DE10225058A1 - Selective transient earth fault detection in polyphase networks involves comparing displacement voltage and null current curves, evaluating proportionality factor to detect earth leakage and direction - Google Patents

Abstract

The method involves continuous sampling and storage of a displacement voltage and null current in an output, detecting the first time the voltage exceeds a threshold in the event of an earth leakage, storing the displacement voltage and null current until the defined earth leakage, computing the null current summation curve, determining the gradient and the offset between the stored curves and evaluating the proportionality factor. The method involves continuous sampling and storage of a displacement voltage and null current in an output, detecting the first time the voltage exceeds a threshold in the event of an earth leakage, searching in the memory after the null crossing and before the first earth leakage, storing the displacement voltage and null current until the defined earth leakage, computing the null current summation curve, determining the gradient and the offset between the stored curves and evaluating the proportionality factor to detect an earth leakage and its direction.

Description

The present invention relates to the selective detection of earth leakage wipers in three-phase networks isolated star point or with a high-resistance grounded star point.

In the technical report "Examination of the compensation processes at Earth fault in power supply networks ", appeared in power engineering, Vol 15, Issue 10, October 1965, page 469ff, describes Prof. Hans Pundt in great detail the transient processes when the earth fault occurs. Among other things, it does the calculation of the expected charging current in terms of amplitude and frequency derived.

On the basis of this technical report, the published patent application DT1 463 729 presents a method in which the first half-wave of the charge-reversal current after being converted into a square-wave pulse of defined width is compared in terms of polarity with the displacement voltage converted into square-wave pulses. With this method it is assumed that the charge reversal current is very large compared to the stationary charge current I _{CE of} the line. On the one hand, this requires transition resistances of only a few ohms and, on the other hand, an ignition timing that is not close to the zero crossing of the phase-earth voltage. These two boundary conditions, especially the low-resistance contact resistance at the fault location of a few ohms, cannot usually be met. The difficult separation of high-frequency discharge vibrations from the medium-frequency charging processes also leads to incorrect displays.

In the published application DE 27 13 000 the charge-reversal current is not compared with the zero sequence voltage but with a charge-reversal current of a network simulation of the network-earth capacitances. In this method, it is assumed that both recharging processes are damped discharge vibrations of approximately the same frequency. The high-frequency discharge vibrations are also taken into account in this described method and can lead to incorrect displays. Furthermore, it can be shown that the necessary increase in the response threshold for the charging currents to suppress the high-frequency discharge vibrations can result in an incorrect phase position and thus a faulty display of the relay at different charging frequencies at the time of the evaluation. The relays would therefore always have to be adapted to the respective network after switching networks.

The object of the invention is the earth fault wiper method in relation to the evaluation of the transient To improve the process so that this method for both higher impedance Fault resistances up to in the range of a few kOhm as well as for ignitions near the zero crossing the conductor-earth voltages is suitable.

The advantage of the new method is that the size of the contact resistance Z _{F is} already taken into account indirectly via the size of the zero sequence voltage. With the new method, a directional decision from the transient process is possible down to a few kOhm. So far, direction detection could only be carried out in the range of a few ohms with wiper relays. Most earth faults, however, already have contact resistances of a few hundred ohms.

Another advantage of the process is that the ignition timing the earth fault does not matter. The procedure works even with ignition in the zero crossing of the conductor-earth voltage, which e.g. through a mechanical destruction of the cable can be done. Previous wiper relays only recognized Ignition processes in the Near the Voltage maximum.

Another advantage is that earth faults too with a burn time in the ms range can be detected, even if this earth fault via a high-resistance resistor he follows.

Another advantage of the method is that the response sensitivity of the relay is set by two parameters:
u _ERD Required _zero sequence voltage for an earth fault
C _0min Required cable length of the healthy network in A.

With the previous relays, a tripping current is specified which is a function of Z _F , ignition timing, transformer impedance, etc. The actual tripping had to be recalculated for each earth fault.

• – Kontinuierliches Abtasten und Speichern der Verlagerungsspannung u₀ und des Nullstromes i_0ΣX in einem Abgang.
• – Im Falle eines Erdschlusses überschreitet die Verlagerungsspannung u₀ erstmalig einen eingestellten Grenzwert u_ERD zum Zeitpunkt t₁
• – Vorzugsweise erfolgt eine Suche im Speicher nach dem Zeitpunkt t₀ des Nulldurchganges der Verlagerungsspannung u₀ vor dem oben definierten erstmaligen Erdschlusseintritt zum Zeitpunkt t₁. Es kann aber auch ein anders definierter Bereich verwendet werden z.B. 10 ms oder der Zeitpunkt des dritten Nulldurchganges vor dem definierten Erdschlusseintritt.
• – Speichern der Verlagerungsspannung u₀ und des Nullstromes i_0ΣX ab diesem Zeitpunkt t₀ bis zum definierten Erdschluss t₁.
• – Zur genaueren Berechnung können zusätzlich die Messwerte, die einige ms nach dem Zeitpunkt t, gemessen worden sind, berücksichtigt werden.
• – Berechnung der Nullstrom-Summenkurve i_{0ΣA_SUMME} durch aufsummieren des Nullstromes i_0ΣX bei zeitäquidistanter Abtastung, andernfalls muss die Zeit zwischen den Abtastpunkten berücksichtigt werden bzw. ein anderes numerisches Integrationsverfahren verwendet werden.
• – Ermittlung der Steigung und des Offsets zwischen den beiden gespeicherten Kurven i_{0ΣX_SUMME} und u₀ nach bekannten Verfahren der Mathematik wie z.B. Regressionsanalyse, Korrelationsanalyse, Polynominterpolation, Pseudoinverse usw.
• – Liegt die Abweichung des Proportionalitätsfaktors innerhalb eines vorgegebenen Toleranzbereiches, so wird der Erdschluss in Richtung Sammelschiene erkannt.
• – Ist die Abweichung des Proportionalitätsfaktors größer als ein vorgegebener zweiter Toleranzbereich, so wird der Erdschluss in Richtung Leitung erkannt.

According to the invention, a method for detecting the direction of the earth fault from the charging oscillation is characterized by the following steps:

• - Continuous sampling and storage of the zero sequence voltage u ₀ and the zero current i _0ΣX in one outflow.
• - In the event of an earth fault, the _{zero sequence voltage} u ₀ exceeds a set limit value u _ERD for the first time at time t ₁
• - A search is preferably carried out in the memory for the time t _{0 of} the zero crossing of the zero sequence voltage u ₀ before the first earth fault occurrence defined at the time t ₁ . However, a differently defined range can also be used, for example 10 ms or the time of the third zero crossing before the defined earth fault occurs.
• - Saving the zero sequence voltage u ₀ and the zero current i _0ΣX from this point in time t ₀ to the defined earth fault t ₁ .
• - For more precise calculation, the measured values that were measured a few ms after time t can also be taken into account.
• - Calculation of the zero current sum curve i _{0ΣA_SUMME} by _adding up the zero current i _0ΣX with time-equidistant sampling, otherwise the time between the sampling points must be taken into account or another numerical integration method must be used.
• - Determination of the slope and the offset between the two stored curves i _{0ΣX_SUMME} and u ₀ using known mathematical methods such as regression analysis, correlation analysis, polynomial interpolation, pseudo inverse etc.
• - If the deviation of the proportionality factor lies within a specified tolerance range, the earth fault in the direction of the busbar is detected.
• - If the deviation of the proportionality factor is greater than a specified second tolerance range, the earth fault in the direction of the line is detected.

In the following the invention is described by way of example with reference to Figure 1. Figure 1 shows a network with three outgoers and an assumed earth fault in the conductor 1 shown in outlet A.

• – Entladung des schadhaften Leiters in allen Abgängen über die Erde
• – Aufladung der gesunden Leiter in allen Abgängen über die Erde
• – Stationär eingeschwungener Zustand

The earth fault consists of the following three processes:

• - Discharge of the defective conductor in all outlets via the earth
• - Charging the healthy conductors in all outlets via the earth
• - Stationary steady state

The effects of these three processes overlap immediately with the beginning of the earth fault.

• – Größe der Kapazitäten des Leiters 1 aller Abgänge gegen Erde
• – Ladezustand der Kapazitäten des Leiters 1 aller Abgänge
• – Leitungsimpedanz Z_L1, zu und in den andern Abgängen
• – Impedanz der Fehlerstelle Z_F

Decisive for the course of the discharge are:

• - Size of the leader's capacities 1 all exits to earth
• - State of charge of the capacities of the conductor 1 all exits
• - Line impedance Z _L1 , to and in the other outlets
• - Impedance of the fault location Z _F

The transformer and the loads a very high impedance, essentially inductance, and are for to consider the discharge process as a high-resistance termination. As well the influence of the two healthy leaders can be neglected. The conductor L1 can be as at the ends of the outlets and on the busbar high-impedance terminated chain conductors can be considered. The healthy ones Stub lines represent a parallel connection, which means the equivalent longitudinal impedance the healthy conductor has a lower resistance and the capacity increases.

The very high-frequency unloading process is of the conductor lengths dependent and is the higher frequency the shorter the lines are and are usually in the range> 10 kHz

• – Kapazität der Leiter 2 und 3 aller Abgänge gegen Erde
• – aktueller Ladezustand der Leiter-Erde Kapazitäten aller Abgänge
• – Größe der Aufladespannungen
• – Streuinduktivitäten des Einspeisetransformators
• – Leitungsimpedanz Z_L1, von der Fehlerstelle bis zum Einspeisetransformator
• – Impedanz der Fehlerstelle Z_F

Decisive for the course of the charging process of the two healthy conductors are:

• - Capacity of the ladder 2 and 3 all exits to earth
• - Current state of charge of the conductor-earth capacities of all outlets
• - Size of the charging voltages
• - leakage inductances of the infeed transformer
• - Line impedance Z _L1 , from the fault location to the infeed transformer
• - Impedance of the fault location Z _F

The distribution transformers or the Loads are only considered with very high impedance and can in the first approximation neglected become. The load essentially acts as an additional attenuation the charging oscillation. If the distribution transformer is not loaded, so the line is essentially with the very high main inductance of the transformer loaded.

Due to the earth fault, the conductor 1 brought to earth potential by discharge via the impedance at the fault location. The conductor-earth voltage of the two healthy conductors is on the √ after the charging process has ended 3 times raised.

Summenstrom im Abgang x:

The following applies to the instantaneous values for each point in time:
Displacement voltage on the busbar:

Total current in the outlet x:

The displacement voltage is determined using known methods such as measurement via the "open delta winding" or measurement of the three individual voltages and subsequent numerical addition in the device.

• – Messung der drei Einzelströme und elektrischer Addition der drei Ströme ( Holmgreen-Schaltung )
• – Messung des Summenstromes mit Hilfe eines Kabelumbauwandlers.
• – Messung der drei Einzelströme und numerische Addition nach der Wandlung im Gerät.

The total current is determined using known methods, such as, for example

• - Measurement of the three individual currents and electrical addition of the three currents (Holmgreen circuit)
• - Measurement of the total current using a cable conversion transformer.
• - Measurement of the three individual currents and numerical addition after conversion in the device.

The total current is three times the zero current i _0X .

For the considered frequency range of the recharging process are the longitudinal impedances of the cables negligible or only reduce the zero capacities determined below low.

It is essential that the displacement voltage u _{0 is determined} according to its definition in relation to the earth.

In picture 1 the following types of disposals can be distinguished:

A) Healthy finish without Ground Fault

Neglecting the longitudinal impedances of the lines, only the zero capacitances to earth are effective in the healthy output for the zero system. The zero capacity of an outlet can be calculated according to the known methods such as the symmetrical components to:

bzw. für die Verlagerungsspannung:

The following applies to the total zero current i _0ΣX in a healthy outlet:

or for the displacement voltage:

The integration can be approximated numerically by forming a sum. With a time-equidistant sampling of the displacement currents and the zero currents, the integration for an addition of the sampling values is simplified. For the case u ₀ (0) = 0, the time profile of u ₀ differs from the cumulative curve only by a proportionality factor.

B) Earth fault leaving

On the other hand, in the faulty outlet The following components are included in the zero current:

• a) flowing into the outlet:
• - Charging current of the faulty outlet A
  
• b) Flowing out of the outlet:
• - Sum of the currents of the healthy outlets including the earthed outlet
  
• - current through the Petersen coil (L _P // R _p )
  

The total zero current of the outgoing circuit A is:

In isolated networks, the current i _{0ΣX corresponds to} the sum of the zero currents of the healthy feeders, however with the opposite sign.

In the case of deleted networks, the settled value is i _0ΣX , depending on the size of the Petersen _coil and the curve shape can even be very similar to that of a healthy outlet.

At the beginning of the earth fault is the Low proportion of the current through the Petersen coil.

From A) it can be seen that from the proportionality factor the zero capacity the line can be calculated. So there is the possibility the trigger sensitivity of the relay in the form of the zero capacitance of the healthy cable length.

Since only the charging oscillation is considered, low-pass filters can be _connected upstream of the measurement of u ₀ and i _0ΣX .

Claims (9)

Method for the selective detection of earth leakage wipers in three-phase networks characterized by the following steps: - Continuous sampling and storage of the zero sequence voltage u ₀ and the zero current i _0ΣX in one outlet; Detection of the first time that the displacement voltage u _{0 exceeds} a set limit value u _ERD at time t ₁ in the event of an earth fault; - Search in the memory for the time t _{0 of} the zero crossing of the zero sequence voltage u ₀ before the first earth fault entry defined at time t ₁ ; - Saving the zero sequence voltage u ₀ and the zero current i _0ΣX from this point in time t ₀ to the defined earth fault t ₁ . - Calculation of the zero current sum curve i _{0ΣA_SUMME} by _adding up the zero current i _0ΣX with time-equidistant sampling, otherwise the time between the sampling points must be taken into account or another numerical integration method must be used. - Determination of the slope and the offset between the two stored curves i _{0ΣX_SUMME} and u ₀ using known mathematical methods such as regression analysis, correlation analysis, polynomial interpolation, pseudo inverse, etc .; - Detection of an earth fault in the direction of the busbar if the proportionality factor deviates within a specified tolerance range; - Detection of an earth fault in the direction of the line with a greater deviation of the proportionality factor than a predetermined second tolerance range. A method according to claim 1, characterized in that for more precise calculation, the measured values, which are a few ms after the point in time t _{1 have been} measured. Method according to Claim 1, characterized in that a differently defined range is used for determining the time t ₀ , for example 10 ms or the time of the third zero crossing before the defined earth fault occurs. A method according to claim 1, characterized in that the zero voltage u _{0 is} calculated by measuring the three voltages u _{1 E} u _2E and u _3E or is measured directly via an open delta winding. A method according to claim 1, characterized in that the Determination of the total current using methods such as - Measurement of the three individual streams and electrical addition of the three currents (Holmgreen circuit) - Measurement of the total current with the help of a cable conversion converter. - Measurement of the three individual streams and numerical addition takes place after the conversion in the device A method according to claim 1, characterized in that threshold values for the Deviation of the proportionality factors introduced become. A method according to claim 1, characterized in that if necessary, low-pass filters are _connected upstream of the measurement of u ₀ and i _0ΣX . Method according to Claim 1, characterized in that the _{zero capacitance of} the outflow is calculated from the proportionality factor of the curves u ₀ and i _0ΣX . A method according to claim 1, characterized in that a further trigger threshold for the Emp sensitivity of the relay is set by specifying the zero capacitance of the outlet.