FI121612B - Method for protecting the electricity distribution network - Google Patents

Description

Method for protecting the electricity distribution network

Background of the Invention

The invention relates to the protection of the electricity distribution network and in particular to the protection of the medium voltage network against a broken or transient earth fault.

5 There are many different types of earth faults in the electricity distribution network.

In underground cable networks, earth faults are usually permanent and of low fault resistance. Such failures require disconnection of the faulty line section from the rest of the network. The overhead transmission network, in turn, is characterized by arc earth gaps which can be eliminated by means of reconnections.

10 A failure case other than the above is an intermittent earth fault in which the ground contact is interrupted and the fault duration is considerably shorter than the period. Such a breaking transient-type earth fault can be caused, for example, by a failure in the cable insulation, and is typically a problem with a compensated cable network. In a compensated network, the likelihood of an earth fault extinguishing 15 is high, but once the insulation failure occurs, the fault is likely to re-ignite. The fault then manifests itself as repeated strong earth-fault current peaks and corresponding rapid changes in the instantaneous value of the zero voltage. The exact moment of ignition is random due to the general nature of the breakthrough phenomenon.

20 The degree of network compensation affects the recurrence of transient fault.

In a fully compensated network, the zero voltage dims over several cycles, and the fault may recur every few hundred milliseconds (Figures 2 and 3). If the compensation is tuned to the side, the fault may recur every half cycle because the zero voltage does not have time to damp down between the faults (Figures 4, 5, 6, 7 and 8).

25 A transient earth fault in this context is an earth fault in which the connection to the earth is very short-term, so that the earth fault disappears quickly. The resulting current pulse is similar to that of a broken earth fault.

An intermittent or transient earth fault may also occur in a ground-isolated or grounded network.

Failure to selectively detect a breaking earth fault may result in the tripping of a large mains section, such as tripping of an electrical station supply cell circuit breaker instead of disconnecting only the faulty wiring output. Typically, such a problem results from the fact that the value of the zero voltage used as a precondition for shielding remains high for clearly longer pulses.

2

In the case of a transient earth fault, protection is not necessary, but selective tripping of a transient earth fault can prevent the occurrence of a permanent fault with potentially serious consequences.

BRIEF DESCRIPTION OF THE INVENTION An object of the invention is to provide a method for solving the above-mentioned problems of a broken earth fault and a temporary earth fault. The object of the invention is achieved by a method characterized by what is stated in the claim.

The invention is based on measuring the zero voltage U0 (2) from the network (substation medium voltage-10 busbar) and the zero current lo (3 and 7) from the substation outputs (4 and 8). When the instantaneous value of l0 exceeds the setpoint (11), and during the next period, the Uo base wave component exceeds the setpoint (12), the protection relay algorithm is triggered. The U0 signal is differentiated (13), the 10 components (15 and 14) close to the network frequency are removed from the 10 and A0o / At signals, and the sum of the inputs (16) of simultaneous samples of the filtered signals is calculated. It describes the similarity and simultaneity of the transients of the lo and Uo signals. In the case of a transient fault, the sum of the fault output (4) is negative (Figs. 2, 4, 6 and 8, ref. 16), since the processed lo and AU0 / At signals (15 and 14) are ideally identical to the others, but of opposite signs. Similarly, at healthy outputs (8), the sum is positive (Figures 3, 5 and 7, ref. 16), since the processed 10O and AU0 / At signals (15 and 14) are ideally similar. It is essential that the fundamental frequency signal is filtered out before the sum is calculated, since it occurs in the same phase in both the fault output and the healthy outputs. The entire procedure is repeated at intervals. When the total duration of the fault burst (Figures 4 and 6) or of individual faults (Figure 2) exceeds the set operating time, a trip command is given to the circuit breaker.

An advantage of the method according to the invention is that the invention makes it possible to identify the line output at which a breaking or transient earth fault 30 occurs. Particular advantages over other methods are the simplicity of the method and the fact that the method does not require a high sampling rate and thus no separate equipment for the operation in question.

3

BRIEF DESCRIPTION OF THE DRAWINGS

The invention will now be further described in connection with preferred embodiments with reference to the accompanying drawings, in which:

Fig. 1 shows a compensated or switched off power network in which several outputs (4 and 5) are output from 5 substations and an intermittent earth fault (5) occurs at the output (4) in step L3. Output (8) represents all other outputs. The compensation coil (1) is tuned to or near the operating frequency in resonance, according to the network compensation degree, with the ground capacitances (6 and 9) of the network. Zero voltage U0 (2) and output lo (3 and 7) are measured .; Figure 2 shows the defective output signals when the network compensation ratio is 100%. The faulty phase voltage (10) is slowly restored, so in this example, the new fault will not start until after about 200 ms. The measured 10 signal (11) and the U0 signal (12) after the device anti-crease filter are shown during a typical transient earth fault. Also, the processed signals 15 according to the algorithm, AU0 / At (13), filtered AU0 / At (14) and lo (15) and the sum of the inputs (16) of simultaneous samples of the filtered signals are shown .;

Figure 3 shows healthy output signals in the same error as Figure 2;

Figure 4 shows the same faulty output signals as Figure 2, but the network 20 has a compensation rate of 30%. The faulty phase voltage (1) recovers rapidly and transient earth faults are likely to occur at very high density compared to the 100% compensation situation.

Figure 5 shows healthy output signals in the same error as Figure 4; Figure 6 shows the same faulty output signals as Figure 2, but the network compensation rate is 170%. The faulty phase voltage (1) recovers rapidly and transient earth faults are likely to occur very dense compared to full compensation in the grid situation.

Figure 7 shows the healthy output signals in the same error as Figure 30;

Figures 2 to 7 are simulation results of a real network model. Figure 8 shows the measured fault output I0 and U0 signals (11 and 12) in a spontaneous fault of an actual network. Also, the sum of the inputs (16) of the simultaneous samples of the processed signals according to the algorithm, AU0 / At (13), filtered AU0 / At (14) and 10 (15) and 35 are shown.

4

DETAILED DESCRIPTION OF THE INVENTION

The invention is based on measuring the zero voltage U0 (2) from the network (the substation medium voltage busbar) and the zero current 10 (3 and 7) from the substation outputs. When the instantaneous value of 10 exceeds the setpoint (11), and during the next 5 cycles, the base wave component of the U0 (12) exceeds the setpoint, the protection relay algorithm starts.

Take samples of parasites (11) and Joy (12) in the work buffer for the period. During the fault transient, currents are mainly capacitive, so l0 is proportional to the derivative of U0. Differentiating U0 gives an approximation of the derivative 10 of U0 and the differentiated waveform (13) corresponds to the waveform (11) of 10. In the event of a fault, the directions of the currents measured from the position of the fault output (4) and the healthy outputs (8) are reversed. Thus, at the fault output, the loin (11) waveform during transient is similar to AU0 / At (13) but of different sign (Figures 2, 4, 6 and 8). At healthy outputs, the 10 waveform has the same shape 15 and the same sign as AU0 / A (Figures 3, 5 and 7).

The measured zero voltage is differentiated as follows: The difference in U0 (n) corresponding to the sample lo (n) of the lo signal is calculated symmetrically with respect to l0: AU0 (n) / At = U0 (n + 1) - U0 (n-1). the accuracy of the result will be improved.

20 When the earth fault is over, the whole network begins to vibrate to return to equilibrium. The frequency of the oscillation is determined by the degree of network compensation, i.e. the resonance frequency of the quin-resonance circuit formed by the quench coil (1) and the total network capacitance (6 and 9). If the compensation ratio is 100%, the frequency is 50 Hz (Figures 2 and 3). Also, in the case of over- and under-compensation, the frequency 25 is close to 50 Hz (Figures 4, 5, 6, 7 and 8). Since the fault is already over, this attenuating signal of about 50 Hz is identical to the lo output of the fault output and the healthy outputs, thus making it difficult to identify the fault output. Therefore, the fundamental frequency component is attenuated from both 10 and differentiated U ues.

Next, the sum of the inputs of 30 simultaneous samples of these filtered signals (15 and 14) is calculated as follows: k = Σ (Ιο χ AUo / At) (16). In the fault output, the sum is negative (16 in Figures 2, 4, 6 and 8) and in healthy outputs the sum is positive (16 in Figures 3, 5 and 7). If the sum is negative, the protection function delay time is incremented.

Only in the case of transient faults, if no current peaks have occurred within the set recovery time of the weapon, or when the sum has been continuously zero or positive during the set recovery time, the running time counter is reset.

5

When the compensation rate deviates clearly from one hundred percent, the voltage of the fault phase reaches the likely breakdown voltage already during the same period (Figures 4 and 6), whereby faults usually occur every period. When the fault counter reaches the set delay, the fault output breaker is tripped.

With the compensation degree close to one hundred percent, the amplitude of the fault phase voltage slowly returns to its original value (Figure 2), and a new fault is likely to occur only after hundreds of milliseconds. However, the protection is triggered by the first fault, and if a new fault occurs within the set recovery time, a trip is triggered when a sufficient number of faults have occurred.

10 Depending on the distance from the substation to the substation and the size of the network, the duration of the fault can vary from less than 100 microseconds to a few milliseconds. When the amplitude of the lo transient is very large, and because the amplitude of the short transient is greater than that of the longer transient and because the protective relay anti-fade filter prepares a narrow lo transient while spreading it 15, used in the figures below.

When the network compensation rate deviates from 100%, the fault phase voltage may return to such a high het-20 clock value during the next half-cycle that a new break-through may occur within the same or next half-period, or after a few cycles (Figures 4 and 6). In this case, the breakthrough strength of the fault site may change from one defect to the next, and the intensity of the 10 current pulse may decrease compared to the first breakthroughs (Figure 8). Due to the strong attenuation of the high-frequency frequencies constituting the sharp current pulses, the low-amplitude lo-current pulses may not be detected. However, even small current pulses maintain a continuous U0 signal. By triggering a time-out timer from the detected 10 pulse that triggers the trip if the U0 signal continuously and sufficiently long exceeds the trigger threshold, rapid triggering can also be guaranteed even when the amplitude of the dense repetitive lo-transient has fallen below the trigger threshold.

Claims (4)

1. Menetelmä sayhkönjakeluverkon suojaamiseksi catkeilevalta tai ohimenevältä maasululta, tunnettu siitä, one menetelmässä: - Only lo: n hetkellisarvo ja Uo: n 50 Hz componentin amplitudi ylittä-ja- näyttet jakson ajalta työpuskuriin - Lasketaan lo-signaalin (11) työpuskurin näytteitä lo (n) vastaavat U0: n difference site (13) symmetry path: AU0 (n) / At = Uo (n + 1) -Uo (n-1) - Mitatusta zero signal signal list lo (11) yes mitatun zero signal signal and difference signal signal ΔΙΙ0 / Δΐ (13) vaimennetaan perustaajuinen component (15 yes 14) - Lasketaan no processojeno l0- yes Δυ0 / Δί signalin 15 (14 yes) Σ (Ι0χ Δυ0 / Δί) (16) - Jos k on negatiivinen (cuviot 2, 4, 6 ja 8), kyseessä on vikaantuut 15 käyntiin lähtenyt toiminta-aikala skuri. - Jos havaitun negatiivisen k: n jälkeen Uo-signaali jokaisella tarkastelujaksolla ylittää asetellun rajan, lisätään Uo-aikakatkolaskurin arvoa. - Only toiminta-aikalaskuri tai U0-aikakatkolaskuri saavuttaa käyttäjän asetteleman arvon, erotetaan vikaantunut johtolähtö (4) cost. claim 1. A method which protects electricity distribution networks from earth faults of a transient nature, characterized by the method: 5. When the instantaneous value of 10 and the amplitude of the 50 Hz component of Uo exceeds the values set by the user, sampling values of the measured 10 and Uo signals (11 and 12) are read for a period to the work buffer. - the differences are calculated by the sampling values of the LO signal (11) in the working buffer 10 (n) corresponding to the differences of U0 (13) symmetrically: AUo (n) / At = Uo (n + 1) -U0 (n-1) - Of the the measured zero sequence current signal lo (11) and of the difference signal of the measured zero sequence voltage AUo / At (13), the basic frequency component (15 and 14) is attenuated. 15. If the dependent factor of the processed Lo and AUo / At signals (15 and 14) is calculated as k = ((Ι0 x AUo / At) (16) - If k is negative (Figures 2, 4, 6 and 8), then the issue of incorrect output management and the function timer are extended. - When k has been continuously positive for the time set, 20 the started function time counter is reset. - If the Uo signal, after a negative k detected, exceeds the setting in each observation period, the value of the Uo interrupt timer is increased. - When the function timer or the U0 interrupt timer has reached the values set by the user, the incorrect output line (4) is isolated from the rest of the network.