CN111398798A - Circuit breaker energy storage state identification method based on vibration signal interval characteristic extraction - Google Patents
Circuit breaker energy storage state identification method based on vibration signal interval characteristic extraction Download PDFInfo
- Publication number
- CN111398798A CN111398798A CN202010147845.9A CN202010147845A CN111398798A CN 111398798 A CN111398798 A CN 111398798A CN 202010147845 A CN202010147845 A CN 202010147845A CN 111398798 A CN111398798 A CN 111398798A
- Authority
- CN
- China
- Prior art keywords
- energy storage
- feature
- circuit breaker
- vibration signal
- storage state
- 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
- 238000004146 energy storage Methods 0.000 title claims abstract description 113
- 238000000034 method Methods 0.000 title claims abstract description 56
- 238000000605 extraction Methods 0.000 title claims abstract description 27
- 239000013598 vector Substances 0.000 claims abstract description 43
- 238000012216 screening Methods 0.000 claims abstract description 10
- 238000007689 inspection Methods 0.000 claims abstract description 8
- 238000004364 calculation method Methods 0.000 claims description 22
- 239000011159 matrix material Substances 0.000 claims description 22
- 238000012549 training Methods 0.000 claims description 16
- 238000013507 mapping Methods 0.000 claims description 10
- 238000011156 evaluation Methods 0.000 claims description 4
- 238000005070 sampling Methods 0.000 claims description 4
- 230000009466 transformation Effects 0.000 claims description 4
- 238000001514 detection method Methods 0.000 claims description 3
- 238000001276 Kolmogorov–Smirnov test Methods 0.000 claims description 2
- 238000012544 monitoring process Methods 0.000 abstract description 8
- 238000011160 research Methods 0.000 abstract description 5
- 238000003745 diagnosis Methods 0.000 abstract description 4
- 230000008569 process Effects 0.000 description 16
- 238000012706 support-vector machine Methods 0.000 description 8
- 238000010586 diagram Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 230000002159 abnormal effect Effects 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 238000005315 distribution function Methods 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 230000009467 reduction Effects 0.000 description 2
- 230000005236 sound signal Effects 0.000 description 2
- OAICVXFJPJFONN-UHFFFAOYSA-N Phosphorus Chemical compound [P] OAICVXFJPJFONN-UHFFFAOYSA-N 0.000 description 1
- 230000001174 ascending effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 239000004744 fabric Substances 0.000 description 1
- 238000000338 in vitro Methods 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 238000012545 processing Methods 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
- G01R31/327—Testing of circuit interrupters, switches or circuit-breakers
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M13/00—Testing of machine parts
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/08—Feature extraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Arc-Extinguishing Devices That Are Switches (AREA)
- Breakers (AREA)
Abstract
The invention discloses a circuit breaker energy storage state identification method based on vibration signal interval characteristic extraction, which comprises the following steps: extracting an energy storage vibration signal of the circuit breaker, and determining a starting point of the energy storage vibration signal; performing KS inspection on the extracted energy storage vibration signal of the breaker to obtain a feature vector; screening the feature vectors to obtain an optimal feature subset; and carrying out state identification on the preferred feature subset based on a KFCM-SVM algorithm to obtain an identification result of the energy storage state of the circuit breaker. In the implementation of the method, the characteristic extraction only needs 0.2s on the premise of ensuring the accuracy, and the method has important research value on the on-line monitoring of the circuit breaker; the method is a novel method for practical diagnosis of the circuit breaker under the field operation environment, shortens the diagnosis time on the premise of not sacrificing the accuracy, and has obvious advantages compared with the traditional circuit breaker state identification method.
Description
Technical Field
The invention relates to the technical field of electrical equipment fault diagnosis, in particular to a circuit breaker energy storage state identification method based on vibration signal interval characteristic extraction.
Background
There are a large number of scholars who have studied circuit breaker operating mechanisms, but there is a relative lack of status identification in the energy storage process. The high-voltage circuit breaker is used as a closing switch, power is provided for the opening and closing actions of the circuit breaker through the energy storage spring, faults such as voltage fluctuation, jamming of a transmission mechanism, falling of the energy storage spring, failure of a limit switch and the like frequently occur in the energy storage process due to the fact that connection among components and the control process are complex, and the high-voltage circuit breaker has important research value for accurately monitoring the energy storage state.
The existing breaker fault diagnosis method mainly comprises coil current, sound signals and vibration signals, wherein the coil current is difficult to comprehensively reflect various mechanical faults, the sound signals are convenient to collect but need complex signal preprocessing to influence the calculation speed, the vibration signals belong to in-vitro monitoring, the signal-to-noise ratio is high and rich in state information, the sensor is convenient to install and can collect signals without auxiliary power supply, and the field real-time monitoring is facilitated.
Disclosure of Invention
The invention aims to overcome the defects of the prior art, and provides a circuit breaker energy storage state identification method based on vibration signal interval feature extraction.
In order to solve the technical problem, an embodiment of the present invention provides a method for identifying an energy storage state of a circuit breaker based on vibration signal interval feature extraction, where the method includes:
extracting an energy storage vibration signal of the circuit breaker, and determining a starting point of the energy storage vibration signal;
performing KS inspection on the extracted energy storage vibration signal of the breaker to obtain a feature vector;
screening the feature vectors to obtain an optimal feature subset;
and carrying out state identification on the preferred feature subset based on a KFCM-SVM algorithm to obtain an identification result of the energy storage state of the circuit breaker.
Optionally, the extracting an energy storage vibration signal of the circuit breaker, and determining a starting point of the energy storage vibration signal includes:
calculating to obtain the envelope of the energy storage vibration signal of the circuit breaker;
dividing the envelope of the energy storage vibration signal of the circuit breaker into a plurality of continuous intervals, and calculating to obtain the kurtosis of the plurality of continuous intervals;
comparing kurtosis values of envelopes of each interval, and determining a rough time segment signal according to the kurtosis values;
performing wavelet transformation on the rough time period signal to obtain a modulus maximum line;
and determining a starting point of the energy storage vibration signal based on the modulus maximum value line.
Optionally, in the envelope of the energy storage vibration signal of the circuit breaker obtained through calculation, a specific calculation formula is as follows:
wherein, y (t) represents Hilbert transform, x (t) represents an energy storage vibration signal of the breaker, t represents t time, tau represents time delay, z (t) represents an analysis signal, and m (t) represents an envelope of the signal.
Optionally, in the step of obtaining the kurtosis of the consecutive intervals through calculation, a specific calculation formula is as follows:
where k denotes a kurtosis, e (x) denotes an expected value of a corresponding vibration signal, μ denotes an envelope mean, and σ denotes a standard deviation.
Optionally, the extracting the energy storage vibration signal of the circuit breaker is performed with KS inspection, and obtaining the feature vector includes:
marking the energy storage vibration signal of the extraction breaker through KS detection to obtain a marked signal;
analyzing the marked signal to obtain an obvious envelope amplitude difference interval;
and summing the envelope mean values of the interval with obvious envelope amplitude difference to obtain a feature vector.
Optionally, the screening the feature vectors to obtain an optimal feature subset includes:
performing initial selection on the feature vector based on a Relieff algorithm to obtain a feature vector after the initial selection;
and removing irrelevant feature quantities from the initially selected feature vectors based on an SFS algorithm to obtain an optimal feature subset.
Optionally, the initially selecting the feature vector based on the ReliefF algorithm to obtain the initially selected feature vector includes:
selecting one sample R from the training set of the feature vectors each time based on a Relieff algorithm;
respectively finding out k adjacent R of the sample R from the samples of the same type and different types as the sample R1、R2And calculating to obtain the weight of each feature a;
and obtaining the initially selected feature vector based on the weight of each feature a.
Optionally, in the obtaining of the weight of each feature through calculation, a specific calculation formula is as follows:
wherein diff (a, R)1) Represents R and R1The distance on the characteristic a, c represents the class of the R, P (c) represents the prior probability of c, m represents the random sampling frequency, j represents the jth number, R represents the sample R, k represents k adjacent R of the sample R respectively found from the samples of the same class and different classes with the sample R1、R2A represents each feature, w (a) represents a weight of each feature, R (a) represents the features a, R in the sample R1(a) Represents a sample R1The feature a in (1), max (a), represents the maximum value of the feature a, and min (a), represents the minimum value of the feature a.
Optionally, the performing state identification on the preferred feature subset based on the KFCM-SVM algorithm to obtain the identification result of the energy storage state of the circuit breaker includes:
pre-classifying the optimal feature subset through KFCM, and establishing membership mapping between the optimal feature subset and the fault category;
after the mapping is established, training is carried out through an SVM, and a training result is obtained through comprehensive evaluation;
and obtaining the identification result of the energy storage state of the circuit breaker based on the training result.
Optionally, the pre-classifying the optimal feature subset by the KFCM, and establishing a membership degree mapping between the optimal feature subset and the fault category includes:
setting fuzzy coefficients, the number of categories, function types and parameters and target function precision based on the optimal feature subset;
initializing a membership matrix based on the fuzzy coefficient, the number of classes, the function type and parameters and the target function precision;
after initialization is completed, a clustering center is obtained through calculation, and an updated membership matrix is obtained;
comparing the updated membership matrix with the membership matrix;
stopping iteration if convergence occurs;
if not, returning to the initial state, obtaining a clustering center through calculation, and obtaining an updated membership matrix.
In the implementation of the invention, a circuit breaker energy storage state identification method based on vibration signal interval feature extraction is characterized by firstly detecting the initial point of a circuit breaker energy storage state according to kurtosis-wavelet modulus maximum to extract an energy storage signal, then marking an interval with obvious envelope amplitude difference by KS (K-class-boundary) inspection on the vibration signal, extracting a signal envelope and using the signal envelope as a feature vector, screening and reducing the dimension of the feature by adopting a Relieff-SFS (robust Fourier transform-sparse-State) method to obtain an optimal feature subset, finally performing pre-classification on the feature by fuzzy C-means clustering (KFCM) to obtain an optimal hyperplane with minimum risk, and establishing a training model by using a Support Vector Machine (SVM) to perform state identification; on the premise of ensuring the accuracy, the characteristic extraction only needs 0.2s, and the method has important research value for on-line monitoring of the circuit breaker.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
Fig. 1 is a schematic flow chart of a method for identifying an energy storage state of a circuit breaker based on vibration signal interval feature extraction in the implementation of the present invention;
FIG. 2 is a time domain waveform amplitude diagram of a normal energy storage state, a high voltage energy storage state, a low voltage energy storage state, and a spring-off energy storage state in an implementation of the present invention;
FIG. 3 is an envelope graph of the normal energy storage state, the high voltage energy storage state, the low voltage energy storage state, and the spring-off energy storage state in an implementation of the present invention;
fig. 4 is a flow chart of KFCM-SVM diagnostics in the practice of the invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Examples
Referring to fig. 1, fig. 1 is a schematic flow chart illustrating a method for identifying an energy storage state of a circuit breaker based on vibration signal interval feature extraction in an implementation of the present invention.
As shown in fig. 1, a method for identifying an energy storage state of a circuit breaker based on vibration signal interval feature extraction includes:
s11: extracting an energy storage vibration signal of the circuit breaker, and determining a starting point of the energy storage vibration signal;
in a specific implementation process of the present invention, the extracting an energy storage vibration signal of a circuit breaker and determining a starting point of the energy storage vibration signal includes: calculating to obtain the envelope of the energy storage vibration signal of the circuit breaker; dividing the envelope of the energy storage vibration signal of the circuit breaker into a plurality of continuous intervals, and calculating to obtain the kurtosis of the plurality of continuous intervals; comparing kurtosis values of envelopes of each interval, and determining a rough time segment signal according to the kurtosis values; performing wavelet transformation on the rough time period signal to obtain a modulus maximum line; and determining a starting point of the energy storage vibration signal based on the modulus maximum value line.
Specifically, the method for determining the starting point of the energy storage vibration signal comprises the following steps:
firstly, the energy storage vibration signal x (t) is used for solving an envelope, and the calculation formula is as follows:
set y (t) as its Hilbert Transform (HT), we can get:
by HT, an analytic signal z (t) is obtained, whose modulus m (t) is the envelope of the signal, as shown below:
dividing the envelope module value into N continuous intervals, calculating the kurtosis of each interval, comparing the kurtosis value of the signal envelope of each interval, and searching for the interval with obvious kurtosis value change difference so as to determine the rough time of the change; the kurtosis calculation formula is as follows:
wherein E (x) corresponds to the expected value of the vibration signal, mu is the envelope mean value, and sigma is the standard deviation;
performing wavelet transformation on the rough time period signal determined by the kurtosis, wherein a modulus maximum value on each scale is related to a singular point; searching layer by layer from a high scale to a low scale by adopting an adhoc algorithm to obtain a modulus maximum value line; taking the obtained maximum value point of the mode on the minimum scale as a singular point, wherein the corresponding moment is the starting point of the vibration signal; in order to correct the singular point position deviation phenomenon, searching a module maximum value point with the same symbol in upper and lower adjacent scales, and taking two points at the left end and the right end as candidate catastrophe points; the selected starting point is discriminated according to the following equation:
Q1≥(1+λ)Q2；
wherein Q1And Q2The method is characterized in that the lambda is an adjusting parameter, and the lambda is a greater number and a smaller number in the difference values before and after the initial point of the stored energy.
After the starting point is determined, the duration of the monitoring signal can judge whether the energy storage process is performed. The general duration of the opening and closing process does not exceed 0.5s, and the time from the moment that the switch energy storage motor is pressed down to the moment that the energy storage and the pawl energy are kept is more than 1.5s in the energy storage process, which is the basis of judging time.
S12: performing KS inspection on the extracted energy storage vibration signal of the breaker to obtain a feature vector;
in a specific implementation process of the present invention, the extracting the energy storage vibration signal of the circuit breaker and performing KS inspection to obtain the feature vector includes: marking the energy storage vibration signal of the extraction breaker through KS detection to obtain a marked signal; analyzing the marked signal to obtain an obvious envelope amplitude difference interval; and summing the envelope mean values of the interval with obvious envelope amplitude difference to obtain a feature vector.
Specifically, the KS test extracts features which are insensitive to noise, so that a complex denoising process can be omitted; the empirical distribution functions of samples X and Y are set to f (X) and g (X), respectively, and the test problem is as follows:
according to glifnko, Smimov test statistics were used:
wherein m, n represent the number of samples, X (i) and Y (i) represent the order statistics of X and Y, respectively, H0D corresponds to a significance level α represented by the reliability distribution function Rs:
wherein
If D is greater than the significance level α, it indicates that the two samples are differentiatedAnd (3) cloth.
S13: screening the feature vectors to obtain an optimal feature subset;
in a specific implementation process of the present invention, the screening the feature vectors to obtain an optimal feature subset includes: performing initial selection on the feature vector based on a Relieff algorithm to obtain a feature vector after the initial selection; and removing irrelevant feature quantities from the initially selected feature vectors based on an SFS algorithm to obtain an optimal feature subset.
When it needs to be explained, the initially selecting the feature vector based on the ReliefF algorithm to obtain the initially selected feature vector includes: selecting one sample R from the training set of the feature vectors each time based on a Relieff algorithm; respectively finding out k adjacent R of the sample R from the samples of the same type and different types as the sample R1、R2And calculating to obtain the weight of each feature a; and obtaining the initially selected feature vector based on the weight of each feature a.
Specifically, in the step of obtaining the weight of each feature through calculation, a specific calculation formula is as follows:
wherein diff (a, R)1) Represents R and R1The distance on the characteristic a, c represents the class of the R, P (c) represents the prior probability of c, m represents the random sampling frequency, j represents the jth number, R represents the sample R, k represents k adjacent R of the sample R respectively found from the samples of the same class and different classes with the sample R1、R2A represents each feature, w (a) represents a weight of each feature, R (a) represents the features a, R in the sample R1(a) Represents a sample R1The feature a in (1), max (a), represents the maximum value of the feature a, and min (a), represents the minimum value of the feature a.
In addition, the feature evaluation criterion function is set to:
in the formula: m is the number of sample classes, μiAnd mujThe mean of the intra-class feature vectors for the class i and class j samples,
and
the within-class variances of the i-th class and the j-th class of samples are represented, respectively.
S14: and carrying out state identification on the preferred feature subset based on a KFCM-SVM algorithm to obtain an identification result of the energy storage state of the circuit breaker.
In a specific implementation process of the present invention, the performing state identification on the preferred feature subset based on a KFCM-SVM algorithm to obtain an identification result of the energy storage state of the circuit breaker includes: pre-classifying the optimal feature subset through KFCM, and establishing membership mapping between the optimal feature subset and the fault category; after the mapping is established, training is carried out through an SVM, and a training result is obtained through comprehensive evaluation; and obtaining an identification result of the energy storage state of the circuit breaker based on the training result.
It should be noted that the pre-classifying the optimal feature subset by the KFCM, and establishing the membership degree mapping between the optimal feature subset and the fault category includes: setting fuzzy coefficients, the number of categories, function types and parameters and target function precision based on the optimal feature subset; initializing a membership matrix based on the fuzzy coefficient, the number of classes, the function type and parameters and the target function precision; after initialization is completed, a clustering center is obtained through calculation, and an updated membership matrix is obtained; comparing the updated membership matrix with the membership matrix; stopping iteration if convergence occurs; if not, returning to the initial state, obtaining a clustering center through calculation, and obtaining an updated membership matrix.
Specifically, for a data set X composed of n vectors, the number c of classes to be classified is set, and a membership matrix is defined, the process is as follows:
(1) setting a fuzzy coefficient m, a category number c, a kernel function type and parameters, wherein the target function precision is as follows;
(2) initializing the membership degree matrix to accord with normalization;
(3) calculating a clustering center:
in the formula, ciFor the ith group of fuzzy clustering centers, uijIs xjIn the membership value of the i-th class, u is not less than 0ijLess than or equal to 1; using a gaussian kernel, K (X, Y) ═ exp (— | | X-Y | | non-conducting phosphor2/σ2)；
Updating the membership degree matrix:
and comparing the iterated membership degree matrix according to the matrix norm, stopping iteration if the iterated membership degree matrix is converged, and returning to the previous step if the iterated membership degree matrix is not converged.
In the specific implementation, the energy storage experiment is carried out on a ZN65-12 type breaker in a laboratory, and the specific implementation mode of the invention is as follows:
And (2) detecting the starting point by adopting the method in the step (1), and judging whether the energy storage process is performed or not by monitoring the signal duration after the starting point is determined. The general duration of the opening and closing process does not exceed 0.5s, and the time from the moment that the switch energy storage motor is pressed down to the moment that the energy storage and the pawl energy are kept is more than 1.5s in the energy storage process, which is the basis of judging time.
And 3, extracting features by adopting the method provided by the step 2, setting the interval width to be 5ms, and setting the energy storage time to be not more than 5s, so that the number of the intervals is set to be 1000, the significance level α is set to be 0.01, the distribution condition of the sample envelope amplitude in each state is respectively checked, the intervals with different envelope distributions are marked, the number of the marking intervals of the four states is shown in table 1, the probability densities of the normal state and the abnormal state are obviously different, and the normal state and the abnormal state can be used as the preferred feature quantity.
TABLE 1 four energy storage State flag intervals
| Status type | Number of marked intervals/number |
| Normal state | 78 |
| The energy storage voltage is higher | 196 |
| Low energy storage voltage | 310 |
| Energy storage bulletSpring drop-out | 243 |
And 4, the SFS evaluates the original feature set features of the function processing (the target subset is empty initially) according to the established rule, and selects the added target feature set with the maximum function value by comparing the feature function values. The non-selected features are matched with the selected features and criterion function values FDR for the matched combined features are calculated (see step 3) and arranged in ascending order. And continuously adding and comparing the feature subsets primarily screened by the Relieff algorithm, and repeating the steps until a set value is reached.
And selecting iteration times m and neighbor number k by adopting a grid search method, iterating the Relieff for m times to obtain feature table weight, setting m and k as 100 and 10 respectively, setting a Relieff threshold η as 0.1, carrying out primary feature screening, carrying out dimension reduction on the selected feature set according to SFS, and taking intersection from the labeled feature interval after dimension reduction as a final feature vector.
And 5, pre-classifying the feature set through KFCM, establishing membership degree mapping between the feature set and the fault category, training through SVM on the basis, and comprehensively evaluating to obtain a final result. Using a clustering validity index lambdaKFCMThe checking is carried out, and the checking is carried out,
uic is the membership degree of the C sample to the i class, C is the number of clusters, and N is the number of samples) algorithm flow is shown in fig. 4, and fig. 4 shows a KFCM-SVM diagnostic flow chart in the implementation of the invention.
The clustering number C is set to 4, the SVM (C-SVC) adopts the RBF radial basis function, the penalty factor C is set to 3.8, and the kernel function parameter g is set to 0.15. Setting a normal state sample characteristic value label as 1(1-10 groups), setting the energy storage voltage higher as 2(11-20 groups), setting the energy storage voltage lower as 3(21-30 groups), setting the spring falling as 4(31-40 groups), carrying out state identification, and verifying the rapidity of the method.
In the implementation of the invention, a circuit breaker energy storage state identification method based on vibration signal interval feature extraction is characterized by firstly detecting the initial point of a circuit breaker energy storage state according to kurtosis-wavelet modulus maximum to extract an energy storage signal, then marking an interval with obvious envelope amplitude difference by KS (K-class-boundary) inspection on the vibration signal, extracting a signal envelope and using the signal envelope as a feature vector, screening and reducing the dimension of the feature by adopting a Relieff-SFS (robust Fourier transform-sparse-State) method to obtain an optimal feature subset, finally performing pre-classification on the feature by fuzzy C-means clustering (KFCM) to obtain an optimal hyperplane with minimum risk, and establishing a training model by using a Support Vector Machine (SVM) to perform state identification; on the premise of ensuring the accuracy, the characteristic extraction only needs 0.2s, and the method has important research value for on-line monitoring of the circuit breaker.
In addition, the method for identifying the energy storage state of the circuit breaker based on the vibration signal interval feature extraction provided by the embodiment of the invention is described in detail, a specific embodiment is adopted to explain the principle and the implementation mode of the invention, and the description of the embodiment is only used for helping to understand the method and the core idea of the invention; meanwhile, for a person skilled in the art, according to the idea of the present invention, there may be variations in the specific embodiments and the application scope, and in summary, the content of the present specification should not be construed as a limitation to the present invention.
Claims (10)
1. A circuit breaker energy storage state identification method based on vibration signal interval feature extraction is characterized by comprising the following steps:
extracting an energy storage vibration signal of the circuit breaker, and determining a starting point of the energy storage vibration signal;
performing KS inspection on the extracted energy storage vibration signal of the breaker to obtain a feature vector;
screening the feature vectors to obtain an optimal feature subset;
and carrying out state identification on the preferred feature subset based on a KFCM-SVM algorithm to obtain an identification result of the energy storage state of the circuit breaker.
2. The method for identifying the energy storage state of the circuit breaker based on the interval feature extraction of the vibration signal as claimed in claim 1, wherein the extracting the energy storage vibration signal of the circuit breaker and determining the starting point of the energy storage vibration signal comprises:
calculating to obtain the envelope of the energy storage vibration signal of the circuit breaker;
dividing the envelope of the energy storage vibration signal of the circuit breaker into a plurality of continuous intervals, and calculating to obtain the kurtosis of the plurality of continuous intervals;
comparing kurtosis values of envelopes of each interval, and determining a rough time segment signal according to the kurtosis values;
performing wavelet transformation on the rough time period signal to obtain a modulus maximum line;
and determining a starting point of the energy storage vibration signal based on the modulus maximum value line.
3. The method for identifying the energy storage state of the circuit breaker based on the vibration signal interval feature extraction as claimed in claim 2, wherein in the envelope of the energy storage vibration signal of the circuit breaker obtained through calculation, a specific calculation formula is as follows:
wherein, y (t) represents Hilbert transform, x (t) represents an energy storage vibration signal of the breaker, t represents t time, tau represents time delay, z (t) represents an analysis signal, and m (t) represents an envelope of the signal.
4. The method as claimed in claim 2, wherein the kurtosis of the consecutive intervals is obtained by calculation according to the following formula:
where k denotes a kurtosis, e (x) denotes an expected value of a corresponding vibration signal, μ denotes an envelope mean, and σ denotes a standard deviation.
5. The method for identifying the energy storage state of the circuit breaker based on the vibration signal interval feature extraction as claimed in claim 1, wherein the step of performing a KS test on the extracted energy storage vibration signal of the circuit breaker to obtain the feature vector comprises:
marking the energy storage vibration signal of the extraction breaker through KS detection to obtain a marked signal;
analyzing the marked signal to obtain an obvious envelope amplitude difference interval;
and summing the envelope mean values of the interval with obvious envelope amplitude difference to obtain a feature vector.
6. The method for identifying the energy storage state of the circuit breaker based on the vibration signal interval feature extraction as claimed in claim 1, wherein the screening the feature vectors to obtain the optimal feature subset comprises:
performing initial selection on the feature vector based on a Relieff algorithm to obtain a feature vector after the initial selection;
and removing irrelevant feature quantities from the initially selected feature vectors based on an SFS algorithm to obtain an optimal feature subset.
7. The method for identifying the energy storage state of the circuit breaker based on the vibration signal interval feature extraction as claimed in claim 6, wherein the performing the initial selection on the feature vector based on the Relieff algorithm to obtain the initially selected feature vector comprises:
selecting one sample R from the training set of the feature vectors each time based on a Relieff algorithm;
respectively finding out k adjacent R of the sample R from the samples of the same type and different types as the sample R1、R2And obtaining the value of each feature a by calculationA weight;
and obtaining the initially selected feature vector based on the weight of each feature a.
8. The method for identifying the energy storage state of the circuit breaker based on the vibration signal interval feature extraction as claimed in claim 7, wherein in the step of obtaining the weight of each feature through calculation, a specific calculation formula is as follows:
wherein diff (a, R)1) Represents R and R1The distance on the characteristic a, c represents the class of the R, P (c) represents the prior probability of c, m represents the random sampling frequency, j represents the jth number, R represents the sample R, k represents k adjacent R of the sample R respectively found from the samples of the same class and different classes with the sample R1、R2A represents each feature, w (a) represents a weight of each feature, R (a) represents the features a, R in the sample R1(a) Represents a sample R1The feature a in (1), max (a), represents the maximum value of the feature a, and min (a), represents the minimum value of the feature a.
9. The method for identifying the energy storage state of the circuit breaker based on the interval feature extraction of the vibration signal as claimed in claim 1, wherein the performing state identification on the preferred feature subset based on the KFCM-SVM algorithm to obtain the identification result of the energy storage state of the circuit breaker comprises:
pre-classifying the optimal feature subset through KFCM, and establishing membership mapping between the optimal feature subset and the fault category;
after the mapping is established, training is carried out through an SVM, and a training result is obtained through comprehensive evaluation;
and obtaining the identification result of the energy storage state of the circuit breaker based on the training result.
10. The method for identifying the energy storage state of the circuit breaker based on the vibration signal interval feature extraction as claimed in claim 9, wherein the pre-classifying the optimal feature subset through KFCM and establishing the membership degree mapping between the optimal feature subset and the fault category comprises:
setting fuzzy coefficients, the number of categories, function types and parameters and target function precision based on the optimal feature subset;
initializing a membership matrix based on the fuzzy coefficient, the number of classes, the function type and parameters and the target function precision;
after initialization is completed, a clustering center is obtained through calculation, and an updated membership matrix is obtained;
comparing the updated membership matrix with the membership matrix;
stopping iteration if convergence occurs;
if not, returning to the initial state, obtaining a clustering center through calculation, and obtaining an updated membership matrix.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010147845.9A CN111398798B (en) | 2020-03-05 | 2020-03-05 | Circuit breaker energy storage state identification method based on vibration signal interval feature extraction |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010147845.9A CN111398798B (en) | 2020-03-05 | 2020-03-05 | Circuit breaker energy storage state identification method based on vibration signal interval feature extraction |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111398798A true CN111398798A (en) | 2020-07-10 |
| CN111398798B CN111398798B (en) | 2022-08-19 |
Family
ID=71428671
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010147845.9A Active CN111398798B (en) | 2020-03-05 | 2020-03-05 | Circuit breaker energy storage state identification method based on vibration signal interval feature extraction |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111398798B (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111855178A (en) * | 2020-07-23 | 2020-10-30 | 贵州永红航空机械有限责任公司 | Diagnosis method for running state of rotary product |
| CN112380932A (en) * | 2020-11-02 | 2021-02-19 | 上海三菱电梯有限公司 | Vibration signal characteristic value selection method and elevator health state evaluation or fault diagnosis method |
| CN112557004A (en) * | 2020-11-23 | 2021-03-26 | 广西电网有限责任公司电力科学研究院 | Method for detecting mechanical fault of high-voltage circuit breaker |
| CN113405788A (en) * | 2021-05-28 | 2021-09-17 | 广西电网有限责任公司电力科学研究院 | On-load tap-changer mechanical state monitoring method based on waveform trend information |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103336243A (en) * | 2013-07-01 | 2013-10-02 | 东南大学 | Breaker fault diagnosis method based on separating/closing coil current signals |
| CN105891707A (en) * | 2016-05-05 | 2016-08-24 | 河北工业大学 | Opening-closing fault diagnosis method for air circuit breaker based on vibration signals |
| CN106646205A (en) * | 2015-10-30 | 2017-05-10 | 国网山西省电力公司电力科学研究院 | Random big-disturbance signal removing algorithm for analyzing circuit breaker fault through sound and vibration combination |
| KR101881411B1 (en) * | 2017-01-18 | 2018-07-24 | (주)다쓰테크 | Arc detection and shutdown method in pv system and pv energy storage system |
| CN109164315A (en) * | 2018-08-03 | 2019-01-08 | 河海大学 | The Reactor Fault diagnostic method of KFCM and SVM based on particle group optimizing |
| CN109444728A (en) * | 2018-09-21 | 2019-03-08 | 国网河南省电力公司济源供电公司 | A kind of circuit breaker failure diagnostic method based on dynamic weighting Hybrid Clustering Algorithm |
-
2020
- 2020-03-05 CN CN202010147845.9A patent/CN111398798B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103336243A (en) * | 2013-07-01 | 2013-10-02 | 东南大学 | Breaker fault diagnosis method based on separating/closing coil current signals |
| CN106646205A (en) * | 2015-10-30 | 2017-05-10 | 国网山西省电力公司电力科学研究院 | Random big-disturbance signal removing algorithm for analyzing circuit breaker fault through sound and vibration combination |
| CN105891707A (en) * | 2016-05-05 | 2016-08-24 | 河北工业大学 | Opening-closing fault diagnosis method for air circuit breaker based on vibration signals |
| KR101881411B1 (en) * | 2017-01-18 | 2018-07-24 | (주)다쓰테크 | Arc detection and shutdown method in pv system and pv energy storage system |
| CN109164315A (en) * | 2018-08-03 | 2019-01-08 | 河海大学 | The Reactor Fault diagnostic method of KFCM and SVM based on particle group optimizing |
| CN109444728A (en) * | 2018-09-21 | 2019-03-08 | 国网河南省电力公司济源供电公司 | A kind of circuit breaker failure diagnostic method based on dynamic weighting Hybrid Clustering Algorithm |
Non-Patent Citations (5)
| Title |
|---|
| MEI FEI: "Classification and recognition of voltage sags based on KFCM - SVM", 《2017 IEEE 13TH INTERNATIONAL CONFERENCE ON ELECTRONIC MEASUREMENT & INSTRUMENTS》 * |
| XIAOFEI XIA: "Energy Storage State Identification Of Circuit Breaker Based On Fast Extraction Of Interval Feature Of Vibration Signal", 《EARTH AND ENVIRONMENTAL SCIENCE》 * |
| 夏小飞: "基于振动信号区间特征快速提取的断路器储能状态辨识方法", 《中国电力》 * |
| 杨元威: "基于声音信号的高压断路器机械故障诊断方法", 《中国电机工程学报》 * |
| 王亚潇: "基于声振信号联合分析的高压断路器状态识别方法研究", 《中国优秀博硕士学位论文全文数据库(硕士)工程科技Ⅱ辑》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111855178A (en) * | 2020-07-23 | 2020-10-30 | 贵州永红航空机械有限责任公司 | Diagnosis method for running state of rotary product |
| CN112380932A (en) * | 2020-11-02 | 2021-02-19 | 上海三菱电梯有限公司 | Vibration signal characteristic value selection method and elevator health state evaluation or fault diagnosis method |
| CN112557004A (en) * | 2020-11-23 | 2021-03-26 | 广西电网有限责任公司电力科学研究院 | Method for detecting mechanical fault of high-voltage circuit breaker |
| CN112557004B (en) * | 2020-11-23 | 2022-06-14 | 广西电网有限责任公司电力科学研究院 | Method for detecting mechanical fault of high-voltage circuit breaker |
| CN113405788A (en) * | 2021-05-28 | 2021-09-17 | 广西电网有限责任公司电力科学研究院 | On-load tap-changer mechanical state monitoring method based on waveform trend information |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111398798B (en) | 2022-08-19 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111398798B (en) | Circuit breaker energy storage state identification method based on vibration signal interval feature extraction | |
| Liu et al. | Classifying transformer winding deformation fault types and degrees using FRA based on support vector machine | |
| CN107505133B (en) | The probability intelligent diagnosing method of rolling bearing fault based on adaptive M RVM | |
| EP3832553A1 (en) | Method for identifying energy of micro-energy device on basis of bp neural network | |
| CN101907681B (en) | Analog circuit dynamic online failure diagnosing method based on GSD-SVDD | |
| CN110110768A (en) | Fault Diagnosis of Roller Bearings based on Concurrent Feature study and multi-categorizer | |
| CN103077402A (en) | Transformer partial-discharging mode recognition method based on singular value decomposition algorithm | |
| CN106405384A (en) | Simulation circuit health state evaluation method | |
| CN111881781B (en) | Transformer winding deformation classification method based on scanning impedance method and support vector machine | |
| CN114781458B (en) | Power distribution network initial fault identification method based on adaptive wavelet and attention mechanism time-frequency memory neural network | |
| CN115905832A (en) | Transformer fault diagnosis method based on vibration signal clustering identification | |
| CN115098962A (en) | A Prediction Method for Remaining Life of Mechanical Equipment in Degraded State Based on Hidden Semi-Marve Model | |
| CN117195077A (en) | An unsupervised detection method for power transformer voiceprint signal faults | |
| CN109580260A (en) | A kind of inferior health diagnostic method of track vehicle door system | |
| CN112560674B (en) | Method and system for detecting sound signal quality | |
| CN109598048A (en) | A kind of lubrication degradation prediction technique of track vehicle door system | |
| CN112880726A (en) | Sensor fault diagnosis method and device based on variational modal decomposition sample entropy | |
| CN111507534A (en) | Student learning data-based predictive analysis algorithm for knowledge point mastering conditions | |
| Lv et al. | A weight recognition method for movable objects in sealed cavity based on supervised learning | |
| CN109840386A (en) | Damnification recognition method based on factorial analysis | |
| Chen et al. | Probabilistic neural network based apple classification prediction | |
| CN119357562B (en) | Integrated circuit package sealing evaluation method and system under high temperature environment | |
| Lv et al. | Recognition Method for Electronic Component Signals Based on LR-SMOTE and Improved Random Forest Algorithm | |
| Soulsbury et al. | A rules-based approach to classification of chemical acoustic emission signals | |
| Vyas et al. | Anomaly Detection in Electric Powertrain System Software Behaviour |
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 |