CN103675605A - Small-current earth fault line selection method based on fault signal transient state correlation analysis - Google Patents

Abstract

The invention relates to a small-current earth fault line selection method based on fault signal transient state correlation analysis, and belongs to the technical field of power distribution network automation. The technical problems in power distribution network single-phase earth fault line selection are solved. The small-current earth fault line selection method is characterized in that a feeder three-phase current is led in, the two-phase current difference of each feeder is obtained, correlation analysis is carried out on a feeder zero-sequence current and respective two-phase current differences, and a comprehensive correlation coefficient matrix E is formed; rhomax and rhomin in the E are selected, and if rhomax-rhomin>=rhoset, it is judged that the system has a single-phase earth fault, and fault lines are the feeders corresponding to rhomax; if rhomax-rhomin< rhoset, it is judged that the single-phase earth fault happens at the position of a busbar. The small-current earth fault line selection method is suitable for 35kV and 10kV voltage-class neutral-point non-effectiveness earthed line-cable mixed power distribution networks. The small-current earth fault line selection method has the advantages of being good in instantaneity, high in adaptability and high in accuracy rate, and the project practical value is achieved.

Description

Low-current ground fault line selection method based on the correlation analysis of fault-signal transient state

Technical field

The invention belongs to electrical technology field, relate to a kind of low-current ground fault line selection method based on the correlation analysis of fault-signal transient state.

Background technology

Domestic medium voltage distribution network is most of adopts small current neutral grounding, isolated neutral, through high resistance ground or through grounding through arc.When there is singlephase earth fault, fault current is less, and electric arc is also unstable, and the various failure line selection technology based on steady-state quantity are undesirable actual Use Limitation fruit.While there is singlephase earth fault, have an obvious transient state process, when transient state earthing capacitance current is often than its stable state at short notice large several times to tens times, this process comprises abundant fault signature, for failure line selection provides advantage.

The method of traditional small current earthing wire-selecting that utilizes correlation analysis has: the correlativity of the zero-sequence current between each outlet in one-period after occurring by fault relatively, forms correlation matrix, and define coefficient of colligation and select faulty line; Or utilize the failure line selection correlation analysis method of WAVELET PACKET DECOMPOSITION, after adopting suitable band width to singlephase earth fault, the zero-sequence current of each outlet carries out wavelet decomposition, and select characteristic spectra by the principle of energy maximum, according to faulty line with perfect the most weak principle of circuit similarity in characteristic spectra, by the WAVELET PACKET DECOMPOSITION coefficient to characteristic spectra, carry out correlation analysis and realize failure line selection.Above-mentioned correlativity route selection is all the correlation analysis based between each feeder line zero-sequence current.When mixing by cable and overhead transmission line the feeder line generation singlephase earth fault forming, joint line easily produces vibration, when it is directly carried out to correlation analysis with the zero-sequence current that does not produce the circuit of vibration, just easily causes route selection inaccurate; In the power distribution network of neutral by arc extinction coil grounding, the compensation of inductance current can affect amplitude and the phase place of faulty line zero-sequence current, makes traditional correlation analysis may occur route selection mistake.

In power distribution network for the power distribution network containing mixing feeder line and employing grounding through arc, the problem of failure line selection difficulty, the selection method that the present invention proposes, not between each feeder current, but between the composite signal by every three-phase current of feeder line own, carry out correlation analysis, avoid the deficiency of classic method, guaranteed the accuracy of route selection.

Summary of the invention

The present invention proposes a kind of low-current ground fault line selection method based on the correlation analysis of fault-signal transient state.

Technical solution of the present invention is as follows: a kind of low-current ground fault line selection method based on the correlation analysis of fault-signal transient state, comprises following step:

Step 1: with sample frequency 8000Hz, the residual voltage of power distribution network bus is carried out to real-time sampling, if continuous three sampled value u of residual voltage ₀(n), u ₀(n+1), u ₀(n+2) be all greater than 0.15 times of bus rated voltage, judge on this bus or this bus and have feeder line that singlephase earth fault has occurred; Otherwise, judge that power distribution network is in normal operating condition, continue sampling and monitor residual voltage sampled value;

Step 2: singlephase earth fault has occurred power distribution network if judge, with sample frequency 8000Hz sampling record trouble, there is three-phase current and the zero-sequence current of all feeder lines on the interior bus three-phase voltage of a rear power frequency period (20 milliseconds) and bus, obtain respectively corresponding sampled value sequence u _a(n), u _b(n), u _c(n), i _a.k(n), i _b.k(n), i _c.k(n), i _0.k(n), n=1,2,3 ..., N, k=1,2,3 ... M; N is sampled point sequence number, and N=160 is a sampling number in power frequency period, and k is each feeder line sequence number, and M is the number of feeder line on bus;

Step 3: the effective value U that calculates as follows bus three-phase voltage after fault _a, U _b, U _c:

$U_A=\sqrt{\frac{1}{N}\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{\left[u_A\left(n\right)\right]}^2},$

$U_B=\sqrt{\frac{1}{N}\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{\left[u_B\left(n\right)\right]}^2},$

$U_C=\sqrt{\frac{1}{N}\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{\left[u_C\left(n\right)\right]}^2};$

Step 4: compare bus three-phase voltage U _a, U _b, U _csize, get minimum voltage value corresponding mutually for fault phase;

Step 5: the zero-sequence current of each feeder line collecting and three-phase current are carried out to digital notch processing, and elimination is 50Hz power frequency component wherein;

Step 6: according to the selected fault phase of step 4, calculate the poor and related coefficient of the biphase current of each feeder line:

If fault phase is A phase, ask for i _aB.k(n)=i _a.k(n)-i _b.kand i (n) _aC.k(n)=i _a.k(n)-i _c.k(n), and respectively calculate the poor i of biphase current _aB.k(n), i _aC.k(n) with corresponding feeder line zero-sequence current i _0.k(n) dependency number:

${\mathrm{\&rho;}}_1\left(k\right)=\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i}_{\mathrm{AB}.k}\left(n\right)i_{0.k}\left(n\right){\left[\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i_{\mathrm{AB}.k}}^2\left(n\right)\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i_{0.k}}^2\left(n\right)\right]}^{-1/2},$

${\mathrm{\&rho;}}_2\left(k\right)=\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i}_{\mathrm{AC}.k}\left(n\right)i_{0.k}\left(n\right){\left[\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i_{\mathrm{AC}.k}}^2\left(n\right)\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i_{0.k}}^2\left(n\right)\right]}^{-1/2};$

If fault phase is B phase, ask for i _bA.k(n)=i _b.k(n)-i _a.kand i (n) _bC.k(n)=i _b.k(n)-i _c.k(n), and respectively calculate the poor i of biphase current _bA.k(n), i _bC.k(n) with corresponding feeder line zero-sequence current i _0.k(n) related coefficient:

${\mathrm{\&rho;}}_1\left(k\right)=\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i}_{\mathrm{BA}.k}\left(n\right)i_{0.k}\left(n\right){\left[\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i_{\mathrm{BA}.k}}^2\left(n\right)\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i_{0.k}}^2\left(n\right)\right]}^{-1/2},$

${\mathrm{\&rho;}}_2\left(k\right)=\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i}_{\mathrm{BC}.k}\left(n\right)i_{0.k}\left(n\right){\left[\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i_{\mathrm{BC}.k}}^2\left(n\right)\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i_{0.k}}^2\left(n\right)\right]}^{-1/2};$

If fault phase is C phase, ask for i _cA.k(n)=i _c.k(n)-i _a.kand i (n) _cB.k(n)=i _c.k(n)-i _b.k(n), and respectively calculate the poor i of biphase current _cA.k(n), i _cB.k(n) with corresponding feeder line zero-sequence current i _0.k(n) related coefficient:

${\mathrm{\&rho;}}_1\left(k\right)=\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i}_{\mathrm{CA}.k}\left(n\right)i_{0.k}\left(n\right){\left[\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i_{\mathrm{CA}.k}}^2\left(n\right)\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i_{0.k}}^2\left(n\right)\right]}^{-1/2},$

${\mathrm{\&rho;}}_2\left(k\right)=\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i}_{\mathrm{CB}.k}\left(n\right)i_{0.k}\left(n\right){\left[\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i_{\mathrm{CB}.k}}^2\left(n\right)\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i_{0.k}}^2\left(n\right)\right]}^{-1/2};$

Wherein, k=1,2,3 ... M;

Step 7: the integrated correlation coefficient that calculates each feeder line

k=1,2,3 ... M, forms integrated correlation coefficient matrix E=[ρ (1), ρ (2) ..., ρ (M)];

Step 8: the correlation coefficient threshold ρ of setting _set=0.5, the integrated correlation coefficient of each feeder line that comparison step 7 obtains, finds maximal value ρ wherein _maxwith minimum value ρ _min, and compare ρ _max-ρ _minwith ρ _setsize, if meet ρ _max-ρ _min>=ρ _set, in decision-making system, singlephase earth fault has occurred, and faulty line is integrated correlation coefficient ρ _maxcorresponding feeder line; If meet ρ _max-ρ _min< ρ _set, be judged to be bus place singlephase earth fault occurred.

The present invention has following beneficial effect:

The present invention propose selection method there is following characteristics:

1) strong adaptability.Problem for cable joint line and arc suppression coil earthing system failure line selection difficulty, corresponding innovative approach has been proposed: the three-phase current of introducing feeder line, the biphase current that obtains fault phase and healthy phases is poor, the zero-sequence current of itself and circuit self is carried out to correlation analysis, can eliminate the impact that impact between the circuit that joint line brings and arc suppression coil bring to faulty line zero-sequence current.

2) route selection accuracy rate is high.Selection method is can be in different fault types (line fault, bus-bar fault, electric arc fault), and different stake resistances, can both correctly select faulty line when different fault close angles and fault distance;

3) antijamming capability is strong.The selection method that the present invention proposes utilizes that the biphase current of same feeder line is poor carries out correlation analysis with zero-sequence current, can eliminate well the noise in sampled signal;

4) real-time.The data of one-period (20 milliseconds) after the selection method that the present invention proposes only need to utilize fault to occur, and select that algorithm is simple, calculated amount is little, can in the time very short after fault occurs, select faulty line, provide actuating signal excision fault, ensure the safe operation of power distribution network.

In a word, the small electric current grounding system of distribution network fault line selection method for single-phase-to-ground fault calculated amount that the present invention proposes is little, and real-time is good, strong adaptability, and accuracy rate is high, has engineering practical value.

Accompanying drawing explanation

Fig. 1 is the process flow diagram of neutral by arc extinction coil grounding system failure route selection;

Fig. 2 is small current neutral grounding system Case Simulation model;

Fig. 3 is each circuit zero-sequence current of Case Simulation and the poor figure of biphase current;

Embodiment

Technical conceive of the present invention:

The basic thought of selection method of the present invention is, bus three-phase voltage, bus residual voltage, each feeder line three-phase current and zero sequence current signal are sampled, and obtains voltage, current sampling signal sequence; The time occurring by fault generation judging unit detection failure, three-phase current and the zero sequence current signal of each feeder line collecting in one-period from fault occurs to start to choose fault constantly and occurs; According to bus three-phase voltage, by Novel Faulty Phase Selector, select the phase that fault occurs, and it is poor according to the fault phase of selecting, to calculate the biphase current of fault phase electric current and all the other healthy phases; Adopt the method for the digital signal processing of digital notch partly to carry out filtering to the 50Hz power frequency in the zero-sequence current obtaining and biphase current difference signal, the transition function of digital trap is $H\left(z\right)=\frac{B\left(z\right)}{A\left(z\right)}=\frac{(z-e^{2\mathrm{\&pi;}{f}_0})(z-e^{{-j2\mathrm{\&pi;f}}_0})}{(z-{\mathrm{ae}}^{2\mathrm{\&pi;}{f}_0})(z-{\mathrm{ae}}^{{-j2\mathrm{\&pi;f}}_0})},$ Wherein a is and the depth parameter of trapper, and a is larger, and notch depth is darker, gets a=0.96, f in the design ₀for treating the frequency size of elimination, get 50Hz.Finally zero-sequence current and the poor correlation analysis that carries out of biphase current to each feeder line after digital notch is processed, selects faulty line.The specific works principle of method is as follows:

The principle of 1 correlation analysis

Correlation analysis is important instrument during signal is processed, and can be used for studying the similarity of two signals, or signal self similarity after time delay, thereby realizes detection, the identification and extraction of signal.

The related function of digital signal x (n), y (n) is defined as:

$r_{\mathrm{xy}}\left(j\right)=\frac{1}{N}\underset{n=0}{\overset{N-1}{\mathrm{\&Sigma;}}}x\left(n\right)y(n+j)=\frac{1}{N}\underset{n=0}{\overset{N-1}{\mathrm{\&Sigma;}}}y\left(n\right)x(n-j)=r_{\mathrm{yx}}(-j)---\left(1\right)$

In formula, the sampling number that N is coherent signal, the time difference between two signals is jt _s, t _sfor sampling time interval (j=0,1 ...).

The auto-correlation of digital signal x (n) is defined as

$r_{\mathrm{xx}}\left(j\right)=\frac{1}{N}\underset{n=0}{\overset{N-1}{\mathrm{\&Sigma;}}}x\left(n\right)x(n+j)---\left(2\right)$

Signal x (n) and y (n) are carried out related calculation, and signal Ax (n) and By (n) are made to related operation (A and B are coefficient), twice computing be not because the correlation operation value that the difference of amplitude obtains can be identical.Therefore, for the correlativity of more different amplitude signal waveforms, just should get rid of the impact of signal amplitude, must make normalized to related operation for this reason.

The root-mean-square value of signal x (n), y (n) is

${\mathrm{\&delta;}}_x={\left[\frac{1}{N}\underset{n=0}{\overset{N-1}{\mathrm{\&Sigma;}}}{x}^2\left(n\right)\right]}^{1/2},{\mathrm{\&delta;}}_y={\left[\frac{1}{N}\underset{n=0}{\overset{N-1}{\mathrm{\&Sigma;}}}{y}^2\left(n\right)\right]}^{1/2}---\left(3\right)$

Therefore, signal correction computing can be expressed as in normalization

${\mathrm{\&rho;}}_{\mathrm{xy}}=\frac{r_{\mathrm{xy}}}{{\mathrm{\&delta;}}_x{\mathrm{\&delta;}}_y}---\left(4\right)$

${\mathrm{\&rho;}}_{\mathrm{xy}}=\underset{n=0}{\overset{N-1}{\mathrm{\&Sigma;}}}x\left(n\right)y\left(n\right){\left[\underset{n=0}{\overset{N-1}{\mathrm{\&Sigma;}}}{x}^2\left(n\right)\underset{n=0}{\overset{N-1}{\mathrm{\&Sigma;}}}{y}^2\left(n\right)\right]}^{-1/2}---\left(5\right)$

Normalized correlation ρ _xybe called cross-correlation coefficient, its span is [1 ,+1], and+1 represents two signal 100% positive correlations;-1 represents two signal 100% negative correlation, but the just the same phase place of shape is just in time contrary; 0 value is called as zero correlation, in other words two signals completely independent, have no relation.

2 route selection principles based on correlation analysis

During neutral by arc extinction coil grounding system generation singlephase earth fault, the transient current of earth point has comprised all sound and non-transient state capacitance current in circuit and transient state inductive currents of arc suppression coil place branch road of perfecting.With the transient current that the fault phase transient current of each feeder line deducts respectively two other healthy phases, obtain the poor i of transient state biphase current of circuit _f1and i _f2.For fault feeder, i _f1and i _f2the transient fault electric current and the fault phase transient state capacitive earth current sum that are actually earth point deduct respectively two other healthy phases transient state capacitive earth current; Transient state earth-fault current direction is by line flows to bus, and healthy phases transient state capacitance current flows to circuit by bus, and healthy phases transient state capacitive earth current is with respect to trouble spot transient state ground current i _fsmaller.Comprehensive above 2 considerations, the i of faulty line _f1and i _f2than i _f, wave form varies is very little.So i _f1and i _f2meet following relation and (suppose that A is fault phase mutually, now

$\left\{\begin{array}{c}{i}_{f_{\mathrm{AB}}}=i_{f_A}-i_{f_B}=i_f+i_{f_{C.A}}-i_{f_{C.B}}\\ i_{f_{\mathrm{AC}}}=i_{f_A}-i_{f_C}=i_f+i_{f_{C.A}}-i_{f_{C.C}}\end{array}\right.---\left(6\right)$

I in formula _ffor trouble spot transient state ground current,

be respectively the three-phase fault electric current of faulty line.

In taking into account system, three-phase load is symmetrical, thereby the transient zero-sequence current of faulty line is:

${3i}_{{\mathrm{<figure-callout\; id="0"\; label="f"\; filenames="DEST\_PATH\_HDA0000455321720000012.png,DEST\_PATH\_HDA0000455321720000021.png"\; state="\{\{state\}\}">f</figure-callout>}}_0}=i_f+i_{f_{C.A}}+i_{f_{C.B}}+i_{f_{C.C}}---\left(7\right)$

be respectively faulty line three-phase transient state capacitive earth current;

transient zero-sequence current for faulty line.On faulty line zero-sequence current numerical value, equal all transient state inductive current sums that perfect circuit transient state capacitive earth current and arc suppression coil.

Formula (6) is become to the form that contains zero-sequence current as follows:

$\left\{\begin{array}{c}{i}_{f_{\mathrm{AB}}}={3i}_{f_0}-{2i}_{f_{C.B}}-i_{f_{C.C}}\\ i_{f_{\mathrm{AC}}}={3i}_{f_0}-i_{f_{C.B}}-{2i}_{f_{C.C}}\end{array}\right.---\left(8\right)$

By the known i of analysis above _fwith

opposite direction, and on numerical value the former much larger than the latter.Consider

with just mutually there is very low-angle phase differential, might as well use

substitute

with

$i_{f_{\mathrm{AB}}}=i_{f_{\mathrm{AC}}}={3i}_{f_0}-i_{f_{C.A}}-i_{f_{C.B}}-i_{f_{C.C}}={3i}_{f_0}-{3i}_0---\left(9\right)$

In now can supposing the system, have increased consistent with a faulty line parameter feeder line that perfects, after fault occurs, the zero-sequence current producing on this virtual feeder line is 3i ₀, in faulty line, biphase current is poor so

be equivalent to the poor of zero-sequence current on the zero-sequence current of faulty line and virtual feeder line.

The transient zero-sequence current of non-fault line is:

${3i}_{{\mathrm{<figure-callout\; id="0"\; label="h"\; filenames="DEST\_PATH\_HDA0000455321720000012.png,DEST\_PATH\_HDA0000455321720000021.png"\; state="\{\{state\}\}">h</figure-callout>}}_0}=i_{h_{C.A}}+i_{h_{C.B}}+i_{h_{C.C}}---\left(10\right)$

Wherein,

be respectively the three-phase transient state capacitive earth current that perfects circuit.For its i of non-fault line _f1and i _f2, because feeder line three-phase transient state capacitance current size differences is little, just there is phase differential, so i in poor for fault phase transient state capacitive earth current and healthy phases transient state capacitive earth current _f1and i _f2meet following relation (suppose that A is mutually for fault phase,

$\left\{\begin{array}{c}{i}_{h_{\mathrm{AB}}}=i_{h_{C.A}}-i_{h_{C.B}}\\ i_{h_{\mathrm{AC}}}=i_{h_{C.A}}-i_{h_{C.C}}\end{array}---\left(11\right)\right.$

Therefore, can realize failure line selection by the poor correlativity of the zero-sequence current of each feeder line and the biphase current of corresponding feeder line relatively, if fault occurs in A phase, the transient zero-sequence current of faulty line so

poor with the biphase current of faulty line

with correlativity close to 1, and perfect the transient zero-sequence current i of circuit _l0poor with the transient state biphase current that perfects circuit

with

correlativity is close to 0.

With this, form new fault-line selecting method: when fault occurs, first the moment t that failure judgement occurs, choosing 1 cycle after fault is data window, and the data in data window, by the power frequency component containing in digital notch elimination data, and are obtained to fault phase by phase selection element.If fault phase is A phase,, to the zero-sequence current of all feeder lines and the corresponding poor correlation analysis that carries out of biphase current, obtain relative coefficient ρ ₁(k), ρ ₂(k), define comprehensive relative coefficient form integrated correlation coefficient matrix E=[ρ (1), ρ (2) ..., ρ (M)].Find the maximal value ρ in integrated correlation coefficient matrix _maxwith minimum value ρ _min, when meeting ρ _max-ρ _min>=0.5, in decision-making system, singlephase earth fault has occurred, and faulty line is integrated correlation coefficient ρ _maxcorresponding feeder line; If ρ _max-ρ _min<0.5, judges that fault has occurred at bus place, k=1 wherein, 2..., N.

Below with reference to the drawings and specific embodiments, the present invention is described in further details:

Embodiment 1:

Consider that in actual distribution network, feeder line is more, and there is the situation that cable mixes.The small current neutral grounding system of having set up 35kV herein under MATLAB/SIMULINK environment is as Fig. 2.Wherein G is infinitely great power supply; T is main-transformer, and no-load voltage ratio is 110kV/35kV, and connection group is YN/d11; T _zfor Z-shaped transformer; L is arc suppression coil; R is the damping resistance of arc suppression coil.Circuit 1 is JC1 type overhead transmission line, and line length is 15km; Circuit 2 is YJV23-35/95 cable, and line length is 6km; Circuit 3 is JM1 pole line, and length is 18km; Circuit 4 is JS1 pole line and YJV23-35/95 type cable hybrid line, and length is respectively 12km and 5km; Circuit 5 is JS1 type overhead transmission line, and length is 30km; Circuit 6 is YJV23-35/95 type cable line, and length is 8km.During normal operation, adopt 10% over-compensation, after fault occurs, adopt full compensation; R cancels 10% of arc coil induction reactance; Sample frequency is got 8KHz.

As shown in Fig. 1 in accompanying drawing explanation, when A phase singlephase earth fault occurs apart from bus 2km place for the feeder line 1 of electrical network, fault close angle is 30 ° and transition resistance size while being 20 Ω, and the concrete steps of the distribution network fault line selection method based on transient signal correlation analysis of the present embodiment are as follows:

Step 1):

With sample frequency 8000Hz, the residual voltage of power distribution network bus is carried out to real-time sampling, if continuous three sampled value u of residual voltage ₀(n), u ₀(n+1), u ₀(n+2) be all greater than 0.15 times of bus rated voltage, judge on this bus or this bus and have feeder line that singlephase earth fault has occurred; Otherwise, judge that power distribution network is in normal operating condition, continue sampling and monitor residual voltage sampled value;

Step 2):

If judge there is singlephase earth fault in power distribution network, record now corresponding moment t=0.04167s, if now corresponding sampled point is initial sampled point, and with sample frequency 8000Hz sampling record trouble, there is three-phase current and the zero-sequence current of all feeder lines on the interior bus three-phase voltage of a rear power frequency period (20 milliseconds) and bus, obtain respectively corresponding sampled value sequence u _a(n), u _b(n), u _c(n), i _a.k(n), i _b.k(n), i _c.k(n), i _0.k(n), n=1,2,3 ..., N, k=1,2,3 ..., 6; N is sampled point sequence number, and N=160 is a sampling number in power frequency period, and k is each feeder line sequence number;

Step 3):

Calculate as follows the effective value U of bus three-phase voltage after fault _a, U _b, U _c:

$U_A=\sqrt{\frac{1}{N}\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{\left[u_A\left(n\right)\right]}^2},$

$U_B=\sqrt{\frac{1}{N}\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{\left[u_B\left(n\right)\right]}^2},$

$U_C=\sqrt{\frac{1}{N}\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{\left[u_C\left(n\right)\right]}^2};$

Step 4):

Compare bus three-phase voltage U _a, U _b, U _csize, obtain U _a<U _b=U _c, get minimum voltage value U _acorresponding A is fault phase mutually;

Step 5):

The zero-sequence current of collect 6 feeder lines and three-phase current are carried out respectively to digital notch processing, and elimination is 50Hz power frequency component wherein;

Step 6):

Because the selected fault phase of step 4 is A phase, good route selection flow process is carried out asking for of the poor and related coefficient of biphase current: i according to a preconcerted arrangement _aB.k(n)=i _a.k(n)-i _b.kand i (n) _aC.k(n)=i _a.k(n)-i _c,k(n), and respectively calculate

The poor i of feeder line biphase current _aB.k(n), i _aC.k(n) with this feeder line zero-sequence current i _0.k(n) related coefficient:

${\mathrm{\&rho;}}_1\left(k\right)=\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i}_{\mathrm{AB}.k}\left(n\right){\mathrm{<figure-callout\; id="0"\; label="i"\; filenames="DEST\_PATH\_HDA0000455321720000012.png,DEST\_PATH\_HDA0000455321720000021.png"\; state="\{\{state\}\}">i</figure-callout>}}_{0.k}\left(n\right){\left[\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i_{\mathrm{AB}.k}}^2\left(n\right)\underset{n=1}{\overset{\mathrm{<figure-callout\; id="0"\; label="N\; i"\; filenames="DEST\_PATH\_HDA0000455321720000012.png,DEST\_PATH\_HDA0000455321720000021.png"\; state="\{\{state\}\}">N</figure-callout>}}{\mathrm{\&Sigma;}}}{i_{}}^{}{{}_{0.k}}^2\left(n\right)\right]}^{-1/2},$

${\mathrm{\&rho;}}_2\left(k\right)=\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i}_{\mathrm{AC}.k}\left(n\right){\mathrm{<figure-callout\; id="0"\; label="i"\; filenames="DEST\_PATH\_HDA0000455321720000012.png,DEST\_PATH\_HDA0000455321720000021.png"\; state="\{\{state\}\}">i</figure-callout>}}_{0.k}\left(n\right){\left[\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i_{\mathrm{AC}.k}}^2\left(n\right)\underset{n=1}{\overset{\mathrm{<figure-callout\; id="0"\; label="N\; i"\; filenames="DEST\_PATH\_HDA0000455321720000012.png,DEST\_PATH\_HDA0000455321720000021.png"\; state="\{\{state\}\}">N</figure-callout>}}{\mathrm{\&Sigma;}}}{i_{}}^{}{{}_{0.k}}^2\left(n\right)\right]}^{-1/2},$

Wherein, k=1,2,3 ..., 6, N=160, calculates respectively to such an extent that the related coefficient of every feeder line is as follows:

ρ ₁(k=1)=0.9792，ρ ₂(k=1)=0.9814；

ρ ₁(k=2)=-0.1019，ρ ₂(k=2)=-0.0753；

ρ ₁(k=3)=0.1571，ρ ₂(k=3)=0.1562；

ρ ₁(k=4)=-0.0729，ρ ₂(k=4)=-0.0679；

ρ ₁(k=5)=-0.0088，ρ ₂(k=5)=-0.0597；

ρ ₁(k=6)=0.0199，ρ ₂(k=6)=-0.0027；

Step 7):

The integrated correlation coefficient of asking for each feeder line is as follows:

ρ (k=1)=0.9803, ρ (k=2)=-0.0088, ρ (k=3)=0.1567, ρ (k=4)=-0.0704, ρ (k=5)=-0.0343, ρ (k=6)=0.0086; Forming integrated correlation coefficient matrix is: E=[0.9803 ,-0.0088,0.1567 ,-0.0704 ,-0.0343,0.0086];

Step 8):

The integrated correlation coefficient matrix that each circuit that comparison step 7 obtains forms, finds maximal value and minimum value wherein and compares ρ _max-ρ _minwith ρ _setsize, meet ρ _max-ρ _min=1.1507>0.5=ρ _set, in decision-making system there is singlephase earth fault in feeder line, and faulty line is ρ _maxcorresponding circuit 1;

In the situation that carrying out emulation to the present embodiment, different fault types, faulty line, fault close angle, fault distance, fault phase and transition resistance obtain this patent algorithm simulating result as table 1:

Line selection algorithm result under table 1 different faults condition

The method has very strong antijamming capability, route selection that equally can be correct in the situation that of noise, herein by add the white Gaussian noise of 15db in current signal, verified that circuit 1, apart from bus 2km, transition resistance, singlephase earth fault occurs while being 20 Ω, when different fault close angles, route selection result is as table 2:

Table 2 adds route selection result after white Gaussian noise

θ	Related coefficient	Result
			0	[0.8114 -0.0568 -0.0433 0.0931 0.0092 -0.1055]	Correctly
30	[0.9076 0.0020 0.0762 0.0226 -0.0520 0.0364]	Correctly
			60	[0.947 -0.0027 -0.0106 -0.0296 -0.0543 0.0742]	Correctly
90	[0.9768 0.1006 -0.0405 0.1032 0.1749 0.0904]	Correctly

Claims (1)

1. the low-current ground fault line selection method based on the correlation analysis of fault-signal transient state, is characterized in that, comprises following step:

Step 1: with sample frequency 8000Hz, the residual voltage of power distribution network bus is carried out to real-time sampling, if continuous three sampled value u of residual voltage ₀(n), u ₀(n+1), u ₀(n+2) be all greater than 0.15 times of bus rated voltage, judge on this bus or this bus and have feeder line that singlephase earth fault has occurred; Otherwise, judge that power distribution network is in normal operating condition, continue sampling and monitor residual voltage sampled value;

Step 2: singlephase earth fault has occurred power distribution network if judge, with sample frequency 8000Hz sampling record trouble, there is three-phase current and the zero-sequence current of all feeder lines on the interior bus three-phase voltage of a rear power frequency period (20 milliseconds) and bus, obtain respectively corresponding sampled value sequence u _a(n), u _b(n), u _c(n), i _a.k(n), i _b.k(n), i _c.k(n), i _0.k(n), n=1,2,3 ..., N, k=1,2,3 ... M; N is sampled point sequence number, and N=160 is a sampling number in power frequency period, and k is each feeder line sequence number, and M is the number of feeder line on bus;

Step 3: the effective value U that calculates as follows bus three-phase voltage after fault _a, U _b, U _c:

$U_A=\sqrt{\frac{1}{N}\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{\left[u_A\left(n\right)\right]}^2},$

$U_B=\sqrt{\frac{1}{N}\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{\left[u_B\left(n\right)\right]}^2},$

$U_C=\sqrt{\frac{1}{N}\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{\left[u_C\left(n\right)\right]}^2};$

Step 4: compare bus three-phase voltage U _a, U _b, U _csize, get minimum voltage value corresponding mutually for fault phase;

Step 5: the zero-sequence current of each feeder line collecting and three-phase current are carried out to digital notch processing, and elimination is 50Hz power frequency component wherein;

Step 6: according to the selected fault phase of step 4, calculate the poor and related coefficient of the biphase current of each feeder line:

If fault phase is A phase, ask for i _aB.k(n)=i _a.k(n)-i _b.kand i (n) _aC.k(n)=i _a.k(n)-i _c.k(n), and respectively calculate the poor i of biphase current _aB.k(n), i _aC.k(n) with corresponding feeder line zero-sequence current i _0.k(n) dependency number:

${\mathrm{\&rho;}}_1\left(k\right)=\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i}_{\mathrm{AB}.k}\left(n\right)i_{0.k}\left(n\right){\left[\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i_{\mathrm{AB}.k}}^2\left(n\right)\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i_{0.k}}^2\left(n\right)\right]}^{-1/2},$

${\mathrm{\&rho;}}_2\left(k\right)=\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i}_{\mathrm{AC}.k}\left(n\right)i_{0.k}\left(n\right){\left[\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i_{\mathrm{AC}.k}}^2\left(n\right)\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i_{0.k}}^2\left(n\right)\right]}^{-1/2};$

If fault phase is B phase, ask for i _bA.k(n)=i _b.k(n)-i _a.kand i (n) _bC.k(n)=i _b.k(n)-i _c.k(n), and respectively calculate the poor i of biphase current _bA.k(n), i _bC.k(n) with corresponding feeder line zero-sequence current i _0.k(n) related coefficient:

${\mathrm{\&rho;}}_1\left(k\right)=\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i}_{\mathrm{BA}.k}\left(n\right)i_{0.k}\left(n\right){\left[\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i_{\mathrm{BA}.k}}^2\left(n\right)\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i_{0.k}}^2\left(n\right)\right]}^{-1/2},$

${\mathrm{\&rho;}}_2\left(k\right)=\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i}_{\mathrm{BC}.k}\left(n\right)i_{0.k}\left(n\right){\left[\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i_{\mathrm{BC}.k}}^2\left(n\right)\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i_{0.k}}^2\left(n\right)\right]}^{-1/2};$

If fault phase is C phase, ask for i _cA.k(n)=i _c.k(n)-i _a.kand i (n) _cB.k(n)=i _c.k(n)-i _b.k(n), and respectively calculate the poor i of biphase current _cA.k(n), i _cB.k(n) with corresponding feeder line zero-sequence current i _0.k(n) related coefficient:

${\mathrm{\&rho;}}_1\left(k\right)=\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i}_{\mathrm{CA}.k}\left(n\right)i_{0.k}\left(n\right){\left[\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i_{\mathrm{CA}.k}}^2\left(n\right)\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i_{0.k}}^2\left(n\right)\right]}^{-1/2},$

${\mathrm{\&rho;}}_2\left(k\right)=\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i}_{\mathrm{CB}.k}\left(n\right)i_{0.k}\left(n\right){\left[\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i_{\mathrm{CB}.k}}^2\left(n\right)\underset{n=1}{\overset{N}{\mathrm{\&Sigma;}}}{i_{0.k}}^2\left(n\right)\right]}^{-1/2};$

Wherein, k=1,2,3 ... M;

Step 7: the integrated correlation coefficient that calculates each feeder line

k=1,2,3 ... M, forms integrated correlation coefficient matrix E=[ρ (1), ρ (2) ..., ρ (M)];

Step 8: the correlation coefficient threshold ρ of setting _set=0.5, the integrated correlation coefficient of each feeder line that comparison step 7 obtains, finds maximal value ρ wherein _maxwith minimum value ρ _min, and compare ρ _max-ρ _minwith ρ _setsize, if meet ρ _max-ρ _min>=ρ _set, in decision-making system, singlephase earth fault has occurred, and faulty line is integrated correlation coefficient ρ _maxcorresponding feeder line; If meet ρ _max-ρ _min< ρ _set, be judged to be bus place singlephase earth fault occurred.

Images