CN117872052A - Defect identification method and device based on partial discharge high-frequency current pulse - Google Patents
Defect identification method and device based on partial discharge high-frequency current pulse Download PDFInfo
- Publication number
- CN117872052A CN117872052A CN202311800140.2A CN202311800140A CN117872052A CN 117872052 A CN117872052 A CN 117872052A CN 202311800140 A CN202311800140 A CN 202311800140A CN 117872052 A CN117872052 A CN 117872052A
- Authority
- CN
- China
- Prior art keywords
- signal
- target
- partial discharge
- discharge high
- frequency
- 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
- 230000007547 defect Effects 0.000 title claims abstract description 76
- 238000000034 method Methods 0.000 title claims abstract description 35
- 238000001228 spectrum Methods 0.000 claims abstract description 50
- 238000010586 diagram Methods 0.000 claims abstract description 40
- 230000009466 transformation Effects 0.000 claims abstract description 33
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 14
- 238000007781 pre-processing Methods 0.000 claims description 27
- 238000012545 processing Methods 0.000 claims description 19
- 230000005540 biological transmission Effects 0.000 claims description 12
- 238000000605 extraction Methods 0.000 claims description 10
- 230000000630 rising effect Effects 0.000 claims description 10
- 238000001514 detection method Methods 0.000 claims description 9
- 238000013527 convolutional neural network Methods 0.000 claims description 7
- 230000003321 amplification Effects 0.000 claims description 6
- 238000004458 analytical method Methods 0.000 claims description 6
- 238000003199 nucleic acid amplification method Methods 0.000 claims description 6
- 238000004364 calculation method Methods 0.000 claims description 5
- 230000007613 environmental effect Effects 0.000 claims description 4
- 238000012935 Averaging Methods 0.000 claims description 3
- 230000002159 abnormal effect Effects 0.000 claims description 3
- 238000001914 filtration Methods 0.000 claims description 3
- 230000001174 ascending effect Effects 0.000 claims description 2
- 239000012212 insulator Substances 0.000 description 3
- 230000001131 transforming effect Effects 0.000 description 2
- 230000001052 transient effect Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000002950 deficient Effects 0.000 description 1
- 230000005684 electric field Effects 0.000 description 1
- 238000009499 grossing Methods 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 230000007257 malfunction Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000035772 mutation Effects 0.000 description 1
- 238000012549 training Methods 0.000 description 1
Landscapes
- Testing Relating To Insulation (AREA)
Abstract
The invention discloses a defect identification method and device based on partial discharge high-frequency current pulse, which relate to the technical field of partial discharge defect identification, wherein a target period is obtained by carrying out short-time Fourier transform on a preprocessed signal, and the preprocessed signal is split into a plurality of first signals according to the target period, so that mutual interference among the plurality of signals is avoided; the instantaneous characteristic quantity of each signal is obtained by carrying out empirical mode decomposition and Hilbert transformation on each first signal, the time-frequency plane transformation is carried out on each characteristic quantity to obtain a Hilbert spectrum, and the target local characteristic and the corresponding frequency distribution diagram are extracted according to the Hilbert spectrum, so that the defect type of the local discharge high-frequency current pulse is determined, and the signal defect type of the local discharge high-frequency current pulse is more accurately identified.
Description
Technical Field
The invention belongs to the technical field of partial discharge defect identification, and particularly relates to a defect identification method and device based on partial discharge high-frequency current pulses.
Background
Partial discharge high-frequency current pulses are a special form of partial discharge, a rapid discharge phenomenon in the insulator of high-voltage power equipment due to the local concentration of an electric field. The rise time of the partial discharge high-frequency current pulse is very short, usually on the order of nanoseconds, so that electromagnetic waves of high frequency can be excited in an extremely short time. The frequency of these electromagnetic waves can range from hundreds of megahertz to several gigahertz and even higher. When the voltage of the power equipment is too high, the insulator is defective or aged, a partial discharge high-frequency current pulse may be generated, and the discharge phenomenon may damage the insulator and may cause malfunction or damage of the power equipment.
Through defect detection and analysis of the partial discharge high-frequency current pulse, potential faults and defects of the power equipment can be found in time, so that corresponding measures are taken for maintenance and replacement, and safe and stable operation of the power equipment is ensured.
However, when a plurality of defects occur at the same position, the generated partial discharge high-frequency current pulse signals affect each other, so that the signals generated by the partial discharge cannot be detected normally, and defect detection and analysis cannot be performed.
Disclosure of Invention
The invention aims to solve the problems that when a plurality of defects appear in the same position, a plurality of partial discharge high-frequency current pulse signals are mutually influenced, so that the signals generated by partial discharge cannot be detected normally and defect detection and analysis cannot be carried out, and provides a defect identification method and device based on the partial discharge high-frequency current pulse.
In a first aspect of the present invention, a defect identification method based on partial discharge high-frequency current pulses is first provided, the method comprising:
preprocessing a partial discharge high-frequency current pulse signal to obtain a preprocessed signal, and performing short-time Fourier transform on the preprocessed signal to obtain a target period of the preprocessed signal;
splitting the preprocessing signal according to the target period to obtain a plurality of first signals;
for each first signal, performing empirical mode decomposition on the first signal to obtain a modal function IMF component set of the first signal, performing Hilbert transform on each IMF component in the IMF component set to obtain instantaneous characteristic quantities of each IMF component in the IMF component set, and recording the instantaneous characteristic quantities as a characteristic quantity set;
performing time-frequency plane transformation on each instantaneous characteristic quantity in the characteristic quantity set to obtain a Hilbert spectrum, and obtaining a target local characteristic of the first signal and a corresponding target frequency distribution diagram according to the Hilbert spectrum;
and obtaining the defect types of the partial discharge high-frequency current pulses according to the target local characteristics of each first signal and the corresponding target frequency distribution diagram.
Optionally, before the preprocessing of the partial discharge high-frequency current pulse signal to obtain a preprocessed signal, the method further comprises:
collecting a target cable transmission signal, and subtracting a preset standard transmission signal to obtain the partial discharge high-frequency current pulse signal; the preset standard transmission signal is a signal detected under the condition that the target cable is not leaked.
Optionally, preprocessing the partial discharge high-frequency current pulse signal to obtain a preprocessed signal includes:
acquiring the partial discharge high-frequency current pulse signal, and eliminating the environmental noise of the partial discharge high-frequency current pulse signal through a filter to obtain a first signal;
and carrying out signal amplification processing on the first signal through an amplifying circuit, and removing the low-quality signal, the value missing signal and the abnormal value signal in the signal amplification process to obtain a preprocessed signal.
Optionally, performing short-time fourier transform on the pre-processed signal to obtain a target period of the pre-processed signal specifically includes:
dividing the pre-processed signal into a plurality of time windows of equal length; the length of the time window is the power of 2;
performing window boundary processing on the time window by using a window function to obtain a target time window; the window boundary processing multiplies the time window by the application window function in the time domain;
performing Fourier transform on the target window, and converting the preprocessing signal from a time domain to a frequency domain to obtain a target frequency spectrum;
according to the amplitude and phase information of the target frequency spectrum, main frequency components of the target frequency spectrum are obtained;
and obtaining the target period of the preprocessing signal according to the main frequency component.
Optionally, performing empirical mode decomposition on the first signal to obtain a set of modal functions IMF components of the first signal, specifically including:
step one: initializing a preset model by taking the first signal as an initial signal; the preset model is obtained by carrying out extreme point detection and envelope calculation on a large number of signals through a convolutional neural network CNN;
step two: detecting and calculating upper and lower extreme points of an initial signal through a preset model to obtain an upper envelope curve and a lower envelope curve corresponding to the initial signal; the upper envelope line is formed for the upper extreme point; the lower envelope line is formed for the lower extreme point;
step three: summing and averaging the average value of the upper envelope curve and the average value of the lower envelope curve to obtain an average value envelope curve signal;
step four: obtaining a difference value between the initial signal and a corresponding mean envelope signal to obtain an intermediate signal;
step five: if the intermediate signal does not meet the preset condition, the intermediate signal is used as an initial signal, and the steps two to four are repeated until the intermediate signal meets the preset condition;
step six: if the intermediate signal meets the preset condition, the intermediate signal is marked as a first IMF component; subtracting the first IMF component from the initial signal to serve as a new initial signal, repeating the second step to the fourth step, and the like to obtain a plurality of IMF components until the residual error value of the preprocessing signal is smaller than a preset threshold value; and if the residual value is smaller than the preset threshold value, the current IMF component cannot be extracted, and the process of empirical mode decomposition is finished.
Optionally, the preset condition is that the number of the upper extreme point and the lower extreme point of the initial signal and the number of the zero crossing point of the initial signal cannot differ by more than one in a data segment of the initial signal; the preset condition is that the upper envelope curve and the lower envelope curve corresponding to the initial signal are locally symmetrical relative to a time axis.
Optionally, hilbert transformation is performed on each IMF component in the IMF component set to obtain instantaneous feature quantities of each IMF component in the IMF component set, which specifically includes:
acquiring each IMF component in the IMF component set, and converting each IMF component in the IMF component set from a time domain to a complex frequency domain through Hilbert transformation to obtain an instantaneous characteristic quantity; the instantaneous characteristic quantity is complex, the real part is the amplitude of the current IMF component, and the imaginary part is the phase of the current IMF component.
Optionally, performing time-frequency plane transformation on each transient feature in the feature set to obtain a hilbert spectrum, and obtaining a target local feature of the first signal and a target frequency distribution diagram corresponding to the target local feature according to the hilbert spectrum, which specifically includes:
performing wavelet transformation and other time-frequency plane transformation on the instantaneous characteristic quantity, and converting the instantaneous characteristic quantity to a time-frequency plane by selecting different window functions and local characteristic parameters to obtain a Hilbert spectrum; the local characteristic parameters are peak rising time, peak falling time, peak voltage, valley falling time, valley rising time, valley voltage, average voltage, voltage crest factor, voltage profile coefficient and equivalent time length;
filtering and frequency domain analysis are carried out on the Hilbert spectrum;
the method comprises the steps of obtaining peak rising time, peak falling time and peak voltage of the peak maximum value of the Hilbert spectrum under equivalent time length to obtain peak characteristics;
obtaining valley value descending time, valley value ascending time and valley value voltage at the minimum value of the valley value under the equivalent time length in the Hilbert spectrum to obtain valley value characteristics;
obtaining signal intensity through average voltage under equivalent time length in the Hilbert spectrum;
acquiring the signal shape of the Hilbert spectrum passing through a voltage crest factor and a voltage shape coefficient under the equivalent time length;
and obtaining the target local characteristics of the first signal and the corresponding target frequency distribution diagram according to the peak value characteristics, the valley value characteristics, the signal strength and the signal shape.
Optionally, obtaining the defect type of the partial discharge high-frequency current pulse according to the target local feature of each first signal and the corresponding target frequency distribution diagram specifically includes:
searching a preset frequency distribution diagram set in a comparison database according to the target local characteristics of the first signal; the control database is a frequency distribution diagram obtained by the known partial discharge defects under different local characteristics;
if a target frequency distribution diagram corresponding to the target local feature of the first signal exists, determining the defect type of the target frequency distribution diagram as the target defect type of the first signal;
and obtaining the target defect types corresponding to the first signals, and obtaining the partial discharge high-frequency current pulse defect results by combining and collecting the target defect types.
In a second aspect of the present invention, a defect recognition device based on partial discharge high frequency current pulses is provided, comprising: the device comprises a preprocessing module, a signal splitting module, a signal characteristic extraction module, a signal characteristic processing module and a defect type determining module:
the pretreatment module is used for carrying out pretreatment on the partial discharge high-frequency current pulse signal to obtain a pretreatment signal, and carrying out short-time Fourier transformation on the pretreatment signal to obtain a target period of the pretreatment signal;
the signal splitting module is used for splitting the preprocessing signal through the target period to obtain a plurality of first signals;
the signal characteristic extraction module is used for carrying out empirical mode decomposition on each first signal to obtain a mode function IMF component set of the first signal, carrying out Hilbert transformation on each IMF component in the IMF component set to obtain instantaneous characteristic quantities of each IMF component in the IMF component set, and marking the instantaneous characteristic quantities as characteristic quantity sets;
the signal characteristic processing module is used for carrying out time-frequency plane transformation on each instantaneous characteristic quantity in the characteristic quantity set to obtain a Hilbert spectrum, and obtaining a target local characteristic of the first signal and a corresponding target frequency distribution diagram according to the Hilbert spectrum;
the defect type determining module is used for obtaining the defect type of the partial discharge high-frequency current pulse according to the target local characteristics of each first signal and the corresponding target frequency distribution diagram.
The invention has the beneficial effects that:
the invention provides a defect identification method based on partial discharge high-frequency current pulses, which comprises the steps of obtaining a target period by carrying out short-time Fourier transform on a preprocessed signal, splitting the preprocessed signal into a plurality of first signals according to the target period, obtaining instantaneous characteristic quantity of each signal by carrying out empirical mode decomposition and Hilbert transform on each first signal, carrying out time-frequency plane transform on each characteristic quantity to obtain a Hilbert spectrum, extracting target local characteristics and corresponding frequency distribution diagrams thereof according to the Hilbert spectrum, and determining defect types of the partial discharge high-frequency current pulses; by splitting and transforming the signals, mutual interference among a plurality of signals is avoided, and the split signals are subjected to feature extraction and time-frequency plane transformation to obtain target local features and corresponding frequency distribution diagrams thereof, so that defect types are determined, and the defect types in the partial discharge high-frequency current pulse signals are more accurately identified and classified.
Drawings
The invention is further described below with reference to the accompanying drawings.
FIG. 1 is a flow chart of a defect identification method based on partial discharge high frequency current pulses according to an embodiment of the present invention;
fig. 2 is a block diagram of a defect recognition device based on partial discharge high-frequency current pulses according to an embodiment of the present invention.
Detailed Description
The following description of the embodiments of the present invention will be made clearly and completely with reference to the accompanying drawings, in which it is apparent that the embodiments described are only some embodiments of the present invention, but not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
The embodiment of the invention provides a defect identification method based on partial discharge high-frequency current pulses. Referring to fig. 1, fig. 1 is a flowchart of a defect identification method based on partial discharge high-frequency current pulses according to an embodiment of the present invention. The method may comprise the steps of:
s101, preprocessing a local discharge high-frequency current pulse signal to obtain a preprocessed signal, and performing short-time Fourier transform on the preprocessed signal to obtain a target period of the preprocessed signal.
S102, splitting the preprocessing signals according to the target period to obtain a plurality of first signals.
S103, performing empirical mode decomposition on each first signal to obtain a modal function IMF component set of the first signal, performing Hilbert transform on each IMF component in the IMF component set to obtain instantaneous characteristic quantities of each IMF component in the IMF component set, and recording the instantaneous characteristic quantities as characteristic quantity sets.
S104, performing time-frequency plane transformation on each instantaneous characteristic in the characteristic quantity set to obtain a Hilbert spectrum, and obtaining a target local characteristic of the first signal and a corresponding target frequency distribution diagram according to the Hilbert spectrum.
S105, obtaining the defect type of the partial discharge high-frequency current pulse according to the target local characteristics of each first signal and the corresponding target frequency distribution diagram.
According to the electronic government information safety protection method provided by the embodiment of the invention, through splitting and transforming the signals, mutual interference among a plurality of signals is avoided, and the characteristics of the split signals are extracted and the time-frequency plane and frequency plane are transformed to obtain the target local characteristics and the frequency distribution diagram corresponding to the target local characteristics, so that defect types are determined, and the defect types in the partial discharge high-frequency current pulse signals are more accurately identified and classified.
In one implementation manner, preprocessing a partial discharge high-frequency current pulse signal to obtain a preprocessed signal, specifically, obtaining the partial discharge high-frequency current pulse signal, eliminating environmental noise of the partial discharge high-frequency current pulse signal through a filter to obtain a target signal, and compressing the target signal in a dynamic range to obtain the preprocessed signal.
In one implementation, preprocessing the partial discharge high-frequency current pulse signal to obtain a preprocessed signal, specifically, obtaining the partial discharge high-frequency current pulse signal, converting the partial discharge high-frequency current pulse signal into discrete sample data, and processing the sample data through a filter to obtain the preprocessed signal.
In one implementation manner, the preprocessed signals are split to obtain a plurality of first signals, the moving range of the target period can be determined according to at least 2 target circumferences of the target period, the target period is moved in the moving range until the length from the left peak to the starting point position of the target period and the length from the right peak to the end point position of the target period are the same as the length from the left peak to the end point position of the target period and the length from the right peak, and then the starting point position and the end point position are taken as splitting conditions of the preprocessed signals.
In one implementation, the hilbert spectrum is obtained by performing time-frequency plane transformation on each instantaneous characteristic quantity in the characteristic quantity set, so that local changes and mutation in the signal can be better captured, and the signal can be accurately analyzed.
In one embodiment, the method further comprises the following steps before preprocessing the partial discharge high-frequency current pulse signal to obtain a preprocessed signal:
collecting a target cable transmission signal, and subtracting a preset standard transmission signal to obtain a partial discharge high-frequency current pulse signal; the preset standard transmission signal is a signal detected under the condition that the target cable is not leaked.
In one implementation, a current transformer is placed on the transmission line of the target cable to collect the target cable transmission signal.
In one implementation, the target cable transmission signal is subtracted by a signal processing software such as MATLAB, labVIEW, signalPad, so as to obtain a partial discharge high-frequency current pulse signal.
In one embodiment, preprocessing the partial discharge high frequency current pulse signal to obtain a preprocessed signal includes:
obtaining a partial discharge high-frequency current pulse signal, and eliminating the environmental noise of the partial discharge high-frequency current pulse signal through a filter to obtain a first signal;
and carrying out signal amplification processing on the first signal through an amplifying circuit, and removing the low-quality signal, the value missing signal and the abnormal value signal in the signal amplification process to obtain a preprocessed signal.
In one implementation, the filter is one of a band pass filter, a high pass filter, a notch filter, and an all pass filter.
In one embodiment, performing short-time fourier transform on the pre-processed signal to obtain the target period of the pre-processed signal specifically includes:
dividing the preprocessed signal into a plurality of time windows of equal length;
performing window boundary processing on the time window by applying a window function to obtain a target time window; window boundary processing is to multiply a time window in the time domain by applying a window function;
performing Fourier transform on the target window, and converting the preprocessing signal from a time domain to a frequency domain to obtain a target frequency spectrum;
according to the amplitude and phase information of the target frequency spectrum, main frequency components of the target frequency spectrum are obtained;
the target period of the pre-processed signal is derived from the dominant frequency component.
In one implementation, the time window length is a power of 2.
In one implementation, the application window function may be any of a hanning window, a hamming window, a smoothing window, and a chebyshev window.
In one embodiment, performing empirical mode decomposition on the first signal to obtain a set of modal function IMF components of the first signal, including:
step one: initializing a preset model by taking the first signal as an initial signal; the preset model is obtained by carrying out extreme point detection and envelope calculation on a large number of signals through a convolutional neural network CNN;
step two: detecting and calculating upper and lower extreme points of an initial signal through a preset model to obtain an upper envelope curve and a lower envelope curve corresponding to the initial signal; the upper envelope is formed by an upper extreme point; the lower envelope is formed by the lower extreme point;
step three: summing and averaging the average value of the upper envelope curve and the average value of the lower envelope curve to obtain an average value envelope curve signal;
step four: obtaining a difference value between the initial signal and a corresponding mean envelope signal to obtain an intermediate signal;
step five: if the intermediate signal does not meet the preset condition, the intermediate signal is used as an initial signal, and the steps two to four are repeated until the intermediate signal meets the preset condition;
step six: if the intermediate signal meets the preset condition, marking the intermediate signal as a first IMF component; and subtracting the first IMF component from the initial signal to serve as a new initial signal, repeating the second step to the fourth step, and the like to obtain a plurality of IMF components until the residual error value of the preprocessing signal is smaller than a preset threshold value.
In one implementation, the extreme point detection and the envelope calculation are performed on the signals through the convolutional neural network, so that noise and other interference factors can be effectively processed, and more accurate extreme point detection and envelope calculation results are obtained.
In one implementation, the processing is performed in an iterative manner, so that the optimal solution can be quickly converged, and the efficiency of signal processing is improved.
In one implementation, if the residual value is smaller than the preset threshold value, the current IMF component cannot be extracted, and the process of empirical mode decomposition is ended.
In one implementation, the preset condition is that the number of the upper extreme points and the lower extreme points of the initial signal and the number of the zero crossing points cannot differ by more than one in the data section of the initial signal; the preset condition is that the upper envelope curve and the lower envelope curve corresponding to the initial signal are locally symmetrical relative to the time axis.
In one embodiment, performing hilbert transformation on each IMF component in the IMF component set to obtain instantaneous feature quantities of each IMF component in the IMF component set, specifically including:
and obtaining each IMF component in the IMF component set, and converting each IMF component in the IMF component set from a time domain to a complex frequency domain through Hilbert transformation to obtain the instantaneous characteristic quantity.
In one implementation, the instantaneous feature is complex, the real part is the amplitude of the current IMF component, and the imaginary part is the phase of the current IMF component.
In one implementation, by performing hilbert transformation on each IMF component in the IMF component set, more comprehensive, accurate and fine signal characteristic information can be obtained, which is helpful to improve the accuracy and efficiency of signal processing.
In one embodiment, performing time-frequency plane transformation on each transient feature in the feature set to obtain a hilbert spectrum, and obtaining a target local feature of the first signal and a target frequency distribution diagram corresponding to the target local feature according to the hilbert spectrum, which specifically includes:
performing wavelet transformation and other time-frequency plane transformation on the instantaneous characteristic quantity, and converting the instantaneous characteristic quantity to a time-frequency plane by selecting different window functions and local characteristic parameters to obtain a Hilbert spectrum; the local characteristic parameters are peak rising time, peak falling time, voltage at peak, valley falling time, valley rising time, voltage at valley, average voltage, voltage crest factor, voltage profile factor and equivalent time length;
filtering and frequency domain analysis are carried out on the Hilbert spectrum;
obtaining peak rising time, peak falling time and peak voltage of a peak maximum value of the Hilbert spectrum under equivalent time length to obtain peak characteristics;
obtaining valley value falling time, valley value rising time and valley value voltage at the minimum value of a valley value under the equivalent time length in the Hilbert spectrum to obtain valley value characteristics;
obtaining signal intensity by average voltage under equivalent time length in the Hilbert spectrum;
acquiring the signal shape passing through the voltage crest factor and the voltage shape factor in the Hilbert spectrum under the equivalent time length;
and obtaining the target local characteristic of the first signal and a corresponding target frequency distribution diagram according to the peak value characteristic, the valley value characteristic, the signal strength and the signal shape.
In one implementation, local features and variations of a signal can be better described by extracting peak rise time, peak fall time, voltage at peak, valley fall time, valley rise time, valley voltage, average voltage, voltage crest factor, voltage profile factor, equivalent time length local feature parameters.
In one implementation, the window function and the local feature parameters can be adjusted and optimized according to specific application scenes and requirements, so that the flexibility and applicability of the method are improved.
In one embodiment, the method for obtaining the defect type of the partial discharge high-frequency current pulse according to the target local feature of each first signal and the corresponding target frequency distribution diagram specifically includes:
searching a preset frequency distribution diagram set in a comparison database according to the target local characteristics of the first signal;
if a target frequency distribution map corresponding to the target local feature of the first signal exists, determining the defect type of the target frequency distribution map as the target defect type of the first signal;
and obtaining the target defect types corresponding to the first signals, and obtaining a partial discharge high-frequency current pulse defect result by combining the target defect types.
In one implementation, the control database is a histogram of known partial discharge defects obtained under different local characteristics.
In one implementation manner, the defect type can be determined more accurately and reliably by comparing the target local characteristic of the actual signal with a preset frequency distribution diagram set, and the preset frequency distribution diagram set is obtained through a large amount of data actual detection training, so that the preset frequency distribution diagram set can cover various possible signal characteristics, and the defect identification accuracy is improved.
The embodiment of the invention also provides a defect identification device based on partial discharge high-frequency current pulses based on the same inventive concept. Referring to fig. 2, fig. 2 is a block diagram of a defect recognition device based on partial discharge high-frequency current pulses according to an embodiment of the present invention, including: the device comprises a preprocessing module, a signal splitting module, a signal characteristic extraction module, a signal characteristic processing module and a defect type determining module:
the pretreatment module is used for carrying out pretreatment on the partial discharge high-frequency current pulse signal to obtain a pretreatment signal, and carrying out short-time Fourier transformation on the pretreatment signal to obtain a target period of the pretreatment signal;
the signal splitting module is used for splitting the preprocessed signals through a target period to obtain a plurality of first signals;
the signal feature extraction module is used for carrying out empirical mode decomposition on each first signal to obtain a mode function IMF component set of the first signal, carrying out Hilbert transformation on each IMF component in the IMF component set to obtain instantaneous feature quantity of each IMF component in the IMF component set, and marking the instantaneous feature quantity as a feature quantity set;
the signal characteristic processing module is used for carrying out time-frequency plane transformation on each instantaneous characteristic quantity in the characteristic quantity set to obtain a Hilbert spectrum, and obtaining a target local characteristic of the first signal and a corresponding target frequency distribution diagram according to the Hilbert spectrum;
and the defect type determining module is used for obtaining the defect type of the partial discharge high-frequency current pulse according to the target local characteristics of each first signal and the corresponding target frequency distribution diagram.
According to the electronic government information safety protection device provided by the embodiment of the invention, the signal splitting module splits and transforms the signals, so that mutual interference among a plurality of signals is avoided, the split signals are subjected to signal characteristic extraction through the signal characteristic extraction module, and the signal characteristic extraction module performs frequency plane transformation to obtain the target local characteristics and the frequency distribution diagram corresponding to the target local characteristics, so that defect types are determined, and the defect types in the partial discharge high-frequency current pulse signals are more accurately identified and classified.
The foregoing describes one embodiment of the present invention in detail, but the disclosure is only a preferred embodiment of the present invention and should not be construed as limiting the scope of the invention. All equivalent changes and modifications within the scope of the present invention are intended to be covered by the present invention.
Claims (10)
1. A defect identification method based on partial discharge high-frequency current pulses, the method comprising:
preprocessing a partial discharge high-frequency current pulse signal to obtain a preprocessed signal, and performing short-time Fourier transform on the preprocessed signal to obtain a target period of the preprocessed signal;
splitting the preprocessing signal according to the target period to obtain a plurality of first signals;
for each first signal, performing empirical mode decomposition on the first signal to obtain a modal function (I MF) component set of the first signal, performing Hilbert transform on each I MF component in the I MF component set to obtain instantaneous characteristic quantities of each I MF component in the I MF component set, and recording the instantaneous characteristic quantities as a characteristic quantity set;
performing time-frequency plane transformation on each instantaneous characteristic quantity in the characteristic quantity set to obtain a Hilbert spectrum, and obtaining a target local characteristic of the first signal and a corresponding target frequency distribution diagram according to the Hilbert spectrum;
and obtaining the defect types of the partial discharge high-frequency current pulses according to the target local characteristics of each first signal and the corresponding target frequency distribution diagram.
2. The defect identifying method based on partial discharge high frequency current pulse according to claim 1, further comprising, before preprocessing the partial discharge high frequency current pulse signal to obtain a preprocessed signal:
collecting a target cable transmission signal, and subtracting a preset standard transmission signal to obtain the partial discharge high-frequency current pulse signal; the preset standard transmission signal is a signal detected under the condition that the target cable is not leaked.
3. The defect identifying method based on partial discharge high frequency current pulse according to claim 1, wherein preprocessing the partial discharge high frequency current pulse signal to obtain a preprocessed signal comprises:
acquiring the partial discharge high-frequency current pulse signal, and eliminating the environmental noise of the partial discharge high-frequency current pulse signal through a filter to obtain a first signal;
and carrying out signal amplification processing on the first signal through an amplifying circuit, and removing the low-quality signal, the value missing signal and the abnormal value signal in the signal amplification process to obtain a preprocessed signal.
4. The defect recognition method based on partial discharge high-frequency current pulse according to claim 1, wherein performing short-time fourier transform on the pre-processed signal to obtain a target period of the pre-processed signal specifically comprises:
dividing the pre-processed signal into a plurality of time windows of equal length; the length of the time window is the power of 2;
performing window boundary processing on the time window by using a window function to obtain a target time window; the window boundary processing multiplies the time window by the application window function in the time domain;
performing Fourier transform on the target window, and converting the preprocessing signal from a time domain to a frequency domain to obtain a target frequency spectrum;
according to the amplitude and phase information of the target frequency spectrum, main frequency components of the target frequency spectrum are obtained;
and obtaining the target period of the preprocessing signal according to the main frequency component.
5. The defect identification method based on partial discharge high-frequency current pulse according to claim 1, wherein the empirical mode decomposition is performed on the first signal to obtain a mode function MF component set of the first signal, and specifically comprising:
step one: initializing a preset model by taking the first signal as an initial signal; the preset model is obtained by carrying out extreme point detection and envelope calculation on a large number of signals through a convolutional neural network CNN;
step two: detecting and calculating upper and lower extreme points of an initial signal through a preset model to obtain an upper envelope curve and a lower envelope curve corresponding to the initial signal; the upper envelope line is formed for the upper extreme point; the lower envelope line is formed for the lower extreme point;
step three: summing and averaging the average value of the upper envelope curve and the average value of the lower envelope curve to obtain an average value envelope curve signal;
step four: obtaining a difference value between the initial signal and a corresponding mean envelope signal to obtain an intermediate signal;
step five: if the intermediate signal does not meet the preset condition, the intermediate signal is used as an initial signal, and the steps two to four are repeated until the intermediate signal meets the preset condition;
step six: if the intermediate signal meets a preset condition, marking the intermediate signal as a first IMF component; subtracting the first I MF component from the initial signal to serve as a new initial signal, repeating the second step to the fourth step, and the like to obtain a plurality of I MF components until the residual value of the preprocessing signal is smaller than a preset threshold value; and if the residual error value is smaller than the preset threshold value, the current I MF component cannot be extracted, and the process of empirical mode decomposition is finished.
6. The defect recognition method based on partial discharge high-frequency current pulses according to claim 5, wherein the preset condition is that the number of upper and lower extreme points of the initial signal and the number of zero crossing points within a data segment of the initial signal differ by no more than one; the preset condition is that the upper envelope curve and the lower envelope curve corresponding to the initial signal are locally symmetrical relative to a time axis.
7. The defect identification method based on partial discharge high-frequency current pulse according to claim 1, wherein performing hilbert transformation on each of the I MF components in the I MF component set to obtain instantaneous feature values of each of the I MF components in the I MF component set, specifically comprising:
obtaining each I MF component in the I MF component set, and converting each I MF component in the I MF component set from a time domain to a complex frequency domain through Hilbert transformation to obtain instantaneous characteristic quantity; the instantaneous characteristic quantity is complex, the real part is the amplitude of the current I MF component, and the imaginary part is the phase of the current I MF component.
8. The method for identifying defects based on partial discharge high-frequency current pulses according to claim 1, wherein the performing time-frequency plane transformation on each instantaneous feature in the feature set to obtain a hilbert spectrum, and obtaining the target local feature of the first signal and the corresponding target frequency distribution diagram thereof according to the hilbert spectrum specifically comprises:
performing wavelet transformation and other time-frequency plane transformation on the instantaneous characteristic quantity, and converting the instantaneous characteristic quantity to a time-frequency plane by selecting different window functions and local characteristic parameters to obtain a Hilbert spectrum; the local characteristic parameters are peak rising time, peak falling time, peak voltage, valley falling time, valley rising time, valley voltage, average voltage, voltage crest factor, voltage profile coefficient and equivalent time length;
filtering and frequency domain analysis are carried out on the Hilbert spectrum;
the method comprises the steps of obtaining peak rising time, peak falling time and peak voltage of the peak maximum value of the Hilbert spectrum under equivalent time length to obtain peak characteristics;
obtaining valley value descending time, valley value ascending time and valley value voltage at the minimum value of the valley value under the equivalent time length in the Hilbert spectrum to obtain valley value characteristics;
obtaining signal intensity through average voltage under equivalent time length in the Hilbert spectrum;
acquiring the signal shape of the Hilbert spectrum passing through a voltage crest factor and a voltage shape coefficient under the equivalent time length;
and obtaining the target local characteristics of the first signal and the corresponding target frequency distribution diagram according to the peak value characteristics, the valley value characteristics, the signal strength and the signal shape.
9. The defect identification method based on partial discharge high frequency current pulses according to claim 1, wherein the defect type of the partial discharge high frequency current pulses is obtained according to the target local features of each first signal and the corresponding target frequency distribution diagram thereof, specifically comprising:
searching a preset frequency distribution diagram set in a comparison database according to the target local characteristics of the first signal; the control database is a frequency distribution diagram obtained by the known partial discharge defects under different local characteristics;
if a target frequency distribution diagram corresponding to the target local feature of the first signal exists, determining the defect type of the target frequency distribution diagram as the target defect type of the first signal;
and obtaining the target defect types corresponding to the first signals, and obtaining the partial discharge high-frequency current pulse defect results by combining and collecting the target defect types.
10. A defect recognition device based on partial discharge high frequency current pulses, the device comprising: the device comprises a preprocessing module, a signal splitting module, a signal characteristic extraction module, a signal characteristic processing module and a defect type determining module:
the pretreatment module is used for carrying out pretreatment on the partial discharge high-frequency current pulse signal to obtain a pretreatment signal, and carrying out short-time Fourier transformation on the pretreatment signal to obtain a target period of the pretreatment signal;
the signal splitting module is used for splitting the preprocessing signal through the target period to obtain a plurality of first signals;
the signal characteristic extraction module is used for carrying out empirical mode decomposition on each first signal to obtain a modal function (I MF) component set of the first signal, carrying out Hilbert transformation on each I MF component in the I MF component set to obtain instantaneous characteristic quantity of each I MF component in the I MF component set, and marking the instantaneous characteristic quantity as a characteristic quantity set;
the signal characteristic processing module is used for carrying out time-frequency plane transformation on each instantaneous characteristic quantity in the characteristic quantity set to obtain a Hilbert spectrum, and obtaining a target local characteristic of the first signal and a corresponding target frequency distribution diagram according to the Hilbert spectrum;
the defect type determining module is used for obtaining the defect type of the partial discharge high-frequency current pulse according to the target local characteristics of each first signal and the corresponding target frequency distribution diagram.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311800140.2A CN117872052A (en) | 2023-12-26 | 2023-12-26 | Defect identification method and device based on partial discharge high-frequency current pulse |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311800140.2A CN117872052A (en) | 2023-12-26 | 2023-12-26 | Defect identification method and device based on partial discharge high-frequency current pulse |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117872052A true CN117872052A (en) | 2024-04-12 |
Family
ID=90594043
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311800140.2A Pending CN117872052A (en) | 2023-12-26 | 2023-12-26 | Defect identification method and device based on partial discharge high-frequency current pulse |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117872052A (en) |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| RU2207581C2 (en) * | 2001-04-17 | 2003-06-27 | Марийский государственный университет | Method determining state of line insulation of distribution networks and location of point of its failure |
| US20120278012A1 (en) * | 2011-04-29 | 2012-11-01 | General Electric Company | Device and method for detecting and locating defects in underground cables |
| CN111044865A (en) * | 2019-12-30 | 2020-04-21 | 杭州国洲电力科技有限公司 | Partial discharge wavelet denoising method for direct-current gas insulation electrical equipment |
| CN112858852A (en) * | 2021-01-15 | 2021-05-28 | 国网山东省电力公司德州供电公司 | Switch cabinet partial discharge ultrasonic monitoring device and map filtering signal-noise separation method |
| CN114690003A (en) * | 2022-03-21 | 2022-07-01 | 广州友智电气技术有限公司 | EEMD-based partial discharge signal noise reduction method |
| CN115144705A (en) * | 2022-06-28 | 2022-10-04 | 广东电网有限责任公司 | Partial discharge detection method, device, equipment and medium based on propagation distortion compensation |
| CN115166446A (en) * | 2022-07-28 | 2022-10-11 | 国网安徽省电力有限公司淮南供电公司 | Transformer partial discharge identification method |
| CN116223985A (en) * | 2022-12-08 | 2023-06-06 | 国网上海市电力公司 | Diagnostic method and system for partial discharge defect of electrical equipment |
| CN117031194A (en) * | 2023-07-27 | 2023-11-10 | 中电鼎润(广州)科技发展有限公司 | Ultrasonic hidden danger detection method and system for power distribution network |
-
2023
- 2023-12-26 CN CN202311800140.2A patent/CN117872052A/en active Pending
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| RU2207581C2 (en) * | 2001-04-17 | 2003-06-27 | Марийский государственный университет | Method determining state of line insulation of distribution networks and location of point of its failure |
| US20120278012A1 (en) * | 2011-04-29 | 2012-11-01 | General Electric Company | Device and method for detecting and locating defects in underground cables |
| CN111044865A (en) * | 2019-12-30 | 2020-04-21 | 杭州国洲电力科技有限公司 | Partial discharge wavelet denoising method for direct-current gas insulation electrical equipment |
| CN112858852A (en) * | 2021-01-15 | 2021-05-28 | 国网山东省电力公司德州供电公司 | Switch cabinet partial discharge ultrasonic monitoring device and map filtering signal-noise separation method |
| CN114690003A (en) * | 2022-03-21 | 2022-07-01 | 广州友智电气技术有限公司 | EEMD-based partial discharge signal noise reduction method |
| CN115144705A (en) * | 2022-06-28 | 2022-10-04 | 广东电网有限责任公司 | Partial discharge detection method, device, equipment and medium based on propagation distortion compensation |
| CN115166446A (en) * | 2022-07-28 | 2022-10-11 | 国网安徽省电力有限公司淮南供电公司 | Transformer partial discharge identification method |
| CN116223985A (en) * | 2022-12-08 | 2023-06-06 | 国网上海市电力公司 | Diagnostic method and system for partial discharge defect of electrical equipment |
| CN117031194A (en) * | 2023-07-27 | 2023-11-10 | 中电鼎润(广州)科技发展有限公司 | Ultrasonic hidden danger detection method and system for power distribution network |
Non-Patent Citations (1)
| Title |
|---|
| 蔡君懿: "变压器局部放电源超声波定位方法研究", 《中国博士学位论文全文数据库 (工程科技Ⅱ辑)》, no. 2, 15 February 2023 (2023-02-15), pages 042 - 190 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110175508B (en) | Eigenvalue extraction method applied to ultrasonic partial discharge detection | |
| CN104459397B (en) | Power quality disturbance recognizing method with self-adaptation multi-resolution generalized S conversion adopted | |
| CN105044566B (en) | A kind of GIS partial discharge fault detection method of feature based ultra-high frequency signal | |
| CN113325277A (en) | Partial discharge processing method | |
| CN114325256A (en) | Power equipment partial discharge identification method, system, equipment and storage medium | |
| CN108957175B (en) | Electric energy quality disturbance identification method based on improved HHT algorithm | |
| CN112036450B (en) | High-voltage cable partial discharge mode identification method and system based on transfer learning | |
| CN110954779A (en) | Voltage sag source feature identification method based on S transformation and multidimensional fractal | |
| CN110794264B (en) | Generator partial discharge type identification method based on time domain pulse characteristics | |
| CN113608073A (en) | Cable partial discharge pulse separation method under variable frequency resonance | |
| CN117872052A (en) | Defect identification method and device based on partial discharge high-frequency current pulse | |
| CN112345885A (en) | Fault detection method for high-voltage direct-current transmission line | |
| CN104089778A (en) | Water turbine vibration fault diagnosis method | |
| CN116773961A (en) | Transmission line corrosion detection method based on vibration signal high-frequency characteristic analysis | |
| CN114034492B (en) | Hilbert-Huang transform-based rapid compression method for load spectrum of automobile part | |
| Seo et al. | A novel signal extraction technique for online partial discharge (PD) measurement of transformers | |
| CN114609258B (en) | Petroleum molecule characterization data processing method based on liquid chromatogram-high resolution mass spectrum | |
| CN114755529A (en) | Multi-feature fusion single-phase earth fault type identification method based on deep learning | |
| CN116280628A (en) | Partial discharge test method for high-voltage power equipment withstand voltage test | |
| CN109917245B (en) | Ultrasonic detection partial discharge signal mode identification method considering phase difference | |
| CN116008735A (en) | Partial discharge signal extraction method and system based on density | |
| CN114036973A (en) | Series arc fault identification method of extreme learning machine based on dynamic online sequence | |
| CN113589119A (en) | Partial discharge fault detection method based on soft threshold function | |
| CN111913078A (en) | Power transmission line fault identification method based on operation | |
| CN116626454B (en) | Oil paper insulation UHF partial discharge signal anti-interference identification and positioning method and device based on correction time-frequency clustering |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |