select 65

75

FERRORESONANCE AS A SOURCE OF DISTURBANCES AND FAILURES IN MEDIUM VOLTAGE DISTRIBUTION GRIDS

Rafał Tarko / AGH University of Science and Technology in KrakówWiesław Nowak / AGH University of Science and Technology in Kraków

Waldemar Szpyra / AGH University of Science and Technology in KrakówMariusz Benesz / AGH University of Science and Technology in Kraków

Andrzej Makuch / AGH University of Science and Technology in Kraków

1. INTRODUCTION

Ferroresonance phenomenon occurs when a ferromagnetic core inside an electrical device – primarily a voltage transformer or an unloaded power transformer – operates in the saturated condition, under which inductance has become a non-linear element. In practice ferroresonance can be triggered even by a temporary introduction of core into saturation, e.g. resulting from switching operations or a change in voltage resulting from an earth fault. Although this phenomenon is known in electric power engineering since the 1930s, neither effective criteria for diagnosing the possibility of its occurrence, nor means to counteract it have been specified to date [1, 2].

Ferroresonance as a source of disturbances and failures in medium voltage distribution networks, e.g. [3], is dangerous in its consequences for the following two main reasons:

• significant saturation of the core, which can lead to, for instance, thermal damage of voltage transformer’s primary winding

• development of (often lengthy) ferroresonance overvoltages.Furthermore, the neutral point’s increased potential also makes the zero sequence voltage appear in the

system, which can falsify the operation of ground fault protection [4].The paper presents an analysis of ferroresonance that occurred in a 6 kV distribution grid and disrupted

its operation. The analysis was based on grid system models developed for the EMTP-ATP programme, and on results of simulation studies designed to determine the ferroresonance occurrence conditions and consequences, as well as means to eliminate it.

2. ANALYSED SYSTEM CHARACTERISTICS – PROBLEM ORIGINS

Subject to the analysis is a part of 6kV power grid supplied form 110/6 kV substation (MSP – Main Supply Points) and cogeneration plant (CHP). A simplified diagram of MV switchgear is shown in Figure 1. The MSP is powered by two 16 MVA transformers: 115 ± 10% / 6.3 kV (TR-1) and 115 ± 10% / 6.6 kV (TR-2). The CHP is equipped with a 11. 4 MVA generator. The total length of MV cable lines in the system is over 60 km. The analysed grid operates with an insulated neutral point, and the ground fault current in the grid, in the normal regime supplied from transformer TR1 is Ic1 = 52.05 A, whereas in the grid supplied from transformer TR2 the ground fault current is Ic2 = 40.99 A.

Ferroresonance as a Source of Disturbances and Failures in Medium Voltage Distribution Grids

Abstract

The article reports a medium voltage power grid analysis carried out to identify the grid’s operating conditions in the aspect of ferroresonance occurrence. A documented ferroresonance instance is presented, which led to the voltage transformers’ damage. The analysis

was based on an original grid system model and on results of simulation studies designed to determine the ferroresonance occurrence conditions and consequences, as well as means to eliminate it.

76

Fig. 1. Simplified diagram of MV switchgear in the grid

A disturbance occurred in the 6 kV network that led to a failure (explosion) of voltage transformers in the measurement bays (bays 0 and 16) of the 6 kV MSP distribution substations and in the CHP switchgear (bays 12 and 17).

The disturbance was caused by a short circuit in the cable line powered from the MSP (bay 24) and opening of the circuit breaker in the bay resulting from tripping of the over-current and ground fault protection. A few minutes after switching off the line transformer No. 2 was switched off as a result of tripping of the ground fault and over-current protection. The grid was switched over by the Automatic Transfer Switch Equipment (ATSE) to supply from transformer TR1.

On the basis of this string of events a hypothesis may be formulated that in the initial line-to-earth short circuit subsequently developed to the two-phase-to-earth fault. Immediately after the damaged cable line’s switch off in the grid an unsuppressed (sustained) ferroresonance developed, which caused severe overload of the voltage transformers’ grounded primary windings, and consequently their damage. The voltage transformers rupture in the MSP substation’s 6 kV switchgear (a few minutes after disconnecting the damaged cable line) led to a short circuit of the buses in the measurement bay, and to actuation of the ATSE automatics, which changed the switchgear’s operating regime from normal to emergency (supply from transformer TR-1).

The presented analysis of simulation calculations results shows that the above hypothesis of the origins of the 6 kV grid failure is true.

3. COMPUTER MODEL OF THE ANALYSED 6 KV GRID

The model of the analysed 6 kV grid was developed in the simulation EMTP-ATP ElectroMagnetic Transients Program. Since the preliminary analysis of the MSP disturbance pointed to a ferroresonance phenomenon as the cause of the voltage transformers damage, therefore in the model’s development a non-linear dependence was taken into account of the voltage transformer core magnetizing currents on voltage. On the basis of the actual grid’s details and on the basis of prepared models of the voltage transformers, a model of the 6 kV system was developed (Fig. 2), consisting of:

• MSP• 6 kV cable lines outgoing from MSP• 6 kV switchgear of CHP power plant.

open circuit breaker

closed circuit breaker

open disconnector

oclosed disconnector

Legend

Rafał Tarko, Wiesław Nowak, Waldemar Szpyra, Mariusz Benesz, Andrzej Makuch / AGH University of Science and Technology in Kraków

77

Fig. 2. ATPDraw model of analysed 6 kV grid

Voltage measurement

Voltage measurement

Voltage measurement

Voltage measurement

Ferroresonance as a Source of Disturbances and Failures in Medium Voltage Distribution Grids

78

The mapped system components’ parameters were implemented on the basis of detailed plans of the power grid, catalogue cards of the devices installed, and laboratory measurements of the voltage transformer in the analyzed MV grid. Accomplishment of the analysis’ objective required accurate mapping of all grid elements, which might have affected the voltage and current waveforms in transient states, and thus were likely to cause the voltage transformer failure. The system components mapped in the model include in particular:

• 110 kV power system• 110/6 kV transformers installed in the MSP and mapped in BCTRAN procedure• cable lines that connect the individual substations supplied from the MSP and mapped in CABLE CON-

STANTS procedure• voltage transformers• ferroresonance suppression system.Based on the results of the 6 kV voltage transformers’ measurements a computer model was developed

as the basis for further analysis of the grid operation and simulation studies. For the purpose of the model development measurements were taken of the magnetization characteristics and short-circuit voltages of a 6 kV voltage transformer (of the same type as the transformers installed in the substation) with the following rated specification:

• primary winding rated voltage U1: 6000/√3V• secondary winding rated voltage U: 100/√3V• additional winding rated voltage U2n: 100/3 V• class: 0,5• rated power: 50 VA.Tab. 1 presents measurement results of a test of short circuit between pairs of windings. The voltage

transformer’s magnetization characteristics measured from the secondary winding is presented in Fig. 3.

Tab. 1. 6 kV voltage transformer short-circuit test results

Measurement winding terminals A-N

Measurement winding terminals a-n

Additional winding terminals da-dn

U, V I, A P , W cos

shorted powered open 5. 48 2.93 14.18 0.96

open powered shorted 6.88 3.36 22 0.88

shorted open powered 2.78 1.93 4.76 0.89

Fig. 3. Current-voltage transformer magnetization characteristics of 6 kV voltage transformer (measured on the secondary side)

The voltage transformer model diagram is presented in Fig. 4. The model consists of the following components:

• primary winding dissipation impedance ZH

0

1000

2000

3000

4000

5000

6000

0 20 40 60 80 100 120 140

I, m

A

U, V

Rafał Tarko, Wiesław Nowak, Waldemar Szpyra, Mariusz Benesz, Andrzej Makuch / AGH University of Science and Technology in Kraków

79

• secondary winding dissipation impedance Z’T transferred to the primary side• additional winding dissipation impedance Z’L transferred to the primary side• resistance RFe representing losses in voltage transformer core• nonlinear reactance Xm mapping magnetization characteristics• ideal transformers TI1, TI2.

Fig. 4. Voltage transformer model diagram Fig. 5. ATPDraw diagram of three voltage transformers’ three-phase system

The above components’ parameters determined on the basis of the measurements are presented in Tab. 1 and in Fig. 3. The ATPDraw diagram of three voltage transformers’ three-phase system is presented in Fig. 5.

The system consists of three VT units, in which were implemented the voltage transformer models presented in Fig. 4.

4. SIMULATION STUDIES OF ANALYSED GRID

The simulation results include current and voltage waveforms relevant to the operation of protective automation and the threat of ferroresonance to the 6 kV grid components. Because of activation (stimulation) during the short circuit of the earth fault as well as overcurrent protections, it was assumed that there was initially a one-phase to earth fault, which eventually evolved into a two-phase to earth fault. The disturbance condition, protection tripping at the MSP, and ferroresonance development were reproduced in the model system subject to the following assumptions:

• the first disturbance (L3 phase earth fault in bay 24) occurs in 15 ms after simulation start• the second disturbance (L2 phase earth fault and the subsequent phase-to-phase short-circuit in bay

24) occurs in 60 ms after simulation start• the circuit breaker in bay 24, where the double fault occurred, opens in 100 ms after simulation start.The resulting waveforms of phase voltages and zero sequence voltage at the 6 kV switchgear busbars in

the MSP substation are shown in Fig. 6. On the other hand in Fig. 7 the waveforms of the currents in the voltage transformers’ primary sides are presented.

VT2A

I U

VT2BI U

VT2CI U

OT2P

OT2K

U

Ferroresonance as a Source of Disturbances and Failures in Medium Voltage Distribution Grids

80

Fig. 6. Voltage waveforms after switching off the short circuit: a) UL1; b) UL2; c) UL3; d) 3U0

Fig. 7. Current waveforms in voltage transformer primary sides after switching off the short circuit: a) IL1; b) IL2; c) IL3

0,0 0,1 0,2 0,3 0,4 0,5[s]-10

-5

0

5

10[kV]

v:TR2B 0,0 0,1 0,2 0,3 0,4 0,5[s]

-10

-5

0

5

10[kV]

0,0 0,1 0,2 0,3 0,4 0,5[s]-9000

-4500

0

4500

9000[V]

0,0 0,1 0,2 0,3 0,4 0,5[s]-200

-100

0

100

200[V]

0,0 0,1 0,2 0,3 0,4 0,5[s]-10

-5

0

5

10[A]

0,0 0,1 0,2 0,3 0,4 0,5[s]-10

-5

0

5

10[A]

0,0 0,1 0,2 0,3 0,4 0,5[s]-10

-5

0

5

10[A]

a)

b)

c)

d)

a)

b)

c)

Rafał Tarko, Wiesław Nowak, Waldemar Szpyra, Mariusz Benesz, Andrzej Makuch / AGH University of Science and Technology in Kraków

81

After the short circuit switch off in bay 24 permanent ferroresonance developed. The occurrence of this phenomenon is evidenced by the appearance of the characteristic phase voltage waveforms (Fig. 6a÷6c) and of the zero sequence voltage (Fig. 6d). This is accompanied by a large increase in the voltage transformer primary side currents (Fig. 7). After the ferroresonance occurrence the peak currents reach 10 A. Such a large current is undoubtedly a serious threat to the voltage transformers and could cause damage to them.

Ferroresonance suppression capabilities in the analyzed grid were checked by adding an additional resistance to the voltage transformer additional windings connected in open triangle. The additional resistance was connected in 500 ms after simulation start (Fig. 8). Variants were analysed that assumed adding the following resistances: 5 Ω, 10 Ω, 20 Ω and 50 Ω.

Fig. 8. Voltage waveforms 3U0 after adding the resistor: a) R = 5 Ω; b) R = 10 Ω; c) R = 20 Ω; d) R = 50 Ω

In the analyzed range of added suppression resistances the possibility of the ferroresonance phenomenon suppression in the analyzed grid is revealed. The added resistance’s value, however, affects the time after which ferroresonance is suppressed – the higher the resistance, the later the suppression, and at significant values (R > 50 Ω) ferroresonance may not be suppressed at all. The analysis shows that the optimum added resistance is 10 Ω (suppression time ca. 1 s).

0,5 0,7 0,9 1,1 1,3 1,5[s]-100

-50

0

50

100[V]

0,5 0,7 0,9 1,1 1,3 1,5[s]-200

-100

0

100

200[V]

0,5 0,7 0,9 1,1 1,3 1,5[s]-100

-50

0

50

100[V]

0,5 0,7 0,9 1,1 1,3 1,5[s]-100

-50

0

50

100[V]

a)

b)

c)

d)

Ferroresonance as a Source of Disturbances and Failures in Medium Voltage Distribution Grids

82

REFERENCES

5. SUMMARY

Ferroresonance phenomena involving voltage transformers are most common in grids with an insulated neutral point, so the voltage transformers operation in medium voltage distribution grids entails the risk of their damage due to increased current in the primary winding.

The 6 kV grid model developed for the study allowed the simulation testing and analysis, that led to the following conclusions:

• ferroresonance can develop resulting from such disturbance conditions as short-circuits,• ferroresonance can be persistent, eventually resulting in voltage transformers’ damage,• ferroresonance may be suppressed by the use of an appropriate device.Effective ferroresonance suppression with a suppression resistor requires selection of a very small

resistor. Such resistance is often too small from the standpoint of the required voltage transformer immunity to long-term ground fault in the grid. Therefore, in practice resistors in the range of 20 Ωare used that ensure ferroresonance suppression in most typical conditions, but are not 100% effective. In order to solve the problem some manufacturers offer suppression devices, whereby conventional transistors are replaced with systems the resistances of which actively adjust to the actual operating conditions.

Such devices operate in the following way [3]: when zero sequence voltage is small (resulting from asymmetry in normal grid conditions) the device’s resistance is very high. When zero sequence voltage appears in excess of the device’s insensitive zone, the resistance falls to a level that effectively suppresses the ferroresonance condition. When a zero sequence voltage in the open triangle circuit persists for a long time, the device automatically switches over to the high-resistance regime without posing undue burden on voltage transformers. When the cause of asymmetry disappears, the device automatically returns to its initial state.

1. Irvani M.R. et al., Modeling and analysis guidelines for slow transients – Part III, The study of ferroresonance, IEEE Trans. on PWRD, 2000, vol. 15, no 1, pp. 255–265

2. Ben-Tal A., Kirk V., Wake G., Banded chaos in power systems, IEEE Trans. on PWRD, 2001, vol. 16, no 1, pp. 105–110.

3. Piasecki W., Florkowski M., Fulczyk M., Mahonen P. , Luto M., Nowak W., Mitigating Ferroresonance in Voltage Trans-formers in Ungrounded MV Networks, IEEE Trans. on PWRD, 2007, vol. 22, no 4, pp. 2362–2369.

4. Moskwa S., Nowak W., Tarko R., Modelowanie i analiza układu sieci średniego napięcia dla oceny warunków i skutków występowania ferrorezonansu oraz sposobów jego eliminacji, Zeszyty Naukowe Wydziału Elektrotechniki i Automatyki Politechniki Gdańskiej, 2009, No. 26, pp. 101–104.

Rafał Tarko, Wiesław Nowak, Waldemar Szpyra, Mariusz Benesz, Andrzej Makuch / AGH University of Science and Technology in Kraków