CN117851758A - Cable partial discharge signal processing method and computer storage medium - Google Patents
Cable partial discharge signal processing method and computer storage medium Download PDFInfo
- Publication number
- CN117851758A CN117851758A CN202410052105.5A CN202410052105A CN117851758A CN 117851758 A CN117851758 A CN 117851758A CN 202410052105 A CN202410052105 A CN 202410052105A CN 117851758 A CN117851758 A CN 117851758A
- Authority
- CN
- China
- Prior art keywords
- partial discharge
- discharge signal
- noise
- signal
- cable
- 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
- 238000003860 storage Methods 0.000 title claims abstract description 12
- 238000003672 processing method Methods 0.000 title abstract description 19
- 230000009467 reduction Effects 0.000 claims abstract description 69
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 47
- 238000012545 processing Methods 0.000 claims abstract description 39
- 238000000034 method Methods 0.000 claims description 81
- 239000011159 matrix material Substances 0.000 claims description 31
- 230000006870 function Effects 0.000 claims description 13
- 238000001228 spectrum Methods 0.000 claims description 11
- 238000013135 deep learning Methods 0.000 claims description 10
- 238000004364 calculation method Methods 0.000 claims description 6
- 238000004590 computer program Methods 0.000 claims description 5
- 230000035772 mutation Effects 0.000 claims description 5
- 238000003909 pattern recognition Methods 0.000 claims description 3
- 230000004044 response Effects 0.000 claims description 3
- 238000010586 diagram Methods 0.000 description 21
- 230000000694 effects Effects 0.000 description 17
- 238000004458 analytical method Methods 0.000 description 11
- 238000004088 simulation Methods 0.000 description 9
- 238000002474 experimental method Methods 0.000 description 6
- 238000005259 measurement Methods 0.000 description 6
- 230000010355 oscillation Effects 0.000 description 6
- 238000009413 insulation Methods 0.000 description 5
- 230000008569 process Effects 0.000 description 5
- 230000001629 suppression Effects 0.000 description 5
- 238000004422 calculation algorithm Methods 0.000 description 4
- 230000007547 defect Effects 0.000 description 4
- 230000000737 periodic effect Effects 0.000 description 4
- 238000005070 sampling Methods 0.000 description 3
- 238000012360 testing method Methods 0.000 description 3
- 238000013016 damping Methods 0.000 description 2
- 238000009826 distribution Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 238000012216 screening Methods 0.000 description 2
- 238000012549 training Methods 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 230000032683 aging Effects 0.000 description 1
- 230000003190 augmentative effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000011109 contamination Methods 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000010891 electric arc Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000003631 expected effect Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000005764 inhibitory process Effects 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000013178 mathematical model Methods 0.000 description 1
- 239000002184 metal Substances 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 239000013307 optical fiber Substances 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 238000011946 reduction process Methods 0.000 description 1
- 230000001052 transient effect Effects 0.000 description 1
Landscapes
- Testing Relating To Insulation (AREA)
Abstract
The invention provides a cable partial discharge signal processing method and a computer storage medium. The processing method comprises the following steps: collecting partial discharge signals of the cable; performing successive variation modal decomposition on the partial discharge signal, and adaptively decomposing a plurality of intrinsic modal components; calculating the cross-correlation coefficient of each eigenmode component and the partial discharge signal; taking the intrinsic mode component with the cross-correlation coefficient larger than the set threshold value as an effective mode component, and taking the intrinsic mode component with the cross-correlation coefficient smaller than the set threshold value as a noise mode component; and respectively carrying out data reconstruction on the effective modal component and the noise modal component to obtain a partial discharge signal and a noise signal after preliminary noise reduction. The scheme of the invention has good signal noise reduction capability and improves the signal processing precision.
Description
Technical Field
The present invention relates to a signal processing technology of an electric power system, and in particular, to a method for processing a cable partial discharge signal and a computer storage medium.
Background
Along with the rapid network development of the power cable, the cable is used as a key component for transmitting electric energy, plays an important role in the industrial field and the urban power grid, and ensures normal production activities by virtue of the characteristics of high efficiency, safety and reliability. However, as the operational age increases, the insulation of the cabling may suffer a degree of ageing damage during operation and may even lead to line faults. The research of the cable line insulation state monitoring has important significance for early warning of potential faults of the cable line and ensuring stable operation of the cable line. Partial discharge (Partial Discharge, PD for short) detection and analysis has been employed as a predictive test to characterize and evaluate the state of the cable in advance. In an actual working field, signals detected by partial discharge are easily affected by white noise of the surrounding environment and periodic narrow-band interference, so that workers cannot monitor the insulation condition of the cable well. Thus, it is necessary to perform noise reduction processing on the detected PD signal.
In recent years, scholars at home and abroad have proposed a number of methods for reducing noise interference, such as: fourier transform thresholding (Fast Fourier Transform, FFT for short), wavelet packet variance (Wavelet Packet Transform, WPT for short), empirical mode decomposition (Empirical Mode Decomposition, EMD for short), variance mode decomposition (Variational Mode Decomposition, VMD for short), and the like.
The fourier transform thresholding method performs signal processing in the frequency domain, and can convert the signal into a frequency spectrum representation, so that the frequency characteristics are better analyzed and processed, and has the defect that the selection of the thresholding method is difficult to determine, the time domain characteristics and the frequency domain characteristics of the signal cannot be distinguished, and noise and high-frequency parts of the signal can be filtered out together, so that signal information is lost.
The wavelet packet denoising method not only decomposes the low-frequency signal but also subdivides the high-frequency signal, so that the signal can be better analyzed, but the decomposition level, the threshold value and other parameter choices have great influence on the result, and certain experience and trial and error are needed, so that the denoising effect is poor.
The empirical mode decomposition method is used for processing the noise-containing signal, so that a plurality of inherent mode functions with different frequencies can be obtained, signal characteristic information and noise can be effectively distinguished, effective information in the signal can be extracted, and the noise reduction effect is reduced due to the problems of mode aliasing and end-point effect.
The variational modal decomposition method is established on a strict mathematical framework, overcomes the defects of an EMD method, has strong time domain and frequency domain analysis capability, does not have the problems of wavelet packet parameter selection and the like, has good nonlinear non-stationary signal processing effect, but has the decomposition effect influenced by parameters K and alpha and needs to be selected according to experience.
Therefore, the method for processing the partial discharge signal aiming at the cable line has poor denoising effect, and the processed signal can lose effective information and can not meet the requirement of further analysis and processing.
Disclosure of Invention
The invention aims to accurately filter noise in partial discharge signals of cables and improve noise reduction effect.
A further object of the invention is to meet the requirements for further analysis of the cable partial discharge signal.
The invention further aims to realize analysis and identification of the partial discharge signals of the cable and improve the safety of the power system.
In particular, the invention provides a method for processing a cable partial discharge signal, which comprises the following steps:
collecting partial discharge signals of the cable;
performing successive variation modal decomposition on the partial discharge signal, and adaptively decomposing a plurality of intrinsic modal components;
calculating the cross-correlation coefficient of each eigenmode component and the partial discharge signal;
taking the intrinsic mode component with the cross-correlation coefficient larger than the set threshold value as an effective mode component, and taking the intrinsic mode component with the cross-correlation coefficient smaller than the set threshold value as a noise mode component;
and respectively carrying out data reconstruction on the effective modal component and the noise modal component to obtain a partial discharge signal and a noise signal after preliminary noise reduction.
Optionally, the step of performing successive variation modal decomposition on the partial discharge signal includes:
establishing constraint criteria according to the formulas (1), (2) and (3):
establishing a constraint minimization model according to the formula (4):
wherein,
solving the lagrangian function according to equation (6) to obtain a plurality of eigenmode components:
in the formulae (1) to (6), J 1 、J 2 、J 3 Respectively representing a first constraint, a second constraint and a third constraint,represents the partial derivative with respect to time t, delta (t) is a dirac function, x is a convolution symbol, ω n Represents the center frequency, beta, of the nth mode i (t) and beta n (t) represents impulse response of the setting filter, alpha is a preset balance parameter, f (t) represents partial discharge signal, u n (t) and f r (t) two decomposed signals of f (t), u n (t) corresponds to the nth eigenmode component, f r (t) is the residual of f (t); f (f) r (t) the first n-1 modes and the unprocessed part f of the signal u (t) and lambda (t) is the Lagrangian operator.
Optionally, the preset balance parameter α is set to α=2000.
Optionally, the cross-correlation coefficient is calculated according to equation (7):
the calculation formula of the set threshold is calculated according to the formula (8):
wherein x is n Representing the effective modal component, y n Which is indicative of the partial discharge signal,data mean value representing partial discharge signal, +.>A data mean representing the effective modal components.
Optionally, after obtaining the partial discharge signal after preliminary noise reduction, the method for processing the cable partial discharge signal further includes:
and performing singular value decomposition noise reduction on the partial discharge signal after primary noise reduction to obtain a partial discharge signal with secondary noise reduction.
Optionally, the step of performing singular value decomposition noise reduction on the partial discharge signal after preliminary noise reduction includes:
converting the partial discharge signal after preliminary noise reduction into a Hankel matrix;
singular value decomposition is carried out on the Hankel matrix;
performing matrix reconstruction according to the decomposed singular values to obtain a reconstructed matrix;
and calculating according to the reconstruction matrix to obtain a partial discharge signal with secondary noise reduction.
Optionally, the step of reconstructing the matrix according to the decomposed singular values includes:
drawing a singular value entropy increment curve spectrum, and determining a singular value effective order according to the mutation position of the singular value entropy increment curve spectrum;
reserving a second preset number of singular values according to the effective orders of the singular values, and setting the rest singular values to zero;
and calculating by using singular value inverse decomposition to obtain a reconstruction matrix.
Optionally, after the step of obtaining the partial discharge signal with the secondary noise reduction, the method for processing the cable partial discharge signal further includes:
extracting characteristic quantity of the partial discharge signal with the secondary noise reduction;
and inputting the characteristic quantity of the partial discharge signal into a pre-trained first deep learning network model to perform cable partial discharge pattern recognition.
Optionally, after obtaining the noise signal, the method for processing the cable partial discharge signal further includes:
extracting characteristic quantity of the noise signal;
and inputting the characteristic quantity of the noise signal into a pre-trained second deep learning network model to perform noise type recognition.
According to another aspect of the present invention, there is also provided a computer storage medium having stored thereon a computer program for implementing the method of processing a cable partial discharge signal of any one of the above, when the computer program is executed.
The invention provides a cable partial discharge signal processing method and a computer storage medium, and provides a cable partial discharge signal noise reduction algorithm based on successive variation modal decomposition (Successive Variational Modal Decomposition, SVMD for short). The method applies SVMD to adaptively decompose a plurality of intrinsic mode components (Intrinsic Mode Function, IMF for short) of the partial discharge signal; the effective component and the noise component are distinguished by the cross-correlation coefficient, thereby preliminarily removing the noise component.
Furthermore, the processing method of the cable partial discharge signal also applies a singular value decomposition (Singular Value Decomposition, SVD for short) to remove residual noise from the partial discharge signal after primary noise reduction, and can obtain a partial discharge signal with secondary noise reduction. Simulation and actual measurement experiment results show that the method has good noise reduction capability, improves the signal processing precision, and compared with a VMD-wavelet threshold method, the average signal-to-noise ratio is improved by 17.98%, and the noise suppression capability is excellent.
Furthermore, the processing method of the cable partial discharge signal can reduce noise to the greatest extent and can well keep the characteristic information of the signal.
Furthermore, the cable partial discharge signal processing method can also utilize the deep learning network model to perform cable partial discharge mode identification and noise type identification on the partial discharge signals and noise signals, can better analyze the cable partial discharge signals, and provides an analysis means for the safety guarantee of an electric power system.
The above, as well as additional objectives, advantages, and features of the present invention will become apparent to those skilled in the art from the following detailed description of a specific embodiment of the present invention when read in conjunction with the accompanying drawings.
Drawings
Some specific embodiments of the invention will be described in detail hereinafter by way of example and not by way of limitation with reference to the accompanying drawings. The same reference numbers will be used throughout the drawings to refer to the same or like parts or portions. It will be appreciated by those skilled in the art that the drawings are not necessarily drawn to scale. In the accompanying drawings:
FIG. 1 is a schematic diagram of a method of processing a cable partial discharge signal according to one embodiment of the invention;
FIG. 2 is an application flow chart of a method of processing a cable partial discharge signal according to one embodiment of the invention; and
FIG. 3 is a waveform diagram of a noiseless partial discharge signal of a simulation experiment to which a method of processing a cable partial discharge signal according to an embodiment of the present invention is applied;
FIG. 4 is a waveform diagram of partial discharge signals after adding noise signals by a simulation experiment of a cable partial discharge signal processing method according to an embodiment of the present invention;
FIG. 5 is a waveform diagram of SVMD decomposition results of a simulation experiment of a cable partial discharge signal processing method employing an embodiment of the present invention;
FIG. 6 is a waveform diagram of a reconstructed partial discharge signal of a simulation experiment to which a method of processing a cable partial discharge signal according to an embodiment of the present invention is applied;
FIG. 7 is a schematic diagram of a singular value distribution of a simulation experiment of a processing method of a cable partial discharge signal to which an embodiment of the present invention is applied;
FIG. 8 is a waveform diagram of a final noise reduced signal of a simulation experiment to which a cable partial discharge signal processing method according to an embodiment of the present invention is applied;
FIG. 9 is a waveform diagram of a comparison result signal obtained using the EMD-ICA method;
FIG. 10 is a waveform diagram of a comparison result signal obtained using the VMD-wavelet thresholding method;
FIG. 11 is a schematic diagram of an actual measurement platform of a cable partial discharge signal processing method according to an embodiment of the present invention;
FIG. 12 is a waveform diagram of an originally acquired partial discharge signal of an actual measurement experiment to which a cable partial discharge signal processing method according to an embodiment of the present invention is applied;
FIG. 13 is a waveform diagram of a final noise-reduced signal of an actual measurement experiment to which a cable partial discharge signal processing method according to an embodiment of the present invention is applied;
FIG. 14 is a waveform diagram of a comparison result signal obtained using the EMD-ICA method;
fig. 15 is a waveform diagram of a comparison result signal obtained using the VMD-wavelet thresholding method.
Detailed Description
The embodiment provides a cable partial discharge signal processing method and a computer storage medium. Fig. 1 is a schematic diagram of a method of processing a cable partial discharge signal according to one embodiment of the invention, which may generally include the steps of:
step S101, collecting partial discharge signals of a cable;
step S102, performing successive variation modal decomposition on a local discharge signal, and adaptively decomposing a plurality of intrinsic modal components;
step S103, calculating the cross-correlation coefficient of each intrinsic mode component and the partial discharge signal;
step S104, taking the intrinsic mode component with the cross-correlation coefficient larger than the set threshold value as an effective mode component, and taking the intrinsic mode component with the cross-correlation coefficient smaller than the set threshold value as a noise mode component;
step S105, respectively carrying out data reconstruction on the effective modal component and the noise modal component to obtain a partial discharge signal and a noise signal after preliminary noise reduction.
The cable partial discharge signal processing method provides a cable partial discharge signal noise reduction algorithm based on successive variation modal decomposition (Successive Variational Modal Decomposition, SVMD for short). The method applies SVMD to adaptively decompose a plurality of intrinsic mode components (Intrinsic Mode Function, IMF for short) of the partial discharge signal; the effective component and the noise component are distinguished by the cross-correlation coefficient, thereby preliminarily removing the noise component.
The step S102 of performing successive variation modal decomposition on the partial discharge signal may include:
establishing constraint criteria according to the formulas (1), (2) and (3):
establishing a constraint minimization model according to the formula (4):
wherein,
solving the lagrangian function according to equation (6) to obtain a plurality of eigenmode components:
in the formulae (1) to (6), J 1 、J 2 、J 3 Respectively representing a first constraint, a second constraint and a third constraint,represents the partial derivative with respect to time t, delta (t) is a dirac function, x is a convolution symbol, ω n Represents the center frequency, beta, of the nth mode i (t) and beta n (t) represents impulse response of the setting filter, alpha is a preset balance parameter, f (t) represents partial discharge signal, u n (t) and f r (t) two decomposed signals of f (t), u n (t) corresponds to the nth eigenmode component, f r (t) is the residual of f (t); f (f) r (t) the first n-1 modes and the unprocessed part f of the signal u (t) and lambda (t) is the Lagrangian operator.
The above calculation process converts the signal decomposition into a variation solution problem, and finds an optimal solution for the variation model to perform the signal decomposition, and the eigenvalue components are obtained after the decomposition, and the eigenvalue components after the decomposition not only satisfy that the sum of the components is equal to the partial discharge signal, but also require that the sum of the estimated bandwidths of each modal function is minimum.
And (3) converting the formula (4) into an unconstrained optimization problem according to the solving thought of the model in the VMD algorithm, and establishing an augmented Lagrange function to obtain an optimal solution.
As with the VMD method, the Alternate Direction Multiplier Method (ADMM) is used to iteratively solve the minimization problem described above, with the final iterative updateAnd lambda (lambda) n+1 。
The SVMD can adaptively resolve the appropriate number of modes without worrying about the choice of parameter K (number of IMF components). Thus, setting of the parameter α is critical, an excessively high value of α may lead to some erroneous modes, and a small value of α may lead to a mixing problem regarding two or more modes as one mode. Through a great deal of researches of the inventor, the preset balance parameter alpha is set to be alpha=2000, the noise reduction effect is most ideal, and the noise reduction characteristics of the cable partial discharge signals are more met.
The cross-correlation coefficient can be calculated according to equation (7):
the calculation formula of the set threshold (which may be referred to as a cross-correlation threshold) is calculated according to equation (8):
wherein x is n Representing the effective modal component, y n Which is indicative of the partial discharge signal,data mean value representing partial discharge signal, +.>A data mean representing the effective modal components.
IMF components decomposed using SVMD methods can be categorized into an effective component and a noise component. In order to be able to distinguish effectively, the cross correlation coefficient method is used for distinguishing. And judging the modal component exceeding the threshold value as an effective component according to the cross-correlation coefficient between each IMF component and the noise-containing partial discharge signal and the set threshold value, and judging the modal component exceeding the threshold value as a noise component on the contrary. The cross correlation coefficient is used to express the linear correlation strength between two random variables. The closer the cross correlation coefficient is to 1, the stronger the correlation between the two variables.
After the preliminary noise-reduced partial discharge signal is obtained through steps S101 to S105, the method for processing a cable partial discharge signal may further include:
and performing singular value decomposition noise reduction on the partial discharge signal after primary noise reduction to obtain a partial discharge signal with secondary noise reduction. Specifically, the step of performing singular value decomposition noise reduction on the partial discharge signal after preliminary noise reduction comprises the following steps: converting the partial discharge signal after preliminary noise reduction into a Hankel matrix; singular value decomposition is carried out on the Hankel matrix; performing matrix reconstruction according to the decomposed singular values to obtain a reconstructed matrix; and calculating according to the reconstruction matrix to obtain a partial discharge signal with secondary noise reduction.
The step of performing matrix reconstruction from the decomposed singular values may include: drawing a singular value entropy increment curve spectrum, and determining a singular value effective order according to the mutation position of the singular value entropy increment curve spectrum; reserving a second preset number of singular values according to the effective orders of the singular values, and setting the rest singular values to zero; and calculating by using singular value inverse decomposition to obtain a reconstruction matrix.
The SVD method is a matrix decomposition method that provides an efficient way to understand and process complex high-dimensional data information from which important features and useful information are extracted. After SVMD processing and cross-correlation coefficient screening, partial discharge signals still contain a small amount of white noise, and the SVD method can effectively inhibit noise signals.
For discrete signals x= [ X (1), X (2), X (N) ], an m×n order matrix is constructed as shown in formula (9):
in the formula (9), 1 < N, m=n-n+1.
The H matrix can be obtained through singular value decomposition:
H＝USV T (10)
in the formula, U, V is a feature vector matrix, and s= [ diag (σ) 1 ,σ 2 ,...,σ k )]，σ 1 ,σ 2 ,...,σ k Known as the singular values of the H matrix. The useful signal can be reflected by the first several larger singular values, the rest reflects the noise signal, and the singular value of the noise signal takes zero. And selecting a plurality of larger singular values to carry out selective reconstruction, so as to realize noise signal suppression.
The method of this embodiment may employ singular value entropy increment Qu Xianpu to determine the choice of reconstruction order. The noise reduction effects of the obtained signals are different due to different singular value noise reduction orders, so that the effective singular value order can be determined by drawing a singular entropy increment graph and searching a singular value mutation position. Namely, the singular spectrum order when the singular entropy starts to be reduced to an asymptotic value is selected as the signal singular spectrum noise reduction order.
And (3) obtaining a reconstructed matrix A by utilizing singular value inverse decomposition, and obtaining a real signal y (i) by averaging opposite angle lines of the matrix A, wherein i=1, 2 and 3 … N.
Where p=max (1, i-m+1), q=min (n, i). And (3) applying a Hankel matrix dimension-increasing to the reconstructed effective signal, then carrying out singular value decomposition, drawing a singular value increment graph, searching a singular value mutation position, removing Gaussian white noise, and carrying out effective signal reconstruction to obtain a noise-reduced signal.
After the step of obtaining the partial discharge signal with the secondary noise reduction, the method for processing the cable partial discharge signal further comprises the following steps: extracting characteristic quantity of the partial discharge signal with the secondary noise reduction; and inputting the characteristic quantity of the partial discharge signal into a pre-trained first deep learning network model to perform cable partial discharge pattern recognition.
The process of extracting the characteristic quantity of the partial discharge signal with the secondary noise reduction specifically comprises the following steps: drawing a PRPD spectrogram (Phase Resolved Partial Discharge, PRPD) of the partial discharge signal, wherein the extracted characteristic quantities comprise: skewness, steepness, cross-correlation coefficient, shannon entropy, and the like. The extracted feature quantity is input as an input vector to the first deep learning network model. The first deep learning network model may be obtained in advance by training the local discharge signal and the discharge mode thereof, and the obtained cable local discharge mode identification may include, but is not limited to: creepage of the outer semiconducting layer, internal air gaps, surface scratches, metal contamination.
In addition, after obtaining the noise signal, the cable partial discharge signal processing method further includes: extracting characteristic quantity of the noise signal; and inputting the characteristic quantity of the noise signal into a pre-trained second deep learning network model to perform noise type recognition. The noise signal separated by the successive variation modal decomposition and the mutual correlation coefficient mainly includes power frequency noise (noise generated by the operating frequency (e.g., 50Hz or 60 Hz) of the power system), high frequency noise (high frequency interference and transient noise such as overvoltage and oscillation caused by switching operation, lightning strike, arc discharge, etc.), thermal noise (noise generated by thermal motion of the internal resistor of the electronic device), flicker noise. The different noise reflects the operation environment of the cable, so that the operation environment of the cable can be known through analysis of the noise signal, and noise sources (such as internal noise, external interference and system operation noise of equipment) can be identified. The second deep learning network model may be obtained in advance by training a noise signal common to the power system and the type thereof.
In the prior art, the analysis of the partial discharge signal mainly focuses on the processing of filtering noise and the discharge signal, but the method of the embodiment further analyzes the noise signal, combines with the analysis result of the partial discharge signal, more comprehensively learns the running state of the cable, and provides an analysis means for further preventing the occurrence of fault functions.
FIG. 2 is an application flow chart of a method of processing a cable partial discharge signal according to one embodiment of the invention; the method for processing the cable partial discharge signal in the embodiment may include the following steps when applied specifically:
step S201, the original collected partial discharge signals are decomposed by SVMD, and K optimal IMF components are obtained in a self-adaptive mode;
step S202, solving a cross correlation coefficient of each IMF component and a partial discharge signal;
step S203, calculating a cross-correlation threshold value, and accurately identifying an effective component and a noise component according to the cross-correlation threshold value;
step S204, reconstructing the effective component and the noise component respectively to obtain a partial discharge signal and a noise signal after preliminary noise reduction;
step S205, converting the partial discharge signal after preliminary noise reduction into a Hankel matrix, and selecting the first a larger singular values for reconstruction through singular value decomposition and singular value differential spectrum calculation to obtain a partial discharge signal after secondary noise reduction;
and S206, analyzing the partial discharge signal and the noise signal of the secondary noise reduction respectively, determining the cable partial discharge mode and the noise signal type, and comprehensively analyzing the results and outputting.
The analysis of the partial discharge signals and the noise signals is beneficial to understanding the influence on the performance of the power system, and effective inhibition and control measures are adopted to ensure the safe, stable and efficient operation of the power system.
A great deal of literature and actual measurement experiments show that the partial discharge signals are mainly expressed by four mathematical models. The characteristics of the partial discharge signal are mostly oscillation damping, and two oscillation damping models are selected for simulation, as shown in the formula (12) and the formula (13):
single exponential oscillation decay model:
y 1 (t)＝A 1 e -t/τ sin(2πf c t) (12)
double exponential oscillation decay model:
y 2 (t)＝A 2 (e -1.3t/τ -e -2.2t/τ )sin(2πf c t) (13)
in the formula, A 1 、A 2 Is the local signal amplitude, A 1 ＝3mVA 2 ＝5mV；f c Is the oscillation frequency f c =3 MHZ; τ is the decay coefficient, τ=0.8 μs.
The number of sampling points is set to 2000 and the sampling frequency is set to 50MHz. Fig. 3 shows a waveform diagram of a simulated clean (noiseless) partial discharge signal. To simulate a real partial discharge signal, gaussian white noise (signal to noise ratio of-3 dB) and periodic narrowband interference are added to the partial discharge signal. The periodic narrowband interference is shown in formula (14): PD signals after different noise are added.
Wherein A is i Is the amplitude; f is the frequency;for the initial phase angle, the values are shown in table 1:
table 1 narrowband interference parameter settings
As can be seen from fig. 4, the partial discharge signal has been completely submerged in noise, and then the noise reduction process is performed by applying the method of this embodiment, to obtain a waveform close to that of the noiseless partial discharge signal. The partial discharge signal containing noise is decomposed by using the SVMD method, the parameter alpha=2000 is selected, the self-adaptive decomposition result is shown in fig. 5, and each curve shows waveforms of IMF1, IMF2, IMF3, IMF4, IMF5 and IMF6 respectively. Table 2 is a table of cross-correlation coefficient values for each IMF and the partial discharge signal containing noise.
TABLE 2 Cross-correlation coefficient of IMF components with noisy PD signals
| Component(s) | Cross correlation coefficient |
| IMF1 | 0.1863 |
| IMF2 | 0.7593 |
| IMF3 | 0.0317 |
| IMF4 | 0.0115 |
| IMF5 | 0.0205 |
| IMF6 | 0.0086 |
Calculating a threshold according to a cross-correlation coefficient threshold formulaAs is clear from table 2, the cross-correlation coefficient values of IMF1 and IMF2 are larger than 0.1653, and belong to the effective components, and the cross-correlation coefficient values of IMF3 to IMF6 are smaller than 0.1653, and belong to the noise components. The effective component is reconstructed and the noise component is removed. The reconstructed signal is shown in fig. 6.
The partial discharge signal after the effective component reconstruction still contains noise, and the effect is not ideal through cross-correlation coefficient screening noise reduction treatment after SVMD decomposition. Further noise reduction was performed using SVD. The reconstructed PD signal is converted into a Hankel matrix, and after singular value decomposition, a series of singular values are obtained, the distribution of which is shown in FIG. 7. And carrying out singular value difference spectrum calculation on the signals, and selecting the first 55 larger singular values to reconstruct as the PD signals after noise reduction.
In order to examine the superiority and feasibility of the method, the processing results of the method to which the present embodiment is applied are compared with EMD-ICA, VMD-wavelet thresholds.
FIG. 8 is a waveform diagram of a final noise reduced signal of a simulation experiment to which a cable partial discharge signal processing method according to an embodiment of the present invention is applied; FIG. 9 is a waveform diagram of a comparison result signal obtained using the EMD-ICA method; fig. 10 is a waveform diagram of a comparison result signal obtained using the VMD-wavelet thresholding method. By comparison, the following can be concluded: the EMD-ICA method has poor noise reduction effect, residual noise seriously affects signal identification, and partial waveform is distorted; the VMD-wavelet threshold method has better noise reduction effect than the former, but the noise is not completely filtered; compared with the first two methods, the noise-reduced signal obtained by processing in the embodiment has most of noise filtered, no waveform distortion phenomenon, and smoother and tidier waveform.
The noise reduction of the partial discharge signal should follow two main principles: on the one hand, the algorithm can effectively extract the signal, namely the signal to noise ratio of the noise reduction signal is high. On the other hand, the distortion of the noise reduction signal is small. The noise-reduced simulated partial discharge signal is evaluated by using a signal-to-noise ratio (SNR) and a Root Mean Square Error (RMSE). Since the measured experiment cannot acquire a clean PD signal, the noise reduction performance of the measured partial discharge signal is analyzed using a noise suppression ratio (NRR). The definition is shown in formulas (15) to (17).
Wherein y is i Representing the partial discharge signal after noise reduction, x i A noise-free partial discharge signal is represented, and n is a sampling point; sigma (sigma) 1 Representing standard deviation, sigma of signal before noise reduction 2 Representing the standard deviation of the signal after noise reduction.
To further verify the superiority of the method in noise reduction performance under different noise environments, white noise and periodic narrow-band interference with different signal-to-noise ratios are added to the noiseless partial discharge signals. As can be seen from Table 3, the noise reduction index of the method of the embodiment is superior to that of the other two methods, and compared with the VMD-wavelet threshold method with better noise reduction effect, the average signal to noise ratio is improved by 17.98%, and the average mean square error is reduced by 48.58%.
Table 3 comparison of noise reduction effects of simulated partial discharge signals
In order to further verify the noise reduction capability and the application prospect of the method of the embodiment, the inventor also processes and analyzes the actually measured partial discharge signals. Fig. 11 shows a schematic circuit diagram of the measured platform. The partial discharge test platform includes: transformer 111, protection resistor 112, 10kV cable strike 113, cable termination 114, oscilloscope 115, bandpass filter 118, high-frequency current sensor 117, capacitive voltage divider 116. And the partial discharge test platform performs a striking partial discharge experiment on the 10kV high-voltage cable to simulate the real cable insulation defect. A high frequency current sensor (HFCT) is used to extract the partial discharge signal at the cable insulation defect, and the actual measurement of the partial discharge signal is shown in fig. 12.
As seen from fig. 12, the partial discharge signal is submerged in noise, and the feature information cannot be effectively recognized. In order to highlight the noise reduction performance of the method herein, the measured partial discharge signals were noise reduced using the methods of the present embodiment, EMD-ICA, VMD-wavelet thresholds, respectively, the results of which are shown in FIGS. 13, 14, 15.
It is obvious from comparison that the effect obtained by the method is overall better than that of the other two methods, and the characteristic information of the signals can be well reserved while noise is reduced to the greatest extent. As shown in table 4, the noise reduction ability of these 3 methods was evaluated using NRR, and it can be seen that the noise suppression ability of the methods herein was strongest.
Table 4 comparison of the denoising effect of the partial discharge signals measured
| Noise reduction method | EMD-ICA | VMD-wavelet threshold | Method of the present embodiment |
| NRR | 5.765 | 7.317 | 8.684 |
Compared with the EMD-ICA and VMD-wavelet threshold methods, the method of the embodiment can effectively retain the characteristic information of local signals, has no obvious distortion of waveforms, has superior noise suppression capability, achieves the expected effect and has good application prospect.
Embodiments of the present invention also provide a schematic diagram of a computer storage medium having a computer program stored thereon, the computer program when executed being configured to implement a method for processing a cable partial discharge signal according to any one of the above.
It should be noted that the logic and/or steps represented in the flowcharts or otherwise described herein, for example, may be considered as a ordered listing of executable instructions for implementing logical functions, and may be embodied in any computer-storage medium for use by or in connection with an instruction execution system, apparatus, or device, such as a computer-based system, processor-containing system, or other system that can fetch the instructions from the instruction execution system, apparatus, or device and execute the instructions.
For the purposes of this description of embodiments, a computer-storage medium can be any apparatus that can contain, store, communicate, propagate, or transport the program for use by or in connection with the instruction execution system, apparatus, or device. More specific examples (a non-exhaustive list) of the computer storage medium would include the following: an electrical connection (electronic device) having one or more wires, a portable computer diskette (magnetic device), a Random Access Memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM).
It should be understood that the various processes of the present embodiment 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.
By now it should be appreciated by those skilled in the art that while a number of exemplary embodiments of the invention have been shown and described herein in detail, many other variations or modifications of the invention consistent with the principles of the invention may be directly ascertained or inferred from the present disclosure without departing from the spirit and scope of the invention. Accordingly, the scope of the present invention should be understood and deemed to cover all such other variations or modifications.
Claims (10)
1. A method of processing a cable partial discharge signal, comprising:
collecting partial discharge signals of the cable;
performing successive variation modal decomposition on the partial discharge signal, and adaptively decomposing a plurality of intrinsic modal components;
calculating the cross-correlation coefficient of each intrinsic mode component and the partial discharge signal;
taking the intrinsic mode component with the cross-correlation coefficient larger than a set threshold value as an effective mode component, and taking the intrinsic mode component with the cross-correlation coefficient smaller than the set threshold value as a noise mode component;
and respectively carrying out data reconstruction on the effective modal component and the noise modal component to obtain a partial discharge signal and a noise signal after preliminary noise reduction.
2. The method of processing a partial discharge signal of a cable according to claim 1, wherein the step of subjecting the partial discharge signal to successive variation modal decomposition comprises:
establishing constraint criteria according to the formulas (1), (2) and (3):
establishing a constraint minimization model according to the formula (4):
wherein,
solving the lagrangian function according to equation (6) to obtain the plurality of eigenmode components:
in the formulas (1) to (6), J 1 、J 2 、J 3 Respectively representing a first constraint, a second constraint and a third constraint,represents the partial derivative with respect to time t, delta (t) is a dirac function, x is a convolution symbol, ω n Representing the nth modeCenter frequency of state beta i (t) and beta n (t) respectively represents impulse response of the setting filter, alpha is a preset balance parameter, f (t) represents the partial discharge signal, u n (t) and f r (t) two decomposed signals of f (t), u n (t) corresponds to the nth eigenmode component, f r (t) is the residual of f (t); f (f) r (t) the first n-1 modes and the unprocessed part f of the signal u (t) composition.
3. The method for processing a partial discharge signal of a cable according to claim 2, wherein,
the preset balance parameter α is set to α=2000.
4. The method for processing a partial discharge signal of a cable according to claim 2, wherein,
the cross-correlation coefficient is calculated according to the formula (7):
the calculation formula of the set threshold is calculated according to the formula (8):
wherein x is n Representing the effective modal component, y n Representing the partial discharge signal in question,data mean value representing said partial discharge signal, is->A data mean representing the effective modal component.
5. The method for processing a cable partial discharge signal according to claim 1, further comprising, after obtaining the preliminary noise-reduced partial discharge signal:
and performing singular value decomposition noise reduction on the partial discharge signal after the preliminary noise reduction to obtain a partial discharge signal with secondary noise reduction.
6. The method for processing a cable partial discharge signal according to claim 5, wherein the step of performing singular value decomposition noise reduction on the preliminary noise reduced partial discharge signal comprises:
converting the partial discharge signal after preliminary noise reduction into a Hankel matrix;
singular value decomposition is carried out on the Hankel matrix;
performing matrix reconstruction according to the decomposed singular values to obtain a reconstructed matrix;
and calculating according to the reconstruction matrix to obtain the partial discharge signal with the secondary noise reduction.
7. The method for processing a cable partial discharge signal according to claim 6, wherein the step of performing matrix reconstruction from the decomposed singular values comprises:
drawing a singular value entropy increment curve spectrum, and determining a singular value effective order according to the mutation position of the singular value entropy increment curve spectrum;
reserving a second preset number of singular values according to the effective orders of the singular values, and setting the rest singular values to zero;
and calculating to obtain the reconstruction matrix by using singular value inverse decomposition.
8. The method for processing a cable partial discharge signal according to claim 5, further comprising, after the step of obtaining the secondary noise-reduced partial discharge signal:
extracting characteristic quantity of the partial discharge signal with the secondary noise reduction;
and inputting the characteristic quantity of the partial discharge signal into a pre-trained first deep learning network model to perform cable partial discharge pattern recognition.
9. The method for processing a cable partial discharge signal according to claim 1, further comprising, after obtaining the noise signal:
extracting a feature quantity from the noise signal;
and inputting the characteristic quantity of the noise signal into a pre-trained second deep learning network model to perform noise type recognition.
10. A computer storage medium having stored thereon a computer program which, when executed, implements the method of processing a cable partial discharge signal according to any one of claims 1 to 9.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410052105.5A CN117851758A (en) | 2024-01-12 | 2024-01-12 | Cable partial discharge signal processing method and computer storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410052105.5A CN117851758A (en) | 2024-01-12 | 2024-01-12 | Cable partial discharge signal processing method and computer storage medium |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117851758A true CN117851758A (en) | 2024-04-09 |
Family
ID=90546022
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410052105.5A Pending CN117851758A (en) | 2024-01-12 | 2024-01-12 | Cable partial discharge signal processing method and computer storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117851758A (en) |
-
2024
- 2024-01-12 CN CN202410052105.5A patent/CN117851758A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108804832B (en) | VMD-based interval threshold partial discharge denoising method | |
| CN108845250B (en) | On-load tap-changer fault identification method based on vibration signal feature extraction | |
| CN109975665B (en) | Power equipment partial discharge signal discharge type identification method | |
| CN112014692A (en) | Partial discharge ultrahigh frequency signal blind source separation denoising method based on principal component analysis | |
| CN110717472B (en) | Fault diagnosis method and system based on improved wavelet threshold denoising | |
| Wang et al. | Noise suppression of corona current measurement from HVdc transmission lines | |
| CN114492538B (en) | Urban medium-voltage distribution cable partial discharge signal denoising method | |
| Ashtiani et al. | Feature-oriented de-noising of partial discharge signals employing mathematical morphology filters | |
| CN113269082A (en) | Partial discharge threshold denoising method based on improved variational modal decomposition | |
| CN107356843B (en) | Transformer partial discharge fault diagnosis method based on hierarchical threshold synchronous extrusion wavelet | |
| Yan et al. | An automatic tool for partial discharge de-noising via short-time Fourier transform and matrix factorization | |
| CN110954779A (en) | Voltage sag source feature identification method based on S transformation and multidimensional fractal | |
| Chan et al. | Hybrid method on signal de‐noising and representation for online partial discharge monitoring of power transformers at substations | |
| CN113486750A (en) | Oil-immersed transformer partial discharge signal denoising method based on improved VMD algorithm and wavelet packet denoising algorithm | |
| CN108089100A (en) | The detection method of small current neutral grounding system arc light resistance earth fault | |
| CN112528853B (en) | Improved dual-tree complex wavelet transform denoising method | |
| Zhang et al. | Suppression of UHF partial discharge signals buried in white-noise interference based on block thresholding spatial correlation combinative de-noising method | |
| CN111239565B (en) | Oil-filled casing partial discharge pulse signal processing method and system based on layered denoising model | |
| CN111044865B (en) | Partial discharge wavelet denoising method for direct-current gas insulation electrical equipment | |
| Xu et al. | Novel denoizing method for partial discharge signals using singular value decomposition and spectral subtraction | |
| CN117851758A (en) | Cable partial discharge signal processing method and computer storage medium | |
| CN114861722B (en) | Partial discharge narrow-band interference suppression method based on time-frequency spectrogram separation | |
| Mitiche et al. | An evaluation of total variation signal denoising methods for partial discharge signals | |
| Li | Research on extraction of partial discharge signals based on wavelet analysis | |
| CN112485616A (en) | Cable insulation aging detection method and device, storage medium and processor |
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 |