EP1904862A1

EP1904862A1 - A method of earth fault identification and location in three-phase electrical network

Info

Publication number: EP1904862A1
Application number: EP06769357A
Authority: EP
Inventors: Andronis Linas Markevicius; Saulius Gudzius; Alfonsas Morkvenas; Linas Markevicius
Original assignee: UAB "Energijos Srautas"
Current assignee: UAB "Energijos Srautas"
Priority date: 2005-07-20
Filing date: 2006-07-18
Publication date: 2008-04-02
Also published as: LT2005067A; WO2007011196A1; LT5337B; WO2007011196B1; RU2008105543A

Abstract

The invention relates to the area of electrical power network and may be applied to a method of permanent or short-term earth fault identification in a three-phase electrical distribution network and distance determination to the fault place. Better accuracy for the earth fault place location is achieved by performing registrations not longer than 100 µs and with time intervals smaller than 100 ns, thus decreasing the influence of noises of lower than 10 kHz frequency. The distance between the beginning of the feeder's section and earth fault place is defined by minimising the goal function f2, which is the function of the difference of registered and simulated processes: (I).

Description

A METHOD OF EARTH FAULT IDENTIFICATION AND LOCATION IN THREE-PHASE ELECTRICAL NETWORK

The invention is intended for the area of electrical power network and may be applied to a method of the earth fault place location and identification in a three-phase electrical distribution network.

The main faults of electrical distribution networks are wire breaks and/or earth faults. Fault detection methods in electrical distribution networks are described in patents SU 1762279 and SU 1698849. Fault place location in patent SU 1762279 is based on useful reflection from the place of fault of signal identification and its usage for the fault location. A method of phase earth fault location is described in patent SU 1698848, where the first half-wave of the discharge current is fixed in a place of short circuit, two- component amplitudes of the discharge current of a half- wave spectrum are measured and under a respective formula distance to the earth fault place is founded.

The shortcoming of both inventions is significantly complicated and not very precise earth fault location. In addition, none of the inventions fix short-term earth faults.

The invention presents a new method of earth fault location in a three-phase electrical network. It differs from the well-known methods by registration not only permanent but also short-term self-disappearing earth faults and more precise location of former or present earth faults between the phase and the earthed object in one section of a three-phase feeder substation connected to the bus section of the feeder substation. It permits the supervisory services of the electrical network to efficiently eliminate the occurred earth fault and/or to prevent possible faults, where for some reasons (the spark, touch of the tree branches, etc.) short-term earth faults occur.

A method is characterised by its registration of the initial processes of zero-phase sequence and voltage variations after earth fault, which are registered at the voltage transformer connected to the bus section and in the feeder current transformer of a zero- phase sequence current. The better accuracy of the earth fault place location is achieved by not longer than 100 μs performance and smaller than 100 iis time intervals, decreasing by this noise influence, which frequency is lower than 10 IdHz, by special algebra separating the processes running in other feeders and loading them into a model, where the simulated process is compared with the registered one. The place of earth fault and model element parameters are found by minimising goal function which is the function of the difference between the registered and simulated processes.

This invention will be described more thoroughly with reference to the enclosed drawings, where:

Fig. 1 presents a block-diagramme of the operational device for this method realisation;

Fig. 2 presents a scaling filter scheme of a high frequency current process.

For the earth fault place location this invention uses a device schematically described in table 1, consisting of a zero-phase sequence current process identification and separation of analogue link block 1, identification of a zero-phase sequence voltage process and separation of analogue link block 2, transducers of analogue code and memory block 3, converting measured within earth fault analogue zero sequence current and voltage processes into a buffer memory, controllers block 4, co-ordinating identification of the registered process performing its analysis, modelling, parameters optimisation, fault identification and being in connection with the outer systems, interface and indication block 5 demonstrating the estimated distance to the place of earth fault

Under this invention earth fault between a phase and the earthed object in one three-phase feeder connected to the feeder substation bus section is located by registering in an analogue method after earth fault zero-phase sequence initial current I₀ and voltage U₀ variations processes in due blocks 1 and 2 obtained from a voltage transformer connected to a bus section and from the feeder's zero-phase sequence high frequency current scaling filter 6 (Fig. 2) equipped in block 1. Parameters of a high frequency current process scaling filter 6 depend on the feeder's section and configuration. Use of scaling filter 6 permits registering initial processes after earth fault of a zero-phase sequence high frequency current and locating faults more precisely. Also better accuracy of earth fault location due this invention is achieved by registering faults no longer than 100 μs and in smaller than 100 ns time intervals reducing by this noise influence whose frequency is smaller than 10 kHz . Registered analogue signal block 3 converts into a digital one. In Unit 4 by special algebra processes operating in other feeders are defined and loaded into a model where the simulated process is compared with the registered one. The place of earth fault and model element parameters, i.e. the simulated voltage and current transient characteristics in a model as well as the feeder's wave parameters and node capacity are found by minimising goal function, which is the function of the difference between the registered and simulated process.

The method of this invention is composed of an algorithm, the processor of which is available in controller block 4, locates earth fault in the feeders' sections in the following stages.

1) At the first stage it is checked whether earth fault occurred in the device feeder. Earth fault is supposed to occur in the device feeder if number (I) has a positive sign:

P = Σ ¹J^UJ (D

M

Where z) -current j registration; U_j - voltage value j registration; n_\ - number of part of current values of all N registrations: n_\<N.

2) If the first stage confirms earth fault occurrence in the feeder, i.e. if number (I) is positive - it fulfils the second stage, during which by minimising goal function f_\ the parameters of transitional characteristics of current and voltage measuring transformers are defined _:

"2 fi ~ Σ |^rem(^i ' derem(l ~ derem(l_λ , U₁ ),1 + derejn(l_l , U₁ )))) = min (II)

Where I₁ - vector: I₁ = [z^' ₁₅/₂,...,z^',_!2] , U₁ - vector: U₁ = [M₁₉M_29^9M₁₁₂], F₁ - «₂ value vector of a definite form impulse function (7?₂<iV), 1 = [l, l,..., l]_n2 ; rem and derem — algebra operations for equal clearance vectors characterised by following formulas:

(z_k )_n = rem((x_k)_H , (y_k )„ ) (III)

^zk ΛJ; ⁽I^V)

(x_k )„ = derem{{z_k )„ , (y_k )„ ) (V)

3) The third stage by minimising goal function /₂ determines the feeder's section where earth fault occurred and distance in meters from the beginning of the section to the earth fault place:

Λ = ∑hm(F₂ Λ)^~rem(F₂,S_k)| = ¹^^{11 (}VI⁾

A=I Where F₂ -M values vector of a definite form impulse function (N = LM , L ≥ l), L- positive integer, S and S - vectors obtained from registrations and found in a model are calculated using the following formulas:

s_k = _ [ ψS MM((kIc--lI))+₊jJ ) /_AM . a. ¹Af(A-I)+! ' M(k-\)₊2 'M(U-V)₊M (VII) a_h

S M(U-X)₊ j ~ - ^wiM(k-V)₊j (VIII)

Where ύ and ϊ - voltage and current values obtained in registrations and calculated in a model; w - non-negative rational number. Derivative results are described in indication block 5 where except distance from the beginning of the section to the earth fault place, section number and identification reliability where earth fault occurred are also indicated.

Claims

1. Earth fault detection method between the phase and earthed object place isolated from the earth in one section of a three-phase feeder by a device consisting of a zero-phase sequence current process identification and separation of analogue link block (1), zero-phase sequence voltage process identification and separation of analogue link block (2), analogue code converter block (3), controllers block (4), interface and indication block (5), characterised by:

- registering in blocks (1) and (2) after earth fault initial zero-phase sequence current I₀ and voltage U₀ transient processes obtained from the voltage transformer connected to the bus section and feeder's scaling filter (6) of a zero- phase sequence high frequency current arranged in block (1);

- checking if earth fault occurred in the feeder, the occurrence of which shows a positive number (1) sign:

where i) - current j registration ; U_j -voltage value j registration; n\ - number of the part of currency values of all N registrations: n\<N; in case the number (I) is positive — determining the parameters of transitional characteristics of current and voltage measuring transformers by minimising goal function/], _:

It₁

/i ⁼ ∑|^rem(^Fi ' derem(l - derem(l_x ,U₁ ),! + derem(l_λ ,U₁)))! = min (II)

7=1 where I₁ - vector. I₁ = [i_x ,i₂ ,...,i_n2] , U₁ - vector: U₁ = {u_λ ,u₂,...,u_n2 ] , F₁ - n₂ values vector of a definite form impulse function («₂< N ); l = [l, 1,..., l]_n2 ; rent and derem — algebra actions for equal clearance vectors are defined by the following formulas:

(z_k)_n = rem((x_k)_n,(y_k)_n ) (III)

^zk [°> JW₂V₅ Jv₁ ]; (iv)

(x_k ) „ = derem((z_k ) „ , (y_k )„); (V); determining the feeder's section where the earth fault occurred and calculating the distance in meters from the beginning of the feeder's section to the earth fault place by minimising goal function /₂ :

/₂ = ∑|«w(F_a,S₄ )- «m(F_a,S_k)| = πώi (VI) k=\ where F₂ - M values vector of a definite form impulse function (N = LM ,

L ≥ l), L - positive integer; S and S - vectors obtained from registrations and found in a model are calculated using the following formulas:

s_k ³M(k-l)+l >M(k-l)+2 Wn₊ J (VII)

^SM(k~l)₊j = ^UM(k-l)₊j ~ ^WiM(le-»+j (VIH) where ύ and i — values of voltage and current obtained in registrations or calculated in a model; w— non-negative rational number.

2. The method according to claim 1, characterised by registrating of initial earth fault zero-phase sequence current and voltage transient processes no longer than 100 μs and smaller than 100 ns time intervals reducing by this noise influence with the frequency lower than 10 IcHz.

3. The method according to any previous claims, characterised by registrating of short- term earth faults.

4. The method according to any previous claims, characterised by registrating of one industrial frequency half period and longer duration earth faults.