CN105093062A - Comprehensive fault positioning method of transmission line - Google Patents

Abstract

The invention relates to a comprehensive fault positioning method of a transmission line. When it is detected that a single-phase grounding fault occurs in a power transmission system, Prony fitting is carried out on zero-sequence currents of detection points within T/4 after the fault, a dominant prony component of the transient zero-sequence currents is extracted, and the amplitude, phase, frequency and attenuation factor of the dominant component are uploaded to a system main station; the relative entropy of adjacent detection points is calculated, and a fault section is positioned according to the characteristic that the waveforms of zero-sequence currents of two points at the same side of the fault point is high in similarity and the waveforms of zero-sequence currents of two points at the two sides of the fault point is low in similarity; and in the positioned fault section, the fault distance is measured by utilizing an injection method. Thus, the upload data bulk is reduced, main characteristics of current signals are not lost, the communication pressure is reduced, and the fault positioning precision is improved.

Description

A kind of transmission line fault synthesized positioning method

Technical field

The present invention relates to a kind of power system failure diagnostic technology, particularly a kind of transmission line fault synthesized positioning method.

Background technology

Extensively adopt the small current neutral grounding method of operation in China 3 ~ 35kV distribution network system, when there is singlephase earth fault, transient characteristic is not obvious, makes to orient position that fault occurs or interval becomes abnormal comparatively difficulty.Existing Fault Locating Method can be divided into fault locating methods and fault section location method.

Fault locating methods passes through to system Injection Signal (S injection method), or the transient state travelling wave utilizing trouble spot to produce (traveling wave method) is finally oriented fault and is occurred in particular location on power transmission line, but when power transmission line distance is longer or wave impedance changes, precision of its location is not high.Fault section location method utilizes the multiple line feed terminals (feederterminalunit installed on the line, FTU) the transient state electric parameters of detection line, system main website analyzes the transient state electric parameters of each check point that FTU uploads again, finally determines that the check point that fault occurs is interval.According to the difference of the transient state electric parameters utilized, fault section location method can be divided into zero sequence power method, zero sequence drying method etc. again.Zero sequence power method needs to install zero sequential potential transformer additional on the line, and a large amount of zero sequential potential transformers works simultaneously, easily produces ferroresonance.Existing zero sequence drying method, needs the zero-sequence current sampled value of each check point to be uploaded to system main website, and volume of transmitted data is large, and wants the retention time synchronous, higher to communication system requirements.

Summary of the invention

The present invention be directed to existing transmission line fault localization method Problems existing, propose a kind of transmission line fault synthesized positioning method, the data volume uploaded can be reduced, the principal character of current signal can not be lost again, alleviate communication pressure, improve the precision of localization of fault.

Technical scheme of the present invention is: a kind of transmission line fault synthesized positioning method, first, after transmission system generation singlephase earth fault being detected, Prony matching is carried out to the zero-sequence current in T/4 after each check point fault, extract transient zero-sequence current Prony dominant component, the amplitude of dominant component, phase place, frequency and decay factor four parameter values are uploaded to system main website; Then, ask for the relative entropy of adjacent check point, high according to trouble spot homonymy 2 zero-sequence current wave-form similarities, and the feature that 2, both sides, trouble spot similarity is low, orient fault section; Finally, in the fault section oriented, injection method is utilized to carry out fault distance mensuration.

It is described that to orient fault section concrete steps as follows:

1) set the sampling period of zero sequence current mutual inductor as Δ t, as system zero sequence voltage U ₀when () is greater than the bus rated voltage of 0.15 times t, can think that system there occurs singlephase earth fault, count if N is sampled data, within the T/4 time, zero sequence current mutual inductor acquires N secondary data, each check point gathers, after fault occurs, obtain N number of sampled data in T/4, then have: N Δ t=T/4;

2) adopt the transient zero-sequence current of Prony algorithm difference matching m check point, select the transient zero-sequence current dominant component I of each check point _k(k=1,2 ... m), line feed terminals FTU is by I _kamplitude, phase place, frequency and decay factor 4 parameter values be uploaded to system main website;

3) the dominant component parameter of m check point uploaded according to FTU of system main website, ask for the relative entropy of adjacent check point, concrete computation process is as follows:

Each check point has N number of sampled data, the n-th data acquisition sampling point corresponding moment t=n Δ t (n=1,2 ... N), 1. for kth (k=1,2 ... m) individual check point, calculating moment t=n Δ t (n=1,2 ... N) transient state dominant component I time _k,taccount for whole system Q _tproportion q _k,t:

$q_{k,t}=\frac{I_{k,t}}{Q_t}(k=\mathrm{1,2},...m),$

Wherein, Q _tfor the zero-sequence current dominant component sum of m check point on t feeder line, be also

$Q_t=\underset{k=1}{\overset{m}{\mathrm{\Σ}}}{I}_{k,t};$

2. calculating check point k relative to the Prony relative entropy of check point k+1 is:

$M_{k,k+1}=\underset{n=1}{\overset{N}{\mathrm{\Σ}}}\left|q_{k,t}\ln \frac{q_{k,t}}{q_{k+1,t}}\right|(k=\mathrm{1,2},...m)$

In formula, t=n Δ t (n=1,2 ... N);

4) M is found out _{k, k+1}(k=1,2 ... m-1) maximal value in, then fault is just between check point k and k+1 of correspondence.

The described injection method that utilizes carries out fault distance mensuration, and calculate the distance of trouble spot and bus bar side, concrete steps are as follows:

A: utilize voltage transformer (VT) to judge fault phase, and at bus place, fault phase is excised;

B: utilize the voltage transformer (VT) of fault section upstream detection point k to inject the voltage signal of 220Hz to fault phase;

C: the steady state voltage of measuring-signal decanting point, current value the distance L of trouble spot to signal injection point is calculated according to following formula:

$\frac{\stackrel{\·}{U}}{\stackrel{\·}{I}}=(R_0+R_L)+{\mathrm{jX}}_L\≈R_0+{\mathrm{jX}}_L,$

In formula, R ₀for fault ground resistance, R _lfor signal injection point is to the line resistance of trouble spot, X _lfor signal injection point is to the fault phase circuit induction reactance of trouble spot, x _lfor circuit unit length induction reactance value, U _m, I _mfor electric current and voltage peak value, for the electric current and voltage phase angle difference of Injection Signal;

4) set a kth check point and line bus lateral extent as L _k, then trouble spot and bus bar side distance, i.e. fault distance L _total=L+L _k.

Beneficial effect of the present invention is: transmission line fault synthesized positioning method of the present invention, combine fault section location method and fault locating methods advantage separately, accurately can orient the section that fault occurs, the particular location of trouble spot in fault section can be recorded further again, improve the precision of localization of fault, accelerate data rate.First stage of the present invention, only need the zero-sequence current information recording check point, and the data volume that each FTU uploads to system main website is only 4 parameter values of this transient zero-sequence current dominant component, communications burden is little, and does not need strictly to keep synchronous; Subordinate phase, only has the voltage transformer (VT) work of fault section upstream detection point, and to fault phase Injection Signal, the voltage transformer (VT) of other check points does not work, and can not cause ferroresonance.

Accompanying drawing explanation

Zero-sequence network equivalent circuit diagram when Fig. 1 is the present invention one neutral by arc extinction coil grounding system generation earth fault;

Fig. 2 is the comprehensive positioning flow figure of transmission line fault of the present invention;

Fig. 3 is fault phase equivalent circuit diagram of the present invention;

Fig. 4 is small current neutral grounding system realistic model figure of the present invention;

Fig. 5 is check point A zero-sequence current dominant component of the present invention and measured current signal comparison diagram.

Embodiment

Transient zero-sequence current feature:

Zero sequence equivalent network figure as shown in Figure 1 during a neutral by arc extinction coil grounding system generation earth fault, wherein arc suppression coil switch is in off-state.In figure: L is arc suppression coil, S is generator windings.A, B, C, D are circuit S ₁on 4 current detecting points.I _0S, i ₀₂, i ₀₃be respectively and flow through generator, circuit S ₂, circuit S ₃zero-sequence current instantaneous value.I _0A, i _0B, i _0C, i _0Dbe respectively the zero-sequence current instantaneous value flowing through A, B, C, D point.I _c1, i _c2be respectively AB section, CD section zero sequence capacitive earth current.Then:

i _0A+i _0S+i ₀₂+i ₀₃＝0(1)

i _0A＝i _C1+i _0B(2)

If fault occurs in circuit S ₁f point, instant of failure, to be equivalent in trouble spot an additional zero sequence virtual voltage source U _0f, trouble spot flows to as F → B → A to the actual current of bus bar side, and trouble spot flows to as F → C → D, so i to the actual current of load side _0Band i _0Cpolarity contrary, different wave shape is larger.I in formula (1) _0Afor non-fault line capacitive earth current summation, due to the close together of generally two check points, i _c1relative to i _0Aproportion is very little, can ignore, so formula (2) can be reduced to i _0A≈ i _0B, the two waveform is substantially identical, in like manner known i _0Cand i _0Dwaveform is also substantially identical.In a word, transient zero-sequence current wave-form similarity is high for trouble spot homonymy two point (A, B or C, D), and both sides, trouble spot two point (B, C) similarity is low, can carry out fault section location according to this feature.

The Prony dominant component of zero-sequence current:

Prony algorithm is that the linear superposition of the sinusoidal signal exponentially decayed with a series of (p) carrys out matching one time signal y (t), each sinusoidal signal exponentially decayed has separately independently amplitude, phase place, frequency and decay factor, and the present invention is referred to as Prony basis function q (t).Be formulated as:

$y\left(t\right)=\underset{i=1}{\overset{p}{\Σ}}{q}_i\left(t\right)=\underset{i=1}{\overset{p}{\Σ}}{A}_i{e}^{{\mathrm{\α}}_{i}t}cos(2{\mathrm{\πf}}_{i}t+{\mathrm{\θ}}_i)---\left(3\right)$

In formula, q _it () is basis function, A _ifor amplitude, θ _ifor initial phase, α _i<0 is decay factor, f _ifor frequency.

Prony algorithm is fast to the processing speed of live signal, and time delay is little, the signal that applicable analysis exponentially decays, and can disclose the principal character in signal exactly.After carrying out prony matching to the transient zero-sequence current of each check point, signal energy mainly concentrates on low frequency component, and high fdrequency component proportion is minimum, defines the Prony dominant component that the maximum basis function of energy is zero-sequence current.Simulation result shows, the peak value of dominant component and measured current overlaps substantially, and Major Variation is similar, so, replace actual current signal by Prony dominant component, the principal character of current signal can not be lost.Compared to directly practical measurement of current value being uploaded to system main website, four of dominant component fitting parameters are only uploaded by each check point, just decrease the amount of uploading of data, alleviate communication pressure.

Relative entropy:

In information theory, relative entropy is in order to measure two probability distribution χ={ χ ₁, χ ₂χ _m, λ={ λ ₁, λ ₂λ _mdegree of closeness, can be expressed as:

$L(\mathrm{\&chi;},\mathrm{\&lambda;})=\underset{\mathrm{\&eta;}=1}{\overset{m}{\&Sigma;}}{\mathrm{\&chi;}}_{\mathrm{\&eta;}}\ln \frac{{\mathrm{\&chi;}}_{\mathrm{\&eta;}}}{{\mathrm{\&lambda;}}_{\mathrm{\&eta;}}}---\left(4\right)$

In engineer applied, relative entropy can be used to the otherness of tolerance two signals.Relative entropy is less, shows that the similarity of two signals is higher; Otherwise, then show that the similarity of two signals is lower.Homonymy 2 transient zero-sequence current wave-form similarities in trouble spot are high, then relative entropy is little; And 2, both sides, trouble spot similarity is low, then relative entropy is large.The present invention Prony relative entropy theory carries out fault section location, reduces the scope that fault distance measures, improves the precision of location.

A kind of transmission line fault synthesized positioning method comprises two stages: the first stage, after transmission system generation singlephase earth fault being detected, Prony matching is carried out to the zero-sequence current in T/4 after each check point fault, extract transient zero-sequence current Prony dominant component, four of dominant component parameter values are uploaded to system main website; Then, ask for the relative entropy of adjacent check point, high according to trouble spot homonymy 2 zero-sequence current wave-form similarities, and the feature that 2, both sides, trouble spot similarity is low, orient fault section; Subordinate phase, in the fault section oriented, utilizes injection method to carry out fault distance mensuration.

The comprehensive positioning flow figure of transmission line fault as shown in Figure 2, if a certain feeder line is provided with altogether m line feed terminals FTU (feederterminalunit), namely there is m check point, and in each FTU, embedded in the program of carrying out Prony matching and extracting Prony dominant component.As system zero sequence voltage U ₀when () is greater than the bus rated voltage of 0.15 times t, can think that system there occurs singlephase earth fault, now, the zero sequence current mutual inductor of each check point starts sampling.The concrete steps of first stage are as follows:

Step one:

If the sampling period of zero sequence current mutual inductor is Δ t, after fault occurs, obtaining N number of sampled data in T/4, (N is that sampled data is counted, namely within the T/4 time, current transformer acquires N secondary data, each check point gathers, all obtain N number of data, namely each check point has N number of sampled data), then have: N Δ t=T/4.

Step 2:

Adopt the transient zero-sequence current of each check point of Prony algorithm difference matching, select the transient zero-sequence current dominant component I of each check point _k(k=1,2 ... m), FTU is by I _k4 parameter values (i.e. amplitude, phase place, frequency and decay factor) be uploaded to system main website.

Step 3:

The dominant component parameter of m the check point that system main website uploads according to FTU, ask for the relative entropy of adjacent check point, concrete computation process is as follows:

Each check point has N number of sampled data, the n-th data acquisition sampling point corresponding moment t=n Δ t (n=1,2 ... N)

1. for kth (k=1,2 ... m) individual check point, calculating moment t=n Δ t (n=1,2 ... N) transient state dominant component I time _k,taccount for whole system Q _tproportion q _k,t:

$q_{k,t}=\frac{I_{k,t}}{Q_t},(k=1,2,...m)---\left(5\right)$

Wherein, Q _tfor the zero-sequence current dominant component sum of m check point on t feeder line, be also

$Q_t=\underset{k=1}{\overset{m}{\&Sigma;}}{I}_{k,t}---\left(6\right)$

2. calculating check point k relative to the Prony relative entropy of check point k+1 is:

$M_{k,k+1}=\underset{n=1}{\overset{N}{\&Sigma;}}\left|q_{k,t}ln\frac{q_{k,t}}{q_{k+1,t}}\right|,(k=1,2,...m)---\left(7\right)$

In formula, t=n Δ t (n=1,2 ... N).

Step 4:

Find out M _{k, k+1}(k=1,2 ... m-1) maximal value in, then fault is just between check point k and k+1 of correspondence.

First stage defines the section that fault occurs, and on this basis, utilizes injection method to carry out fault distance mensuration, calculates the distance of trouble spot and bus bar side.After there is singlephase earth fault, faulted phase voltage is 0, and healthy phases voltage is upgraded to line voltage.According to this feature, voltage transformer (VT) is utilized to determine Earth Phase.Then, at bus place, this fault phase is logged off, but do not affect the operation of healthy phases.Now, utilize the voltage transformer (VT) of trouble spot upstream detection point (the B point as in Fig. 1) to inject high-frequency voltage signal to fault phase, Injection Signal forms current path through fault phase circuit to trouble spot.Equivalent electrical circuit is fault phase equivalent circuit diagram as shown in Figure 3. for Injection Signal.Subordinate phase concrete steps are as follows:

1) utilize voltage transformer (VT) to judge fault phase, and at bus place, fault phase is excised.

2) voltage transformer (VT) of fault section upstream detection point (check point k) is utilized to inject the voltage signal of 220Hz to fault phase.

3) steady state voltage, the current value of measuring-signal decanting point the distance L of trouble spot to signal injection point is calculated according to formula (8,9).

$\frac{\stackrel{\&CenterDot;}{U}}{\stackrel{\&CenterDot;}{I}}=(R_0+R_L)+{\mathrm{jX}}_L\&ap;R_0+{\mathrm{jX}}_L---\left(8\right)$

In formula, R ₀for fault ground resistance, R _lfor signal injection point is to the line resistance of trouble spot, X _lfor signal injection point is to the fault phase circuit induction reactance of trouble spot, x _lfor circuit unit length induction reactance value, U _m, I _mfor electric current and voltage peak value, for electric current and voltage phase angle difference (fault phase excises, and phase angle difference is the electric current and voltage phase angle difference of Injection Signal) herein;

4) set a kth check point and line bus lateral extent as L _k, then trouble spot and bus bar side distance, i.e. fault distance

L _total＝L+L _k(10)

Accompanying drawing 4 is small current neutral grounding system realistic model described in the embodiment of the present invention.As shown in Figure 4, in the present embodiment, transmission line length S ₁=20km, S ₂=15km, S ₃=24km, S ₄=30km, S ₅=16km, S ₆=30km.Positive order parameter: R ₁=0.17 Ω/km, L ₁=1.2mH/km, C ₁=9.697nF/km; Zero sequence parameter: R ₀=0.23 Ω/km, L ₀=5.48mH/km, C ₀=6nF/km.Sample frequency f _s=1000kHz.Transformer connected mode is Y/ Δ, and no-load voltage ratio is 220kV/35kV.A, B, C, D are circuit S ₁on 4 current detecting points, be installed on 8.5km, 9.5km, 10.5km, 11.5km place of circuit respectively, singlephase earth fault occurs in circuit S ₁10km place.

Carry out segmentation Prony matching to the transient zero-sequence current of T/4 after check point A fault, obtain a series of fitting parameter, arrange from small to large by frequency, table 1 lists the Prony fitting parameter of the A point transient zero-sequence current of front 5 group components.

Table 1

Can draw according to table 1 data, after carrying out segmentation prony matching to the transient zero-sequence current of check point A, signal energy mainly concentrates on the low frequency component (i.e. the Prony dominant component of zero-sequence current) of the 1st group, and high fdrequency component proportion is minimum.Dominant component and zero-sequence current measured signal are more as shown in Figure 5.

As seen from Figure 5, the peak value of dominant component and measured current overlaps substantially, and both Major Variations are similar, so, replace actual current signal by Prony dominant component, only four of dominant component fitting parameters are uploaded to system main website, both decrease the amount of uploading of data, alleviate communication pressure, the principal character of current signal can not be lost again.

Under initial phase angle 90 °, stake resistance are 5 Ω conditions, the fitting parameter of each check point transient zero-sequence current dominant component is as shown in table 2.As shown in Table 2, four fitting parameters of the zero-sequence current dominant component of A, B (or C, D) 2 are very close, B and C 2 then differs greatly.

Table 2

Fixing initial phase angle be 90 ° constant, change stake resistance time, adjacent check point dominant component relative entropy during different grounding resistance as shown in table 3; Solid ground resistance is that 5 Ω are constant, when changing initial phase angle, and adjacent check point dominant component relative entropy during corresponding different voltage initial angle as shown in table 4.As can be seen from table 3-4 data, under different grounding resistance, different initial phase angle condition, the relative entropy M of B, C point-to-point transmission _b,Call much larger than M _a,Band M _c,D, can determine that fault section is between B, C check point, conforms to actual conditions accordingly.

Table 3

Table 4

After orienting fault section, inject 100V, 220Hz voltage signal at B point, carry out the mensuration of fault distance.According to the fault condition of table 3-4, fault distance during as different with table 6 in the fault distance during table 5 different grounding resistance voltage initial angle of the fault distance obtained.As can be seen from data in table, under different grounding resistance, different initial phase angle condition, the difference of the fault distance calculated and physical fault distance, namely error distance is all less than 0.3km, and the absolute value of range error is less than 3%, can meet actual requirement of engineering.

Table 5

Table 6

Claims (3)

1. a transmission line fault synthesized positioning method, it is characterized in that, first, after transmission system generation singlephase earth fault being detected, Prony matching is carried out to the zero-sequence current in T/4 after each check point fault, extract transient zero-sequence current Prony dominant component, the amplitude of dominant component, phase place, frequency and decay factor four parameter values are uploaded to system main website; Then, ask for the relative entropy of adjacent check point, high according to trouble spot homonymy 2 zero-sequence current wave-form similarities, and the feature that 2, both sides, trouble spot similarity is low, orient fault section; Finally, in the fault section oriented, injection method is utilized to carry out fault distance mensuration.

2. transmission line fault synthesized positioning method according to claim 1, is characterized in that, described in orient fault section concrete steps as follows:

1) set the sampling period of zero sequence current mutual inductor as Δ t, as system zero sequence voltage U ₀when () is greater than the bus rated voltage of 0.15 times t, can think that system there occurs singlephase earth fault, count if N is sampled data, within the T/4 time, zero sequence current mutual inductor acquires N secondary data, each check point gathers, after fault occurs, obtain N number of sampled data in T/4, then have: N Δ t=T/4;

2) adopt the transient zero-sequence current of Prony algorithm difference matching m check point, select the transient zero-sequence current dominant component I of each check point _k(k=1,2 ... m), line feed terminals FTU is by I _kamplitude, phase place, frequency and decay factor 4 parameter values be uploaded to system main website;

3) the dominant component parameter of m check point uploaded according to FTU of system main website, ask for the relative entropy of adjacent check point, concrete computation process is as follows:

Each check point has N number of sampled data, the n-th data acquisition sampling point corresponding moment t=n Δ t (n=1,2 ... N), 1. for kth (k=1,2 ... m) individual check point, calculating moment t=n Δ t (n=1,2 ... N) transient state dominant component I time _k,taccount for whole system Q _tproportion q _k,t:

$q_{k,t}=\frac{I_{k,t}}{Q_t},(k=1,2,...m),$

Wherein, Q _tfor the zero-sequence current dominant component sum of m check point on t feeder line, be also

$Q_t=\underset{k=1}{\overset{m}{\&Sigma;}}{I}_{k,t};$

2. calculating check point k relative to the Prony relative entropy of check point k+1 is:

$M_{k,k+1}=\underset{n=1}{\overset{N}{\&Sigma;}}\left|q_{k,t}ln\frac{q_{k,t}}{q_{k+1,t}}\right|,(k=1,2,...m)$

In formula, t=n Δ t (n=1,2 ... N);

4) M is found out _{k, k+1}(k=1,2 ... m-1) maximal value in, then fault is just between check point k and k+1 of correspondence.

3. transmission line fault synthesized positioning method according to claim 2, it is characterized in that, the described injection method that utilizes carries out fault distance mensuration, and calculate the distance of trouble spot and bus bar side, concrete steps are as follows:

A: utilize voltage transformer (VT) to judge fault phase, and at bus place, fault phase is excised;

B: utilize the voltage transformer (VT) of fault section upstream detection point k to inject the voltage signal of 220Hz to fault phase;

C: the steady state voltage of measuring-signal decanting point, current value the distance L of trouble spot to signal injection point is calculated according to following formula:

$\frac{\stackrel{\&CenterDot;}{U}}{\stackrel{\&CenterDot;}{I}}=(R_0+R_L)+{\mathrm{jX}}_L\&ap;R_0+{\mathrm{jX}}_L,$

In formula, R ₀for fault ground resistance, R _lfor signal injection point is to the line resistance of trouble spot, X _lfor signal injection point is to the fault phase circuit induction reactance of trouble spot, x _lfor circuit unit length induction reactance value, U _m, I _mfor electric current and voltage peak value, for the electric current and voltage phase angle difference of Injection Signal;

4) set a kth check point and line bus lateral extent as L _k, then trouble spot and bus bar side distance, i.e. fault distance L _total=L+L _k.