Aalborg Universitet

Methods to Minimize Zero-Missing Phenomenon

da Silva, Filipe Miguel Faria; Bak, Claus Leth; Gudmundsdottir, Unnur Stella; Wiechowski,
W.; Knardrupgård, M. R.
Published in:
I E E E Transactions on Power Delivery

DOI (link to publication from Publisher):

10.1109/TPWRD.2010.2045010

Publication date:
2010

Document Version
Accepted author manuscript, peer reviewed version

Link to publication from Aalborg University

Citation for published version (APA):

da Silva, F. M. F., Bak, C. L., Gudmundsdottir, U. S., Wiechowski, W., & Knardrupgård, M. R. (2010). Methods to
Minimize Zero-Missing Phenomenon. I E E E Transactions on Power Delivery, 25(4), 2923-2930.
https://doi.org/10.1109/TPWRD.2010.2045010

General rights
Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners
and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

? Users may download and print one copy of any publication from the public portal for the purpose of private study or research.
? You may not further distribute the material or use it for any profit-making activity or commercial gain
? You may freely distribute the URL identifying the publication in the public portal ?
Take down policy
If you believe that this document breaches copyright please contact us at vbn@aub.aau.dk providing details, and we will remove access to
the work immediately and investigate your claim.

Downloaded from vbn.aau.dk on: junio 09, 2021

Page 1 of 8 IEEE PES Transactions on Power Delivery
TPWRD-00683-2009 1

1
2
3
4 Methods to Minimize Zero-Missing
5
6
7
Phenomenon
8
9 F. Faria da Silva, C. L. Bak, U. S. Gudmundsdottir, W. Wiechowski and M. R. Knardrupgård
10
11
12 this happens, the current passing through the circuit breaker
13 Abstract—With the increasing use of high-voltage AC cables at does not cross zero until the DC component becomes smaller
14 transmission levels, phenomena such as current zero-missing start than the AC component.
15 to appear more often in transmission systems. During its energization the cable is unloaded, and the
16 Zero-missing phenomenon can occur when energizing cable
resistance of the system (cable+shunt reactor(s)) is very small.
17 lines with shunt reactors. This may considerably delay the
18 opening of the circuit breaker, leaving the system unprotected As a result, the DC component may take several seconds to be
19 and vulnerable to failures. damped, period during which the circuit breaker cannot be
20 Methods to prevent zero-missing phenomenon are still being opened.
21 studied and compared in order to identify effective This paper is an extension of [3], and for a 50 Hz power
22 countermeasures. This paper contributes to these efforts, by frequency it describes zero-missing phenomenon and presents
presenting several countermeasures that can be applied to reduce countermeasures that can be used to avoid it. The
23
the hazards of zero-missing phenomenon.
24 countermeasures are divided into two types: Cases where the
The authors discovered that this phenomenon can be
25 eliminated, merely by using an extra circuit breaker or a pre- shunt reactor is directly connected to the cable and cases
26 insertion resistor. where the shunt reactor is connected to the cable via a circuit
27 breaker.
28 Index Terms—Power Cables, AC Circuit Breaker, Shunt
29 Reactor, Switching Transients, Level-crossing problems II. ZERO-MISSING PHENOMENON AND SWITCHING
30 OVERVOLTAGES
31 I. INTRODUCTION

Z
32 ERO-MISSING phenomenon is defined as an AC current A. Basic Circuit: Inductor in Parallel with a Capacitor
33 An easy way of understand zero-missing phenomenon is by
not crossing zero value during several cycles. If a current
34
does not cross zero value it is not possible to open the circuit analysing an inductor in parallel with a capacitor of equal
35
breaker without risk of damage, except if the circuit breaker is impedance. In this situation the currents in the capacitor and
36
designed to interrupt DC currents or open at a non-zero current inductor have equal amplitude and are in phase opposition.
37
38 value [1][2]. The current in the inductor can also have a DC component,
39 Because of the large capacitive reactive power of HVAC whose value depends on the voltage at moment of connection.
40 cables, shunt reactors are needed for power compensation. For If the inductor is connected at peak voltage the current at
41 unloaded cable systems, the shunt reactor current is almost in t(0+) is zero, if it is connected for zero voltage the current has a
42 phase opposition to the current in the cable, reducing the peak value at t(0+). As the current in the inductor must
43 amplitude of the resultant AC component through the circuit maintain its continuity, and it was zero at t(0-), if the voltage is
44 breaker. As the current in the shunt reactor has a transient DC not at a peak value during connection, a DC component will
45 component, the resulting current in the circuit breaker may appear in the inductor current to maintain its continuity. The
46 have a DC component larger than its AC component. When DC component is equal to minus the value of the AC
47 component at t(0-) [4].
48 Manuscript received September 07, 2009. If there is no resistance in the system, the DC component is
49 This work was supported in part by the Danish Transmission System not damped and it will be maintained infinitely. In reality there
50 Operator, Energinet.dk. is always some resistance and the DC component disappears
51 F. F. da Silva is a PhD student at the Institute of Energy Technology,
Aalborg University, 9220 Aalborg, Denmark (e-mail:ffs@iet.aau.dk). after some time.
52
C. L. Bak is with the Institute of Energy Technology, Aalborg University, Fig. 1 shows an inductor in series with a resistor, both of
53 9220 Aalborg, Denmark (e-mail: clb@iet.aau.dk). them in parallel with a capacitor. The resistance is 100 times
54 U. S. Gudmundsdottir is a PhD student at the Institute of Energy
Technology, Aalborg University, 9220 Aalborg, Denmark (e-mail: smaller than the inductor reactance, which is equal to the
55
56 usg@iet.aau.dk). capacitor reactance. Fig. 2 shows a simulation of Fig. 1. The
W. Wiechowski is a Senior System Analyst at the Planning Department of circuit breaker closes when the voltage is crossing zero, and
57 Energinet.dk (e-mail: wwi@energinet.dk)
58 M. R. Knardrupgård is with the Planning Department of Energinet.dk (e-
therefore the DC component in the inductor is maximum. The
59 mail: mra@energinet.dk) inductive and capacitive AC components cancel out (IL and IC
60 .
 IEEE PES Transactions on Power Delivery Page 2 of 8
TPWRD-00683-2009 2

1
have equal amplitude and are in phase opposition) and the
2 The system of Fig. 3 is described by (1). The circuit breaker
current I1 contains only the decaying DC component
3 current I1, equal to IS+IC+I3, is obtained from (1) and shown in
4 (2).
5
 dI
6
 V cos ( ωt ) = Ls s + Rs I s
7 dt

8  1
 V cos (ωt ) = ∫ I c dt (1)
9 C

10  dI 3 1
11 Fig. 1 Equivalent scheme of an inductor in series with a resistor, both in V cos (ωt ) = RI 3 + L dt + C ∫ I 3 dt
12 parallel with a capacitor 
40
13 30
 
R
14 V  ωL −t s
20
I1 ( t ) = ⋅ cos  ω t − arctan  s  +
  sAI ⋅ e Ls
+
Current [kA]

15 
Rs2 + (ω Ls )
2
16 10   Rs 
17 0
  1 
18 -10
V   ω L − ωC 
19 -20 + ⋅ cos  ω t − arctan   + (2)
2
  
0 0.02 0.04 0.06 0.08 0.1
 1   R
20 Time [s]
R2 + ω L − 
Fig. 2 Current in the inductor (IL, dashed line), in the capacitor (IC, dotted    
21 line) and the sum (I1=IL+IC, solid line)  ωC 
22
−
 R
 R  
2
 π
23 t
⋅ cos  ω0 2 − 
The behaviour of a system consisting of a shunt reactor and + I 3 aux e  t  + V ⋅ cos  ω t +  ⋅ ωC
2L
24
a cable is not very different from the one depicted in Fig. 1.   2 L    2
25 
The shunt reactor can be modelled as an inductor in series with Where IsA is the initial value of the DC component and is
26
27 a resistor, and the cable is mainly a capacitive shunt element calculated by (3). I3aux is related with the energization of the
28 [5]. There are, however differences between the behaviour of a inductor and capacitor and is calculated by (4). Both depend
29 simple RLC circuit and a physical cable/shunt reactor system, on the moment when the shunt reactor is connected.
30 as for instance switching overvoltages.
V   ωL 
31 Switching overvoltages occur due to the charging of the I sA = − ⋅ cos  ωt − arctan  s   (3)
 
Rs2 + (ω Ls )
2
32 cable's capacitance and energy oscillation between the cable's   Rs  
33 capacitance and inductance [6]. As for zero-missing
  1 
34 phenomenon, the value of switching overvoltages depends on
V   ω L − ωC  
35 the voltage value when connecting the cable/reactor system. I 3 aux = − ⋅ cos  − arctan    (4)
2
36 However, to avoid zero-missing phenomenon the connection  1   

R  
R2 +  ω L − 
37     
should be made when the voltage is at its peak, whereas the  ωC 
38 opposite applies when it comes to avoiding switching The equations can be confirmed using EMTDC/PSCAD to
39 overvoltages [2][7].
40 simulate the system. Fig. 4 shows the current in the circuit
According to [8], when the shunt reactor is directly breaker using both EMTDC/PSCAD and (2) for a single phase
41
connected to the cable it is necessary to choose between cable, and a 400 kV voltage source. The parameters used in
42
avoiding either zero-missing phenomenon or switching the simulation are based on a 400 kV, 50 km cable from
43
overvoltages. This paper will show that it is possible to avoid SAGEM [9]: C=3.929 µF; L=25.7 mH; R=15.05 Ω; for the
44
45 both by applying specific countermeasures. shunt reactor the parameters are: LS=1.29 H; RS=0.3003 Ω.
46 B. System using Cable's Pi-model
47
For analysis purposes the shunt reactor had been modeled as
48
an inductor in series with a resistor and the cable is
49
represented by its equivalent pi-model (see Fig. 3).
50
51
52
53
54
55 Fig. 4 Current I1 during 0.05s for 100% reactive power compensation
56
57 In order to have more accurate results for the simulations
58 Fig. 3 Equivalent scheme of a shunt reactor and a cable: V1-Voltage Source; instead of using a pi-model the cable will be simulated by a
RS-Shunt reactor resistor; LS-Shunt reactor inductor; R-Cable's series resistor;
59
L-Cable's series inductor; C-Cable's shunt capacitor
PSCAD-EMTDC frequency dependent phase model, as it is
60
Page 3 of 8 IEEE PES Transactions on Power Delivery
TPWRD-00683-2009 3

1
currently the most precise method to simulate cables [10]. The 10
2
electrical cable parameters are equal to the ones previous
3 5
mentioned for the pi-model.

Current [kA]
4
5 As expected there are small differences between the two 0

6 simulations, mainly in the moments following the connection

-5
7 of the cable/shunt reactor.
8 C. Circuit Breaker
-10
0 0.02 0.04 0.06 0.08 0.1
Time [s]
9
Two different circuit breakers types can be used. A first type
600
10
11 that closes all the phases at the same time, and a second type 400

called single-pole mode, which closes the three phases at 200

Voltage [kV]
12
13 different moments [11]. 0

14 A circuit breaker operating in single-pole mode is similar to -200

15 three different single-phase circuit breakers operating -400

16 independently. In Fig. 5 an example is shown of the closing of -600

0 0.02 0.04 0.06 0.08 0.1
a circuit breaker operating in this mode, phase R closes at 0
Time [s]
17
Fig. 7 Current I1 and voltage at the sending end of the cable in a three-phase
18 ms, phase T at 3.3 ms and phase S 6.7 ms after phase R. system for a circuit breaker closing the three phases at the same time (phase
19 R: dotted line; phase S: solid line; phase T: dashed line)
20
21 III. COUNTERMEASURES WHEN THE SHUNT REACTOR IS
22 CONNECTED VIA A CIRCUIT BREAKER
23 Depending on the system load, it may be necessary to
24
connect and disconnect shunt reactors [12]. In such situations
25
the shunt reactors will be connected to the cable via a circuit
26
27 breaker instead of being directly connected. In this situation it
28 Fig. 5 Example of a single-pole mode operation [11] is possible to eliminate/reduce the initial DC component
29 simply by controlling the circuit breaker closing.
30 The type of circuit breaker used has an influence on the A. Shunt Reactor Connection for a Voltage Peak
31 energizing transient. Fig. 6 and Fig. 7 show the differences
As explained before, the DC component is zero for a shunt
32 between the two modes. In Fig. 6 all the phases close when the
reactor energized at voltage peak. For direct connected shunt
33 voltage is zero in the respective phase, therefore the DC
reactors, the cable and shunt reactor are energized
34 component is maximum and the switching transient minimum
simultaneously, and it is not possible to use this method
35 in all the phases. In Fig. 7 all the three phases are closed at the
36 because of switching overvoltage. But if the shunt reactors are
same time, with phase R having a maximum DC component
37 connected via circuit breakers, it is possible to connect the
and the other phases having a lower DC components but also
38 cable when the voltage is zero, and after a short period connect
larger switching overvoltages.
39 1.5 the shunt reactors when the voltage is at a peak value.
40 1 For this countermeasure to be completely effective, the
41 0.5 circuit breaker associated to the shunt reactor must operate in
Current [kA]

42 0 single-pole mode, or it would not be possible to connect all the

43 phases at peak voltage.
-0.5
44
45 -1 B. Disconnecting the Shunt Reactor before the Cable
46 -1.5
0 0.01 0.02
Time [s]
0.03 0.04 0.05 Zero-missing phenomenon only becomes problematic if it is
47 400 necessary to disconnect the cable shortly after its energization.
48 So instead of eliminating zero-missing phenomenon, a possible
200
49 countermeasure could be to find a method of opening the
Voltage [kV]

50 0 circuit breaker when desired.

51 Even for a maximum DC component, the current on the
52 -200
shunt reactor (IS) always has a zero crossing. It is therefore
53
always possible to disconnect the shunt reactor from the cable.
54 -400
0 0.01 0.02 0.03 0.04 0.05
Time [s] When the shunt reactor is disconnected both the DC
55 Fig. 6 Current I1 and voltage at the sending end of the cable in a three-phase component and inductive AC component disappear. As the
56 system for a circuit breaker operating in single-pole mode (phase R: dotted
57 line; phase S: solid line; phase T: dashed line) cable's circuit breaker current now only has an AC component
58 the circuit breaker can be disconnect when desired.
59 One advantage of this countermeasure is that it does not
60
 IEEE PES Transactions on Power Delivery Page 4 of 8
TPWRD-00683-2009 4

1
require a circuit breaker operating in single-pole mode. system, and only the DC component of the second shunt
2
Fig. 8 shows an example of countermeasure effectiveness. reactor remains. This DC component value is half of what it
3
4 The shunt reactor circuit breaker opens at 0.05 s and the cable would have been if only one shunt reactor had been used,
5 circuit breaker at 0.08 s. As a circuit breaker can only open whereas the AC component continues to have the same
6 when the current is crossing zero, a delay is required between amplitude, and as such the current crosses zero sooner.
7 the two opening orders. But as can be observed, after the 10

8 disconnection of the shunt reactor there is only AC current 5

9 without DC offset, and therefore the cable can be disconnected

Current [kA]
10 without any risks for the circuit breaker. 0
1
11
-5
12 0.5
13
Current [kA]

-10
0 5 10 15 20 25 30 35
14 0 Time [ms]

15 Fig. 9 Currents for a maximum DC component for a shunt reactor

-0.5 compensation half of the cable's reactive power: Solid line: Current in the
16 cable; Dotted line: Current in shunt reactor; Dashed line: Current in the
17 -1
circuit breaker (I1)
0.04 0.05 0.06 0.07 0.08 0.09 0.1
18 Time [s]
19 Fig. 8 Current I1 when the shunt reactor is disconnected before the cable Fig. 10 and Fig. 11 show the current I1 before and after the
20 connection of the first and the second shunt reactor. Note that
21 as expected the DC current in Fig. 11 is only half of the
C. Use of Several Shunt Reactors
22 indicated in Fig. 8, as was expected.
23 If more than one shunt reactor is used to compensate for the 1

24 reactive power, it is possible to synchronize their energization

25 thus reducing the risks of zero-missing phenomenon occurring. 0.5

26 For zero-missing phenomenon to occur, a shunt reactor has to Current [kA]

0
27 compensate more than 50% of the reactive power generated by
28 the cable. When a cable is compensated by several shunt -0.5

29 reactors, all with the same compensation level, it is necessary

30 to have more than one connected to have zero-missing -1
0.08 0.1 0.12
Time [s]
0.14 0.16 0.18

31 phenomenon (unless if it is necessary to compensate more than Fig. 10 Current I1 in the circuit breaker after the connection of the first shunt
32 100% of the generated reactive power). reactor (phase R: dotted line; phase S: solid line; phase T: dashed line)
33 This countermeasure is demonstrated for a system with two
34 shunt reactors, each compensating half of the reactive power.
0.5

35
The reasoning is similar for systems with more shunt reactors.
36
This countermeasure involves connecting the second shunt
Current [kA]

37
reactor after the DC component caused by the first one has 0
38
39 been completely damped.
40 When a shunt reactor compensates 50% of a cable's reactive
41 power, the amplitude of its AC component is half of the -0.5
13.96 13.98 14 14.02 14.04 14.06
Time [s]
42 amplitude of the cable's AC component. As these two currents Fig. 11 Current I1 in the circuit breaker after the connection of the second
43 are in phase opposition, the resulting current I1 has an AC shunt reactor (phase R: dotted line; phase S: solid line; phase T: dashed line)
44 component whose amplitude is equal to that of the shunt
45 reactor's current. The initial DC component is never larger This countermeasure does not eliminate zero-missing
46 than the amplitude of the shunt reactor AC component, and in phenomenon, it merely reduces its duration. It is therefore still
47 the worst case scenario the minimum peak value of the current not possible to open the cable circuit breaker in the moments
48 I1 is therefore zero. following the energizing of the second circuit breaker. On the
49 Fig. 9 gives an example of the currents in the shunt reactor, other hand when a cable is energized, problems are more likely
50 cable and circuit breaker, for a situation where the initial DC to occur immediately after the connection of the cable. As this
51 component is maximum. countermeasure results in no zero-missing phenomenon until
52 After the DC component caused by the first shunt reactor is the second shunt reactor is connected, the risks to the system
53 are reduced.
damped, the second shunt reactor is connected. Because there
54
already is one shunt reactor connected, the amplitude of I1
55 IV. COUNTERMEASURES FOR A SHUNT REACTOR DIRECTLY
56 prior to the connection of the second shunt reactor is half that
of the cable's current and equal to the current in the shunt CONNECTED TO THE CABLE
57
58 reactors (see Fig. 9). Therefore, when the second shunt reactor The drawback of the countermeasures presented in the
59 is connected its AC current cancels out the AC current of the previous section is that the shunt reactor must be connected to
60
Page 5 of 8 IEEE PES Transactions on Power Delivery
TPWRD-00683-2009 5

1
the cable via a circuit breaker. As a circuit breaker is an depend on the connection moment and are unknown. For
2
expensive piece of switchgear, it will only be installed if it is 100% reactive power compensation, the AC components of the
3
4 absolutely necessary. A common design choice is to connect shunt reactor and cable's currents cancel each other out, and at
5 the shunt reactor direct to the cable and energize both cable the moment of connection the current I1 is equal to ISDC,
6 and shunt reactor at the same time. whereas both should ideally be zero after 10ms.
7 This section presents two countermeasures that can be used Considering that the current I1 decreases linearly (this is an
8 to avoid zero-missing phenomenon when the shunt reactor is approximation, but as Rp is large the error is small), and
9 directly connected to the cable. neglecting RS (which is much smaller than Rp), (7) can be
10 simplified to (8), and the value of Rp is calculated by (10).
A. Pre-insertion Resistor 2
11  I ( 0) 
12 A pre-insertion resistor consists of resistor blocks connected W = 0.01R p  1  (8)
13 in parallel with the circuit breaker’s breaking chamber, and  2 
14 closes the circuit 8-12 ms before the arcing contacts [11] (in 2
 I ( 0) 
2
1 1 1
 = Ls ( I s ) ⇔ 0.01R p   = Ls
this article the time considered is 10 ms). DC 2
15 0.01R p  1 (9)
16 A pre-insertion resistor can damp the entire DC component  2  2 2 2
17 in 10ms, thus eliminating zero-missing phenomenon. But for 2 Ls
18 this happen the pre-insertion resistor value has to be precise. Rp = (10)
0.01
19 The inclusion of this resistor changes the system layout, see Because of the simplifications this method is not always
20 Fig. 12 and compare with Fig. 3. Just this little change is
21 accurate. If the DC component is maximum the error can be
sufficient to change the system behaviour so that it is no longer disregarded, but if the DC component is smaller, the error
22 described by (1), but by (5).
23 increases.
24 The use of differential equations allows a more accurate
25 calculation of the pre-insertion resistor value, but an iterative
26 process is required to calculate the value of Rp. Part of the
27 equation solution is presented in appendix.
28 To perform the iterative process, a small program was
29 written in Matlab. The program increases Rp, until it reaches a
30
Fig. 12 Equivalent scheme of the shunt reactor and the cable when using a
value at which the DC component is damped in 10ms.
31 pre-insertion resistor (RP) To verify that the DC component is damped, the peak value
32 of IS is calculated 10 ms after connection. For that value be
33 Unlike the differential equation in (1), (5) is not analytically equal to the amplitude of the AC component, the DC
34 component must be equal to zero. So when the calculated
solvable in the time domain [13]. It is therefore necessary to
35
do a Laplace transformation in order to solve it. value is equal to (11) plus a small tolerance the iterative
36
 dI process stops.
37 V2 = Ls s + Rs I s
38  dt V2
 I s peak = (11)
Rs + (ω Ls )
2
39  1 2
V2 = ∫ I c dt
40  C
41  (5)
dI 1
42 V2 = R ⋅ I 3 + L 3 + ∫ I 3 dt
43  dt C
 I1 = I 2 + I 3 + I c
44 
45  2 V1 cos(ωt ) − R p ⋅ I1
V =
46
It is though possible to introduce some simplifications in
47
order to obtain a first approximation of the pre-insertion
48
49 resistor value. This is done by calculating the energy that the
50 pre-insertion resistor should dissipate (6).
1
W = Ls ( I sDC )
51 2
(6)
52 2
53 The energy dissipated in the pre-insertion resistor is
54 calculated by the integral in (7), whose limits are the time
55 during which the pre-insertion resistor is connected.
56 0.01
57 W = ∫ Pdt ⇔ ∫ R p I12 dt (7)
58 0

59 The objective is to calculate Rp, where both I1 and ISDC Fig. 13 - Flowchart of the iterative process for the calculation of the pre-
insertion resistor value
60
 IEEE PES Transactions on Power Delivery Page 6 of 8
TPWRD-00683-2009 6

1
value of the DC component after 10 ms for different resistor
2
The value of the pre-insertion resistor depends on the initial values when the cable and the shunt reactor are connected for
3
4 value of the DC component. As this value depends on the zero voltage.
5 connection moment, it was decided to solve the equations for 1.5

1
6 the worst case scenario, maximum DC component. For that 0.5
case the calculated value of Rp is ideal, whereas for the other

Current [kA]
7 0

8 cases the error is small. -0.5

9 Solving the differential equations, using the system -1

10 described before, it is obtained 295 Ω as ideal value for the -1.5

11 pre-insertion resistor. -2
0 0.01 0.02 0.03 0.04 0.05
12 Fig. 14 and Fig. 15 show two simulations of a circuit 400
Time [s]

13 breaker equipped with a pre-insertion resistor. In Fig. 14 the

14 circuit breaker operates in single-pole mode and closes each 200

Voltage [kV]
15 phase for the maximum DC component in the phase. In Fig. 15
16 the circuit breaker closes the three phases at the same time for
0

17 maximum DC component in just one of the phases. -200

18 In Fig. 14 all the three phases are zero during connection. In
19
this situation the pre-insertion resistor completely damps the -400
0 0.01 0.02 0.03 0.04 0.05
20 Title [s]
DC component in 10 ms and there is no zero-missing Fig. 15 Current in the circuit breaker and voltage on the sending end of the
21
22 phenomenon. As the connection in all the phases is made for cable for a situation of maximum switching overvoltage, using a pre-insertion
resistor (phase R: dotted line; phase S: solid line; phase T: dashed line)
23 zero voltage, the switching overvoltage in the three phases is
24 minimum.
The curve in Fig. 16 is non-linear, and for pre-insertion
25 For the situation depicted in Fig. 15 it is not possible to
resistor values close to the ideal, the DC component is very
26 have zero voltage in the three phases when the circuit breaker
small, but for larger differences there is still zero-missing
27 closes, and therefore it is neither possible to eliminate the DC
phenomenon during long periods of time.
28 component nor completely minimize the switching
If, for instance the value of the pre-insertion resistor value
29 overvoltages. In one of the phases though there is neither a DC
was calculated using the energy equations (10) instead of the
30 component nor switching overvoltage; this phase is the one
differential equations, the initial DC current would be about 51
31 where the voltage is zero at the connection moment. When
32 A, which is 16 times lower than the value of the initial DC
comparing Fig. 15 with Fig. 7, it can be seen how the DC
33 component when no pre-insertion resistor is used (800 A). It
component in the other two phases is smaller when the pre-
34 can therefore be concluded that the method applied in the
insertion resistor is used (80 A instead of 400 A) and that there
35 energy equations can be used to obtain a first approximation of
is no switching overvoltage (maximum 1 pu or 326 kV instead
36 the final resistor value.
of 1.58 pu or 516 kV). 800
37
DC component after 10ms [A]

1.5
38 1 600
39 0.5
Current [kA]

40 400
0
41
-0.5 200
42
43 -1
0
0 50 100 150 200 250 300
44 -1.5
0 0.01 0.02 0.03 0.04 0.05 Pre-insertion resistor [Ohm]
Time [s]
45 400
Fig. 16 Initial value of the DC component after bypassing the pre-insertion
46 resistor for different Rp values, for a phase closing when the voltage is zero
47 200
B. Cable Energized from Both Ends
Voltage [kV]

48
0 In the previous countermeasures it was considered that the
49
cable is energized only at one end while the other end was
50
51
-200
open. It is also possible to energize the cable from both ends
52 -400
simultaneously.
When the cable is energized from one end only, there is a 1
0 0.01 0.02 0.03 0.04 0.05
Time [s]
53
Fig. 14 Simulation of cable energizing, for a circuit breaker operating in pu voltage drop between the cable and the ground, and almost
54 single-pole mode and using pre-insertion resistor (phase R: dotted line; phase
55 S: solid line; phase T: dashed line) all the voltage drops are due to the cable capacitance to the
56 ground (see Fig. 3).
57 Even if the value of the resistor is not ideal, the DC When the cable is energized from both ends, the cable
58 component of the shunt reactor current is always reduced, and resistance and inductance become more relevant, and part of
59 it will be lower than a given maximum. Fig. 16 shows the the reactive power is compensated by the grid instead of being
60
Page 7 of 8 IEEE PES Transactions on Power Delivery
TPWRD-00683-2009 7

1
solely compensated by the shunt reactor. impedance, cable length, etc….
2
Therefore the current in the cable is no longer fully As an example Fig. 19 and Fig. 20 show the minimum
3
4 cancelled out by the shunt reactor current, and there is an AC values of the difference between the phase angles and the
5 component in the current going through the circuit breaker. amplitudes required to prevent zero-missing phenomenon. For
6 Fig. 17 shows a simplified model for a cable being instance, a 40 km cable should have an energising voltage with
7 energized from both ends. By Ohm's law the current IC2 only a 0.0156 pu amplitude difference between the two ends or a 2º
8 depends on V2, whereas the current I3 depends more on the phase angle difference.
9 cable resistance and inductance. If, however the cable was 3.5

10 energized from one end only, it would be the cable capacitance 3

2.5

Phase Angle [º]

11 to define most of the current I3 value.
2
12 The current I1 has the same DC component as before, but 1.5
13 now it also has an AC component, reducing the minimum 1
14 value of I1. The current in the second generator (If) has an AC 0.5
15 amplitude that is almost equal to the one of I1 but without the 0
0 10 20 30 40 50
16 DC component (the shunt reactor is in the other cable's end). Length [km]

17 Fig. 19 Minimum difference between phase angles to prevent zero-missing

Therefore it is possible to open the circuit breaker connected phenomenon for different cable lengths
18 to the second generator at any time.
19
In real cables the capacitance is spread along the cable, and 1.025
20
it is not possible to do a so simplistic analysis, but the
21 1.0125
reasoning is similar.

Amplitude [pu]
22
1
23
24 0.9875
25
26 0.975
0 10 20 30 40 50
Length [km]
27
Fig. 20 Minimum difference between the amplitudes to prevent zero-missing
28 phenomenon for different cable lengths (solid line: V2>1pu; dashed line;
29 V2<1pu)
30 Fig. 17 Model for a cable energized from both ends
31 V. CONCLUSIONS
32 Fig. 18 shows the current I1 for a voltage that is the same at
Zero-missing phenomenon is a non-desirable effect that can
33 both terminals terminals. I1 is shown both for cable
occur when energizing cable lines with shunt reactors. The
34 energization from one end and both ends, respectively.
phenomenon can lead to the degradation or even destruction of
35 It can be observed that the DC component is equal in both
the cable circuit breaker. With the increasing use of long high-
36 situations, but for the cable energised from both ends there is
37 voltage cables, this problem becomes more relevant, and
also an AC component whose amplitude is about half that of
38 countermeasures for it are necessary.
the cable's AC current.
39 1.5
This paper presents several countermeasures that can be
Phase R (One End)

40 Phase
Phase
S (One End)
T (One End)
applied for different conditions/systems, reducing/eliminating
1
41
Phase
Phase
R (Two Ends)
S (Two Ends) zero-missing phenomenon. An analysis of all the
Current [kA]

Phase T (Two Ends)

42 0.5
countermeasures proposes that two of them be considered as
43 0 the most effective:
44 -0.5 • Connect the shunt reactor when the voltage is at a
45 peak value (when the shunt reactor is connected to
-1
46 0.03 0.04 0.05 0.06 0.07
Time [s]
0.08 0.09 0.1
the cable through a circuit breaker);
47 Fig. 18 Current in the circuit breaker when the cable is energized from both • Use of a pre-insertion resistor (when the shunt
48 ends and the voltage is equal in both terminals (oscillating lines) or just from
reactor is connected direct to the cable);
49 one end (flat lines)
If correctly applied these two countermeasures completely
50
Zero-missing phenomenon does not disappear by energizing eliminate zero-missing phenomenon, something that it is not
51
52 the cable from both ends. But, if the voltage is not the same in granted when the others countermeasures are applied.
53 the two cable ends, the AC component is larger and at a point As the system parameters change from case to case, it is
54 it becomes even larger than the DC component. The question always necessary to analyse the system to see which
55 is: What should the minimum difference between the sending countermeasure should be applied.
56 and receiving end voltages (amplitude and/or phase angle) be
57 in order to prevent zero-missing phenomenon?
58 There is no definitive answer to this question, since it
59 depends on the system parameters: Voltage level, cable
60
 IEEE PES Transactions on Power Delivery Page 8 of 8
TPWRD-00683-2009 8

1
APPENDIX BIOGRAPHIES
2
3  A B (1) + B ( 2 )  Filipe Faria da Silva was born in Portugal in 1985
4 V1 = V2  +  (12) and received his MSc in Electrical and Computers
N N  Engineering in 2008 from Instituto Superior Técnico
5 (IST), Portugal.
6 He is currently employed with the Danish TSO
7 A = s 5 LLs Cω 2 + s 4 ( Ls RCω 2 + LCRsω 2 ) + (Energinet.dk) and doing a PhD at the Institute of
Energy Technology of Aalborg University, where he
8
9 s 3 ( Lsω 2 + RCRsω 2 + LLs Cω 4 ) + studies high-voltage transmissions systems with
(13) underground cables.
10 + s 2 ( Ls RCω 4 + LCRsω 4 + Rsω 2 ) +
11 Claus Leth Bak was born in Århus in Denmark, on
12 +s ( Lsω 4 + RCRsω 4 ) + Rsω 4 April 13, 1965. He studied at the Engineering
College in Århus, where he received the B.Sc. with
13 honors in Electrical Power Engineering in 1992. He
14 pursued the M.Sc. in Electrical Power Engineering
B (1) = s 4 LCω 2 + s 3 RCω 2 + with specialization in High Voltage Engineering at
15 (14)
16 + s 2 (ω 2 + LCω 4 ) + sRCω 4 + ω 4 the Institute of Energy Technology (IET) at Aalborg
University (AAU), which he received in 1994. After
17 his studies he worked with Electric power
18 transmission and substations with specializations
19 B ( 2 ) = s 6 LC 2 Lsω 2 + s 5 ( LC 2 Rsω 2 + RLs C 2ω 2 ) within the area of power system protection at the NV Net transmission
company. In 1999 he got employed as an assistant professor at IET-AAU,
+ s 4 ( RRs C 2ω 2 + 2CLsω 2 + LLs C 2ω 4 ) +
20
where he is holding an associate professor position today. His main research
21 (15) areas include corona phenomena on overhead lines, power system transient
22 + s 3 ( 2CRsω 2 + LRs C 2ω 4 + RLs C 2ω 4 ) + simulations and power system protection. He is the author/coauthor of app.
23 30 publications and IEEE Senior Member.
24 + s 2 ( RRs C 2ω 4 + 2CLsω 4 ) + s 2 Rs Cω 4
Unnur Stella Gudmundsdottir was born in
25
Reykjavik in Iceland, in 1980. She received her B.Sc.
26
27 N = s 3 LLs Cω 3 + s 2 ( Ls RCω 3 + LCRsω 3 ) + degree in Electrical and Computer engineering in
(16) 2003 from The University of Iceland. She studied for
28 + s ( Lsω 3 + RCRsω 3 ) + Rsω 3 the M.Sc. in Electric Power Systems at the institute of
29 Energy Technology, Aalborg University in Denmark
and received her degree in 2007 with speciality in
30 state estimation and observability analysis. She
31 REFERENCES received an honours prise for her M.Sc. final thesis.
32 [1] GE Power Systems Energy Consulting, "Connecticut Cable Transient She was a guest researcher at SINTEF in Norway in
33 and Harmonics Study for Phase 2: Final Report", November 2003 November 2008 and at Manitoba HVDC Research Centre in Canada during
[2] Tokyo Electric Power Company, "Joint Feasibility Study on the 400kV June-October 2009. Currently she is studying PhD at the Institute of Energy
34 Cable Line Endrup-Idomlund: Final Report", April 2008 Technology, Aalborg University, in cooperation with the Danish TSO
35 [3] F. Faria da Silva, C. L. Bak, U. S. Gudmundsdóttir, W. Wiechowski, M. (Energinet.dk), where she also supervises students pursuing their M.Sc.
36 R. Knardrupgård, "Use of a Pre-Insertion Resistor to Minimize Zero- degree in energy technology. Her PhD studies are focused on modelling of
37 Missing Phenomenon and Switching Overvoltages", IEEE-PES General underground cable system at the transmission level.
Meeting 2009, July 2009
38
[4] J. F. Borges da Silva, "Electrotecnia Teórica - 1ª Parte", 2nd edition, Wojciech Wiechowski received the M.Sc. degree
39 AEIST, 1995 (in Portuguese) from Warsaw University of Technology in 2001 and
40 [5] J. H. R. Enslin, Yi Hu, R. A. Wakefield, "System Considerations and the Ph.D. degree from Aalborg University, Denmark
41 Impacts of AC Cable Networks on Weak High Voltage Transmission in 2006. From 2001 to 2002 he worked for HVDC
42 Networks", Transmission and Distribution Conference and Exhibition, SwePol Link as a Technical Executor. In the period
2005/2006 IEEE PES, May 2006 from 2002 to 2006 he was with the Institute of
43 [6] Alan Greenwood, "Electric Transients in Power Systems", John Wiley & Energy Technology, Aalborg University, first as a
44 Sons, 1st Edition, 1971 PhD Student and later as an Assistant Professor.
45 [7] Y. H. Fu, G. C. Damstra, "Switching Transients During Energizing Since 2006 he has been employed in the Planning
46 Capacitive Load by a Vacuum Circuit Breaker", IEEE Transactions on Department of the Danish TSO Energinet.dk. His
Electrical Insulation, vol.28 No.4, pp. 657-666, August 1993 current responsibilities include various power system analysis tasks related to
47 [8] I. U. S. Hutter, M. Krepela, B. F. Grčić, F. Jakl, “Transients Due to the planning of the transmission network with extensive use of long AC cable
48 Switching of 400kV Shunt Reactor”, International Conference on lines and wind power generation. He is a Senior Member of IEEE.
49 Power Systems Transients (IPST), Brazil, Paper No. 045, June 2001
50 [9] SAGEM, "1200mm2 Al XLPE 400kV datasheet" Martin Randrup Knardrupgård was born in
[10] A. Morched, B. Gustavsen, M. Tartibi, "A universal model for accurate Copenhagen 1978, and received his M.Sc. E.E. 2003
51
calculation of electromagnetic transients on overhead lines and from the Technical University of Denmark.
52 underground cables," IEEE Transactions on Power Delivery, 14(3), From 2003 to 2006 he worked for the Swedish
53 p.1032–1038, July 1999 electric power company Sydkraft/E.ON where he
54 [11] ABB, "Live Tank Circuit Breakers: Buyer's Guide"; 4th edition, May was involved in the planning of the regional
55 2008 transmission grid in southern Sweden. Since 2006
[12] M. V. Escudero, M. Redfern, "Effects of transmission line construction he joined the planning department of the Danish
56 on resonance in shunt compensated EHV lines", IPST05-109, June 2005 TSO Energinet.dk. His current responsibilities
57 [13] S. Schulz, "Four Lectures on Differential-Algebraic Equations", include long term planning of the Danish
58 Humboldt Universität zu Berlin, June 2003 transmission grid, especially interconnections with UCTE and Nordel, aspects
59 and feasibility studies of 400 kV cabling and the connection of the offshore
wind farms Horns Reef 2, Rødsand 2 and Anholt.
60