CN112557966A - Transformer winding looseness identification method based on local mean decomposition and support vector machine - Google Patents

Transformer winding looseness identification method based on local mean decomposition and support vector machine Download PDF

Info

Publication number
CN112557966A
CN112557966A CN202011393866.5A CN202011393866A CN112557966A CN 112557966 A CN112557966 A CN 112557966A CN 202011393866 A CN202011393866 A CN 202011393866A CN 112557966 A CN112557966 A CN 112557966A
Authority
CN
China
Prior art keywords
vector machine
transformer
support vector
decomposition
signal
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN202011393866.5A
Other languages
Chinese (zh)
Inventor
许洪华
张勇
马宏忠
李勇
颜锦
王春宁
刘宝稳
陈寿龙
王立宪
顾仲翔
朱雷
朱昊
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Hohai University HHU
Nanjing Power Supply Co of State Grid Jiangsu Electric Power Co Ltd
Original Assignee
Hohai University HHU
Nanjing Power Supply Co of State Grid Jiangsu Electric Power Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Hohai University HHU, Nanjing Power Supply Co of State Grid Jiangsu Electric Power Co Ltd filed Critical Hohai University HHU
Priority to CN202011393866.5A priority Critical patent/CN112557966A/en
Publication of CN112557966A publication Critical patent/CN112557966A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
    • G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
    • G01R31/50—Testing of electric apparatus, lines, cables or components for short-circuits, continuity, leakage current or incorrect line connections
    • G01R31/62—Testing of transformers
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01H—MEASUREMENT OF MECHANICAL VIBRATIONS OR ULTRASONIC, SONIC OR INFRASONIC WAVES
    • G01H11/00—Measuring mechanical vibrations or ultrasonic, sonic or infrasonic waves by detecting changes in electric or magnetic properties
    • G01H11/06—Measuring mechanical vibrations or ultrasonic, sonic or infrasonic waves by detecting changes in electric or magnetic properties by electric means
    • G01H11/08—Measuring mechanical vibrations or ultrasonic, sonic or infrasonic waves by detecting changes in electric or magnetic properties by electric means using piezoelectric devices
    • 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
    • G01R31/50—Testing of electric apparatus, lines, cables or components for short-circuits, continuity, leakage current or incorrect line connections
    • G01R31/72—Testing of electric windings
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/211—Selection of the most significant subset of features
    • G06F18/2113—Selection of the most significant subset of features by ranking or filtering the set of features, e.g. using a measure of variance or of feature cross-correlation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
    • 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]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V10/00—Arrangements for image or video recognition or understanding
    • G06V10/40—Extraction of image or video features
    • G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/02—Preprocessing
    • 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

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Data Mining & Analysis (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • General Engineering & Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Evolutionary Computation (AREA)
  • Artificial Intelligence (AREA)
  • Power Engineering (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Evolutionary Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Biophysics (AREA)
  • Health & Medical Sciences (AREA)
  • Biomedical Technology (AREA)
  • Multimedia (AREA)
  • Computational Linguistics (AREA)
  • General Health & Medical Sciences (AREA)
  • Molecular Biology (AREA)
  • Computing Systems (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Testing Of Short-Circuits, Discontinuities, Leakage, Or Incorrect Line Connections (AREA)

Abstract

The invention discloses a transformer winding looseness identification method based on local mean decomposition and a support vector machine, which comprises the following steps: step 1: respectively collecting vibration signals of a normal state and a winding loosening state at the moment of closing a transformer; step 2: carrying out variation modal decomposition on the acquired vibration signals to obtain each PF component; and step 3: calculating the energy and singular value of each PF component and the permutation entropy and singular spectrum entropy of the reconstructed signal; and 4, step 4: selecting features with higher precision by a Fisher-Score method to form a feature vector group; and 5: training a support vector machine model for simulating annealing optimization by using a training sample set; step 6: and using the obtained support vector machine model as a classifier to classify and identify the test sample set so as to realize fault diagnosis. The method can identify the loosening state of the transformer winding at the moment of closing the transformer, realizes early warning of the transformer, and provides a new method for extracting the vibration signal characteristics of the transformer and diagnosing faults.

Description

Transformer winding looseness identification method based on local mean decomposition and support vector machine
Technical Field
The invention belongs to the technical field of power transformer mechanical fault diagnosis, and particularly relates to a transformer winding looseness identification method based on local mean decomposition and a support vector machine.
Background
The transformer is in electric power system, and the structure is complicated, and the price is high, has key effect to electric power system's safe and stable operation, in case the transformer breaks down and causes the power failure, will lead to great economic loss. Therefore, the method and the device can monitor and diagnose the running condition of the transformer, find and eliminate hidden troubles of the fault in advance, improve the reliability of the transformer and have important significance for ensuring the reliable running of the transformer.
Due to long operation time, the transformer is easy to have various faults, such as iron core looseness, winding deformation, deformation and the like. At present, the method for monitoring the mechanical fault states of windings, iron cores and the like of the power transformer comprises the following steps: short circuit impedance method, low voltage pulse method, frequency response analysis method. The short-circuit impedance method has long test time and low precision, and needs to invest a large amount of analysis and verification work. The frequency response method has the defects of diagnosis lag, failure in effectively diagnosing the mechanical fault of the transformer and the like. Among them, in recent years, the vibration analysis method has become a hot spot for diagnosing a mechanical failure of a transformer. Compared with other methods, the vibration method is not electrically connected with the transformer, has stronger anti-interference capability and reliably monitors the state of the transformer. Therefore, the invention provides a transformer winding looseness identification method based on local mean decomposition and a support vector machine by using a vibration signal at the switching-on moment of a transformer.
Disclosure of Invention
Aiming at the problems, the invention provides a transformer winding looseness identification method based on local mean decomposition and a support vector machine, and solves the problem of misjudgment of a diagnosis method in the prior art.
In order to solve the technical problems, the technical scheme provided by the invention is as follows: a transformer winding looseness identification method based on local mean decomposition and a support vector machine comprises the following steps:
step 1: respectively collecting vibration signals of a normal state and a winding loosening state at the moment of closing a transformer;
step 2: performing local mean decomposition on the acquired vibration signals to obtain PF components;
vibration signal at the moment of closing power transformer
Figure 395397DEST_PATH_IMAGE002
Decomposition into n PF components and a residual functionRAnd (c) the sum, i.e.:
Figure 843696DEST_PATH_IMAGE004
；
and step 3: calculating the energy value and singular value of each PF component and the permutation entropy and singular spectrum entropy value of the reconstructed signal;
and 4, step 4: selecting features with higher precision by a Fisher-Score method to form a feature vector group;
fisher Score is defined as:
Figure 623433DEST_PATH_IMAGE006
；
in the formula:
Figure 549801DEST_PATH_IMAGE008
as the ith feature of the sample
Figure 364173DEST_PATH_IMAGE010
The mean value of (a);
Figure 717794DEST_PATH_IMAGE012
for the ith feature of the sample
Figure 984828DEST_PATH_IMAGE010
Mean in class k; h is the total number of samples;
Figure DEST_PATH_IMAGE014
the number of kth type samples;
Figure DEST_PATH_IMAGE016
the value of the jth sample for the ith feature;
Figure DEST_PATH_IMAGE018
the smaller, the characteristic
Figure 246045DEST_PATH_IMAGE010
Faults can be identified;
and 5: dividing the vibration signal into a training sample set and a testing sample set, and training a support vector machine model for simulating annealing optimization by using the training sample set;
step 6: and using the obtained support vector machine model as a classifier to classify and identify the test sample set so as to realize fault diagnosis.
The scheme is further improved in that: in the step 2, the decomposition refers to separating a pure frequency modulation signal and an envelope signal from a complex original signal, and multiplying the pure frequency modulation signal and the envelope signal to obtain a PF function component with an instantaneous frequency having a physical meaning.
The scheme is further improved in that: in the step 3, the time series of each PF component after the local mean decomposition is discretized, and the energy E of each PF component is calculatediForming an energy feature vector; forming a component matrix by K PF components as an initial matrix, and performing singular value decomposition on the initial matrix to obtain characteristic singular values of different frequencies; and screening each PF component of the local mean decomposition, and reconstructing a signal. And calculating permutation entropy and singular spectrum entropy by using the construction signal.
The scheme is further improved in that: in the third step, PF component energy value EiThe solution is as follows:
Figure DEST_PATH_IMAGE020
；
in the formula: n is each sampling point; m is the total number of sample points.
The PF component energy values are normalized as follows:
Figure 446082DEST_PATH_IMAGE022
；
the energy feature vector can be obtained as:
Figure 439446DEST_PATH_IMAGE024
；
(2) the PF component singular values are solved as follows:
forming PF components into a feature vector matrix
Figure 724933DEST_PATH_IMAGE026
And carrying out singular value decomposition on the U to obtain a characteristic value capable of reflecting the state of the transformer winding.
(3) The permutation entropy of the reconstructed signal is solved as follows:
setting a threshold value, screening each PF component subjected to local mean decomposition, eliminating components with low correlation degree with the original signal, and recombining the remaining components to form a signal.
The permutation entropy is defined as:
Figure 461945DEST_PATH_IMAGE028
；
in the formula:
Figure 516489DEST_PATH_IMAGE030
is the embedding dimension; piIs the probability of a different sequence in the i-th sequence.
(4) The singular spectrum entropy of the reconstructed signal reflects the energy distribution of the signal, and the more concentrated the energy distribution is, the larger the entropy is; the more discrete the energy distribution, the smaller the entropy; i.e. to characterize different states of the transformer winding.
The scheme is further improved in that: in the step 4, a characteristic quantity with identification power is selected by using a Fisher-Score method, and the variance between the characteristic quantity and the same class sample is as small as possible
The scheme is further improved in that: the fifth step is as follows:
step 5.1: setting the initial value number, penalty factor and upper and lower limits of kernel function parameters of Metropolis;
step 5.2: generating random numbers in the search range of the kernel function and the parameters to form an original structure of a support vector machine, training the initial support vector machine by using a training sample set, classifying a detection sample set by using the vector machine model, and calculating the accuracy;
step 5.3: generating a random disturbance in the range of the kernel function and the parameters, updating the kernel function and the parameters, training the kernel function and the parameters by using a sample set to obtain a new model corresponding to the parameters of the support vector machine, classifying the sample set and calculating the accuracy;
step 5.4: accepting or discarding new kernel functions and parameters according to Metropolis criteria;
step 5.5: repeating the step 5.3, the step 5.4 and the step 5.5 to complete the iteration process;
step 5.6: judging whether the optimal value meets the requirement, if not, gradually reducing the annealing temperature, and repeating the steps 5.3, 5.4 and 5.5 until the condition is met; otherwise, outputting the penalty factor and the kernel function parameter, and supporting the vector machine model and the classification result.
The beneficial results of the invention are: when the power transformer is switched on and excitation inrush current is generated, an acceleration sensor is used for collecting vibration signals at the moment of switching on the transformer, so that the diagnosis sample is sufficient, and the early warning and other advantages are achieved. In addition, the excitation inrush current has complex waveform, and the information of vibration signals after impacting a winding is rich, so that fault analysis is facilitated, and the power transformer is prevented from continuously operating in a fault state;
according to the method, nonlinear and non-stable complex signals at the moment of switching on the transformer can be efficiently decomposed by local mean decomposition, compared with a set empirical mode decomposition algorithm, the method can avoid mode aliasing, and is high in accuracy. The energy and singular value of each PF component, the arrangement entropy and singular spectrum entropy of the reconstructed signal contain fault features of the transformer, and features which can reflect faults most can be selected from the feature vectors by utilizing a Fisher-Score method, so that the fault diagnosis capability is improved;
the method utilizes the simulated annealing algorithm to automatically find the optimal kernel function and the parameters thereof, is applied to the problem of identifying the winding looseness at the moment of switching on the transformer, reduces the difficulty of manually selecting the parameters to a certain extent, and improves the calculation efficiency.
Drawings
FIG. 1 is a flow chart of an embodiment of the present invention.
Fig. 2 is a diagram of an equivalent circuit model of a reactor turn-to-turn short circuit according to an embodiment of the present invention.
Detailed Description
The technical scheme of the invention is explained in detail in the following with the accompanying drawings.
The invention provides a transformer winding loosening fault identification method based on local mean decomposition and a support vector machine, and the specific flow is shown in figure 1, and the method comprises the following steps:
step 1: and respectively acquiring vibration signals of a normal state and a winding loosening state at the moment of closing the transformer.
Step 2: and carrying out local mean decomposition on the acquired vibration signals to obtain each PF component.
Vibration signal at the moment of closing power transformer
Figure 946333DEST_PATH_IMAGE002
Decomposition into n PF components and a residual functionRAnd (c) the sum, i.e.:
Figure 719117DEST_PATH_IMAGE004
；
and step 3: and calculating the energy value and singular value of each PF component and the arrangement entropy and singular spectrum entropy value of the reconstructed signal.
In the third step, PF component energy value EiThe solution is as follows:
Figure 994241DEST_PATH_IMAGE020
；
in the formula: n is each sampling point; m is the total number of sample points.
The PF component energy values are normalized as follows:
Figure DEST_PATH_IMAGE031
；
the energy feature vector can be obtained as:
Figure 700029DEST_PATH_IMAGE024
；
(2) the PF component singular values are solved as follows:
forming PF components into a feature vector matrix
Figure 35195DEST_PATH_IMAGE032
And carrying out singular value decomposition on the U to obtain a characteristic value capable of reflecting the state of the transformer winding.
(3) The permutation entropy of the reconstructed signal is solved as follows:
setting a threshold value, screening each PF component subjected to local mean decomposition, eliminating components with low correlation degree with the original signal, and recombining the remaining components to form a signal.
The permutation entropy is defined as:
Figure 764117DEST_PATH_IMAGE028
；
in the formula:
Figure 374089DEST_PATH_IMAGE030
is the embedding dimension; piIs the probability of a different sequence in the i-th sequence.
(4) The singular spectrum entropy of the reconstructed signal reflects the energy distribution of the signal, and the more concentrated the energy distribution is, the larger the entropy is; the more discrete the energy distribution, the smaller the entropy; i.e. to characterize different states of the transformer winding.
And 4, step 4: and selecting the features with higher precision by a Fisher-Score method to form a feature vector group.
Fisher Score is defined as:
Figure 872067DEST_PATH_IMAGE006
；
in the formula:
Figure 174872DEST_PATH_IMAGE008
as the ith feature of the sample
Figure 125511DEST_PATH_IMAGE010
The mean value of (a);
Figure 539175DEST_PATH_IMAGE012
for the ith feature of the sample
Figure 891658DEST_PATH_IMAGE010
Mean in class k; h is the total number of samples;
Figure 365365DEST_PATH_IMAGE014
the number of kth type samples;
Figure DEST_PATH_IMAGE033
the value of the jth sample for the ith feature;
Figure 334458DEST_PATH_IMAGE018
the smaller, the characteristic
Figure 755075DEST_PATH_IMAGE010
Faults can be identified;
and 5: dividing the vibration signal into a training sample set and a testing sample set, and training a support vector machine model for simulating annealing optimization by using the training sample set;
the fifth step is as follows:
step 5.1: setting the initial value number, penalty factor and upper and lower limits of kernel function parameters of Metropolis;
step 5.2: generating random numbers in the search range of the kernel function and the parameters to form an original structure of a support vector machine, training the initial support vector machine by using a training sample set, classifying a detection sample set by using the vector machine model, and calculating the accuracy;
step 5.3: generating a random disturbance in the range of the kernel function and the parameters, updating the kernel function and the parameters, training the kernel function and the parameters by using a sample set to obtain a new model corresponding to the parameters of the support vector machine, classifying the sample set and calculating the accuracy;
step 5.4: accepting or discarding new kernel functions and parameters according to Metropolis criteria;
step 5.5: repeating the step 5.3, the step 5.4 and the step 5.5 to complete the iteration process;
step 5.6: judging whether the optimal value meets the requirement, if not, gradually reducing the annealing temperature, and repeating the steps 5.3, 5.4 and 5.5 until the condition is met; otherwise, outputting the penalty factor and the kernel function parameter, and supporting the vector machine model and the classification result.
Step 6: and using the obtained support vector machine model as a classifier to classify and identify the test sample set so as to realize fault diagnosis.
The experimental transformer is an SFZ10-31500/110 oil-immersed power transformer, a JF-2020 piezoelectric acceleration sensor is adopted to measure a vibration signal right above the transformer, and the sampling frequency is set to be 10 kHz. In the experimental process, in order to reduce the influence of the excitation inrush current generated at the moment of switching on the result, two faults of incomplete loosening and complete loosening of a transformer winding are set, and signals are measured five times respectively.
And dividing the data into a training sample set and a testing sample set, wherein the training sample set is used for training the support vector machine, and the testing sample set is used for testing the model. Training is performed by using a support vector machine based on simulated annealing optimization, and as shown in fig. 2, the obtained optimal classification model is as follows: the penalty factor is 0.25 and the kernel function parameter is 4. In order to verify the effect of the method, collective empirical mode decomposition is carried out respectively under two fault states of the transformer, the energy and singular value of the transformer and the arrangement entropy and singular value entropy of the reconstructed signal are calculated by utilizing the obtained modal components, the characteristics are selected by using a Fisher-Score method, the optimal characteristics are obtained, a support vector machine for simulating annealing optimization is used for training, and the fault diagnosis condition of the method is tested. In addition, in order to compare the effectiveness of feature selection, all features obtained by calculation after the transformer no-load closing vibration signal is subjected to local mean decomposition are input into a support vector machine for training and testing. The results show that the diagnostic accuracy of the present invention is better than the latter two methods.
The above embodiments are only for illustrating the technical idea of the present invention, and the protection scope of the present invention is not limited thereby, and any modifications made on the basis of the technical scheme according to the technical idea of the present invention fall within the protection scope of the present invention.

Claims (6)

1. A transformer winding looseness identification method based on local mean decomposition and a support vector machine is characterized by comprising the following steps:
step 1: respectively collecting vibration signals of a normal state and a winding loosening state at the moment of closing a transformer;
step 2: performing local mean decomposition on the acquired vibration signals to obtain PF components;
vibration signal at the moment of closing power transformer
Figure 976288DEST_PATH_IMAGE001
Decomposition into n PF components and a residual functionRAnd (c) the sum, i.e.:
Figure 805704DEST_PATH_IMAGE002
；
and step 3: calculating the energy value and singular value of each PF component and the permutation entropy and singular spectrum entropy value of the reconstructed signal;
and 4, step 4: selecting features with higher precision by a Fisher-Score method to form a feature vector group;
fisher Score is defined as:
Figure 786299DEST_PATH_IMAGE003
；
in the formula:
Figure 421679DEST_PATH_IMAGE004
as the ith feature of the sample
Figure 781116DEST_PATH_IMAGE005
The mean value of (a);
Figure 673111DEST_PATH_IMAGE006
for the ith feature of the sample
Figure 750789DEST_PATH_IMAGE005
Mean in class k; h is the total number of samples;
Figure 252177DEST_PATH_IMAGE007
the number of kth type samples;
Figure 731700DEST_PATH_IMAGE008
the value of the jth sample for the ith feature;
Figure 293131DEST_PATH_IMAGE009
the smaller, the characteristic
Figure 858105DEST_PATH_IMAGE005
Faults can be identified;
and 5: dividing the vibration signal into a training sample set and a testing sample set, and training a support vector machine model for simulating annealing optimization by using the training sample set;
step 6: and using the obtained support vector machine model as a classifier to classify and identify the test sample set so as to realize fault diagnosis.
2. The method for identifying the looseness of the winding of the transformer based on the local mean decomposition and the support vector machine according to claim 1, wherein the method comprises the following steps: in the step 2, the decomposition refers to separating a pure frequency modulation signal and an envelope signal from a complex original signal, and multiplying the pure frequency modulation signal and the envelope signal to obtain a PF function component with an instantaneous frequency having a physical meaning.
3. The method for identifying the looseness of the winding of the transformer based on the local mean decomposition and the support vector machine according to claim 1, wherein the method comprises the following steps: in the step 3, the time series of each PF component after the local mean decomposition is discretized, and the energy E of each PF component is calculatediForming an energy feature vector; forming a component matrix by K PF components as an initial matrix, and performing singular value decomposition to obtain features of different frequenciesCharacterizing singular values; and screening each PF component of the local mean decomposition, reconstructing a signal, and calculating the permutation entropy and the singular spectrum entropy by using the constructed signal.
4. The method for identifying the looseness of the winding of the transformer based on the local mean decomposition and the support vector machine according to claim 3, wherein the method comprises the following steps: in the third step, PF component energy value EiThe solution is as follows:
Figure 163184DEST_PATH_IMAGE010
；
in the formula: n is each sampling point; m is the total number of sampling points;
the PF component energy values are normalized as follows:
Figure 293951DEST_PATH_IMAGE011
；
the energy feature vector can be obtained as:
Figure 901650DEST_PATH_IMAGE012
；
(2) the PF component singular values are solved as follows:
forming PF components into a feature vector matrix
Figure 568300DEST_PATH_IMAGE013
Carrying out singular value decomposition on the U to obtain a characteristic value capable of reflecting the state of the transformer winding;
(3) the permutation entropy of the reconstructed signal is solved as follows:
setting a threshold value, screening each PF component subjected to local mean decomposition, eliminating components with low correlation degree with the original signal, and recombining the remaining components to form a signal;
the permutation entropy is defined as:
Figure 286857DEST_PATH_IMAGE014
；
in the formula:
Figure 600027DEST_PATH_IMAGE015
is the embedding dimension; piIs the probability of the different sequences in the i;
(4) the singular spectrum entropy of the reconstructed signal reflects the energy distribution of the signal, and the more concentrated the energy distribution is, the larger the entropy is; the more discrete the energy distribution, the smaller the entropy; i.e. to characterize different states of the transformer winding.
5. The method for identifying the looseness of the winding of the transformer based on the local mean decomposition and the support vector machine according to claim 1, wherein the method comprises the following steps: in the step 4, a characteristic quantity with identification power is selected by using a Fisher-Score method, and the variance between the characteristic quantity and the same class sample is as small as possible.
6. The method for identifying the looseness of the winding of the transformer based on the local mean decomposition and the support vector machine according to claim 1, wherein the method comprises the following steps: the fifth step is as follows:
step 5.1: setting the initial value number, penalty factor and upper and lower limits of kernel function parameters of Metropolis;
step 5.2: generating random numbers in the search range of the kernel function and the parameters to form an original structure of a support vector machine, training the initial support vector machine by using a training sample set, classifying a detection sample set by using the vector machine model, and calculating the accuracy;
step 5.3: generating a random disturbance in the range of the kernel function and the parameters, updating the kernel function and the parameters, training the kernel function and the parameters by using a sample set to obtain a new model corresponding to the parameters of the support vector machine, classifying the sample set and calculating the accuracy;
step 5.4: accepting or discarding new kernel functions and parameters according to Metropolis criteria;
step 5.5: repeating the step 5.3, the step 5.4 and the step 5.5 to complete the iteration process;
step 5.6: judging whether the optimal value meets the requirement, if not, gradually reducing the annealing temperature, and repeating the steps 5.3, 5.4 and 5.5 until the condition is met; otherwise, outputting the penalty factor and the kernel function parameter, and supporting the vector machine model and the classification result.
CN202011393866.5A 2020-12-02 2020-12-02 Transformer winding looseness identification method based on local mean decomposition and support vector machine Pending CN112557966A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011393866.5A CN112557966A (en) 2020-12-02 2020-12-02 Transformer winding looseness identification method based on local mean decomposition and support vector machine

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011393866.5A CN112557966A (en) 2020-12-02 2020-12-02 Transformer winding looseness identification method based on local mean decomposition and support vector machine

Publications (1)

Publication Number Publication Date
CN112557966A true CN112557966A (en) 2021-03-26

Family

ID=75047451

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011393866.5A Pending CN112557966A (en) 2020-12-02 2020-12-02 Transformer winding looseness identification method based on local mean decomposition and support vector machine

Country Status (1)

Country Link
CN (1) CN112557966A (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113252322A (en) * 2021-04-27 2021-08-13 国网安徽省电力有限公司亳州供电公司 Method and device for diagnosing mechanical fault of circuit breaker under noise-containing background
CN113391244A (en) * 2021-06-13 2021-09-14 河海大学 VMD-based transformer switching-on vibration signal characteristic frequency calculation method
CN114088400A (en) * 2021-11-01 2022-02-25 中国人民解放军92728部队 Rolling bearing fault diagnosis method based on envelope permutation entropy
CN114325480A (en) * 2021-11-19 2022-04-12 广东核电合营有限公司 Diode open-circuit fault detection method and device for multiphase brushless exciter
CN116757100A (en) * 2023-08-18 2023-09-15 国科大杭州高等研究院 Thrust prediction model training and thrust prediction method, device, equipment and medium

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107832687A (en) * 2017-10-27 2018-03-23 武汉大学 Fault diagnostic method for transformer winding based on wireless identification sensing
CN109270445A (en) * 2018-09-07 2019-01-25 广西电网有限责任公司电力科学研究院 Breaker spring operating mechanism abnormal state detection method based on LMD
CN109975634A (en) * 2019-03-04 2019-07-05 河南理工大学 A kind of fault diagnostic method for transformer winding based on atom sparse decomposition
CN110031703A (en) * 2019-04-25 2019-07-19 国网冀北电力有限公司检修分公司 A kind of the unusual service condition method of discrimination and system of super extra-high voltage transformer
CN110146268A (en) * 2019-05-28 2019-08-20 河海大学 A kind of OLTC method for diagnosing faults based on mean value decomposition algorithm
CN110927490A (en) * 2019-12-02 2020-03-27 吉林松江河水力发电有限责任公司 Transformer winding fault diagnosis method based on IVMD (integrated virtual machine direction) permutation entropy and CWAA-SVM (continuous wave operation-support vector machine)
CN111579238A (en) * 2020-05-14 2020-08-25 常州信息职业技术学院 Gear crack fault online monitoring method based on MFFOA optimization SVM

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107832687A (en) * 2017-10-27 2018-03-23 武汉大学 Fault diagnostic method for transformer winding based on wireless identification sensing
CN109270445A (en) * 2018-09-07 2019-01-25 广西电网有限责任公司电力科学研究院 Breaker spring operating mechanism abnormal state detection method based on LMD
CN109975634A (en) * 2019-03-04 2019-07-05 河南理工大学 A kind of fault diagnostic method for transformer winding based on atom sparse decomposition
CN110031703A (en) * 2019-04-25 2019-07-19 国网冀北电力有限公司检修分公司 A kind of the unusual service condition method of discrimination and system of super extra-high voltage transformer
CN110146268A (en) * 2019-05-28 2019-08-20 河海大学 A kind of OLTC method for diagnosing faults based on mean value decomposition algorithm
CN110927490A (en) * 2019-12-02 2020-03-27 吉林松江河水力发电有限责任公司 Transformer winding fault diagnosis method based on IVMD (integrated virtual machine direction) permutation entropy and CWAA-SVM (continuous wave operation-support vector machine)
CN111579238A (en) * 2020-05-14 2020-08-25 常州信息职业技术学院 Gear crack fault online monitoring method based on MFFOA optimization SVM

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
杨森等: "基于变分模态分解与特征选择的变压器有载分接开关机械故障诊断", 《南方电网技术》 *
王涛云 等: "基于空载合闸振动信号的变压器绕组松动诊断", 《中国电力》 *
黄辉敏 等: "基于LMD 和SVM 的高压断路器弹簧操动机构状态异常检测方法研究", 《高压电器》 *

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113252322A (en) * 2021-04-27 2021-08-13 国网安徽省电力有限公司亳州供电公司 Method and device for diagnosing mechanical fault of circuit breaker under noise-containing background
CN113252322B (en) * 2021-04-27 2022-05-31 国网安徽省电力有限公司亳州供电公司 Method and device for diagnosing mechanical fault of circuit breaker under noise-containing background
CN113391244A (en) * 2021-06-13 2021-09-14 河海大学 VMD-based transformer switching-on vibration signal characteristic frequency calculation method
CN113391244B (en) * 2021-06-13 2024-01-12 河海大学 VMD-based transformer closing vibration signal characteristic frequency calculation method
CN114088400A (en) * 2021-11-01 2022-02-25 中国人民解放军92728部队 Rolling bearing fault diagnosis method based on envelope permutation entropy
CN114088400B (en) * 2021-11-01 2024-04-09 中国人民解放军92728部队 Rolling bearing fault diagnosis method based on envelope permutation entropy
CN114325480A (en) * 2021-11-19 2022-04-12 广东核电合营有限公司 Diode open-circuit fault detection method and device for multiphase brushless exciter
CN114325480B (en) * 2021-11-19 2023-09-29 广东核电合营有限公司 Diode open-circuit fault detection method and device for multiphase brushless exciter
CN116757100A (en) * 2023-08-18 2023-09-15 国科大杭州高等研究院 Thrust prediction model training and thrust prediction method, device, equipment and medium
CN116757100B (en) * 2023-08-18 2023-12-15 国科大杭州高等研究院 Thrust prediction model training and thrust prediction method, device, equipment and medium

Similar Documents

Publication Publication Date Title
CN112557966A (en) Transformer winding looseness identification method based on local mean decomposition and support vector machine
Liu et al. Classifying transformer winding deformation fault types and degrees using FRA based on support vector machine
CN106443316B (en) Multi-information detection method and device for deformation state of power transformer winding
CN110503004B (en) On-line judging method for operating state of switching power supply
CN105547465A (en) Transformer vibration signal winding state feature extraction method
CN106017879A (en) Universal circuit breaker mechanical fault diagnosis method based on feature fusion of vibration and sound signals
CN106934421A (en) Converting station electric power transformer fault detecting system and detection method based on 2DPCA and SVM
CN110926778A (en) Mechanical fault diagnosis method for gas insulated switchgear assembly based on abnormal vibration
CN112083328A (en) Fault diagnosis method, system and device for high-voltage circuit breaker
CN113297922B (en) High-voltage switch cabinet fault diagnosis method, device and storage medium
CN115877205A (en) Intelligent fault diagnosis system and method for servo motor
CN110068776A (en) Three-level inverter open-circuit fault diagnostic method based on Support Vector Machines Optimized
CN117192347A (en) High-voltage circuit breaker on-line monitoring and fault diagnosis system and method
CN111079647A (en) Circuit breaker defect identification method
CN116796261B (en) Closed switch equipment mechanical characteristic prediction method based on artificial intelligence
CN110929673A (en) Transformer winding vibration signal identification method based on ITD (inverse discrete cosine transformation) permutation entropy and CGWO-SVM (Carrier-support vector machine)
CN111352365A (en) Dustproof ventilation type electric power and electrical equipment cabinet and control method
CN115238733A (en) Method for evaluating operation state of switching-on and switching-off coil of high-voltage circuit breaker and related equipment
CN115906350A (en) Micro-grid fault positioning and diagnosing method
CN108594062A (en) A kind of the shorted-turn fault localization method and system of feature based wave-form similarity
CN211123115U (en) Motor slot insulation electric field impact evaluation device
CN114792108A (en) Fault diagnosis method and device for high-voltage circuit breaker
CN115291156A (en) Online detection system and detection method for error characteristics of voltage transformer
CN110703080B (en) GIS spike discharge diagnosis method, discharge degree identification method and device
Ma et al. GIS mechanical fault diagnosis and test analysis based on S transform-SVM-AlexNet model

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
RJ01 Rejection of invention patent application after publication

Application publication date: 20210326

RJ01 Rejection of invention patent application after publication