CN105203936A - Method for determining power cable partial discharge defect type based on spectral analysis - Google Patents

Abstract

A method for determining a power cable partial discharge defect type based on spectral analysis comprises the following steps: 1) reading a power cable partial discharge amount time-domain waveform which is free of background noise and interference and collected by a power cable partial discharge detecting instrument, performing equivalent time-frequency analysis on the obtained partial discharge time-domain waveform, performing normalization processing on signals, performing time-domain conversion on the normalized signals, converting time-domain characteristic waveform to a frequency-domain waveform to obtain a frequency-domain characteristic pattern of the partial discharge waveform, and extracting a corresponding partial discharge waveform frequency spectrum characteristic value; performing other steps. The method has the benefits that the detection speed and accuracy are high, and the detection procedure is simple; a visualized power cable partial discharge frequency-domain oscillogram is output, so that the power cable defect style can be intuitively assessed by a detector, and the intelligence level of power cable partial discharge detection and the detection efficiency can be improved.

Description

A kind of power cable shelf depreciation defect type method of discrimination based on spectrum analysis

Technical field

The present invention relates to a kind of power cable shelf depreciation defect type method of discrimination, especially relate to a kind of power cable shelf depreciation defect type based on Partial Discharge spectrum analysis and sentence method for distinguishing, belong to power industry detection technique field.

Background technology

When power cable exists insulation fault, often partial discharge phenomenon occurs, and partial discharge phenomenon can cause the expansion of defect and the deterioration of cable insulation.Therefore shelf depreciation is quick on the draw to power cable insulation initial failure, and the Partial Discharge Detection of power cable is the effective means of detection streamer state of insulation, has been widely applied with the putting into operation of cable at present, has safeguarded and in maintenance process.

But it is diversified during the fault of shelf depreciation type and representative, and as the Important Parameters in power cable Partial Discharge Detection, shelf depreciation type and its insulation status have close ties, so to identify and failure modes is necessary the local discharge signal gathered, for ensureing the safe and reliable operation of power cable and instructing the service work of power cable to have very important meaning.

And the differentiation of traditional power cable shelf depreciation defect type, mainly with the time domain waveform of local discharge signal for foundation, carry out subjective judgement by the experience of testing staff.The invention provides a kind of special Partial Discharge frequency domain technique, effectively can improve accuracy and the detection efficiency of shelf depreciation determining defects.

Summary of the invention

The technical problem to be solved in the present invention is: provide a kind of power cable shelf depreciation defect type and sentence method for distinguishing, the method is based on the normalized of signal transacting and the time domain waveform time-frequency conversion to frequency-domain waveform, achieve the Intelligent Recognition of power cable shelf depreciation defect type, judgement, improve intelligent level and the detection efficiency of power cable Partial Discharge Detection.

The object of invention can be achieved through the following technical solutions:

Based on a power cable shelf depreciation defect type method of discrimination for spectrum analysis, the present invention is characterised in that, comprises the following steps:

1) time domain waveform that power cable Partial discharge detector device collects the power cable partial discharge quantity after removing background noise and interference is read, equivalent time frequency analysis is carried out to the shelf depreciation time domain beamformer obtained, signal is normalized, then normalized signal is carried out time-frequency conversion, the signature waveform of time domain is converted to frequency-domain waveform, obtain the frequency domain character collection of illustrative plates of Partial Discharge, and extract corresponding Partial Discharge general characteristic quantity frequently;

2) frequency spectrum graphics of the typical discharges defect fault of office obtained above being put in discharge waveform frequency domain character amount and database is analyzed, according to the defect type of the similarity determination shelf depreciation of measurement target frequency spectrum and all kinds of defect standard frequency spectrum of database;

3) corresponding power cable shelf depreciation defect type and frequency collection of illustrative plates is exported according to comparative result, as gained frequency spectrum and class database all types defect frequency spectrum similarity all undesirable, then the frequency spectrum graphics after output processing judges voluntarily for testing staff.

Step 1 of the present invention) be specially:

First the time domain waveform of the shelf depreciation read is normalized.Normalization is a kind of dimensionless process means, makes the absolute value of physical system numerical value become certain relative value relation, is to simplify to calculate, and reducing the effective way of value, is the necessary condition of the time domain waveform of local discharge signal being carried out time-frequency conversion.

Represent the time-domain signal of the single Partial Discharge collected with s (t), signal s (t) be normalized as follows:

$\hat{s}\left(t\right)=\frac{s\left(t\right)}{\sqrt{\∫s^2\left(t\right)dt}}$

Normalized signal is carried out time-frequency conversion, the signature waveform of time domain is converted to frequency-domain waveform:

$F\left(\mathrm{\ω}\right)=\underset{-\mathrm{\∞}}{\overset{\mathrm{\∞}}{\∫}}\hat{s}\left(t\right)e^{-j\mathrm{\ω}t}dt$

Step 2 of the present invention) be specially:

Different discharge defect type is different due to discharge mechanism, and the frequency domain profile variation that discharges is comparatively large, and spectrum peak is different.Determining defects process is divided into two, first judges the similarity of target waveform frequency spectrum graphics and typical fault frequency spectrum graphics, as obtained a result, judges that the peak Distribution characteristic of spectrogram is put in target office further:

A) frequency spectrum graphics of the typical discharges defect fault processed in the shelf depreciation frequency-domain waveform that obtains and database is analyzed, first the similarity that frequency spectrum graphics and typical fault frequency spectrum graphics are put in target office is contrasted, frequency spectrum is put and a certain typical fault frequency spectrum similarity reaches certain value as target office, then be judged to be this fault type, put frequency spectrum as target office to fail to meet the requirements with any typical fault frequency spectrum similarity, then be judged to be other faults, frequency spectrum is put and multiple typical fault frequency spectrum similarity meets the requirements as target office, then study and judge the peak Distribution characteristic that spectrogram is put in target office further,

B) calculate the difference value of target shelf depreciation spectrogram peaks distribution character and each typical fault spectrogram peaks distribution character further, get wherein difference value reckling, then target shelf depreciation is the type fault.

Step of the present invention a) is specially:

Hypothetical target frequency spectrum graphics is F (ω), and the template frequency spectrum function of six kinds of typical discharges signals is respectively A ₁(ω) ~ A ₆(ω), the different wave shape value of target spectrum and six kinds of standrded fault pattern frequency spectrum functions is asked for respectively,

X _n＝∫|F(ω)-A _n(ω)|dω(n＝1，2，3，4，5，6)

As X ₁~ X ₆middle minimum value X _xlower than judgment threshold X, all the other five different wave shape values are all greater than this threshold value, illustrate that the template frequency spectrum graphics of target spectrum figure and this typical discharges signal extremely coincide, then directly judge that this shelf depreciation is xth kind fault;

As X ₁~ X ₆middle nothing is lower than the value of judgment threshold X, and illustrate that the template frequency spectrum graphics of target spectrum figure and this typical discharges signal is neither identical, failure judgement type is " other ";

As X ₁~ X ₆in have multiple, if X lower than the value of judgment threshold X _a, X _b, X _call be less than judgment threshold X, then the spectrogram peaks distribution character of further evaluating objects waveform.

Step b of the present invention) be specially:

The frequency of getting main peak in target spectrum figure waveform, secondary peak and the 3rd peak is followed successively by Y ₁~ Y ₃, in the template spectrogram of six kinds of typical discharges signals, the frequency at main peak, secondary peak and the 3rd peak is followed successively by A ₁~ A ₃, B ₁~ B ₃, C ₁~ C ₃, D ₁~ D ₃, E ₁~ E ₃, F ₁~ F ₃;

Calculate the spectrum peak difference value of target spectrum and six kinds of typical faults successively;

Z ₁＝|Y ₁-A ₁|+|Y ₂-A ₂|+|Y ₃-A ₃|

Z ₂＝|Y ₁-B ₁|+|Y ₂-B ₂|+|Y ₃-B ₃|

Z ₃＝|Y ₁-C ₁|+|Y ₂-C ₂|+|Y ₃-C ₃|

Z ₄＝|Y ₁-D ₁|+|Y ₂-D ₂|+|Y ₃-D ₃|

Z ₅＝|Y ₁-E ₁|+|Y ₂-E ₂|+|Y ₃-E ₃|

Z ₆＝|Y ₁-F ₁|+|Y ₂-F ₂|+|Y ₃-F ₃|

Choose Z ₁~ Z ₆middle minimum value Z _n, then can judge that target shelf depreciation is as n-th kind of typical fault; Fig. 4 is shown in by whole determination flow block diagram.

Step 3 of the present invention) be specially:

According to the different wave shape value X of target spectrum and six kinds of standrded fault pattern frequency spectrum functions _nminimum value, or the spectrum peak difference value Z of target spectrum and six kinds of typical faults _nminimum value, judges that shelf depreciation is as n-th kind of fault, and exports its spectrogram.

The invention has the beneficial effects as follows:

The present invention's application time-frequency analysis technology, time domain waveform is converted to frequency domain collection of illustrative plates, judge with the frequency characteristic of shelf depreciation, automatically carry out spectrum analysis by system, have detection speed fast, accuracy is high, the simple feature of step, and export the frequency-domain waveform figure of visual power cable office electric discharge, be conducive to testing staff's visual assessment power cable defect type, improve intelligent level and the detection efficiency of power cable Partial Discharge Detection.

Accompanying drawing explanation

Fig. 1 is the FB(flow block) of patent of the present invention;

Fig. 2 is the shelf depreciation time domain beamformer of the impaired fault of cable body;

Fig. 3 is the spectrogram after the shelf depreciation process of the impaired fault of cable body;

Fig. 4 is cable typical fault type decision calculation process block diagram.

Embodiment

Below in conjunction with the drawings and specific embodiments, the present invention is described in detail.

See Fig. 1, Fig. 2, Fig. 3, Fig. 4, a kind of power cable shelf depreciation defect type method of discrimination based on spectrum analysis, the present invention is characterised in that, comprises the following steps:

1) time domain waveform that power cable Partial discharge detector device collects the power cable partial discharge quantity after removing background noise and interference is read, equivalent time frequency analysis is carried out to the shelf depreciation time domain beamformer obtained, signal is normalized, then normalized signal is carried out time-frequency conversion, the signature waveform of time domain is converted to frequency-domain waveform, obtain the frequency domain character collection of illustrative plates of Partial Discharge, and extract corresponding Partial Discharge general characteristic quantity frequently;

2) frequency spectrum graphics of the typical discharges defect fault of office obtained above being put in discharge waveform frequency domain character amount and database is analyzed, according to the defect type of the similarity determination shelf depreciation of measurement target frequency spectrum and all kinds of defect standard frequency spectrum of database;

3) corresponding power cable shelf depreciation defect type and frequency collection of illustrative plates is exported according to comparative result, as gained frequency spectrum and class database all types defect frequency spectrum similarity all undesirable, then the frequency spectrum graphics after output processing judges voluntarily for testing staff.

Step 1 of the present invention) be specially:

First the time domain waveform of the shelf depreciation read is normalized.Normalization is a kind of dimensionless process means, makes the absolute value of physical system numerical value become certain relative value relation, is to simplify to calculate, and reducing the effective way of value, is the necessary condition of the time domain waveform of local discharge signal being carried out time-frequency conversion.

Represent the time-domain signal of the single Partial Discharge collected with s (t), signal s (t) be normalized as follows:

$\hat{s}\left(t\right)=\frac{s\left(t\right)}{\sqrt{\∫s^2\left(t\right)dt}}$

Normalized signal is carried out time-frequency conversion, the signature waveform of time domain is converted to frequency-domain waveform:

$F\left(\mathrm{\ω}\right)=\underset{-\mathrm{\∞}}{\overset{\mathrm{\∞}}{\∫}}\hat{s}\left(t\right)e^{-j\mathrm{\ω}t}dt$

Step 2 of the present invention) be specially:

Different discharge defect type is different due to discharge mechanism, and the frequency domain profile variation that discharges is comparatively large, and spectrum peak is different.Determining defects process is divided into two, first judges the similarity of target waveform frequency spectrum graphics and typical fault frequency spectrum graphics, as obtained a result, judges that the peak Distribution characteristic of spectrogram is put in target office further:

A) frequency spectrum graphics of the typical discharges defect fault processed in the shelf depreciation frequency-domain waveform that obtains and database is analyzed, first the similarity that frequency spectrum graphics and typical fault frequency spectrum graphics are put in target office is contrasted, frequency spectrum is put and a certain typical fault frequency spectrum similarity reaches certain value as target office, then be judged to be this fault type, put frequency spectrum as target office to fail to meet the requirements with any typical fault frequency spectrum similarity, then be judged to be other faults, frequency spectrum is put and multiple typical fault frequency spectrum similarity meets the requirements as target office, then study and judge the peak Distribution characteristic that spectrogram is put in target office further,

B) calculate the difference value of target shelf depreciation spectrogram peaks distribution character and each typical fault spectrogram peaks distribution character further, get wherein difference value reckling, then target shelf depreciation is the type fault.

Step of the present invention a) is specially:

Hypothetical target frequency spectrum graphics is F (ω), and the template frequency spectrum function of six kinds of typical discharges signals is respectively A ₁(ω) ~ A ₆(ω), the different wave shape value of target spectrum and six kinds of standrded fault pattern frequency spectrum functions is asked for respectively,

X _n＝∫|F(ω)-A _n(ω)|dω(n＝1，2，3，4，5，6)

As X ₁~ X ₆middle minimum value X _xlower than judgment threshold X, all the other five different wave shape values are all greater than this threshold value, illustrate that the template frequency spectrum graphics of target spectrum figure and this typical discharges signal extremely coincide, then directly judge that this shelf depreciation is xth kind fault;

As X ₁~ X ₆middle nothing is lower than the value of judgment threshold X, and illustrate that the template frequency spectrum graphics of target spectrum figure and this typical discharges signal is neither identical, failure judgement type is " other ";

As X ₁~ X ₆in have multiple, if X lower than the value of judgment threshold X _a, X _b, X _call be less than judgment threshold X, then the spectrogram peaks distribution character of further evaluating objects waveform.

Step b of the present invention) be specially:

The frequency of getting main peak in target spectrum figure waveform, secondary peak and the 3rd peak is followed successively by Y ₁~ Y ₃, in the template spectrogram of six kinds of typical discharges signals, the frequency at main peak, secondary peak and the 3rd peak is followed successively by A ₁~ A ₃, B ₁~ B ₃, C ₁~ C ₃, D ₁~ D ₃, E ₁~ E ₃, F ₁~ F ₃;

Calculate the spectrum peak difference value of target spectrum and six kinds of typical faults successively;

Z ₁＝|Y ₁-A ₁|+|Y ₂-A ₂|+|Y ₃-A ₃|

Z ₂＝|Y ₁-B ₁|+|Y ₂-B ₂|+|Y ₃-B ₃|

Z ₃＝|Y ₁-C ₁|+|Y ₂-C ₂|+|Y ₃-C ₃|

Z ₄＝|Y ₁-D ₁|+|Y ₂-D ₂|+|Y ₃-D ₃|

Z ₅＝|Y ₁-E ₁|+|Y ₂-E ₂|+|Y ₃-E ₃|

Z ₆＝|Y ₁-F ₁|+|Y ₂-F ₂|+|Y ₃-F ₃|

Choose Z ₁~ Z ₆middle minimum value Z _n, then can judge that target shelf depreciation is as n-th kind of typical fault; Fig. 4 is shown in by whole determination flow block diagram.

Step 3 of the present invention) be specially:

According to the different wave shape value X of target spectrum and six kinds of standrded fault pattern frequency spectrum functions _nminimum value, or the spectrum peak difference value Z of target spectrum and six kinds of typical faults _nminimum value, judges that shelf depreciation is as n-th kind of fault, and exports its spectrogram.

Embodiment

As the FB(flow block) that Fig. 1 is a kind of power cable shelf depreciation defect type method of discrimination based on spectrum analysis of the present invention, once concrete power cable shelf depreciation defect type differentiates that process comprises the steps:

(1) time domain waveform of this power cable partial discharge quantity after the removal background noise and interference that power cable Partial discharge detector device collects is read.

(2) read waveform is normalized, represents the time-domain signal of the single Partial Discharge obtained with s (t), signal s (t) is normalized as follows:

$\hat{s}\left(t\right)=\frac{s\left(t\right)}{\sqrt{\∫s^2\left(t\right)dt}}$

For the impaired shelf depreciation caused of cable body, the shelf depreciation time domain beamformer obtained is shown in Fig. 2.

(3) to obtained shelf depreciation time domain waveform carry out Fourier conversion, the signature waveform of time domain be converted to frequency-domain waveform:

$F\left(\mathrm{\ω}\right)=\underset{-\mathrm{\∞}}{\overset{\mathrm{\∞}}{\∫}}\hat{s}\left(t\right)e^{-j\mathrm{\ω}t}dt$

After the impaired shelf depreciation time domain waveform caused of cable body carries out frequency domain change, Fig. 3 is shown in by its frequency domain picture.And the frequency domain character amounts such as the frequency distribution scope at main peak, secondary peak and the 3rd peak in waveform are extracted by spectrogram.

(4) different discharge defect type is different due to discharge mechanism, and the frequency domain profile variation that discharges is comparatively large, and spectrum peak is different.

The frequency spectrum graphics of the typical discharges defect fault processed in the shelf depreciation frequency-domain waveform that obtains and database is analyzed, first the similarity that frequency spectrum graphics and typical fault frequency spectrum graphics are put in target office is calculated, frequency spectrum is put and a certain typical fault frequency spectrum similarity reaches certain value as target office, then be judged to be this fault type, put frequency spectrum as target office to fail to meet the requirements with any typical fault frequency spectrum similarity, then be judged to be other faults, frequency spectrum is put and multiple typical fault frequency spectrum similarity meets the requirements as target office, then calculate the similarity of target shelf depreciation spectrogram peaks distribution character and remaining typical fault spectrogram peaks distribution character further, get wherein difference value reckling, similarity soprano, be judged to be this fault.

Fig. 4 is shown in by whole determination flow block diagram.

(5) export the associated disadvantages type of shelf depreciation according to result of determination, and draw its spectrogram.

Claims (6)

1., based on a power cable shelf depreciation defect type method of discrimination for spectrum analysis, it is characterized in that, comprise the following steps:

1) time domain waveform that power cable Partial discharge detector device collects the power cable partial discharge quantity after removing background noise and interference is read, equivalent time frequency analysis is carried out to the shelf depreciation time domain beamformer obtained, signal is normalized, then normalized signal is carried out time-frequency conversion, the signature waveform of time domain is converted to frequency-domain waveform, obtain the frequency domain character collection of illustrative plates of Partial Discharge, and extract corresponding Partial Discharge general characteristic quantity frequently;

2) frequency spectrum graphics of the typical discharges defect fault of office obtained above being put in discharge waveform frequency domain character amount and database is analyzed, according to the defect type of the similarity determination shelf depreciation of measurement target frequency spectrum and all kinds of defect standard frequency spectrum of database;

3) corresponding power cable shelf depreciation defect type and frequency collection of illustrative plates is exported according to comparative result, as gained frequency spectrum and class database all types defect frequency spectrum similarity all undesirable, then the frequency spectrum graphics after output processing judges voluntarily for testing staff.

2. a kind of power cable shelf depreciation defect type method of discrimination based on spectrum analysis according to claim 1, is characterized in that, described step 1) be specially:

First the time domain waveform of the shelf depreciation read is normalized;

Represent the time-domain signal of the single Partial Discharge collected with s (t), signal s (t) be normalized as follows:

$\hat{s}\left(t\right)=\frac{s\left(t\right)}{\sqrt{\∫s^2\left(t\right)dt}}$

Normalized signal is carried out time-frequency conversion, the signature waveform of time domain is converted to frequency-domain waveform:

$F\left(\mathrm{\ω}\right)=\underset{-\mathrm{\∞}}{\overset{\mathrm{\∞}}{\∫}}\hat{s}\left(t\right)e^{-j\mathrm{\ω}t}dt.$

3. a kind of power cable shelf depreciation defect type method of discrimination based on spectrum analysis according to claim 1, is characterized in that, described step 2) be specially:

Different discharge defect type is different due to discharge mechanism, and the frequency domain profile variation that discharges is comparatively large, and spectrum peak is different.Determining defects process is divided into two, first judges the similarity of target waveform frequency spectrum graphics and typical fault frequency spectrum graphics, as obtained a result, judges that the peak Distribution characteristic of spectrogram is put in target office further:

A) frequency spectrum graphics of the typical discharges defect fault processed in the shelf depreciation frequency-domain waveform that obtains and database is analyzed, first the similarity that frequency spectrum graphics and typical fault frequency spectrum graphics are put in target office is contrasted, frequency spectrum is put and a certain typical fault frequency spectrum similarity reaches certain value as target office, then be judged to be this fault type, put frequency spectrum as target office to fail to meet the requirements with any typical fault frequency spectrum similarity, then be judged to be other faults, frequency spectrum is put and multiple typical fault frequency spectrum similarity meets the requirements as target office, then study and judge the peak Distribution characteristic that spectrogram is put in target office further,

B) calculate the difference value of target shelf depreciation spectrogram peaks distribution character and each typical fault spectrogram peaks distribution character further, get wherein difference value reckling, then target shelf depreciation is the type fault.

4. a kind of power cable shelf depreciation defect type method of discrimination based on spectrum analysis according to claim 3, it is characterized in that, described step a) is specially:

Hypothetical target frequency spectrum graphics is F (ω), and the template frequency spectrum function of six kinds of typical discharges signals is respectively A ₁(ω) ~ A ₆(ω), the different wave shape value of target spectrum and six kinds of standrded fault pattern frequency spectrum functions is asked for respectively,

X _n＝∫|F(ω)-A _n(ω)|dω(n＝1，2，3，4，5，6)

As X ₁~ X ₆middle minimum value X _xlower than judgment threshold X, all the other five different wave shape values are all greater than this threshold value, illustrate that the template frequency spectrum graphics of target spectrum figure and this typical discharges signal extremely coincide, then directly judge that this shelf depreciation is xth kind fault;

As X ₁~ X ₆middle nothing is lower than the value of judgment threshold X, and illustrate that the template frequency spectrum graphics of target spectrum figure and this typical discharges signal is neither identical, failure judgement type is " other ";

As X ₁~ X ₆in have multiple, if X lower than the value of judgment threshold X _a, X _b, X _call be less than judgment threshold X, then the spectrogram peaks distribution character of further evaluating objects waveform.

5. a kind of power cable shelf depreciation defect type method of discrimination based on spectrum analysis according to claim 3, is characterized in that, described step b) be specially:

The frequency of getting main peak in target spectrum figure waveform, secondary peak and the 3rd peak is followed successively by Y ₁~ Y ₃, in the template spectrogram of six kinds of typical discharges signals, the frequency at main peak, secondary peak and the 3rd peak is followed successively by A ₁~ A ₃, B ₁~ B ₃, C ₁~ C ₃, D ₁~ D ₃, E ₁~ E ₃, F ₁~ F ₃;

Calculate the spectrum peak difference value of target spectrum and six kinds of typical faults successively;

Z ₁＝|Y ₁-A ₁|+|Y ₂-A ₂|+|Y ₃-A ₃|

Z ₂＝|Y ₁-B ₁|+|Y ₂-B ₂|+|Y ₃-B ₃|

Z ₃＝|Y ₁-C ₁|+|Y ₂-C ₂|+|Y ₃-C ₃|

Z ₄＝|Y ₁-D ₁|+|Y ₂-D ₂|+|Y ₃-D ₃|

Z ₅＝|Y ₁-E ₁|+|Y ₂-E ₂|+|Y ₃-E ₃|

Z ₆＝|Y ₁-F ₁|+|Y ₂-F ₂|+|Y ₃-F ₃|

Choose Z ₁~ Z ₆middle minimum value Z _n, then can judge that target shelf depreciation is as n-th kind of typical fault.

6. a kind of power cable shelf depreciation defect type method of discrimination based on spectrum analysis according to claim 1, is characterized in that, described step 3) be specially:

According to the different wave shape value X of target spectrum and six kinds of standrded fault pattern frequency spectrum functions _nminimum value, or the spectrum peak difference value Z of target spectrum and six kinds of typical faults _nminimum value, judges that shelf depreciation is as n-th kind of fault, and exports its spectrogram.