CN112241605A - Method for identifying state of circuit breaker energy storage process by constructing CNN characteristic matrix through acoustic vibration signals - Google Patents
Method for identifying state of circuit breaker energy storage process by constructing CNN characteristic matrix through acoustic vibration signals Download PDFInfo
- Publication number
- CN112241605A CN112241605A CN201910643289.1A CN201910643289A CN112241605A CN 112241605 A CN112241605 A CN 112241605A CN 201910643289 A CN201910643289 A CN 201910643289A CN 112241605 A CN112241605 A CN 112241605A
- Authority
- CN
- China
- Prior art keywords
- energy storage
- cnn
- circuit breaker
- sound
- vibration
- 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.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 58
- 239000011159 matrix material Substances 0.000 title claims abstract description 36
- 238000004146 energy storage Methods 0.000 title claims abstract description 35
- 230000008569 process Effects 0.000 title claims abstract description 31
- 230000008859 change Effects 0.000 claims abstract description 10
- 238000005457 optimization Methods 0.000 claims abstract description 8
- 238000003745 diagnosis Methods 0.000 claims description 9
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000001514 detection method Methods 0.000 claims description 5
- 230000007246 mechanism Effects 0.000 claims description 5
- 230000000739 chaotic effect Effects 0.000 claims description 4
- 230000009466 transformation Effects 0.000 claims description 4
- 230000035772 mutation Effects 0.000 claims description 2
- 230000004075 alteration Effects 0.000 claims 1
- 238000010276 construction Methods 0.000 claims 1
- 230000037430 deletion Effects 0.000 claims 1
- 238000012217 deletion Methods 0.000 claims 1
- 238000013527 convolutional neural network Methods 0.000 abstract description 32
- 241000282461 Canis lupus Species 0.000 abstract description 8
- 238000012706 support-vector machine Methods 0.000 abstract description 8
- 238000012549 training Methods 0.000 abstract description 5
- 238000011176 pooling Methods 0.000 description 8
- 230000005236 sound signal Effects 0.000 description 7
- 239000013598 vector Substances 0.000 description 7
- 238000010586 diagram Methods 0.000 description 6
- 230000006870 function Effects 0.000 description 6
- 238000011160 research Methods 0.000 description 5
- 238000005070 sampling Methods 0.000 description 5
- 230000004913 activation Effects 0.000 description 4
- 210000002569 neuron Anatomy 0.000 description 4
- 238000012544 monitoring process Methods 0.000 description 3
- 238000000605 extraction Methods 0.000 description 2
- 238000009434 installation Methods 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- RTAQQCXQSZGOHL-UHFFFAOYSA-N Titanium Chemical compound [Ti] RTAQQCXQSZGOHL-UHFFFAOYSA-N 0.000 description 1
- 230000001133 acceleration Effects 0.000 description 1
- 238000009825 accumulation Methods 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 230000006399 behavior Effects 0.000 description 1
- 230000000295 complement effect Effects 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000002405 diagnostic procedure Methods 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000005259 measurement Methods 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000035939 shock Effects 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The invention discloses a method for identifying the state of a circuit breaker in the energy storage process by constructing a Convolutional Neural Network (CNN) characteristic matrix by using a sound vibration signal, which comprises the following steps: firstly, a time scale alignment method based on kurtosis and envelope similarity is provided to ensure the synchronism of the sound vibration signals, then a Lyapunov index-wavelet modulus maximum value (L-wavelet) is adopted to detect the initial point of the vibration signals, after the overlapped data expansion is carried out on the data, a pearson correlation coefficient is utilized to construct a sound vibration signal two-dimensional characteristic matrix. And finally, training the feature matrix by using the CNN, optimizing the CNN structure by using a Support Vector Machine (SVM) to replace a Soft-Max classifier, and searching for the optimal parameters of the SVM by using a Hui wolf optimization (GWO). The optimized CNN model is insensitive to the condition of large data change in the energy storage process of the circuit breaker, and the optimized CNN model is used as a new method for identifying the state of the energy storage process of the circuit breaker, so that the accuracy and the generalization of state identification are greatly improved.
Description
Technical Field
The invention relates to the technical field of electrical equipment fault diagnosis, in particular to a method for identifying the energy storage process state of a circuit breaker by constructing a characteristic matrix required by a Convolutional Neural Network (CNN) through combining sound vibration signals.
Background
The circuit breaker is used as an important control and protection device in the power system, and whether reliable action can directly affect the safety and stability of the power system, so that the operation reliability of the circuit breaker is important for the protection and control of a power grid.
At present, the research on fault diagnosis of the circuit breaker is mostly focused on the switching-on and switching-off process: and identifying the mechanical fault by using the control coil current, the displacement of the insulating pull rod and the vibration signal. The research focuses on the problems of the breaker in the operation process, but the research on the problems of the energy storage process is not deep enough, a quantitative judgment basis is lacked, and how to find the faults in the energy storage process and the development and change rules of the faults are worthy of deep research.
The existing fault diagnosis method of the circuit breaker mainly takes a vibration signal as a main part, but in practical application, a saturation phenomenon exists when the amplitude is large, and high-frequency impact failure caused by a charge accumulation effect is easy to generate. The sound signal can effectively avoid the saturation failure phenomenon due to the wide measuring frequency band, the sound pick-up is convenient to install, and the influence of the installation mode on the signal is small. And the sound signal and the vibration signal belong to homologous signals, and are generated by the vibration of the parts of the circuit breaker, so that the homologous complementary characteristics of the sound signal and the vibration signal can be utilized, and the respective advantages are exerted to realize the charged monitoring. However, the difference of the two signals is not considered in the traditional acoustic vibration signal combination method, the acoustic vibration signals are mechanically combined, and although deep learning algorithms such as a convolutional neural network are introduced, the diagnosis accuracy is not high enough and the generalization is poor due to the loss of characteristic information and poor pertinence of a CNN general structure. How to invent a diagnostic method which can independently learn from end to end, has accurate classification and excellent generalization and has important research value.
Disclosure of Invention
The invention aims to provide a method for improving the state identification accuracy and generalization performance of a circuit breaker in the energy storage process, which is used for extracting the acoustic vibration signal characteristics and carrying out characteristic extraction and fault diagnosis by relying on the strong self-learning capability of CNN (CNN), wherein the key points are the structure of a circuit breaker acoustic vibration combined characteristic matrix and the optimization of a CNN model.
In order to achieve the purpose, the invention adopts the following technical scheme:
a circuit breaker energy storage process state identification method for a sound vibration signal to construct a CNN characteristic matrix includes the steps of firstly providing a time scale alignment method based on kurtosis and envelope similarity to guarantee synchronism of the sound vibration signal, then adopting a Lyapunov index-wavelet modulus maximum (L-wavelet) to detect a vibration signal initial point, conducting overlapped data expansion on data, and then constructing a sound vibration signal two-dimensional characteristic matrix by using a Pearson correlation coefficient. And finally, training the feature matrix by using the CNN, optimizing the CNN structure by using a Support Vector Machine (SVM) to replace a Soft-Max classifier, and searching for the optimal parameters of the SVM by using a Hui wolf optimization (GWO). The optimized CNN model is insensitive to the condition of large data change in the energy storage process of the circuit breaker, and the method is used as a new method for identifying the state of the energy storage process of the circuit breaker, so that the accuracy and the generalization of fault identification are greatly improved.
(1) And considering the asynchrony of the sound vibration signals, determining a time period with changing by adopting kurtosis, further calculating the envelope similarity by adopting a Minkowski formula, finding out the time difference with the highest similarity, and finally realizing sound vibration time scale alignment.
(2) In the actual operation process, a Lyapunov exponent-wavelet modulus maximum (L-wavelet) is adopted to detect the starting point of the vibration signal. (since the step (1) carries out time scale contraposition on the sound vibration, the method is also equivalent to the starting point of the sound signal detection)
(3) And the overlapped sample data expansion is adopted to provide a large amount of data required by CNN, and the dimensionality of the characteristic matrix is increased.
(4) And calculating the correlation coefficient of the sound-vibration signal sample by adopting the Pearson product moment correlation coefficient, and constructing a sound-vibration combined characteristic matrix by using the normalized correlation coefficient as a matrix element.
(5) According to the energy storage characteristic of the circuit breaker, an SVM is used for replacing Soft-Max, GWO is introduced for parameter optimization, a GWO-SVM classifier of a CNN full-connection layer is constructed, and a model structure is optimized.
And (2) determining the time period of the change by using kurtosis in the sound vibration time scale alignment in the step (1). The kurtosis is taken as a dimensionless parameter and is particularly sensitive to signal impact, so that the kurtosis can be used for detecting the peak degree of the envelope curve of the sound vibration signal. The kurtosis is calculated as follows:
wherein: x is the instantaneous value of the acoustic vibration envelope, mu is the average value of the envelope,
σ is the standard deviation for probability density.
And (3) calculating by using a Minkowski formula after determining the corresponding time period of the acoustic vibration signal:
wherein: a and b are n-dimensional acoustic vibration signal data points, and q is a distance adjustment parameter. And after the time with the highest similarity is found, the sound starting time minus the vibration starting time is delta T, and the sound signal is advanced by delta T to be aligned with the vibration signal.
In the L-wavelet initial point detection method in the step (2), as each mechanical component starts, moves and stops according to a certain sequence in the operation process of the circuit breaker, a series of shock wave superposed vibration signals are generated. As a nonlinear and non-stationary time sequence, the vibration signal in the energy storage process shows a high chaotic characteristic, at the moment, the Lyapunov exponent is positive, and the formula is defined as follows:
λ is the Lyapunov exponent in the prime mover system. If the energy storage is not carried out, the vibration signal is mainly vibration noise and does not have chaotic characteristics, and the Lyapunov exponent obtained at the moment is negative. Whether energy storage is started or not can be judged through the numerical value symbols, but only the initial time period can be determined, the specific initial time can be detected through the wavelet modulus maximum, the wavelet transformation is used as a variable time window, the amplitude mutation can be detected, and the calculation is as follows:
corresponding to wavelet transform, W1f (s, x) appears as a maximum.
The overlapped data capacity expansion in the step (3): for a signal x with the length of N, setting the sample length as L and the overlap ratio as lambda, and adopting the capacity expansion and division method as follows:
obtaining the maximum divisible sample number under the current signal length:
to round down the operator
Each segmented sample is evaluated. The position of the ith sample in the original signal is represented as follows:
xi＝X[(i-1)×L×(1-λ)+(0：1)×L]，i∈[1，n]
the short sample segmentation length can improve the convergence speed of the model and save the training time, but the loss of nonlinear characteristic information is easy to cause; too long a sample segmentation length may reduce the convergence rate of the model and affect the real-time performance of the diagnosis. The appropriate sample length and overlap ratio is selected.
Constructing a sound and vibration combined characteristic matrix in the step (4): the Pearson product-moment correlation coefficient is used for measuring the change relation between two variables, is independent of the specific values of the variables, and is a non-parameter statistic. If the mean values of the vibro-acoustic signals tend to be greater than or less than their respective mean values at the same time, the correlation coefficient is positive; if they both tend to fall on the opposite side of their mean, the correlation coefficient is negative. The calculation formula is as follows:
wherein: { xiI ═ 1, 2, …, n } and { y ═ yiAnd i ═ 1, 2, …, n } represents the values of the sound and vibration signal sample points, respectively. Pearson's coefficient range of [ -1, 1]Therefore, the directly calculated correlation coefficients are normalized and then filled into the feature matrix.
And (5) the CNN model consists of an input layer, a convolution layer, a pooling layer, a full-connection layer and an output layer. And defining a weight matrix in each layer to be convolved with the characteristic matrix, and outputting the convolution result of the previous layer to be the neuron for constructing the corresponding characteristic of the next layer through the activation function operation.
The convolutional layer performs convolution on input data by a convolution kernel, and constructs a feature vector by a nonlinear activation function. The same convolution kernel shares parameters in the convolution process to obtain a class of characteristics, and the formula of the calculation process is as follows:
wherein
Representing the l-th input, NjRepresents the input feature vector, l represents the l-th layer network, k represents the convolution kernel, and b represents the bias of the convolution kernel. The modified linear unit (ReLU) is often selected as the nonlinear activation function, so that part of neurons can output 0, the interdependence of parameters is reduced, the sparsity of the network is improved, and the overfitting problem is effectively inhibited. The calculation of ReLU is as follows:
wherein:
is that
The value of the activation of (a) is,
representing the output value of the convolution operation.
The pooling layer comprises average pooling and maximum pooling, and input data is subjected to scaling mapping through pooling to realize data reduction and feature extraction. The transformation function is as follows:
wherein: w is the width of the convolution kernel,
is the value of the t neuron in the ith feature of the ith layer,
the value for the l +1 th neuron.
The output layer of the CNN is fully connected with the output of the last pooling layer, and the model is as follows:
O＝f(bo+fvwo)
wherein: boIs a deviation vector, fvIs a feature vector, woIs a weight matrix.
Considering that the circuit breaker has violent vibration at the initial position and the final position in the energy storage process, the energy storage time has difference, which causes larger change of each measurement signal and limited generalization performance. The method of the invention classifies the full-link layer feature set by using the GWO-SVM method, and can optimize the CNN classification performance by fully utilizing the strong generalization capability of the SVM.
GWO As a new group intelligent optimization algorithm for simulating the social level system and the family group hunting behavior of the gray wolf family, 4 types of gray wolfs are defined in sequence from high to low according to the social status as alpha, beta, sigma and omega, the hunting process is guided by the alpha wolf, the beta wolf and the sigma wolf, and the omega wolf is responsible for tracking and hunting the hunting objects, the model is as follows:
in the formula: t represents the current iteration number; xp(t) is the prey position vector, and A and C are coefficient vectors. A and C are as follows:
in the formula: a is a convergence factor and satisfies a epsilon [0, 2 ]]；r1And r2Is [0, 1 ]]A random vector of (1).
Drawings
FIG. 1 is a flow chart of circuit breaker energy storage process state identification
FIG. 2 is a schematic diagram of a circuit breaker vibration leading sound signal
FIG. 3 is a schematic diagram of overlapping data expansion
FIG. 4 is a two-dimensional graph of a joint acoustic-vibration feature matrix
FIG. 5 is a flow chart of the GWO-SVM algorithm
FIG. 6 is a graph illustrating the convergence of curve GWO
FIG. 7 is a CNN model diagnostic diagram
Detailed Description
For the example of the ZN65-12 type breaker, the process for identifying the state of the energy storage process is shown in fig. 1. The specific embodiment of the invention is as follows:
And step 2, sequentially dividing the vibration signals, and setting the division length to be 300. Aiming at each segment of data, calculating Lyapunov indexes of each segment in sequence by using a small data method, carrying out phase space reconstruction on a vibration signal, searching nearest adjacent points of each track in a phase space, estimating the maximum Lyapunov index according to the average divergence rate of each point, and obtaining a final value by least square fitting after cyclic iteration and logarithm taking. The start time period may be determined according to its sign. And then, performing wavelet multi-scale transformation on the signals in the initial time period, searching from a high scale to a low scale, searching a modulus maximum line by using an adhoc algorithm, and taking a modulus maximum point on the minimum scale as a singular point. Test Condition p1≥(1+λ)p2. Wherein p is1、p2The lambda is an adjusting parameter for the analog greater number and the analog smaller number in the difference value before and after the operation starting point.
Dividing the sound vibration signal into N equal parts, calculating the kurtosis of each signal envelope, comparing the kurtosis values of each signal segment, and searching for signal segments with obvious kurtosis value change difference so as to determine the time of the change; then, the similarity is judged by q roots of the sum of the absolute difference q powers of the envelopes of the sound signal and the vibration signal, as shown in the following formula:
and q is 2, namely the plotted Euclidean distance. Finally, after the time with the highest similarity is found, adjusting delta T to realize sound vibration time scale alignment, and a schematic diagram is shown in FIG. 2.
And 4, adopting the overlapped sample data expansion to completely keep the correlation of adjacent samples, avoiding the characteristic loss caused by sample truncation, and simultaneously increasing the dimensionality of the matrix. The expansion schematic is shown in fig. 3. Through the verification of the method, the sample length is set to be 1024, and the overlapping rate lambda is set to be 0.5.
And 5, constructing a sound vibration joint matrix and calculating according to the following formula:
the acoustic vibration signals are respectively used as rows and columns of the matrix, and the correlation coefficients of the samples corresponding to the acoustic vibration signals are calculated according to the formula to form CNN characteristic matrix elements. The time from the torque output of the switch energy storage motor to the energy storage maintaining pawl energy maintaining is not more than 10s, so that the single signal acquisition time is set to be 10 s. 350 groups of data are collected in each state, each group comprises 50000 sampling points, and each group of sampling points form a 96 multiplied by 96 CNN characteristic matrix according to the expansion sampling parameter setting in the step 4 and the calculation formula in the step (3). And selecting 40 sample data continuously acquired under the condition of higher energy storage voltage to calculate a characteristic matrix, and drawing a two-dimensional schematic diagram, which is shown in figure 4.
And 6, the CNN model comprises the following structures and parameters: the number of CNN convolutional layers is 2, the parameters are set to be 32@2 x 2 and 64@2 x 2, the pooling layer adopts the maximum pooling with the size of 2 x 2 and the step length of 2, and the nodes of the two fully-connected layers are set to be 256 and 64. An RMSprop optimizer is adopted, the initial learning rate is set to be 0.03, and the attenuation rate is 0.99; and (5) iteration times are 50, dropout is set for the full connection layer, and the coefficient is 0.5. And (3) classifying the output layer of the CNN by using GWO-SVM, replacing a cost function by using a kernel function, and searching an optimal hyperplane in a high-dimensional space. The model structure is as follows:
TABLE 1 CNN Structure Table
CNN diagnosis environment, the model adopts software Python and Tensorflow; the operating system is Windows10, the processor is Intel (R) core (TM) i7-6850KCPU, the Nvidia Titan Xp GPU is equipped for acceleration, and the running memory is 8 GB.
GWO-SVM algorithm flow is shown in FIG. 5, the number of iterations is set to 100, and the GWO algorithm convergence process is shown in FIG. 6.
After the CNN optimization model is obtained, state recognition is started below. For the type of states mentioned in step 1, 350 sets of data are collected for each type of state, 220 sets of data are used for training, 130 sets of data are used for testing, and a diagnostic diagram of the model is shown in FIG. 7. Along with the increase of the training times, the identification accuracy of the model gradually rises, the loss value gradually falls, and the accuracy is not improved.
To the source and the structure difference of data among the actual circuit breaker monitoring process, verify this model generalization ability: the model ZN65-12 vacuum circuit breaker is used instead, the sampling rate is changed from 40kHz to 30kHz, the model ZD-530 piezoelectric sensor and the model TONY-A2 waterproof sound pickup are used instead, and the installation position of the sensor is changed simultaneously.
Claims (7)
1. The fault diagnosis method for the circuit breaker energy storage process by constructing the CNN characteristic matrix through the acoustic vibration signals is characterized by comprising the following steps of:
(1) and considering the asynchrony of the sound vibration signals, determining a time period with changing by adopting kurtosis, further calculating the envelope similarity by adopting a Minkowski formula, finding out the time difference with the highest similarity, and finally realizing sound vibration time scale alignment.
(2) In the actual operation process, a Lyapunov exponent-wavelet modulus maximum (L-wavelet) is adopted to detect the starting point of the vibration signal.
(3) And the overlapped sample data expansion is adopted to provide a large amount of data required by CNN, and the dimensionality of the characteristic matrix is increased.
(4) And calculating the correlation coefficient of the sound-vibration signal sample by adopting the Pearson product moment correlation coefficient, and constructing a sound-vibration combined characteristic matrix by using the normalized correlation coefficient as a matrix element.
(5) According to the energy storage characteristic of the circuit breaker, an SVM is used for replacing Soft-Max, GWO is introduced for parameter optimization, a GWO-SVM classifier of a CNN full-connection layer is constructed, and a model structure is optimized.
2. The method for diagnosing the fault of the energy storage mechanism of the circuit breaker for constructing the CNN characteristic matrix according to the sound vibration signals, as claimed in claim 1, is characterized by providing a time scale alignment method for kurtosis and envelope similarity. Firstly, dividing the sound vibration signal into N equal parts, and calculating the kurtosis of each signal envelope as follows:
comparing the kurtosis values of each segment of signals, and searching for signal segments with obvious kurtosis value change differences so as to determine the time of occurrence of the change; the similarity is then determined from the q-th root of the sum of the absolute difference q-th powers of the sound and vibration signal envelopes, as shown in the following equation.
3. The method for diagnosing the fault of the circuit breaker energy storage mechanism by constructing the CNN characteristic matrix according to the vibro-acoustic signals, as claimed in claim 1, is characterized in that an L-wavelet initial point detection method is provided. Judging whether the detection signal has chaotic characteristics or not by utilizing the Lyapunov index, wherein the vibration signal shows high chaotic characteristics in the energy storage process, the Lyapunov index is positive at the moment, and the formula is defined as follows:
after the Lyapunov index determines the starting time period, the specific starting time can be detected through a wavelet modulus maximum, and the wavelet transformation is used as a variable time window to detect the amplitude mutation, and the calculation is as follows:
corresponding to wavelet transform, W1f (s, x) appears as a maximum.
4. The method for diagnosing the fault of the circuit breaker energy storage mechanism with the acoustic-vibration signal construction CNN characteristic matrix according to claim 1, characterized in that a nonparametric statistic Pearson product moment correlation coefficient is used for measuring the variation relation of the acoustic-vibration signal, and the formula is as follows:
and the sound and vibration signals are respectively used as the rows and the columns of the matrix, the correlation coefficient of the corresponding sample of the sound and vibration signals is calculated according to the formula, and the normalized correlation coefficient is filled in the characteristic matrix to form CNN characteristic matrix elements.
5. The method for diagnosing the fault of the circuit breaker energy storage mechanism by constructing the CNN characteristic matrix according to the sound vibration signals as claimed in claim 1, is characterized in that the CNN structure is improved according to the characteristics of severe vibration of the starting and ending positions, difference existing in energy storage time and large data fluctuation in the energy storage process of the circuit breaker, an SVM is adopted to replace Soft-Max, meanwhile GWO is introduced to carry out parameter optimization, a GWO-SVM classifier of a CNN full connection layer is constructed, and the model structure is optimized.
6. The diagnosis method as claimed in claim 1, wherein the electrical signal variation of voltage level during the energy storage process can be identified by the non-electrical signal characteristics such as sound vibration, thereby greatly reducing the complexity of the live detection during the energy storage process of the circuit breaker.
7. Other electrical devices are also within the scope of the patent claims, according to the method steps of claim 1. The related adjustment, alteration, addition and deletion of the steps mentioned in the step 1 belong to the patent requirements.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910643289.1A CN112241605B (en) | 2019-07-17 | 2019-07-17 | Breaker energy storage process state identification method for constructing CNN feature matrix by using acoustic vibration signals |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910643289.1A CN112241605B (en) | 2019-07-17 | 2019-07-17 | Breaker energy storage process state identification method for constructing CNN feature matrix by using acoustic vibration signals |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112241605A true CN112241605A (en) | 2021-01-19 |
| CN112241605B CN112241605B (en) | 2023-12-19 |
Family
ID=74167571
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910643289.1A Active CN112241605B (en) | 2019-07-17 | 2019-07-17 | Breaker energy storage process state identification method for constructing CNN feature matrix by using acoustic vibration signals |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112241605B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113239075A (en) * | 2021-05-13 | 2021-08-10 | 中国公路工程咨询集团有限公司 | Construction data self-checking method and system |
| CN113361592A (en) * | 2021-06-03 | 2021-09-07 | 哈尔滨工业大学 | Acoustic event identification method based on public subspace representation learning |
| CN114088364A (en) * | 2021-09-30 | 2022-02-25 | 广西电网有限责任公司电力科学研究院 | Breaker mechanical fault diagnosis method based on SystemVue acoustic signal separation |
| CN114088274A (en) * | 2021-11-19 | 2022-02-25 | 中国直升机设计研究所 | Amplitude-phase comprehensive correlation identification method for bending moment identification of helicopter main shaft |
| CN116627040A (en) * | 2023-05-23 | 2023-08-22 | 滁州市伟博电气有限公司 | Dryer control system and method thereof |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104458297A (en) * | 2014-11-12 | 2015-03-25 | 南京航空航天大学 | Fault detection method for train suspension system sensor based on nonlinearity random model |
| US20180357542A1 (en) * | 2018-06-08 | 2018-12-13 | University Of Electronic Science And Technology Of China | 1D-CNN-Based Distributed Optical Fiber Sensing Signal Feature Learning and Classification Method |
| WO2019090878A1 (en) * | 2017-11-09 | 2019-05-16 | 合肥工业大学 | Analog circuit fault diagnosis method based on vector-valued regularized kernel function approximation |
| CN110926782A (en) * | 2019-12-06 | 2020-03-27 | 国网河南省电力公司三门峡供电公司 | Circuit breaker fault type judgment method and device, electronic equipment and storage medium |
| CN111879397A (en) * | 2020-09-01 | 2020-11-03 | 国网河北省电力有限公司检修分公司 | Fault diagnosis method for energy storage mechanism of high-voltage circuit breaker |
| CN112326210A (en) * | 2019-07-17 | 2021-02-05 | 华北电力大学（保定） | Large motor fault diagnosis method combining sound vibration signals with 1D-CNN |
-
2019
- 2019-07-17 CN CN201910643289.1A patent/CN112241605B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104458297A (en) * | 2014-11-12 | 2015-03-25 | 南京航空航天大学 | Fault detection method for train suspension system sensor based on nonlinearity random model |
| WO2019090878A1 (en) * | 2017-11-09 | 2019-05-16 | 合肥工业大学 | Analog circuit fault diagnosis method based on vector-valued regularized kernel function approximation |
| US20180357542A1 (en) * | 2018-06-08 | 2018-12-13 | University Of Electronic Science And Technology Of China | 1D-CNN-Based Distributed Optical Fiber Sensing Signal Feature Learning and Classification Method |
| CN112326210A (en) * | 2019-07-17 | 2021-02-05 | 华北电力大学（保定） | Large motor fault diagnosis method combining sound vibration signals with 1D-CNN |
| CN110926782A (en) * | 2019-12-06 | 2020-03-27 | 国网河南省电力公司三门峡供电公司 | Circuit breaker fault type judgment method and device, electronic equipment and storage medium |
| CN111879397A (en) * | 2020-09-01 | 2020-11-03 | 国网河北省电力有限公司检修分公司 | Fault diagnosis method for energy storage mechanism of high-voltage circuit breaker |
Non-Patent Citations (6)
| Title |
|---|
| JINBAO ZHANG;YONGQIANG ZHAO;LINGXIAN KONG;MING LIU;: "Morphology Similarity Distance for Bearing Fault Diagnosis Based on Multi-Scale Permutation Entropy", no. 01 * |
| 小GANYMEDES: "闵科夫斯基(minkovski)距离" * |
| 张伟伟;武静;马宏伟;: "基于Lyapunov指数的超声导波检测技术", 振动.测试与诊断, no. 02, pages 60 - 67 * |
| 张思聪;傅攀;蒋恩超;朱奥辉;: "QPSO-WT和QPSO-SVM在滚动轴承故障诊断中的应用", 机械与电子, no. 05, pages 35 - 38 * |
| 李金辉;李杰;余佩倡;王连春;: "Adaptive backstepping control for levitation system with load uncertainties and external disturbances", no. 12, pages 74 - 84 * |
| 赵书涛;王二旭;陈秀新;王科登;李小双;: "声振信号联合1D-CNN的大型电机故障诊断方法", no. 09, pages 116 - 122 * |
Cited By (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN113239075A (en) * | 2021-05-13 | 2021-08-10 | 中国公路工程咨询集团有限公司 | Construction data self-checking method and system |
| CN113361592A (en) * | 2021-06-03 | 2021-09-07 | 哈尔滨工业大学 | Acoustic event identification method based on public subspace representation learning |
| CN113361592B (en) * | 2021-06-03 | 2022-11-08 | 哈尔滨工业大学 | Acoustic event identification method based on public subspace representation learning |
| CN114088364A (en) * | 2021-09-30 | 2022-02-25 | 广西电网有限责任公司电力科学研究院 | Breaker mechanical fault diagnosis method based on SystemVue acoustic signal separation |
| CN114088274A (en) * | 2021-11-19 | 2022-02-25 | 中国直升机设计研究所 | Amplitude-phase comprehensive correlation identification method for bending moment identification of helicopter main shaft |
| CN114088274B (en) * | 2021-11-19 | 2023-06-06 | 中国直升机设计研究所 | Amplitude-phase comprehensive correlation identification method for helicopter main shaft bending moment identification |
| CN116627040A (en) * | 2023-05-23 | 2023-08-22 | 滁州市伟博电气有限公司 | Dryer control system and method thereof |
| CN116627040B (en) * | 2023-05-23 | 2024-04-02 | 滁州市伟博电气有限公司 | Dryer control system and method thereof |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112241605B (en) | 2023-12-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112241605B (en) | Breaker energy storage process state identification method for constructing CNN feature matrix by using acoustic vibration signals | |
| CN113343633B (en) | Thermal runaway fault classification and risk prediction method and system for power lithium battery | |
| US20230385617A1 (en) | Signal processing method for neuron in spiking neural network and method for training said network | |
| CN111487046B (en) | Fault diagnosis method for circuit breaker voiceprint and vibration entropy feature fusion | |
| Liang et al. | Rolling bearing fault diagnosis based on one-dimensional dilated convolution network with residual connection | |
| CN111879397B (en) | Fault diagnosis method for energy storage mechanism of high-voltage circuit breaker | |
| Pan et al. | An approach for HVCB mechanical fault diagnosis based on a deep belief network and a transfer learning strategy | |
| Han et al. | Fault diagnosis system of induction motors based on neural network and genetic algorithm using stator current signals | |
| CN113673346A (en) | Motor vibration data processing and state recognition method based on multi-scale SE-Resnet | |
| CN104569747A (en) | System and method for checking insulativity of power-off cable | |
| CN114254668A (en) | Fault detection method and device based on discharge signal of insulated switchgear | |
| WO2023226355A1 (en) | Dual-ion battery fault detection method and system based on multi-source perception | |
| CN114417948A (en) | Method for predicting thermal runaway of battery based on neural network | |
| CN116226646A (en) | Method, system, equipment and medium for predicting health state and residual life of bearing | |
| CN113203914A (en) | Underground cable early fault detection and identification method based on DAE-CNN | |
| CN114118225A (en) | Method, system, electronic device and storage medium for predicting remaining life of generator | |
| CN114091525A (en) | Rolling bearing degradation trend prediction method | |
| CN109901064B (en) | ICA-LVQ-based high-voltage circuit breaker fault diagnosis method | |
| CN114881106A (en) | Transformer fault diagnosis method and device based on MPA-SVM | |
| CN113987910A (en) | Method and device for identifying load of residents by coupling neural network and dynamic time planning | |
| Aggarwal et al. | A novel hybrid architecture for classification of power quality disturbances | |
| KR20220023545A (en) | Apparatus and method for diagnosing partial discharge | |
| CN116383630A (en) | Probabilistic neural network arc fault detection method based on improved wolf algorithm | |
| CN116298725A (en) | Fault arc detection method, system and storage medium | |
| CN116400168A (en) | Power grid fault diagnosis method and system based on depth feature clustering |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |