a plurality of frames of current signals
And performing frame division and windowing processing on the current signals subjected to the pre-emphasis processing to obtain a plurality of frame current signals.
Optionally, when the mel-frequency cepstrum coefficient of each frame of current signal is calculated, the calculating unit is specifically configured to:
and respectively carrying out fast Fourier transform on each frame of current signal, carrying out feature extraction on a transform result by adopting a preset triangular band-pass filter, and carrying out discrete cosine transform on the feature extracted by each filter to obtain a corresponding Mel frequency cepstrum coefficient.
Optionally, when a two-dimensional convolutional neural network model is constructed and trained to obtain a trained two-dimensional convolutional neural network model, the construction unit is specifically configured to:
constructing a convolutional neural network model and initializing weight parameters;
inputting sample data into a convolutional neural network model, and carrying out forward propagation on the sample data through a convolutional layer, a pooling layer and a full-link layer to obtain an output value;
calculating an error between the output value and a target value;
if the error is larger than the expected value, the error is transmitted back to the convolutional neural network model, the errors of the fully-connected layer, the pooling layer and the convolutional layer are sequentially obtained, the weight of each layer of the convolutional neural network model is adjusted according to the errors of each layer, sample data is input into the convolutional neural network model, and the step that the sample data is transmitted in the forward direction through the convolutional layer, the pooling layer and the fully-connected layer to obtain an output value is executed;
and if the error is less than or equal to the expected value, determining that the training of the convolutional neural network model is finished, and obtaining the trained convolutional neural network model.
Alternatively,
the general structure of the convolutional neural network comprises: one input layer, five convolution layers and pooling layers, and one full-connection layer.
In a third aspect, an embodiment of the present application provides an arc detection apparatus, including:
a memory for storing program instructions;
a processor for calling program instructions stored in said memory and for executing the steps of the method according to any one of the above first aspects in accordance with the obtained program instructions.
In a fourth aspect, the present application further provides a computer-readable storage medium storing computer-executable instructions for causing a computer to perform the steps of the method according to any one of the above first aspects.
In summary, the arc detection method provided in the embodiment of the present application collects current signals, and performs frame windowing on the collected current signals to obtain a plurality of frames of current signals; calculating the Mel frequency cepstrum coefficient of each frame of current signal, and constructing a Mel frequency cepstrum coefficient matrix according to the Mel frequency cepstrum coefficient of each frame of current signal; constructing a convolutional neural network model, and training the convolutional neural network model to obtain a trained convolutional neural network model; and adopting a convolution neural network model which is trained in advance to recognize and classify the Mel frequency cepstrum coefficient matrix, and judging whether the electric arc occurs in the circuit according to the recognition and classification result.
By adopting the arc detection method provided by the embodiment of the application, the Mel frequency cepstrum coefficient matrix characteristic of the current signal is collected and extracted, and the arc fault is identified on line based on the convolutional neural network, so that whether the arc is generated in the circuit can be detected on line quickly and accurately. Therefore, the electric fire disaster can be reduced, personal casualty accidents are avoided, and the serious economic loss of enterprises caused by the electric fire disaster due to electric arcs is avoided.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments of the present application or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments described in the present application, and other drawings can be obtained by those skilled in the art according to the drawings of the embodiments of the present application.
Fig. 1 is a detailed flowchart of an arc detection method according to an embodiment of the present disclosure;
fig. 2 is a schematic flow chart of an MFCC extraction process according to an embodiment of the present disclosure;
FIG. 3 is a schematic structural diagram of an arc detection device according to an embodiment of the present disclosure;
fig. 4 is a schematic structural diagram of another arc detection device according to an embodiment of the present application.
Detailed Description
The terminology used in the embodiments of the present application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. As used in this application and the claims, the singular forms "a", "an", and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It should also be understood that the term "and/or" as used herein is meant to encompass any and all possible combinations of one or more of the associated listed items.
It should be understood that although the terms first, second, third, etc. may be used in the embodiments of the present application to describe various information, the information should not be limited to these terms. These terms are only used to distinguish one type of information from another. For example, first information may also be referred to as second information, and similarly, second information may also be referred to as first information, without departing from the scope of the present application. Depending on the context, moreover, the word "if" as used may be interpreted as "at … …" or "when … …" or "in response to a determination".
Illustratively, referring to fig. 1, a detailed flowchart of an arc detection method provided in an embodiment of the present application is shown, where the method includes the following steps:
step 100: and collecting current signals, and performing frame windowing on the collected current signals to obtain a plurality of frame current signals.
In this embodiment of the application, after the collected current signal and before performing frame windowing on the collected current signal, the arc detection method may further include the steps of:
and carrying out pre-emphasis processing on the acquired current signal by adopting a high-pass filter to obtain the current signal after the pre-emphasis processing.
Then, when performing frame-wise windowing on the acquired current signals to obtain a plurality of frame current signals, a preferred implementation manner is:
and performing frame division and windowing processing on the current signals subjected to the pre-emphasis processing to obtain a plurality of frame current signals.
Step 110: and calculating the Mel frequency cepstrum coefficient of each frame of current signal, and constructing a Mel frequency cepstrum coefficient matrix according to the Mel frequency cepstrum coefficient of each frame of current signal.
Specifically, in the embodiment of the present application, when calculating the mel-frequency cepstrum coefficient of each frame of current signal, a preferred implementation manner is as follows:
and respectively carrying out fast Fourier transform on each frame of current signal, carrying out feature extraction on a transform result by adopting a preset triangular band-pass filter, and carrying out discrete cosine transform on the feature extracted by each filter to obtain a corresponding Mel frequency cepstrum coefficient.
Step 120: and constructing a convolutional neural network model, and training the convolutional neural network model to obtain the trained convolutional neural network model.
In the embodiment of the present application, when a two-dimensional convolutional neural network model is constructed and trained to obtain a trained two-dimensional convolutional neural network model, a preferred implementation manner is:
step 1: and constructing a convolutional neural network model and initializing weight parameters.
Step 2: and inputting the sample data into the convolutional neural network model, and carrying out forward propagation on the sample data through the convolutional layer, the pooling layer and the full-connection layer to obtain an output value.
And step 3: an error between the output value and a target value is calculated.
And 4, step 4: if the error is larger than the expected value, the error is returned to the convolutional neural network model, the errors of the full-link layer, the pooling layer and the convolutional layer are sequentially obtained, the weight of each layer of the convolutional neural network model is adjusted according to the errors of each layer, and the step 2 and the step 3 are executed.
And 5: and if the error is less than or equal to the expected value, determining that the training of the convolutional neural network model is finished, and obtaining the trained convolutional neural network model.
Step 130: and adopting a convolution neural network model which is trained in advance to recognize and classify the Mel frequency cepstrum coefficient matrix, and judging whether the electric arc occurs in the circuit according to the recognition and classification result.
The arc detection method provided by the embodiment of the present application is described in detail below with reference to specific application scenarios.
In practical application, in a process of implementing arc detection, feature extraction is performed on a current signal first, and preferably, the acquired current signal may be represented by a parameter with less redundant information. Then, the collected current signal is subjected to feature extraction (Mel-Frequency Cepstral Coefficients, MFCC) using Mel-Frequency Cepstral Coefficients. Fig. 2 is a schematic flow chart of an MFCC extraction process according to an embodiment of the present disclosure.
The first step is as follows: and collecting a current signal.
The second step is that: and carrying out pre-emphasis processing on the acquired current signals.
In particular, the sampled vibration signal s (z) is passed through a high pass filter. The formula (1) is as follows:
H(z)＝1-uz-1(1)
wherein H (z) is a Gaussian filter function, z is an input signal, and u is a weighting coefficient having a value of 0.9 to 1.0. The purpose of pre-emphasis is to boost the high frequency portion of the signal, flattening the spectrum of the signal, and maintaining it in the entire band from low to high frequencies.
The third step: and performing frame division and windowing processing on the current signal subjected to the pre-emphasis processing.
Specifically, N sampling points are first grouped into one observation unit, which is called a frame. The value of N is usually set to 256 or 512, and in practical applications, the electrical equipment adopts a nominal frequency of 50Hz, so that the covered time unit is 20 ms. In order to avoid signal loss between adjacent 2 frames, an overlap region is provided between 2 adjacent frames during sampling, the overlap region includes M sampling points, and the value of M is usually set to 1/2 or 1/3 of N. In the embodiment of the present application, the signal sampling rate may be set to 48kHz, the number of sampling points is set to 1024, and the corresponding sampling time is about 21 ms. Each frame is then multiplied by a hamming window to increase the continuity of the left and right ends of the frame.
S′(z)＝S(z)×W(z)(2)
Where, S' (z) is the framed signal, and w (z) is the hamming window function.
Different values of a will result in different Hamming windows, typically a being 0.46.
The fourth step: and performing fast Fourier transform on each frame of current signal, and performing feature extraction on a transform result by adopting a preset triangular band-pass filter.
Specifically, it is difficult to distinguish the difference of different signal characteristics by analyzing the change of the signal in the time domain, and in the digital signal processing process, the signal is usually converted from the change in the time domain into the energy distribution in the frequency domain to observe the characteristics of different signals. The signal needs to be fast fourier transformed after windowing to obtain the energy distribution of the signal over the frequency spectrum. And carrying out fast Fourier transform on each frame signal subjected to framing and windowing to obtain the frequency spectrum of each frame. And obtaining the power spectrum of the vibration signal by taking the modulus square of the frequency spectrum of each frame. Let DFT of the current signal be:
in the formula, j is an imaginary unit, and k is a natural integer.
Then, the power spectrum passes through a group of Mel-scale filter banks, a string of triangular filters which are arranged in a cross overlapping manner in a low-frequency region is usually used, due to the nonlinear correspondence of Hz-Mel frequency, the number of filters in the low-frequency region is large, so that the low-order MFCC coefficients weaken the spectrum information in the high-frequency region, and the number K of MFCC filters is usually set to 9 or 13.
According to equation (5):
wherein f is the maximum frequency of the vibration signal, so that the maximum Mel frequency f can be obtainedmelSince the center frequencies of the respective triangular filters are equally spaced linear distributions over the Mel scale. Thus, the distance between the center frequencies of two adjacent triangular filters can be calculated as
The fifth step: and performing discrete cosine transform on the Mel characteristic corresponding to each frame of current signal.
Specifically, the logarithmic energy of each filter bank output is calculated as shown in equation (7). Wherein Hm(h) H is the frequency response of the mth filter and h is a discrete digital variable.
Where the MFCC coefficients are then obtained via Discrete Cosine Transform (DCT).
And substituting the logarithmic energy into discrete cosine transform to obtain the MFCC parameter of L order. In the embodiment of the present application, the value of L is 16.
In order to realize the detection of the arc signal, the current signal generated by the circuit in the experimental process is collected and is divided into frames by taking 20ms as a unit. Mel-frequency cepstral coefficients (MFCCs) of the signals of each frame are calculated, and the Mel-frequency cepstral coefficients (MFCCs) of all frames are formed into a (1872,26) MFCC matrix, as shown below.
And further, inputting the mel frequency cepstrum coefficient feature matrix into a pre-trained convolutional neural network model, identifying and classifying the mel frequency cepstrum coefficient matrix by the convolutional neural network model, and judging whether the electric arc occurs in the circuit according to the identification and classification result.
In practical application, a Convolutional Neural Network (CNN) has a stronger adaptability to a two-dimensional signal with a complex background environment and an insignificant feature. The convolutional neural network is a feedforward neural network formed by a plurality of convolutional layers, pooling layers and full-connection layers. The convolutional layer senses local features of different sizes of data through convolutional kernels of different sizes, the local features are weighted and averaged through convolutional kernel parameters, and then local information is integrated at a higher layer to obtain global signal features. The pooling layer is capable of better aggregating features and reducing the amount of computation by down-sampling high-dimensional data and scaling or reconstructing the sample size. Common pooling methods are maximum pooling and average pooling. And the full connection layer performs weight connection on all neuron nodes between layers, so that local features are fused into global features. Since the MFCC coefficients can represent the characteristics of the current signal frequency domain, the convolutional neural network can extract deeper features on the current signal frequency domain, and the softmax classifier can map high-dimensional features into the probability of the class to which the features belong. In the embodiment of the application, a classification model combining Mel-Frequency Cepstral Coefficients (MFCC) and a convolutional neural network is provided based on the above principle, a MFCC two-dimensional matrix formed by collected current signals is used for classifying whether an arc occurs in a circuit, and whether an arc fault occurs in the circuit is detected according to a classification result.
In the embodiment of the application, the FD-CNN network is provided on the basis of the LD-CNN network. In the embodiment of the application, preferably, the step size of the convolutional layer is set to be 1, the step size of the pooling layer is maximized to be 2, and finally, a full connection layer is used to replace a FeatureSum layer in the LD-CNN. For example, the overall structure of an FD-CNN network includes: one input layer, five convolution layers and pooling layers, and one full-connection layer. The convolution kernel uses a 3 × 3 matrix and the activation function uses a relu function.
In the embodiment of the present application, the training process of the convolutional neural network model is as follows:
(1) constructing a convolutional neural network model and initializing a weight;
(2) the input data is transmitted forward through a convolution layer, a pooling layer and a full-connection layer to obtain an output value;
(3) calculating an error between the output value of the network and a target value;
(4) when the error is larger than the expected value, the error is transmitted back to the network, the errors of the network full-connection layer, the pooling layer and the convolution layer are sequentially obtained, when the error is smaller than or equal to the given expected value, the training is finished, and the step 6 is entered;
(5) adjusting the weight of each layer of the network according to the obtained error, and entering the step 2 again;
(6) and (5) outputting high-level feature data after training is finished.
Based on the same inventive concept as the above method embodiment, for example, referring to fig. 3, a schematic structural diagram of an arc detection apparatus provided in an embodiment of the present application is shown, where the apparatus includes:
the acquisition unit 30 is used for acquiring current signals and performing frame windowing on the acquired current signals to obtain a plurality of frame current signals;
the calculating unit 31 is configured to calculate a mel-frequency cepstrum coefficient of each frame of current signal, and construct a mel-frequency cepstrum coefficient matrix according to the mel-frequency cepstrum coefficient of each frame of current signal;
the building unit 32 is configured to build a convolutional neural network model, and train the convolutional neural network model to obtain a trained convolutional neural network model;
and the judging unit 33 is configured to recognize and classify the mel-frequency cepstrum coefficient matrix by using a convolutional neural network model trained in advance, and judge whether an arc occurs in the circuit according to a recognition and classification result.
Optionally, before performing frame-wise windowing on the acquired current signal, the apparatus further includes:
the preprocessing unit is used for performing pre-emphasis processing on the acquired current signals by adopting a high-pass filter to obtain current signals after the pre-emphasis processing;
the collecting unit 30 is specifically configured to perform frame windowing on the collected current signals to obtain a plurality of frames of current signals
And performing frame division and windowing processing on the current signals subjected to the pre-emphasis processing to obtain a plurality of frame current signals.
Optionally, when the mel-frequency cepstrum coefficient of each frame of the current signal is calculated, the calculating unit 31 is specifically configured to:
and respectively carrying out fast Fourier transform on each frame of current signal, carrying out feature extraction on a transform result by adopting a preset triangular band-pass filter, and carrying out discrete cosine transform on the feature extracted by each filter to obtain a corresponding Mel frequency cepstrum coefficient.
Optionally, when a two-dimensional convolutional neural network model is constructed and trained to obtain a trained two-dimensional convolutional neural network model, the construction unit 32 is specifically configured to:
constructing a convolutional neural network model and initializing weight parameters;
inputting sample data into a convolutional neural network model, and carrying out forward propagation on the sample data through a convolutional layer, a pooling layer and a full-link layer to obtain an output value;
calculating an error between the output value and a target value;
if the error is larger than the expected value, the error is transmitted back to the convolutional neural network model, the errors of the fully-connected layer, the pooling layer and the convolutional layer are sequentially obtained, the weight of each layer of the convolutional neural network model is adjusted according to the errors of each layer, sample data is input into the convolutional neural network model, and the step that the sample data is transmitted in the forward direction through the convolutional layer, the pooling layer and the fully-connected layer to obtain an output value is executed;
and if the error is less than or equal to the expected value, determining that the training of the convolutional neural network model is finished, and obtaining the trained convolutional neural network model.
Alternatively,
the general structure of the convolutional neural network comprises: one input layer, five convolution layers and pooling layers, and one full-connection layer.
The above units may be one or more integrated circuits configured to implement the above methods, for example: one or more Application Specific Integrated Circuits (ASICs), or one or more microprocessors (DSPs), or one or more Field Programmable Gate Arrays (FPGAs), among others. For another example, when one of the above units is implemented in the form of a Processing element scheduler code, the Processing element may be a general-purpose processor, such as a Central Processing Unit (CPU) or other processor capable of calling program code. For another example, these units may be integrated together and implemented in the form of a system-on-a-chip (SOC).
Further, in the arc detection apparatus provided in the embodiment of the present application, from a hardware level, a schematic diagram of a hardware architecture of the arc detection apparatus may be shown in fig. 4, where the arc detection apparatus may include: a memory 40 and a processor 41, which,
Optionally, the present application also provides an arc detection apparatus comprising at least one processing element (or chip) for performing the above-described method embodiments.
Optionally, the present application also provides a program product, such as a computer-readable storage medium, having stored thereon computer-executable instructions for causing the computer to perform the above-described method embodiments.
Here, a machine-readable storage medium may be any electronic, magnetic, optical, or other physical storage device that can contain or store information such as executable instructions, data, and so forth. For example, the machine-readable storage medium may be: a RAM (random Access Memory), a volatile Memory, a non-volatile Memory, a flash Memory, a storage drive (e.g., a hard drive), a solid state drive, any type of storage disk (e.g., an optical disk, a dvd, etc.), or similar storage medium, or a combination thereof.
The systems, devices, modules or units illustrated in the above embodiments may be implemented by a computer chip or an entity, or by a product with certain functions. A typical implementation device is a computer, which may take the form of a personal computer, laptop computer, cellular telephone, camera phone, smart phone, personal digital assistant, media player, navigation device, email messaging device, game console, tablet computer, wearable device, or a combination of any of these devices.
For convenience of description, the above devices are described as being divided into various units by function, and are described separately. Of course, the functionality of the units may be implemented in one or more software and/or hardware when implementing the present application.
As will be appreciated by one skilled in the art, 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, embodiments of 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 flow diagrams and/or block diagrams, and combinations of flows and/or blocks in the flow diagrams 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.
Furthermore, 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 exemplary of the present application and should not be taken as limiting the present application, as any modification, equivalent replacement, or improvement made within the spirit and principle of the present application should be included in the scope of protection of the present application.
Claims (10)
1. A method of arc detection, the method comprising:
collecting current signals, and performing frame windowing processing on the collected current signals to obtain a plurality of frame current signals;
calculating the Mel frequency cepstrum coefficient of each frame of current signal, and constructing a Mel frequency cepstrum coefficient matrix according to the Mel frequency cepstrum coefficient of each frame of current signal;
constructing a convolutional neural network model, and training the convolutional neural network model to obtain a trained convolutional neural network model;
and adopting a convolution neural network model which is trained in advance to recognize and classify the Mel frequency cepstrum coefficient matrix, and judging whether the electric arc occurs in the circuit according to the recognition and classification result.
2. The method of claim 1, wherein prior to performing frame windowing on the acquired current signal, the method further comprises:
carrying out pre-emphasis processing on the acquired current signal by adopting a high-pass filter to obtain a current signal after the pre-emphasis processing;
the step of performing frame windowing on the collected current signals to obtain a plurality of frames of current signals comprises the following steps:
and performing frame division and windowing processing on the current signals subjected to the pre-emphasis processing to obtain a plurality of frame current signals.
3. The method of claim 1 or 2, wherein the step of calculating mel-frequency cepstral coefficients of each frame of the current signal comprises:
and respectively carrying out fast Fourier transform on each frame of current signal, carrying out feature extraction on a transform result by adopting a preset triangular band-pass filter, and carrying out discrete cosine transform on the feature extracted by each filter to obtain a corresponding Mel frequency cepstrum coefficient.
4. The method of claim 1 or 2, wherein the step of constructing a two-dimensional convolutional neural network model and training the two-dimensional convolutional neural network model to obtain a trained two-dimensional convolutional neural network model comprises:
constructing a convolutional neural network model and initializing weight parameters;
inputting sample data into a convolutional neural network model, and carrying out forward propagation on the sample data through a convolutional layer, a pooling layer and a full-link layer to obtain an output value;
calculating an error between the output value and a target value;
if the error is larger than the expected value, the error is transmitted back to the convolutional neural network model, the errors of the fully-connected layer, the pooling layer and the convolutional layer are sequentially obtained, the weight of each layer of the convolutional neural network model is adjusted according to the errors of each layer, sample data is input into the convolutional neural network model, and the step that the sample data is transmitted in the forward direction through the convolutional layer, the pooling layer and the fully-connected layer to obtain an output value is executed;
and if the error is less than or equal to the expected value, determining that the training of the convolutional neural network model is finished, and obtaining the trained convolutional neural network model.
5. The method of claim 4,
the general structure of the convolutional neural network comprises: one input layer, five convolution layers and pooling layers, and one full-connection layer.
6. An arc detection device, characterized in that the device comprises:
the acquisition unit is used for acquiring current signals and performing frame windowing processing on the acquired current signals to obtain a plurality of frame current signals;
the calculation unit is used for calculating the Mel frequency cepstrum coefficient of each frame of current signal and constructing a Mel frequency cepstrum coefficient matrix according to the Mel frequency cepstrum coefficient of each frame of current signal;
the building unit is used for building a convolutional neural network model and training the convolutional neural network model to obtain a trained convolutional neural network model;
and the judging unit is used for identifying and classifying the Mel frequency cepstrum coefficient matrix by adopting a convolutional neural network model which is trained in advance, and judging whether electric arcs occur in the circuit according to the identification and classification result.
7. The apparatus of claim 6, wherein prior to performing frame windowing on the acquired current signal, the apparatus further comprises:
the preprocessing unit is used for performing pre-emphasis processing on the acquired current signals by adopting a high-pass filter to obtain current signals after the pre-emphasis processing;
the acquisition unit is specifically used for performing frame windowing on the acquired current signals to obtain a plurality of frames of current signals
And performing frame division and windowing processing on the current signals subjected to the pre-emphasis processing to obtain a plurality of frame current signals.
8. The apparatus according to claim 6 or 7, wherein the computing unit, when computing mel-frequency cepstral coefficients of each frame of current signal, is specifically configured to:
and respectively carrying out fast Fourier transform on each frame of current signal, carrying out feature extraction on a transform result by adopting a preset triangular band-pass filter, and carrying out discrete cosine transform on the feature extracted by each filter to obtain a corresponding Mel frequency cepstrum coefficient.
9. The apparatus according to claim 6 or 7, wherein, when a two-dimensional convolutional neural network model is constructed and trained to obtain a trained two-dimensional convolutional neural network model, the construction unit is specifically configured to:
constructing a convolutional neural network model and initializing weight parameters;
inputting sample data into a convolutional neural network model, and carrying out forward propagation on the sample data through a convolutional layer, a pooling layer and a full-link layer to obtain an output value;
calculating an error between the output value and a target value;
if the error is larger than the expected value, the error is transmitted back to the convolutional neural network model, the errors of the fully-connected layer, the pooling layer and the convolutional layer are sequentially obtained, the weight of each layer of the convolutional neural network model is adjusted according to the errors of each layer, sample data is input into the convolutional neural network model, and the step that the sample data is transmitted in the forward direction through the convolutional layer, the pooling layer and the fully-connected layer to obtain an output value is executed;
and if the error is less than or equal to the expected value, determining that the training of the convolutional neural network model is finished, and obtaining the trained convolutional neural network model.
10. The apparatus of claim 9,
the general structure of the convolutional neural network comprises: one input layer, five convolution layers and pooling layers, and one full-connection layer.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111237108.9A CN113970680A (en) | 2021-10-22 | 2021-10-22 | Arc detection method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111237108.9A CN113970680A (en) | 2021-10-22 | 2021-10-22 | Arc detection method and device |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN113970680A true CN113970680A (en) | 2022-01-25 |
Family
ID=79588052
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111237108.9A Pending CN113970680A (en) | 2021-10-22 | 2021-10-22 | Arc detection method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113970680A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115273438A (en) * | 2022-07-05 | 2022-11-01 | 广东远景信息科技有限公司 | Forest intelligent fire prevention method, device, system, equipment and medium based on 5G |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101299183B1 (en) * | 2012-06-22 | 2013-08-22 | 성균관대학교산학협력단 | Classification method for quality of power signal using speech processing techniques, monitoring apparatus for quality of power signal using speech processing techniques and monitoring system for quality of power signal using speech processing techniques |
| JP2014509396A (en) * | 2011-02-28 | 2014-04-17 | エスエムエー ソーラー テクノロジー アーゲー | Method and system for detecting arc faults in power circuits |
| WO2019182166A1 (en) * | 2018-03-19 | 2019-09-26 | 한국 전기안전공사 | Electrical load determination apparatus and method using modified mfcc |
-
2021
- 2021-10-22 CN CN202111237108.9A patent/CN113970680A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2014509396A (en) * | 2011-02-28 | 2014-04-17 | エスエムエー ソーラー テクノロジー アーゲー | Method and system for detecting arc faults in power circuits |
| KR101299183B1 (en) * | 2012-06-22 | 2013-08-22 | 성균관대학교산학협력단 | Classification method for quality of power signal using speech processing techniques, monitoring apparatus for quality of power signal using speech processing techniques and monitoring system for quality of power signal using speech processing techniques |
| WO2019182166A1 (en) * | 2018-03-19 | 2019-09-26 | 한국 전기안전공사 | Electrical load determination apparatus and method using modified mfcc |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115273438A (en) * | 2022-07-05 | 2022-11-01 | 广东远景信息科技有限公司 | Forest intelligent fire prevention method, device, system, equipment and medium based on 5G |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110516305B (en) | Intelligent fault diagnosis method under small sample based on attention mechanism meta-learning model | |
| EP3440598A1 (en) | Frequency based audio analysis using neural networks | |
| CN108877783B (en) | Method and apparatus for determining audio type of audio data | |
| CN111785288B (en) | Voice enhancement method, device, equipment and storage medium | |
| CN112801169A (en) | Camouflage target detection method based on improved YOLO algorithm | |
| Meyer et al. | Efficient convolutional neural network for audio event detection | |
| CN116503399B (en) | Insulator pollution flashover detection method based on YOLO-AFPS | |
| Sharma et al. | A modified MFCC feature extraction technique for robust speaker recognition | |
| CN110502659A (en) | The training method of image characteristics extraction and network, device and equipment | |
| CN113970680A (en) | Arc detection method and device | |
| CN115273904A (en) | Angry emotion recognition method and device based on multi-feature fusion | |
| CN114582325A (en) | Audio detection method and device, computer equipment and storage medium | |
| Imran et al. | An analysis of audio classification techniques using deep learning architectures | |
| CN113689383B (en) | Image processing method, device, equipment and storage medium | |
| CN112365335B (en) | Method and device for processing credit data | |
| CN115273814A (en) | Pseudo voice detection method, device, computer equipment and storage medium | |
| CN115019760A (en) | Data amplification method for audio and real-time sound event detection system and method | |
| CN109448287B (en) | Safety monitoring method and device, readable storage medium and terminal equipment | |
| CN109741733B (en) | Voice phoneme recognition method based on consistency routing network | |
| CN112861949A (en) | Face and voice-based emotion prediction method and system | |
| Godoy et al. | Using deep learning for detecting spoofing attacks on speech signals | |
| CN115982757B (en) | Method, device and equipment for determining privacy protection degree of model | |
| CN117198272B (en) | Voice processing method and device, electronic equipment and storage medium | |
| CN116935889B (en) | Audio category determining method and device, electronic equipment and storage medium | |
| Abou-Zleikha et al. | Non-linguistic vocal event detection using online random forest |
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 | ||
| TA01 | Transfer of patent application right | ||
| TA01 | Transfer of patent application right |
Effective date of registration: 20230417 Address after: 33, 4th Floor, Commercial Service Building, No. 19, Yuhuili Second District, Xiaoying Road, Chaoyang District, Beijing 100000 Applicant after: Beijing Smart Power Technology Co.,Ltd. Address before: 100000 3f06, heshenghui office building, No. 22, xidawang Road, Chaoyang District, Beijing Applicant before: Duan Xiaohan Applicant before: Xian Ce Applicant before: Pan Yingchao |