CN105738764A - Power distribution network faulty section positioning method based on transient information full frequency band - Google Patents

Abstract

The present invention discloses a power distribution network faulty section positioning method based on a transient information full frequency band. The method comprises the following steps of acquiring the fault transient signals of the monitoring points of a distribution line; carrying out the denoising filtering on the fault transient signals to obtain the transient filtering signals; determining the fault initial phase angles of the transient filtering signals of the distribution line; if the fault initial phase angles are zero, utilizing an EMD method to decompose the transient filtering signals to obtain the DC components, and utilizing the DC components to position the fault sections; if the fault initial phase angles are not zero, utilizing the adaptive harmonic wavelet packet transform to carry out the full frequency band subdivision on the transient filtering signals to determine the characteristic frequency bands, and extracting the extreme values and the polarities of the wavelet transform coefficients in the characteristic frequency bands corresponding to the monitoring points to position the fault sections. The power distribution network faulty section positioning method of the present invention fully utilizes the transient full frequency band information at the fault moments, and determines the fault sections accurately aiming at different grounding fault initial phase angles.

Description

Fault Section Location of Distribution Network based on transient information Whole frequency band

Technical field

The present invention relates to a kind of Fault Section Location of Distribution Network based on transient information Whole frequency band.

Background technology

Along with the access of distributed energy and carrying out of intelligent power distribution network construction; existing distribution network failure location technology mainly adopts by standing interior protection device or line selection device for low current carries out route selection or direct trip protection; then artificial lookup mode along the line is adopted to carry out location and the failture evacuation of trouble point; owing to power distribution network is mainly close to user side; when trouble point occurs area with a varied topography or fault vestige inconspicuous; relying on artificial line walking can substantially prolongs the service restoration time, inefficiency does not meet the development trend of following power distribution network.

Current traveling wave method is very ripe in the application of transmission line of electricity, but owing to distribution network compares transmission line of electricity, there is substantial amounts of T wiring, and line end is mostly user, and both-end traveling wave ranging device is inapplicable at power distribution network.

In recent years, power distribution network carries out fault location by installing fault detector, when due to fault detector principle simple, sample rate is low, therefore may be only available for the big current failure such as phase fault and two phase ground, during circuit generation singlephase earth fault, fault detector is often reported by mistake, and occurs the probability of singlephase earth fault the highest in power distribution network.

On the basis realizing small current line selection, further determine that fault section, reduce the scope of patrolling and examining, be a focus of research at present.

Compare load current yet with single-phase earth fault current very faint, and ground fault condition is complicated, based on the fault section location method of tradition power frequency amount or harmonic component, be difficult to judge real abort situation.

Chinese patent literature CN103454559B discloses a kind of one-phase earthing failure in electric distribution network Section Location, detects the Current Transformer Secondary synthesis zero-sequence current of installed position in real time including the terminal being arranged on the multiple position of circuit；After the zero-sequence current amplitude of any terminal detection exceedes default initiation value, all terminals are accurately captured the zero-sequence current transient signal in 3 cycles after zero-sequence current exceedes initiation value front 1 cycle and exceedes initiation value at once；Each terminal uses Prony algorithm, Wavelet Packet Algorithm, HHT algorithm and fractal algorithm to be analyzed the zero-sequence current transient signal in 4 cycles calculating, and extracts the zero-sequence current characteristic of each algorithm respectively, and is uploaded to main website；After the zero-sequence current characteristic of each algorithm that each terminal that receives main website transmits, the input and output neuron node number of the BP neutral net number of plies and each layer is determined, it is determined that the network structure of BP neutral net and network parameter according to the zero-sequence current characteristic under each algorithm that each terminal extracts.

Above-mentioned BP neutral net is a kind of prognoses system, bring conveniently to research, without providing multiple zero-sequence current characteristic and fault section response mechanism accurately therebetween, but BP neutral net needs great amount of samples to be trained, and the problem that forecasting inaccuracy is true simultaneously is difficult to eradicate.

Summary of the invention

It is an object of the invention to provide a kind of Fault Section Location of Distribution Network based on transient information Whole frequency band, to be accurately positioned fault section.

For this, the invention provides a kind of Fault Section Location of Distribution Network based on transient information Whole frequency band, comprise the following steps: gather the fault transient signals of each monitoring point of distribution line；Fault transient signals is carried out noise-removed filtering, obtains transient state filtering signal；The fault initial phase angle of the transient state filtering signal of distribution line is judged；If fault initial phase angle is zero, utilizes EMD method to decompose transient state filtering signal, it is thus achieved that DC component, and utilize DC component that fault section is positioned；And if fault initial phase angle is non-zero, utilize adaptive harmonic wavelet packet transform that transient state filtering signal is carried out Whole frequency band segmentation, to determine characteristic spectra, and extract and the extreme value of wavelet conversion coefficient and polarity in the characteristic of correspondence frequency range of each monitoring point, fault section is positioned.

In above-mentioned Fault Section Location of Distribution Network, further comprise the steps of: the transient state filtering signal utilizing the minimum section both sides of fault, adopt both-end traveling wave method localization of faults position or fault branch.

In above-mentioned Fault Section Location of Distribution Network, utilize the fault transient signals of each monitoring point of measuring terminals collection distribution line, wherein, whether the residual voltage that each measuring terminals detects distribution line in real time is out-of-limit, if out-of-limit, after fault initiating moment front 2 cycles and Startup time, 4 cycles are as fault transient signals.

In above-mentioned Fault Section Location of Distribution Network, utilizing RLS sef-adapting filter that fault transient signals is carried out noise-removed filtering, wherein, two cycles before fault are as noise signal, and behind trouble point, 4 cycles are as periodic signal.

In above-mentioned Fault Section Location of Distribution Network, " utilize DC component to carry out fault section location " to include: the DC component of often adjacent two monitoring points subtracted each other, find maximum two the adjacent monitoring points of DC component difference, then between the adjacent monitoring point of the two for fault section.

In above-mentioned Fault Section Location of Distribution Network, " utilizing adaptive harmonic wavelet packet transform to carry out the segmentation of transient state filtering signal Whole frequency band, it is determined that characteristic spectra " comprises the following steps: transient state filtering signal carries out J layer harmonic wavelet packet transform A)；B) each subband generalized harmonic wavelet of jth layer being resolved into three subbands, three subbands are equiband, after decomposing, obtain 3 × 2^jIndividual subband, wherein, j=0,1,1.6,2,2.6,3 ..., J；C) by above-mentioned subband band people to, between harmonic wavelet packet transform jth+1 layer and j+2 layer, forming new harmonic wavelet packet transform；D) jth layer, the s sub belt energy are calculated relative to the normalized energy of all sub belt energy summation of jth layer, wherein, s=1,2,3 ..., 2J:E) calculate each subband spectrum kurtosis and F) according to composing the maximum principle of kurtosis, it is determined that characteristic spectra.

The present invention is compared with prior art, it has the beneficial effects that: take full advantage of the transient state Whole frequency band information of fault moment, for different earth fault initial phase angles, it is utilized respectively high frequency transient capacitive current component and low-frequency range attenuating dc component carries out the judgement of fault section, give multiple zero-sequence current characteristic and fault section response mechanism accurately therebetween, and be suitable for being combined use with both-end traveling wave method.

Except purpose described above, feature and advantage, the present invention also has other purpose, feature and advantage.Below with reference to figure, the present invention is further detailed explanation.

Accompanying drawing explanation

The Figure of description constituting the part of the application is used for providing a further understanding of the present invention, and the schematic description and description of the present invention is used for explaining the present invention, is not intended that inappropriate limitation of the present invention.In the accompanying drawings:

Fig. 1 is the flow chart of Fault Section Location of Distribution Network based on transient information Whole frequency band according to the preferred embodiment of the invention；

Fig. 2 is the schematic diagram for a power distribution network of the Fault Section Location of Distribution Network according to the present invention；

Fig. 3 is according to the theory diagram of the RLS sef-adapting filter of application in the Fault Section Location of Distribution Network of the present invention；And

Fig. 4 is in the distribution network failure section location side according to the present invention to be the processing method in non-zero situation in fault initial phase angle in method.

Detailed description of the invention

It should be noted that when not conflicting, the embodiment in the application and the feature in embodiment can be mutually combined.Describe the present invention below with reference to the accompanying drawings and in conjunction with the embodiments in detail.

The present invention provides a kind of Fault Section Location of Distribution Network based on transient information Whole frequency band, and as depicted in figs. 1 and 2, the method includes the steps of: gather the step S10 of the fault transient signals of each monitoring point of distribution line；Described fault transient signals is carried out noise-removed filtering, obtains the step S20 of transient state filtering signal；The step S30 that the fault initial phase angle of described distribution line is judged；If described fault initial phase angle is non-zero, utilize adaptive harmonic wavelet packet transform that described transient state filtering signal is carried out the segmentation of Whole frequency band, and determine characteristic spectra, extract extreme value and the polarity of wavelet conversion coefficient in the characteristic spectra of each monitoring point, the step S40 that fault section is positioned afterwards；If described fault initial phase angle is zero, utilizes EMD method to decompose described transient state filtering signal, it is thus achieved that DC component, and utilize the step S50 that fault section positions by described DC component；And utilize the terminal filtering data of the minimum section both sides of fault, adopt both-end traveling wave method Db wavelet transformation, calculate the transient state travelling wave wave head moment, with the step S60 of localization of faults position or fault branch.

The present invention is compared with prior art, it has the beneficial effects that: take full advantage of the transient state Whole frequency band information of fault moment, for different earth fault initial phase angles, it is utilized respectively high frequency transient capacitive current component and low-frequency range attenuating dc component carries out the judgement of fault section, and be suitable for being combined use with both-end traveling wave method.

The present invention utilize characteristic quantity (DC component, each monitoring point characteristic spectra in the extreme value of wavelet conversion coefficient and polarity) carry out fault section location, concrete criterion is as follows:

During low current fault ground connection, ground connection transient current can be expressed by following general equation (1):

Wherein, on the right side of equal sign, Section 1 is earth current steady-state component, and equal to the difference of stable state capacity current Yu inductance current, all the other are earth current transient state component, equal to capacitance current transient state component and inductive current transient DC component sum.

As shown from the above formula, when earth fault occurs maximum in phase voltage, namely during pi/2, transient current is mainly transient state capacity current, perception DC component is close to 0, when earth fault occurs when phase voltage near zero-crossing point, transient current is mainly perception DC component and close to maximum, and capacity current is close to zero.Owing to big multiple faults is that phase voltage occurs when maximum, when small probability generation phase voltage is zero.

According to above-mentioned fault signature, if institute's Monitoring Line fault occurs when phase voltage is maximum, then fault section location can be carried out according to the extreme value of wavelet conversion coefficient/mould each in characteristic spectra and polarity；If institute's Monitoring Line breaks down, initial phase angle is zero, then the acquisition DC component according to the EMD method described transient state filtering signal of decomposition carries out fault section location.

The present invention gives the preferred version implementing above-mentioned steps S10, specific as follows: distributed high-density sampling terminal is installed on the distribution line monitored, for gathering outdoor aerial line three-phase voltage and current signal, converted by Karenbauer, transient state filtering data can be decomposed into zero _exit and Aerial mode component, calculates zero-sequence current and zero sequence voltage component by following formula (2) and (3):

$\left\{\begin{array}{c}{I}_0=(I_A+I_B+I_C)/3\\ I_{\mathrm{\α}}=(I_A-I_B)/3\\ I_{\mathrm{\β}}=(I_A-I_C)/3\end{array}\right.---\left(2\right)$

$\left\{\begin{array}{c}{U}_0=(U_A+U_B+U_C)/3\\ U_{\mathrm{\α}}=(U_A-U_B)/3\\ U_{\mathrm{\β}}=(U_A-U_C)/3\end{array}\right.---\left(3\right)$

Detection residual voltage U in real time₀Whether out-of-limit, if out-of-limit, start transient state filtering, the out-of-limit threshold value of residual voltage is (0.15 times) bus rated voltage；Starting transient state filter length is fault initiating moment front 2 cycles, 4 cycles after Startup time.

On circuit 1.～be 10. numbered circuit install 10 measuring terminals, as shown in Figure 2, carry out the acquisition of three-phase voltage and three-phase current respectively, faulty line is achieved the division of faulty section by each two terminal, wherein terminal 3,4,6,7,9,10 is arranged on line end for catching fault transient component of voltage, and all the other catch fault transient state current component.

The present invention gives the preferred version implementing above-mentioned steps S20, often there is the interference factors such as various random noise and non-equilibrium current solving three-phase current in distribution system running, when ground fault resistance is relatively larger, the problem that fault transient component signal to noise ratio is low, the program is specific as follows:

RLS adaptive filter algorithm is adopted to be filtered processing to the signal gathered, as shown in Figure 3, wherein two cycles before fault are as noise signal d (k), behind trouble point, 4 cycles are as periodic signal x (k), this wave filter adopts 8 rank transverse direction FIR filter designs, and utilizing a piece of FPGA to realize, exporting signal after filtering is y (k).

The present invention gives enforcement above-mentioned steps S40 and the fault zero-sequence current transient signal gathered carried out the preferred version of adaptive harmonic wavelet packet transform, as shown in Figure 4, specifically include following steps:

S41) signal carrying out J=5 layer harmonic wavelet packet transform, wavelet basis function adopts Db8 wavelet function；

S42) each subband generalized harmonic wavelet of jth layer being resolved into three subbands, three subbands are equiband, after above-mentioned decomposition, obtain 3 × 2^jIndividual subband；

S43) by these subbands band people to, between harmonic wavelet packet transform jth+1 layer and j+2 layer, forming new harmonic wavelet packet transform；

S44) by following formula (4) be calculated jth layer, the s sub belt energy relative to all sub belt energy summation of jth layer normalized energy Y (j, s):

$\mathrm{\γ}(j,s)=\frac{\underset{n=0}{\overset{N-1}{\mathrm{\Σ}}}{\left|c_j^s\left(n\right)\right|}^2}{\underset{s=1}{\overset{2^j}{\mathrm{\Σ}}}\underset{n=0}{\overset{N-1}{\mathrm{\Σ}}}{\left|c_j^s\left(n\right)\right|}^2},j=0,1,1.6,2,2.6,3,...,J;s=1,2,3,...,2J;---\left(4\right)$

WhereinFor jth layer, the s subband wavelet coefficient, N is the length of each sub-band coefficients；

S45) according to fault transient component feature, set a threshold value λ, calculate each subband spectrum kurtosis by following formula (5):

$k_j^s\left\{\begin{array}{c}\frac{<{\left|c_s^j\left(n\right)\right|}^4>}{<{\left|c_s^j\left(n\right)\right|}^2{>}^2}-2,\mathrm{\&gamma;}(j,s)\&GreaterEqual;\mathrm{\&lambda;}\\ 0,\mathrm{\&gamma;}(j,s)\&GreaterEqual;\mathrm{\&lambda;}\end{array}\right.,j=0,1,1.6,2,2.6,3,\mathrm{...},J;s=1,2,\mathrm{3...},2^j---\left(5\right)$

Wherein,For the spectrum kurtosis that jth layer, the s subband estimate；

S46) according to the spectrum maximum principle of kurtosis, it is determined that optimum frequency range FS is characteristic spectra；And

S47) after determining characteristic spectra, wavelet conversion coefficient modulus maximum and polarity in characteristic spectra is extracted.

The present invention gives the preferred version implementing step S50, specific as follows: to utilize EMD method to decompose transient state filtering signal, (6) calculate the IMF component and residual components that obtain under each frequency according to the following formula:

$y\left(t\right)=\underset{j=1}{\overset{n}{\mathrm{\&Sigma;}}}{c}_j\left(t\right)+r_n\left(t\right)---\left(6\right)$

R in above formula_nT () is the DC component that transient signal decomposes；

The present invention gives the preferred version implementing step S60, specific as follows: to utilize the terminal filtering data of the minimum section both sides of fault, adopt both-end traveling wave method and Db8 wavelet transformation, calculate the transient state travelling wave wave head moment, and further localization of faults position or fault branch, both-end traveling wave computing formula (7) is as follows:

$l_f=\frac{l-v_1(t_2-t_1)}{2}---\left(7\right)$

Wherein l_fFor the distance of distance measurement point, l is the distance between two measuring terminals, v₁For traveling wave speed, t₂、t₁The moment at two ends is arrived for traveling wave.

Below in conjunction with Fig. 1 and Fig. 2, the Section Location of the present invention is discussed in detail:

Assume arrow place between terminal 5 and 7, the trouble point, fault initial phase angle non-zero, terminal 1,2,5,8,9 zero-sequence current on faulty line processes through adaptive harmonic wavelet packet transform, it is determined that has optimum fault frequency range, and calculates such as following table one of the polarity after transient state component wavelet transformation:

Terminal number	①	②	③	⑧	⑨
						Polarity	+	+	+	—	—

As can be seen from the above table on the circuit between terminal 2 and terminal 5 of the trouble point, the transient signal of terminal 2 and terminal 5 is adopted Db8 wavelet transformation, and the calculating of trouble point is carried out according to both-end traveling wave formula, the branch point place that if trouble point is placed exactly in terminal 5,7 and terminal 5,8 circuit crosses, explanation trouble point, side is positioned on fault branch, the measuring terminals that available branches end is installed proceeds both-end traveling wave location, eventually finds position of failure point.

Above-mentioned localization method is to be non-zero situation for fault initial phase angle, namely capacitive zero-sequence current transient state component is obvious, when fault occurs when initial phase angle is zero, section location can be carried out according to the EMD DC component decomposed, namely terminal 1,2,5,8,9 is carried out EMD decomposition, extracts attenuating dc component r_nT (), often adjacent two terminals are subtracted each other, and find two adjacent end that DC component difference is maximum, then the part of path that may determine that between these two terminals is fault section.

The foregoing is only the preferred embodiments of the present invention, be not limited to the present invention, for a person skilled in the art, the present invention can have various modifications and variations.All within the spirit and principles in the present invention, any amendment of making, equivalent replacement, improvement etc., should be included within protection scope of the present invention.

Claims (6)

1., based on the Fault Section Location of Distribution Network of transient information Whole frequency band, comprise the following steps:

(1) fault transient signals of each monitoring point of distribution line is gathered；

(2) described fault transient signals is carried out noise-removed filtering, obtains transient state filtering signal, it is characterised in that further comprising the steps of:

(3) the fault initial phase angle of described distribution line is judged；

(4) if described fault initial phase angle is zero, utilize EMD method to decompose described transient state filtering signal, it is thus achieved that DC component, and utilize described DC component that fault section is positioned；And

(5) if described fault initial phase angle is non-zero, utilize adaptive harmonic wavelet packet transform that described transient state filtering signal is carried out Whole frequency band segmentation, to determine characteristic spectra, and extract and the extreme value of wavelet conversion coefficient and polarity in the characteristic of correspondence frequency range of each monitoring point, fault section is positioned.

2. Fault Section Location of Distribution Network according to claim 1, it is characterised in that further comprise the steps of:

(6) utilize the transient state filtering signal of the minimum section both sides of fault, adopt both-end traveling wave method localization of faults position or fault branch.

3. Fault Section Location of Distribution Network according to claim 1, it is characterized in that, in step (1), utilize the fault transient signals of each monitoring point of measuring terminals collection distribution line, wherein, whether the residual voltage that each described measuring terminals detects described distribution line in real time is out-of-limit, if out-of-limit, after fault initiating moment front 2 cycles and Startup time, 4 cycles are as fault transient signals.

4. Fault Section Location of Distribution Network according to claim 3, it is characterized in that, in step (2), utilize RLS sef-adapting filter that described fault transient signals is carried out noise-removed filtering, wherein, two cycles before fault are as noise signal, and behind trouble point, 4 cycles are as periodic signal.

5. Fault Section Location of Distribution Network according to claim 1, it is characterized in that, in step (4), " utilize described DC component to carry out fault section location " to include: the DC component of often adjacent two monitoring points subtracted each other, find maximum two the adjacent monitoring points of DC component difference, then between the adjacent monitoring point of the two for fault section.

6. Fault Section Location of Distribution Network according to claim 1, it is characterized in that, in step (5), " utilizing adaptive harmonic wavelet packet transform to carry out the segmentation of described transient state filtering signal Whole frequency band, it is determined that characteristic spectra " comprises the following steps:

A) described transient state filtering signal is carried out J layer harmonic wavelet packet transform；

B) each subband generalized harmonic wavelet of jth layer being resolved into three subbands, three subbands are equiband, after decomposing, obtain 3 × 2^jIndividual subband, wherein, j=0,1,1.6,2,2.6,3 ..., J；

C) by above-mentioned subband band people to, between harmonic wavelet packet transform jth+1 layer and j+2 layer, forming new harmonic wavelet packet transform；

D) jth layer, the s sub belt energy are calculated relative to the normalized energy of all sub belt energy summation of jth layer, wherein, s=1,2,3 ..., 2^J:

E) each subband spectrum kurtosis is calculated, and

F) according to the spectrum maximum principle of kurtosis, it is determined that characteristic spectra.