CN117992772A - Method, device and equipment for extracting characteristics of vibration signal of on-load tap-changer - Google Patents
Method, device and equipment for extracting characteristics of vibration signal of on-load tap-changer Download PDFInfo
- Publication number
- CN117992772A CN117992772A CN202410208053.6A CN202410208053A CN117992772A CN 117992772 A CN117992772 A CN 117992772A CN 202410208053 A CN202410208053 A CN 202410208053A CN 117992772 A CN117992772 A CN 117992772A
- Authority
- CN
- China
- Prior art keywords
- decomposition
- vibration signal
- whale
- value
- vmd
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 71
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 104
- 241000283153 Cetacea Species 0.000 claims abstract description 71
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 33
- 238000005457 optimization Methods 0.000 claims abstract description 33
- 230000009467 reduction Effects 0.000 claims abstract description 28
- 239000013598 vector Substances 0.000 claims abstract description 28
- 230000006870 function Effects 0.000 claims description 33
- 238000012545 processing Methods 0.000 claims description 18
- 238000000605 extraction Methods 0.000 claims description 14
- 230000009466 transformation Effects 0.000 claims description 14
- 230000002159 abnormal effect Effects 0.000 claims description 12
- 230000008569 process Effects 0.000 claims description 12
- 238000009826 distribution Methods 0.000 claims description 8
- 238000004590 computer program Methods 0.000 claims description 6
- 238000001914 filtration Methods 0.000 claims description 5
- 230000005540 biological transmission Effects 0.000 claims description 3
- 230000021615 conjugation Effects 0.000 claims description 3
- 238000006073 displacement reaction Methods 0.000 claims description 3
- 230000007246 mechanism Effects 0.000 claims description 3
- 238000001514 detection method Methods 0.000 abstract description 15
- 238000004364 calculation method Methods 0.000 description 8
- 230000000694 effects Effects 0.000 description 6
- 238000004458 analytical method Methods 0.000 description 3
- 230000001629 suppression Effects 0.000 description 3
- 230000002411 adverse Effects 0.000 description 2
- 238000003491 array Methods 0.000 description 2
- 238000003745 diagnosis Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 230000001133 acceleration Effects 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000011478 gradient descent method Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000005312 nonlinear dynamic Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 239000002245 particle Substances 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 238000009827 uniform distribution Methods 0.000 description 1
Landscapes
- Measurement Of Mechanical Vibrations Or Ultrasonic Waves (AREA)
Abstract
The application provides a method, a device and equipment for extracting characteristics of vibration signals of an on-load tap-changer. The method comprises the steps of carrying out wavelet threshold noise reduction on a measured OLTC vibration signal, carrying out parameter optimization on VMD by utilizing an improved whale optimization algorithm, carrying out modal decomposition on the OLTC signal according to the obtained VMD parameter, and obtaining a characteristic value vector based on sample entropy. Therefore, noise signals in the collected vibration signals can be effectively filtered, and the detection result is more accurate when the detection is carried out based on the finally obtained eigenvalue vector.
Description
Technical Field
The application relates to a method, a device and equipment for extracting the characteristics of a vibration signal of an on-load tap-changer, and belongs to the technical field of signal processing.
Background
An On-Load Tap-changer (OLTC) is a switching device for controlling current in a high voltage power system. The main function of the system is to separate and connect circuits in the power system so as to realize maintenance, protection and control of the operation of the power equipment. The on-load tap-changer is used for monitoring the running state of the on-load tap-changer, and plays an important role in maintaining the reliability and safety of power supply of a power grid.
Currently, the detection of an on-load tap-changer mainly comprises off-line detection and on-line detection. The former method needs to be carried out when the OLTC is off-grid, and the method has higher accuracy but influences the normal operation of the power grid. In order to ensure that OLTC fault information can be found in advance without damaging the internal structure of OLTC, on-line detection of OLTC based on a switching action vibration signal is generally selected. However, since OLTC works in a high voltage environment, the collected vibration signal is often accompanied with high-frequency noise interference, and thus the detection result is adversely affected.
Disclosure of Invention
The application provides a method, a device and equipment for extracting the characteristics of a vibration signal of an on-load tap-changer, which are used for solving the problem that the vibration signal is often accompanied with high-frequency noise interference during on-line detection of the on-load tap-changer, so that the detection result is adversely affected.
In order to achieve the above object, the present application provides the following technical solutions:
In a first aspect, an embodiment of the present application provides a method for extracting a feature of a vibration signal of an on-load tap changer, including:
Collecting original vibration signals of the on-load tap-changer in a normal state and an abnormal state;
Performing threshold noise reduction processing on the original vibration signal by utilizing wavelet transformation to obtain a noise-reduced vibration signal;
Optimizing VMD parameters by using an improved whale optimization algorithm to obtain an optimal decomposition parameter combination, and performing VMD decomposition on the vibration signal after noise reduction by using the optimal decomposition parameter combination to obtain a plurality of IMF components; wherein the decomposition parameter combination comprises a decomposition level number and a regularization parameter;
and extracting sample entropy features of the IMF components, and constructing and obtaining a feature value vector.
Based on the above method, optionally, the performing threshold noise reduction processing on the original vibration signal by using wavelet transformation to obtain a noise-reduced vibration signal includes:
dividing the original vibration signal according to frequency bands through wavelet transformation; wherein the wavelet is defined as follows:
Wherein, phi a,b is a wavelet, phi is a mother wavelet, s is an independent variable, a is a scale factor, and b is a displacement factor;
Wavelet decomposition is performed on the original vibration signal using the principle of the above formula:
wherein W f (a, b) represents the decomposition result, Representing the high frequency component of the wavelet decomposition, and/>For/>Conjugation of/>A low frequency component representing wavelet decomposition, f (x) representing an original vibration signal;
and filtering the noise signal based on the decomposition result to obtain a noise-reduced vibration signal.
Based on the above method, optionally, the optimizing the VMD parameter by using the improved whale optimization algorithm to obtain an optimal decomposition parameter combination includes:
Initializing a population by adopting a quasi-reverse learning method;
calculating the value of the fitness function according to the current whale position, comparing the value with a set threshold value, and judging whether the value is smaller than the set threshold value, namely judging whether an optimal decomposition parameter combination is obtained;
If yes, ending iteration; if not, updating the next generation whale group position, recalculating the value of the fitness function, and comparing with a set threshold value until the optimal decomposition parameter combination is obtained.
Based on the above method, optionally, initializing the population by using a quasi-reverse learning method includes:
Setting whale population n=20, dimension 2, and position of i whale as Setting a k value range [1,10], wherein k is an integer, and a penalty factor value range [800, 2300]; the quasi-reverse population is obtained using the following formula:
In the method, in the process of the invention, Representing quasi-reverse population,/>Respectively express/>J=1, 2, i e [1,20];
Randomly generating N initial individuals, calculating fitness functions for the N initial individuals and N quasi-reverse populations, and selecting the whale population with the minimum fitness functions as an initial population.
Based on the above method, optionally, an envelope entropy is adopted as an fitness function of VMD decomposition; the calculating the value of the fitness function according to the current whale group position comprises the following steps:
VMD decomposition is carried out on the vibration signal after noise reduction according to the current whale position;
Carrying out Hilbert demodulation on the decomposed IMF component to obtain an envelope signal, and processing the envelope signal into a probability distribution sequence;
And calculating the entropy value of the probability distribution sequence, and calculating the envelope entropy of each IMF component, namely the value of the fitness function.
Based on the above method, optionally, the improved whale optimization algorithm adopts a nonlinear convergence factor, specifically:
Where α is a convergence factor, t is a current iteration number, and max_iter is a maximum iteration number.
Based on the above method, optionally, the updating the next generation whale group location comprises:
based on the motion characteristics of whales surrounding the prey, introducing a system random probability p to determine the updating mode of whales individuals, and assuming that the optimal position in the current group is the prey, updating the position of whales far away from the prey by using the following formula:
Wherein t is the current iteration number, and X (t) represents the current whale group position vector; x p (t) represents the current optimal solution, A and C both represent coefficient vectors, r 1 and r 2 represent random numbers between 0 and 1, l is a random number between [ -1,1], b is a screw shape constant for whale predation, and p represents a random probability
Based on the above method, optionally, the abnormal state includes: the spring is soft, the spring breaks, the transmission mechanism is blocked and the moving contact and the fixed contact are loose.
In a second aspect, an embodiment of the present application further provides a feature extraction device for a vibration signal of an on-load tap changer, including:
The acquisition module is used for acquiring original vibration signals of the on-load tap-changer in a normal state and an abnormal state;
The noise reduction module is used for carrying out threshold noise reduction processing on the original vibration signal by utilizing wavelet transformation to obtain a noise-reduced vibration signal;
The decomposition module is used for optimizing the VMD parameters by utilizing an improved whale optimization algorithm to obtain an optimal decomposition parameter combination, and performing VMD decomposition on the vibration signals after noise reduction by utilizing the optimal decomposition parameter combination to obtain a plurality of IMF components; wherein the decomposition parameter combination comprises a decomposition level number and a regularization parameter;
And the extraction module is used for extracting sample entropy characteristics of the IMF components and constructing and obtaining characteristic value vectors.
In a third aspect, an embodiment of the present application further provides an electronic device, including a memory and a processor, where the memory stores a computer program, and the processor implements the method for extracting the characteristics of the vibration signal of the on-load tap changer according to any one of the first aspects when the processor invokes and executes the computer program.
According to the feature extraction method, the device and the equipment for the vibration signal of the on-load tap changer, wavelet threshold noise reduction is carried out on the measured OLTC vibration signal, then the VMD is subjected to parameter optimization by utilizing an improved whale optimization algorithm, finally modal decomposition is carried out on the OLTC signal according to the obtained VMD parameter, and a feature value vector is obtained based on sample entropy. Therefore, noise signals in the collected vibration signals can be effectively filtered, and the detection result is more accurate when the detection is carried out based on the finally obtained eigenvalue vector.
Drawings
The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the application and together with the description, serve to explain the principles of the application. Furthermore, these drawings and the written description are not intended to limit the scope of the inventive concept in any way, but to illustrate the inventive concept to those skilled in the art by referring to the specific embodiments.
Fig. 1 is a flow chart illustrating a method for extracting characteristics of a vibration signal of an on-load tap changer according to an embodiment of the application;
FIG. 2 is a schematic structural diagram of a device for extracting a vibration signal of an on-load tap changer according to an embodiment of the application;
fig. 3 is a schematic structural diagram of an electronic device according to an embodiment of the present application.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the present application more apparent, the technical solutions in the embodiments of the present application will be clearly and completely described in the following in conjunction with the embodiments of the present application, and it is apparent that the described embodiments are some embodiments of the present application, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application. The following embodiments and features of the embodiments may be combined with each other without conflict.
As shown in the background art, since the OLTC works in a high voltage environment, the collected vibration signal is often accompanied with high-frequency noise interference, and in order to ensure the accuracy of the detection result, a proper filtering method needs to be selected to perform noise reduction processing on the OLTC. In addition, although the decomposition process is required for nonlinear and non-stationary signals such as OLTC vibration signals, the conventional decomposition method has problems such as difficulty in selecting decomposition parameters, difficulty in evaluating decomposition results, and the like.
Based on the method, the invention provides a characteristic extraction scheme of the vibration signal of the on-load tap-changer, filtering processing is carried out based on wavelet transformation, and a VMD decomposition method is adopted to carry out signal decomposition to improve a whale optimization algorithm to carry out parameter optimization on VMD decomposition, so that the finally obtained characteristics are accurate and effective, and the final detection accuracy is improved. The specific implementations are described below without limitation by way of several examples or embodiments.
First, some embodiments of the present application provide a method for extracting a characteristic of a vibration signal of an on-load tap-changer, and referring to fig. 1, fig. 1 is a flow chart of a method for extracting a characteristic of a vibration signal of an on-load tap-changer according to an embodiment of the present application. In specific implementation, the solution of this embodiment may be a software system configured into a computer or a server, etc. That is, the solution of the present embodiment may be implemented by a software system in the device.
As shown in fig. 1, the method for extracting the characteristics of the vibration signal of the on-load tap changer of the present embodiment includes the following steps:
step S101: and collecting original vibration signals of the on-load tap-changer in a normal state and an abnormal state.
Specifically, in order to detect and identify the state of the on-load tap-changer OLTC, the original vibration signals of the on-load tap-changer OLTC in the normal state and the abnormal state need to be collected.
Wherein the abnormal state includes: the spring is soft, the spring breaks, the transmission mechanism is blocked and the moving contact and the fixed contact are loose. The vibration signal of OLTC in an abnormal state is different from that in a normal state, and there is also a difference in the vibration signal in the different abnormal state.
In practice, the raw vibration signal may be acquired with an IEPE (INTEGRATED ELECTRONICS PIEZO ELECTRIC, piezoelectric integrated circuit) acceleration sensor at a certain sampling frequency.
Step S102: and performing threshold noise reduction processing on the original vibration signal by utilizing wavelet transformation to obtain a noise-reduced vibration signal.
Specifically, OLTC vibration signals have characteristics of non-stationary and non-linear, whereas wavelet transform has a better effect in processing non-stationary and non-linear signals than conventional fourier transform, so that the wavelet transform is adopted for noise reduction processing in the present embodiment. Wavelet transformation breaks a segment of a signal into multiple wavelet components, each wavelet component representing a different frequency component and time domain information of the signal. Based on the above, according to the characteristics of the noise signal, the components with specific frequencies are filtered, so that noise reduction can be realized. .
Further, the specific process of step S102 may include:
step S1021: dividing the original vibration signal according to frequency bands through wavelet transformation; wherein the wavelet is defined as follows:
Wherein, phi a,b is a wavelet, phi is a mother wavelet, s is an independent variable, a is a scale factor, and b is a displacement factor;
Wavelet decomposition is performed on the original vibration signal using the principle of the above formula:
wherein W f (a, b) represents the decomposition result, Representing the high frequency component of the wavelet decomposition, and/>For/>Conjugation of/>A low frequency component representing wavelet decomposition, f (x) representing an original vibration signal;
Step S1022: filtering the noise signal based on the decomposition result to obtain a noise-reduced vibration signal
Specifically, the basic idea of the wavelet threshold noise reduction method is that after the original vibration signal is subjected to wavelet decomposition, coefficients with larger amplitude contain key information of the signal, and the coefficients with uniform distribution and smaller amplitude correspond to high-frequency noise, so that the noise-reduced vibration signal can be obtained after the coefficients containing the noise signal are filtered.
Step S103: optimizing VMD parameters by using an improved whale optimization algorithm to obtain an optimal decomposition parameter combination, and performing VMD decomposition on the vibration signal after noise reduction by using the optimal decomposition parameter combination to obtain a plurality of IMF components; wherein the decomposition parameter combination includes a decomposition level and a regularization parameter.
Specifically, methods such as fourier transform, empirical mode decomposition, ensemble empirical decomposition, empirical wavelet decomposition, and variational mode decomposition (variational mode decomposition, VMD) are mainly used for nonlinear and nonstationary signal processing modes at present. The empirical mode decomposition can adaptively decompose the vibration signal, however, the problem of mode aliasing is very prominent, so that the physical meaning of the decomposed mode components is lost, and the fault cause is difficult to explain. The integrated empirical mode decomposition is improved on the basis of the empirical mode decomposition, white noise signals are added, and the endpoint effect and the modal aliasing phenomenon are restrained to a certain extent, but the computational complexity is increased. The problem of insufficient extraction of high-frequency components of signals is solved on the basis of wavelet transformation by wavelet packet decomposition, however, the decomposition effect is very dependent on manual wavelet base selection, and the actual online diagnosis requirement is not met. The variable-division modal decomposition solves the problem of modal aliasing, and can restore the original signal well, and adaptively decompose the original signal into time sequences with different frequencies and relatively stable time sequences, so that the VMD method is adopted to decompose the original vibration signal. The result after decomposition is an IMF (INTRINSIC MODE FUNCTION, eigenmode function) component.
The basic steps of VMD decomposition are as follows:
1. Defining the number of decomposition levels of the signal.
2. The signal is subjected to a hilbert transform to convert the real signal into a complex signal.
3. The frequencies of the components of the signal are initialized.
4. Each component is reconstructed using the VMD to conform to the frequency constraints of each component.
5. Noise cancellation is performed on each component.
6. Reconstructing the signal.
The essence of the VMD is that the vibration signal is decomposed into k modal components with certain sparsity through iterative search of a variation model, each modal component is decomposed by a center frequency omega k, and finally the constraint variation problem is obtained:
Wherein { u k}＝{u1,…,uk } is the decomposed k modal components; { ω k}＝{ω1,…,ωk } is the frequency center of each modal component after decomposition, δ (t) is the dirac function, s.t represents the constraint function, and f represents the frequency of the original vibration signal that is decomposed. To solve the above equation, the above equation is converted to an unconstrained problem of the form:
Where α is a regularization parameter (or penalty factor); lambda (t) is Lagrange multiplier.
In addition, the specific implementation process of the VMD is as follows:
Step 1: initializing set parameters: n≡0; (the number 1 of the subscript on each parameter represents algebra)
Step 2: repeating n- & gt n+1;
Step 3: the number of decomposition is 1-k, and for all omega not less than 0, updating according to the following formula
The omega k is updated using the following formula,
Where n represents the number of iterations, n+1 per iteration.
Stopping iteration until the following iteration conditions are met:
Wherein epsilon is the preset discrimination precision.
Among them, the parameter selection of VMD decomposition has a large influence on modal decomposition, and thus an appropriate decomposition parameter combination needs to be selected. The decomposition parameter combination includes two parts: the number of decomposition levels k and a regularization parameter (or penalty factor) α. The number of decomposition levels k, i.e. the signal is decomposed into k local frequency modes. Generally, the larger the number of k, the finer the mode frequency after decomposition, but the calculation amount increases. And the regularization parameter alpha has an effect on the smoothness of the decomposition result and the noise suppression capability. The greater α, the higher the smoothness, but the worse the noise suppression capability; the smaller α is, the better the noise suppression capability is, but the lower the smoothness is.
In practice, it is found that during the process of optimizing the decomposition parameters, the parameter selection range is large, and the data volume of the acquired signal of the original vibration signal is too large, so that the difficulty of parameter optimization is increased. Therefore, in this embodiment, the VMD parameters are optimized by using the improved whale optimization algorithm to obtain the optimal decomposition parameter combination.
A whale optimization algorithm (whale optimization algorithm, WOA) whose mathematical modeling process simulates the whale predation process, consisting essentially of three phases: search for food, shrink wrap, and spiral update locations. The whale optimization algorithm does not need to set various control parameter values artificially, so that the use efficiency of the algorithm is improved, and the application difficulty is reduced. In addition, compared with other intelligent optimization algorithms of colony, the WOA algorithm has novel structure and less control parameters, and has better optimizing performance in a plurality of numerical optimization and solution of engineering problems, which is superior to intelligent optimization algorithms such as ant colony algorithm, particle swarm algorithm and the like. Compared with other intelligent optimization algorithms, the method has the advantages of being capable of converging to the globally optimal solution more quickly, and having better robustness and universality.
In some embodiments, in step S103, when optimizing VMD parameters by using the improved whale optimization algorithm to obtain an optimal decomposition parameter combination, the specific process includes:
Step S1031: and initializing the population by adopting a quasi-reverse learning method.
Specifically, the quasi-inverse learning method is an unconstrained optimization algorithm for solving an unconstrained nonlinear optimization problem. Compared with the traditional gradient descent method, the quasi-reverse learning method greatly reduces the computational complexity while maintaining the convergence speed of the algorithm.
In some embodiments, the specific implementation procedure of step S1031 includes:
Setting whale population n=20, dimension 2, and position of i whale as Setting a k value range [1,10], wherein k is an integer, and a penalty factor value range [800, 2300]; the quasi-reverse population is obtained using the following formula:
In the method, in the process of the invention, Representing quasi-reverse population,/>Respectively express/>J=1, 2, i e [1,20];
Randomly generating N initial individuals, calculating fitness functions for the N initial individuals and N quasi-reverse populations, and selecting the whale population with the minimum fitness functions as an initial population.
Wherein the fitness function measures the fitness of each individual in the population, i.e. their degree of contribution in solving a given problem. Thus, the smaller the value of the fitness function, the better the fitness of the corresponding individual.
Step S1032: and calculating the value of the fitness function according to the current whale position, comparing the value with a set threshold value, and judging whether the value is smaller than the set threshold value, namely judging whether the optimal decomposition parameter combination is obtained.
In particular, in practice, the value of the fitness function cannot be infinitely small. Therefore, in this embodiment, the threshold of the fitness is set, and when the value of the fitness function is smaller than the set threshold, the current result is considered to reach the requirement, and the iteration can be stopped.
In some embodiments, the envelope entropy (Envelope Entropy) may be selected as an indicator of the evaluation of the VMD decomposition effect, i.e., the fitness function described above. The envelope entropy is a nonlinear dynamics feature extraction method based on signal envelope analysis, and has wide application in signal processing, machine learning and industrial diagnosis. The core idea of envelope entropy is: and carrying out envelope analysis on the signal, namely extracting the amplitude envelope of the signal, and carrying out feature extraction on the signal by utilizing the concept of entropy. The envelope entropy can be calculated by:
1. Performing Hilbert transformation on the original signal to obtain an analysis signal;
2. Calculating an amplitude envelope of the resolved signal;
3. Dividing the amplitude envelope into a plurality of intervals, the value in each interval being regarded as a symbol;
4. calculating probability distribution of each symbol sequence, and calculating entropy value of the sequence;
5. and averaging the entropy values of all the symbol sequences to obtain envelope entropy.
Accordingly, for step S1032 of the present embodiment, where the value of the fitness function is calculated according to the current whale group position, it may specifically include:
VMD decomposition is carried out on the vibration signal after noise reduction according to the current whale position; the decomposed IMF component is subjected to Hilbert demodulation to obtain an envelope signal, and the envelope signal is processed into a probability distribution sequence; and calculating the entropy value of the probability distribution sequence, and calculating the envelope entropy of each IMF component, namely the value of the fitness function.
Wherein, the calculation formula of the envelope entropy is as follows:
Wherein: e p is an envelope entropy, a (j) is a hilbert-demodulated envelope signal of each IMF component, and p j is a probability distribution sequence of a (j).
Step S1033: if yes, ending iteration; if not, updating the next generation whale group position, recalculating the value of the fitness function, and comparing with a set threshold value until the optimal decomposition parameter combination is obtained.
After the optimal decomposition parameter combination (k, α) is obtained, it can be used for subsequent VMD decomposition.
In addition, when updating the next generation whale group position, considering that the convergence factor of the traditional whale optimization algorithm is reduced in global searching and local development capacity in a linear manner, in some embodiments, the whale optimization algorithm is improved by adopting a nonlinear convergence factor, wherein the nonlinear convergence factor is as follows:
In the formula, t is the current iteration number, and max_iter is the maximum iteration number.
In addition, the whale optimization algorithm includes a search for food (search for prey) phase and a shrink wrap (spiral wrap prey) phase. Different updating strategies are adopted for different stages, specifically, the system judges whether whales are in a hunting stage or a spiral surrounding hunting stage according to the |A|, A is a coefficient vector, the coefficient vector is obtained through calculation (a calculation formula is shown below), when the |A| < 1, the whales are in the hunting stage, and when the |A| > 1, the whales are in the spiral surrounding hunting stage.
Wherein, for the spiral surrounding prey stage, when updating the next generation whale group position:
based on the motion characteristics of whales surrounding the prey, introducing a system random probability p to determine the updating mode of whales individuals, and assuming that the optimal position in the current group is the prey, updating the position of whales far away from the prey by using the following formula:
wherein t is the current iteration number, and X (t) represents the current whale group position vector; x p (t) represents the current optimal solution, A and C both represent coefficient vectors, r 1 and r 2 represent random numbers between 0 and 1, l is a random number between [ -1,1], b is a screw shape constant for whale predation, and p represents a random probability.
And for the hunting phase, when updating the next generation whale group position:
based on the motion characteristics of whale searching for prey, the position is updated using the following formula:
X(t+1)＝Xrand(t)-A·|C*Xrand(t)-X(t)|
Wherein X rand (t) is a randomly extracted whale individual position vector in the current whale group.
In addition, performing a location update requires performing population variations to avoid local optimality, where random differential variations are used to optimize whale population:
X(t+1)＝r1*(Xp(t)-X(t))+r2*(X′(t)-X(t))
wherein X' (t) is a randomly mutated position vector.
Step S104: and extracting sample entropy features of the IMF components, and constructing and obtaining a feature value vector.
Specifically, the sample entropy (Sample Entropy, sampEn) is a method for measuring the complexity of a time sequence by measuring the probability of generating a new pattern in an original signal, and the more the time sequence is complex, the larger the value of the sample entropy is when the noise component is more. Therefore, the method has a good characteristic extraction effect on the vibration signal of the on-load tap-changer with nonlinear characteristics. As an improved algorithm of the approximate entropy, the calculation of the sample entropy reduces the dependence degree of the data length, has better consistency, is simpler and has higher calculation speed. The algorithm for calculating the time series from the sample entropy is as follows:
(1) Sequentially forming a time sequence X (i) with a fixed length of N into m-dimensional vectors, namely:
X(i)＝{u(i)，u(i+1)，...u(i+m-1)}
Wherein: i=1, 2, …, N-m+1.
(2) Calculating the distance between the vector X (i) and the other vector X (j):
dij＝max|u(i+k)-u(j+k)|
wherein: j=1, 2, …, N-m+1; k=0, 1, …, m-1, i+.j.
(3) Given a threshold r (r > 0), the number of d ij < r is counted and denoted as n ij (r). Calculating the ratio of N ij (r) to the total vector number N-m+1, and recording asNamely:
the average value can be calculated:
(4) Increasing the dimension to m+1 while repeating steps (2), (3), it is possible to obtain:
(5) When the length N of the time series X (i) is a finite value, the sample entropy can be defined as:
From the above equation, the values of the parameters m, r, and N are important for the accuracy of the sample entropy calculation. Therefore, in this embodiment, the values m=2, r= (0.1-0.25) δ (δ is the standard deviation of the original data) are determined through experiments to obtain the most effective statistical characteristics of the sample entropy.
Sample entropy is selected to be applied to an OLTC feature extraction process, and the obtained feature value vector can reflect dynamic changes of a time sequence in different states of the OLTC, so that the method can be applied to subsequent state/fault identification.
After determining the VMD optimal parameters (K), decomposing the vibration signals after noise reduction by utilizing the VMD to obtain K decomposition signals, and performing multi-scale sample entropy calculation on the decomposition signals to obtain characteristic value vectors of the vibration signals.
Through the scheme, the wavelet threshold value noise reduction is carried out on the measured OLTC vibration signal, the improved whale optimization algorithm is utilized to carry out parameter optimization on the VMD, finally, modal decomposition is carried out on the OLTC signal according to the obtained VMD parameter, and the eigenvalue vector is obtained based on the sample entropy. Therefore, noise signals in the collected vibration signals can be effectively filtered, and the detection result is more accurate when the detection is carried out based on the finally obtained eigenvalue vector.
In addition, an embodiment of the present application provides a feature extraction device for a vibration signal of an on-load tap-changer, referring to fig. 2, the feature extraction device for a vibration signal of an on-load tap-changer includes:
the acquisition module 21 is used for acquiring original vibration signals of the on-load tap-changer in a normal state and an abnormal state;
The noise reduction module 22 is configured to perform threshold noise reduction processing on the original vibration signal by using wavelet transformation, so as to obtain a noise-reduced vibration signal;
The decomposition module 23 is configured to optimize the VMD parameters by using an improved whale optimization algorithm to obtain an optimal decomposition parameter combination, and perform VMD decomposition on the vibration signal after noise reduction by using the optimal decomposition parameter combination to obtain a plurality of IMF components; wherein the decomposition parameter combination comprises a decomposition level number and a regularization parameter;
The extracting module 24 is configured to perform sample entropy feature extraction on the plurality of IMF components, and construct a feature value vector.
For a specific implementation method of each module of the above-mentioned feature extraction device for a vibration signal of an on-load tap-changer, reference may be made to corresponding content in the foregoing method embodiment, which is not repeated herein.
In addition, an embodiment of the present application provides an electronic device, as shown in fig. 3, including a memory 31 and a processor 32; the memory 31 stores a computer program, and the processor 32 executes and invokes the computer program to implement the method for extracting the characteristics of the vibration signal of the on-load tap changer in any of the above embodiments.
The electronic device may be a computer or a server, among others.
It is to be understood that the same or similar parts in the above embodiments may be referred to each other, and that in some embodiments, the same or similar parts in other embodiments may be referred to.
It should be noted that in the description of the present invention, the terms "first," "second," and the like are used for descriptive purposes only and are not to be construed as indicating or implying relative importance. Furthermore, in the description of the present invention, unless otherwise indicated, the meaning of "plurality" means at least two.
Any process or method descriptions in flow charts or otherwise described herein may be understood as representing modules, segments, or portions of code which include one or more executable instructions for implementing specific logical functions or steps of the process, and further implementations are included within the scope of the preferred embodiment of the present invention in which functions may be executed out of order from that shown or discussed, including substantially concurrently or in reverse order, depending on the functionality involved, as would be understood by those reasonably skilled in the art of the present invention.
It is to be understood that portions of the present invention may be implemented in hardware, software, firmware, or a combination thereof. In the above-described embodiments, the various steps or methods may be implemented in software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, may be implemented using any one or combination of the following techniques, as is well known in the art: discrete logic circuits having logic gates for implementing logic functions on data signals, application specific integrated circuits having suitable combinational logic gates, programmable Gate Arrays (PGAs), field Programmable Gate Arrays (FPGAs), and the like.
Those of ordinary skill in the art will appreciate that all or a portion of the steps carried out in the method of the above-described embodiments may be implemented by a program to instruct related hardware, and the program may be stored in a computer readable storage medium, where the program when executed includes one or a combination of the steps of the method embodiments.
In addition, each functional unit in the embodiments of the present invention may be integrated in one processing module, or each unit may exist alone physically, or two or more units may be integrated in one module. The integrated modules may be implemented in hardware or in software functional modules. The integrated modules may also be stored in a computer readable storage medium if implemented as software functional modules and sold or used as a stand-alone product. The above-mentioned storage medium may be a read-only memory, a magnetic disk or an optical disk, or the like.
In the description of the present specification, a description referring to terms "one embodiment," "some embodiments," "examples," "specific examples," or "some examples," etc., means that a particular feature, structure, material, or characteristic described in connection with the embodiment or example is included in at least one embodiment or example of the present invention. In this specification, schematic representations of the above terms do not necessarily refer to the same embodiments or examples. Furthermore, the particular features, structures, materials, or characteristics described may be combined in any suitable manner in any one or more embodiments or examples.
While embodiments of the present invention have been shown and described above, it will be understood that the above embodiments are illustrative and not to be construed as limiting the invention, and that variations, modifications, alternatives and variations may be made to the above embodiments by one of ordinary skill in the art within the scope of the invention.
Claims (10)
1. A method for extracting characteristics of a vibration signal of an on-load tap changer, comprising:
Collecting original vibration signals of the on-load tap-changer in a normal state and an abnormal state;
Performing threshold noise reduction processing on the original vibration signal by utilizing wavelet transformation to obtain a noise-reduced vibration signal;
Optimizing VMD parameters by using an improved whale optimization algorithm to obtain an optimal decomposition parameter combination, and performing VMD decomposition on the vibration signal after noise reduction by using the optimal decomposition parameter combination to obtain a plurality of IMF components; wherein the decomposition parameter combination comprises a decomposition level number and a regularization parameter;
and extracting sample entropy features of the IMF components, and constructing and obtaining a feature value vector.
2. The method of claim 1, wherein the performing threshold noise reduction processing on the original vibration signal using wavelet transform to obtain a noise reduced vibration signal comprises:
dividing the original vibration signal according to frequency bands through wavelet transformation; wherein the wavelet is defined as follows:
Wherein, phi a,b is a wavelet, phi is a mother wavelet, s is an independent variable, a is a scale factor, and b is a displacement factor;
Wavelet decomposition is performed on the original vibration signal using the principle of the above formula:
wherein W f (a, b) represents the decomposition result, Representing the high frequency component of the wavelet decomposition, and/>For/>Conjugation of/>A low frequency component representing wavelet decomposition, f (x) representing an original vibration signal;
and filtering the noise signal based on the decomposition result to obtain a noise-reduced vibration signal.
3. The method of claim 1, wherein optimizing the VMD parameters using a modified whale optimization algorithm to obtain an optimal combination of decomposition parameters comprises:
Initializing a population by adopting a quasi-reverse learning method;
calculating the value of the fitness function according to the current whale position, comparing the value with a set threshold value, and judging whether the value is smaller than the set threshold value, namely judging whether an optimal decomposition parameter combination is obtained;
If yes, ending iteration; if not, updating the next generation whale group position, recalculating the value of the fitness function, and comparing with a set threshold value until the optimal decomposition parameter combination is obtained.
4. A method according to claim 3, wherein initializing the population using quasi-reverse learning comprises:
Setting whale population n=20, dimension 2, and position of i whale as Setting a k value range [1,10], wherein k is an integer, and a penalty factor value range [800, 2300]; the quasi-reverse population is obtained using the following formula:
In the method, in the process of the invention, Representing quasi-reverse population,/>Respectively express/>J=1, 2, i e [1,20];
Randomly generating N initial individuals, calculating fitness functions for the N initial individuals and N quasi-reverse populations, and selecting the whale population with the minimum fitness functions as an initial population.
5. A method according to claim 3 or 4, characterized in that the envelope entropy is used as a fitness function of the VMD decomposition; the calculating the value of the fitness function according to the current whale group position comprises the following steps:
VMD decomposition is carried out on the vibration signal after noise reduction according to the current whale position;
Carrying out Hilbert demodulation on the decomposed IMF component to obtain an envelope signal, and processing the envelope signal into a probability distribution sequence;
And calculating the entropy value of the probability distribution sequence, and calculating the envelope entropy of each IMF component, namely the value of the fitness function.
6. A method according to claim 3, wherein the improved whale optimization algorithm employs a nonlinear convergence factor, in particular:
In the formula, t is the current iteration number, and max_iter is the maximum iteration number.
7. A method according to claim 3, wherein the updating of next generation whale positions comprises:
based on the motion characteristics of whales surrounding the prey, introducing a system random probability p to determine the updating mode of whales individuals, and assuming that the optimal position in the current group is the prey, updating the position of whales far away from the prey by using the following formula:
Wherein t is the current iteration number, and X (t) represents the current whale group position vector; x p (t) represents the current optimal solution, A and C both represent coefficient vectors, r 1 and r 2 represent random numbers between 0 and 1, l is a random number between [ -1,1], b is a screw shape constant for whale predation, and p represents a random probability.
8. The method of claim 1, wherein the abnormal state comprises: the spring is soft, the spring breaks, the transmission mechanism is blocked and the moving contact and the fixed contact are loose.
9. A device for extracting characteristics of a vibration signal of an on-load tap-changer, comprising:
The acquisition module is used for acquiring original vibration signals of the on-load tap-changer in a normal state and an abnormal state;
The noise reduction module is used for carrying out threshold noise reduction processing on the original vibration signal by utilizing wavelet transformation to obtain a noise-reduced vibration signal;
The decomposition module is used for optimizing the VMD parameters by utilizing an improved whale optimization algorithm to obtain an optimal decomposition parameter combination, and performing VMD decomposition on the vibration signals after noise reduction by utilizing the optimal decomposition parameter combination to obtain a plurality of IMF components; wherein the decomposition parameter combination comprises a decomposition level number and a regularization parameter;
And the extraction module is used for extracting sample entropy characteristics of the IMF components and constructing and obtaining characteristic value vectors.
10. An electronic device comprising a memory and a processor, the memory storing a computer program, the processor implementing a method for extracting the characteristics of the vibration signal of an on-load tap-changer according to any one of claims 1 to 8 when the computer program is invoked and executed.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410208053.6A CN117992772A (en) | 2024-02-26 | 2024-02-26 | Method, device and equipment for extracting characteristics of vibration signal of on-load tap-changer |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410208053.6A CN117992772A (en) | 2024-02-26 | 2024-02-26 | Method, device and equipment for extracting characteristics of vibration signal of on-load tap-changer |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117992772A true CN117992772A (en) | 2024-05-07 |
Family
ID=90888483
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410208053.6A Pending CN117992772A (en) | 2024-02-26 | 2024-02-26 | Method, device and equipment for extracting characteristics of vibration signal of on-load tap-changer |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117992772A (en) |
-
2024
- 2024-02-26 CN CN202410208053.6A patent/CN117992772A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zheng et al. | A review on non-model based diagnosis methodologies for PEM fuel cell stacks and systems | |
| Koskela et al. | Time series prediction using recurrent SOM with local linear models | |
| CN107480386B (en) | Test excitation optimization method based on response aliasing measurement and genetic algorithm | |
| CN108399434B (en) | Analysis and prediction method of high-dimensional time series data based on feature extraction | |
| CN112101554A (en) | Anomaly detection method and device, equipment and computer-readable storage medium | |
| CN114004091B (en) | CEEMDAN-BNs-based wind power variable pitch system fault diagnosis method | |
| WO2012160062A1 (en) | Transition detection method for automatic-setup non-intrusive appliance load monitoring | |
| CN114077846B (en) | Fault current multi-domain identification method based on RF-LSTM and storage medium | |
| CN112085111B (en) | Load identification method and device | |
| CN111931983A (en) | Precipitation prediction method and system | |
| CN111881594A (en) | Non-stationary signal state monitoring method and system for nuclear power equipment | |
| Liao et al. | Data-driven missing data imputation for wind farms using context encoder | |
| CN111723862A (en) | Switch cabinet state evaluation method and device | |
| CN117791570A (en) | Short-term power load prediction system and method | |
| CN111930728B (en) | Method and system for predicting characteristic parameters and failure rate of equipment | |
| CN117992772A (en) | Method, device and equipment for extracting characteristics of vibration signal of on-load tap-changer | |
| CN116449081A (en) | Data acquisition system, device and storage medium with self-adaptive regulation and control function | |
| CN116561927A (en) | Digital twin-driven small sample rotary machine residual life prediction method and system | |
| CN116933119A (en) | Signal data trend removal method based on convolutional neural network | |
| CN113339208B (en) | Method for selecting abnormal vibration segments of wind turbine generator | |
| CN116306217A (en) | Track similarity residual life prediction method based on slow characteristic information gain ratio | |
| CN115238824A (en) | Method for aligning monitoring data of wind generating set | |
| KR20200086548A (en) | Method for compressing and restoring time series data | |
| CN115291103A (en) | Motor fault diagnosis method based on GR-SWPT wavelet packet algorithm embedded with HD-RCF | |
| CN115510925B (en) | Mechanical fault diagnosis method, device and medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |