74 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 30, NO.

1, FEBRUARY 2015

Analytic Approximation of Fault Current

Contributions From Capacitive Components
in HVDC Cable Networks
Matthias K. Bucher, Student Member, IEEE, and Christian M. Franck, Senior Member, IEEE

Abstract—A novel approach for the calculation of transient the semiconductor devices, make the multiterminal HVDC net-
fault current contributions from capacitive network components work a viable option. Its protection in case of a fault is, however,
in HVDC cable networks during pole-to-ground faults is presented still a major obstacle in the realization of such a network. While
in this paper. Analytic approximations considering the skin effect
and the resulting distortion of the waveshape of the fault surge ac-side circuit breakers (CBs) can adequately protect point-to-
are proposed. Only fault current contributions from capacitive point HVDC connections, an interconnected HVDC network re-
components, such as dc capacitors and adjacent cable feeders, are quires dc CBs to selectively isolate a faulty cable, which cannot
taken into account in this paper, since they are dominant during be realized by ac-side CBs due to the forced de-energization of
the first few milliseconds and yield the highest rates of rise of fault the entire system [5]. DC CB concepts can be found in [6]–[8],
current. The results of the proposed expressions are compared
with a benchmark model implemented in PSCAD and exhibit which still have significant drawbacks in terms of onstate losses
an accurate representation of the time development of the fault or speed. To specify the requirements of such dc CBs and pos-
current. The derived approximations may serve as a starting point sible fault clearance support options [9], the maximum and min-
for a short-circuit calculation standard for HVDC networks and imum short-circuit currents as well as overvoltages have to be
the specification of HVDC circuit-breaker requirements. known. For a deeper understanding of the short-circuit current
Index Terms—HVDC transmission, power system faults, power and its influencing parameters, a breakdown of the prospec-
system simulation, PSCAD. tive short-circuit current into its individual contributions from
the different network components has been presented in [10].
Fault feeding network components include dc capacitors, adja-
I. INTRODUCTION
cent feeders at the same busbar as the faulted feeder, and the ac
network, which contributes to the fault current through the free-

M ULTITERMINAL high-voltage dc (HVDC) networks wheeling diodes of the blocked half-bridge-based converters.
are widely recognized as a key component in the future Further parameter variations, including the transmission tech-
power networks, which will have to cope with increasing elec- nology, that is, cables or overhead lines (OHLs), can be found
tricity demand and high penetration of new renewable energy in [11]. All results are based on simulations performed with the
sources. Several initiatives from the academic community, in- PSCAD simulation software using fully detailed, frequency-de-
dustry consortia, and environmental nongovernmental organi- pendent models for the cables and OHLs.
zations (NGOs) [1]–[4] envision the creation of interconnected Standardized calculations of the maximum and minimum
HVDC networks with the expected benefits of higher redun- values of current and voltage, rather than fully detailed sim-
dancy, increased flexibility for power trading, and reduced op- ulations, are required to ease the specification of the network
erational costs. Since the renewable energy sources are often far components, for example, the CBs. Standard calculation pro-
away from the load centers, the transmission distances become cedures exist for short-circuit currents in HVAC systems,
longer and cannot be realized in a technically and economically but not yet for HVDC networks. A summary of technical
reasonable way based on HVAC technology, particularly, in the guidelines and prestandardization studies for such HVDC
case of offshore connections with cables. grids has been published in [12]. Recently, the IEC 61660
Recent advances in HVDC technology, such as the develop- standard for medium-voltage (MV) dc systems [13] has been
ment of voltage-source converters (VSCs) and higher ratings of used for short-circuit current calculations in HVDC systems
[14]. This standard provides calculation rules of the short-cir-
cuit currents in small low-voltage (LV) and MV auxiliary
Manuscript received November 05, 2013; revised March 07, 2014; accepted
May 24, 2014. Date of publication June 18, 2014; date of current version January networks. It simplifies, however, the lines of the network to
21, 2015. This work was supported in part by ABB Switzerland Ltd., in part by lumped frequency-independent series impedances, which is a
Siemens AG, in part by Alstom Grid, and in part by the Swiss Federal Office of
valid assumption for MV networks with short interconnections
Energy (BfE). Paper no. TPWRD-01258-2013.
The authors are with the Power Systems and High Voltage Laboratories, ETH and small line capacitances. In an HVDC network, however,
Zurich, Zurich 8092, Switzerland (e-mail: [email protected]; franck@eeh. the capacitance cannot be neglected. The waveshape of the
ee.ethz.ch).
propagating surges is distorted, particularly, in systems with
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org. long cable interconnections. Hence, the lines have to be rep-
Digital Object Identifier 10.1109/TPWRD.2014.2327132 resented by distributed, frequency-dependent parameters and

0885-8977 © 2014 IEEE. Personal use is permitted, but republication/redistribution requires IEEE permission.
See http://www.ieee.org/publications_standards/publications/rights/index.html for more information.
BUCHER AND FRANCK: ANALYTIC APPROXIMATION OF FAULT CURRENT CONTRIBUTIONS 75

the pole and the sheath of the cable and a ground loop through
the sheath and the next grounding point is established. The cur-
rent through the arc increases rapidly, which likely leads to ex-
plosion and destruction of the cable at the ground fault location.
Subsequently, the arc burns between the pole and the ground,
and a low-ohmic path is established in between.
After the ground fault occurs, the voltage at the fault loca-
tion decreases rapidly and negative voltage surges start to travel
from the fault location into both directions toward the terminals.
Along its way, the distributed cable capacitance is discharged
gradually into the ground fault.
Upon the arrival at the terminals after the traveling time ,
the negative voltage surge is reflected back as a positive surge
due to the capacitive termination of the cable given by the dc
capacitors [16]. DC capacitors include the VSC capacitors and
possible tuned filter capacitors, which are usually installed at the
Fig. 1. IEC results (blue) compared with simulations (red)—top: 10 km, dc side of a VSC in order to reduce the voltage ripple injected
100 F; middle: 100 km, 100 F; bottom: 10 km, 1 F.
by the converter. The converter technology determines the size
of the dc capacitor as explained in [10].
the skin effect has to be taken into account. The IEC 61660 In bipolar HVDC schemes, the midpoint of the dc capaci-
standard yields only good results for HVDC networks under tors is usually grounded to provide a reference voltage to the
the following conditions: the lines are relatively short and the pole voltages [17], [18]. The midpoint is grounded either via a
dc capacitor, as well as the pole reactor are large, so that the low-ohmic connection or through a reactor depending on the re-
lumped elements (capacitor and reactor) become dominant quirement, whether the bipole has to operate in monopolar mode
over the frequency-dependent, distributed line parameters. This or not [19]. The grounded capacitor midpoint and the ground
effect is illustrated in Fig. 1 for different sets of line length and fault form a loop that provokes a discharge of the capacitors.
dc capacitor size. This discharge current is superposed on the reflected, backward
This paper proposes new expressions for the approximation traveling surge. As the surge arrives again at the fault location,
of overcurrents in HVDC cable networks based on individual one part is reflected and the other part is transmitted through the
surges and the planar skin effect. The derived approximations, fault into the opposite section depending on the fault resistance.
however, can also be applied to OHL-based systems. Results are The forward and backward traveling waves result in multiple
only shown for the cable system due to its higher fault current peaks in the fault current waveform.
levels [11]. Approximations are given for the individual contri-
butions from dc capacitors at the dc side of the converters and III. SIMULATION MODEL
neighboring feeders at the same busbar as the faulted cable. The A description of the PSCAD models used in the benchmark
contribution from the ac side, which is present in a half-bridge- simulations for the validation of the approximate calculations
based converter, is neglected in this study, since the main focus of the prospective short-circuit current in the CB is presented in
is on the initial transient period during the first few millisec- this section.
onds after fault occurrence. During this period, the capacitive
contributors are dominant and result in the highest rates of rise A. Network Model
of fault current [10]. The network used in the simulations is illustrated in Fig. 2. A
This paper is structured as follows: Section II explains the solid pole-to-ground fault occurs at a certain distance to the CB
transients in an HVDC system during a pole-to-ground fault and under study. The faulted cable is connected to a dc busbar to-
Section III describes the benchmark simulations and cable mod- gether with an additional feeder representing a meshed dc grid.
eling in PSCAD. In Section IV, the alternative calculations are A dc capacitor comprising the converter capacitors and possible
derived, which are compared with the results from the simula- tuned filter capacitors is installed between the converter and the
tion in PSCAD in Section V. Section VI concludes this paper. dc busbar. The converters can be omitted in the transients sim-
ulations, since they are assumed to have full-bridge configura-
II. TRANSIENTS IN HVDC NETWORKS tion and are blocked immediately after detection of the fault,
Potential sources of transients in an HVDC network include yielding the equivalent circuit diagram shown in Fig. 3. The
surges due to pole-to-ground faults, pole-to-pole faults, the op- contributions to the CB current are the dc capacitor current
eration of switching devices, lightning strokes, and the sudden and the adjacent feeder current . The cables are represented
loss of a terminal and the subsequent change in the dc voltage. by their characteristic impedance .
In the following text, only pole-to-ground faults in underground
cable systems are considered. B. Cable Model
Aging of the cable's main insulation or external damage due The system is modeled in PSCAD-EMTDC and makes
to digging or anchoring in case of sea cables [15] may lead to a use of a detailed frequency-dependent, distributed-parameter
break down of the cable insulation. First, an arc burns between cable model. The general design of the cable cross-section is
76 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 30, NO. 1, FEBRUARY 2015

(8)

with the propagation constant defined as

(9)

The solutions of the second-order differential equations (7)

and (8) are

(10)
Fig. 2. Network layout of the benchmark model.
(11)

which is the superposition of the forward and backward trav-

eling waves with the initial amplitudes and , re-
spectively, and the characteristic surge impedance de-
fined as

(12)

The closed-form solution of the back-transformation into the

time domain for constant, frequency-independent cable parame-
Fig. 3. Simplified equivalent diagram of capacitive fault current contributors. ters can be found in [22]. In the following text, the back-transfor-
mations for frequency-dependent cable parameters considering
the skin effect are derived.
derived from a real 150-kV XLPE VSC-HVDC submarine
cable [15], [20]. The cross-section was scaled up to a 320-kV A. Frequency Dependence of Parameters
cable respecting the diameter of the copper conductor [21], The frequency dependence of the cable parameters is inves-
while keeping the electric-field stress (cold condition) similar. tigated in the frequency range from 1 mHz to 1 MHz using
The material properties and cable cross-section dimensions are the PSCAD Line Constants Program. The shunt parameters of
given in [10]. The shunt conductance of the XLPE insulation the capacitance and the conductance are practically fre-
is set to S/km. quency independent. The resistance reveals the highest
dependence on the frequency over the entire frequency range of
IV. DERIVATION OF NEW EXPRESSIONS this study as illustrated in Fig. 4 (dotted line). This is due to the
The basics for the derivation of analytic expressions of fault skin effect, that is, the displacement of the current to the surface
surges are in the well-known traveling-wave theory, which im- of the inner conductor at high frequencies and corresponding
plies the following partial differential equations for the voltage small penetration depths causing an increase of the effective re-
and current at the point on the cable: sistance. Fig. 4 depicts the cable inductance (solid line),
which decreases rapidly in the low-frequency range below 1 Hz
(1) and decreases very little with increasing frequency above 1 Hz.

B. Short-Circuited, Infinitely Long Cable

(2)
In the first step of the derivation of an analytic expression for
with the cable resistance , inductance , capacitance , and the distortion of the waveshape, an infinitely long cable that is
shunt conductance . The transformation into the Laplace do- short-circuited at one end as depicted in Fig. 5 is analyzed. A
main yields closed-form expression for the timely development of the cur-
rent and the voltage at any point on the cable
(3) shall be derived.
As described in Section IV-A, only the cable resistance ex-
(4) hibits strong frequency dependence, whereas the other parame-
ters remain almost constant over the frequency range of interest.
(5) The ground fault initiates a negative voltage surge with a very
(6) steep wavefront, which results in high frequencies. Even if the
wavefronts are flattened out during the propagation through the
where is the cable impedance and is the cable admit- cable as well as after the reflection at the dc capacitor, the domi-
tance. Rearrangement of these equations results in independent nant frequency content is still above the frequency range, where
expressions for the voltage and current the cable inductance shows a high dependence (Fig. 4). There-
fore, only the resistance is modeled as a frequency-dependent
(7) parameter. Due to the high-frequency content of the traveling
BUCHER AND FRANCK: ANALYTIC APPROXIMATION OF FAULT CURRENT CONTRIBUTIONS 77

domain are computed using (10) and (11), while considering

only the forward traveling wave and assuming an initial
voltage step of magnitude at the fault location

(17)

(18)

The time-domain solutions are derived in [23] and [24], re-

spectively
Fig. 4. Frequency-dependent cable resistance (blue, dotted line) and induc-
tance (red, solid line). (19)

(20)

where is the distortion factor, the trav-

eling-wave delay, the unit step function, and is the
complementary error function [25].
The proposed equations provide a good approximation to the
high-frequency behavior of a coaxial cable, where the planar
Fig. 5. Discharge of an infinitely long line.
skin effect is predominant [26]. Other skin effect models for
lower frequencies are the cylindrical skin effect [27] and the
waves, the penetration depth is much smaller than the diam- -law with as described in [26]. The general
eter of the conductor and it is assumed that the current flows case, including dielectric losses, is discussed in [28] and [26],
through an indefinitely thin layer at the surface of the conductor. and a model considering a nonzero dc resistance of the inner
This so-called planar skin effect impedance is proportional to conductor is presented in [29]. However, the aforementioned
the square root of the frequency [23]. In general, XLPE cables models cannot be represented by closed-form analytic expres-
exhibit very low dc resistance of the inner conductor and neg- sions.
ligible dielectric losses as indicated in Section IV-A and, there-
C. DC Capacitor Contribution
fore, the dc resistance and the shunt conductance can be ne-
glected. Hence, the cable impedance and admittance can be de- The voltage across the dc capacitor is equal to the sum of
scribed in the Laplace domain as the forward and reflected, backward traveling wave. The corre-
sponding expression in the Laplace domain for a ground fault at
(13) distance from the CB is
(14)

where represents the skin effect factor. Expanding the square (21)
root of the propagation constant and the characteristic
impedance by binomial expansion and truncation after with the reflection coefficient
the second term, one obtains the approximations [23]
(22)

which depends on the characteristic impedance and the

(15) impedance of the capacitor with value .
1) Exact Transformation: The exact back-transformation
(16) into the time domain can be found using the transformation
pairs in [30] and the displacement law
with the characteristic cable impedance at high frequency
and the propagation speed .
The equations for the voltage and current surges in the Laplace
78 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 30, NO. 1, FEBRUARY 2015

D. Adjacent Feeder Contribution

The contribution of the adjacent feeder cable at the same
busbar as the faulted cable (Figs. 2 and 3) can be derived in a
similar way as the capacitor contribution described in the pre-
vious paragraph. The incident negative voltage surge initiated at
the ground fault location is transmitted partly through the busbar
into the neighboring feeder, which is consequently discharged
and contributes to the total fault current in the CB. Hence, the
adjacent feeder current is determined in the Laplace domain for
(23)
a ground fault at distance from the CB by
Note that the error function in (23) has a complex argument (28)
and requires the computation of the Faddeeva-Function denoted
as in [25]. where is the incident current surge and is the trans-
Hence, the capacitor current contribution can be derived by mission coefficient. The adjacent feeder is assumed to be infin-
itely long and, therefore, unlike in the equation for the capacitor
(24) voltage (21), there is no backward traveling wave that has to be
considered.
The minus sign in (24) arises from the definition of the 1) Exact Transformation: The exact transformation into the
voltage polarity of as illustrated in Fig. 3. time domain is computed by partial fraction decomposition of
2) Approximation: Instead of using the lengthy equation of the denominator of (28)
the exact time-domain solution (24), an approximation based
on the modification of the argument of the exponential function
in the Laplace domain for the reflected voltage wave in (21) is
proposed. To do so, the frequency is shifted by in the
nominator of

(29)

The first and last term in (29) can be directly transformed into
the time domain using the transformation pairs in [30] and [31],
(25) whereas the nominator and denominator of the second term have
to be multiplied first by to be transformed, which implies
the derivative of the resulting time-domain solution
The error introduced by this frequency shift is compensated
by the factor for the first term in the exponential func-
tion, but is not corrected for the second term due to the square (30)
root of the frequency. This approximation is only valid for large
compared to , where is the time constant of The equation in the time domain is too long to be reproduced
the dc capacitor. The modified equation in the Laplace domain here.
(25) allows for the application of the displacement law when 2) Approximation: A simpler, approximate solution
transforming the equation back into the time domain can be obtained again by applying a frequency shift to
and corresponding correction factors, except
for the square root terms, to make use of the displacement law.
The resulting time-domain equation is
(26)

One obtains the capacitor current contribution using again the

(31)
time derivative of the voltage
This approximation is again valid for .

E. Total CB Current
The total fault current during the first surge is obtained by the
superposition of the individual fault current contributions from
the dc capacitor and the neighboring feeder as
(27)
(32)
BUCHER AND FRANCK: ANALYTIC APPROXIMATION OF FAULT CURRENT CONTRIBUTIONS 79

To consider also the subsequent surges, which have been re-

flected at the fault location, the travel delay in the equations for
the dc capacitor current and the adjacent feeder current
have to be incremented by for each new surge. The CB
current including subsequent surges is then

(33)

Note that no backward traveling surges on the adjacent feeder

cable are considered here due to the assumption of an infinitely
long and, therefore, reflectionless adjacent feeder.

V. COMPARISON AND DISCUSSION

In this section, the results of the analytic calculations and the Fig. 6. Voltage at various locations on the infinitely long cable—solid curve:
benchmark simulations are compared and validated. First, the PSCAD simulations, dashed curve: analytic calculations.
results of a single current surge on a short-circuited, infinitely
long cable are compared and, later, the total CB current is eval-
uated for different distances to the ground fault and dc capacitor
sizes.

A. Parameters
The required input parameters for the analytical cal-
culations are derived from the cable model described in
Section III-B evaluated at 1 MHz using the PSCAD Line
Constants Program. The resulting values are 2.886 km
, 1.531 H/km, and 1.882 F/km. The
skin effect factor is then computed as [24]

(34)

B. Single Surge Fig. 7. Current at various locations on the infinitely long cable—solid curve:
PSCAD simulations, dashed curve: analytic calculations.
Figs. 6 and 7 show the comparison of the results of the bench-
mark model (solid curves) and the analytic calculations (dashed
curves) for a single voltage (19) and current surge (20) at various of the capacitor and feeder contributions (dotted curve) show
distances from the fault location. The wavefronts of the analytic very good agreement with benchmark simulations (solid curve)
expressions exhibit very good agreement with the simulations, as depicted in Fig. 8. The approximate transformation (dashed
whereas an increasing discrepancy can be seen for increasing curve), however, overestimates the CB current peak due to the
time. This is due to the assumptions of constant inductance and frequency shift introduced in the Laplace domain (25). To im-
planar skin effect, which are valid at high frequencies. The wave prove the approximate solution, a simple scaling factor for the
tails, however, contain also lower frequencies and the value of capacitor contribution (27), which is dominant in this case, is
the inductance becomes slightly higher, which damps the cur- proposed and the result for a visually estimated scaling factor
rent amplitudes as seen in Fig. 7. At low frequencies, the skin of 0.88 (dashed-dotted line) is illustrated in the same plot. The
depth increases and the square root law of the planar skin effect adjacent feeder contribution (31) remains unscaled.
is not valid anymore. In general, the error of the analytic cal- Fig. 9 illustrates the results for a large dc capacitor size of 100
culation is negligible during the first few milliseconds for short F. On the one hand, the exact transformation exhibits again
distances to the fault, whereas it increases with increasing dis- good accuracy for the wavefront, but somehow there is a larger
tance and time. discrepancy afterwards due to the smoothing of the surge by
the capacitor and, consequently, there is a downwards frequency
C. Total CB Current
shift. On the other hand, the approximate solution reveals a high
To validate the analytic expression for the CB current derived correlation with the temporal development of the solution of the
in this paper, a variation of the dc capacitor size and distance benchmark simulations, but with a certain offset and slightly ex-
to fault is presented here and plots are shown for the first surge. aggerated slope of the wavefront. The same scale factor for the
1) Variation of DC Capacitor Size: For a low dc capacitor capacitor contribution as in the previous case is applied, and the
size of 1 , the superposition of the exact back-transformations results are plotted (dashed-dotted curve) in the same figure. The
80 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 30, NO. 1, FEBRUARY 2015

TABLE I
RELATIVE ERROR OF IN PERCENT

TABLE II
RELATIVE ERROR OF IN PERCENT

Fig. 8. CB current for 100 km and 1 F —solid line: PSCAD,

dotted line: exact solution, dashed line: approximate solution, dashed-dotted
line: scaled approximate solution.

Fig. 10. CB current for 10 km and 100 F—solid line: PSCAD,

dotted line: exact solution, dashed line: approximate solution, dashed-dotted
line: scaled approximate solution.

results as indicated in Table II and reveal an increasing error to-

ward larger dc capacitors, particularly, for the approximate solu-
tion. In general, the approximate solution matches the wavetails
Fig. 9. CB current for 10 km and 100 F —solid line: PSCAD, better and the exact solution matches the high-frequency wave-
dotted line: exact solution, dashed line: approximate solution, dashed-dotted front better. However, its analytic expression is not as compact
line: scaled approximate solution. as the approximate solution and has a very long form for the ad-
jacent feeder contribution. The performance of the approximate
solution can be considerably improved by a scaling factor, but
scaled approximation shows the best agreement for the large dc this is rather a model fitting factor.
capacitor. 2) Variation of Distance to Fault: A variation of the dis-
The relative error of the peak CB current during the tance to fault only has a little influence on the results as shown
first surge of the analytic expressions (exact and approximate) in Fig. 10, where the distance to fault is reduced to 10 km.
is analyzed and summarized in the first and second column of The overall fault current level is higher compared to the 100-km
Table I for the two capacitor sizes. A clear trend of increasing cable between fault and CB in Fig. 9 given the lower damping.
relative error of the exact solution for increasing capacitor size In the benchmark results in Fig. 10 (solid line), the current in-
is visible, whereas the approximate solution shows the opposite creases again at 0.16 ms due to the arrival of the second surge
trend. A larger capacitor shifts the waves' dominant frequency of the forward and backward traveling wave on the short 10-km
downwards and, hence, deteriorates the accuracy of the exact cable. The relative error of the analytic calculations (exact and
solution based on the planar skin effect, which is only valid at approximate) is smaller compared to the case with a 100-km
high frequencies as explained in Section V-B. The performance cable and 100- F capacitor as indicated in the third columns of
of the approximate solution, however, is improved by a larger Tables I and II.
capacitor, because its time constant is higher and so is
smaller compared to the frequency , reducing the error intro- VI. CONCLUSION
duced in (27) and (31) (cf. Sections IV-C and IV-D). The exact New analytic time-domain expressions for the calculation of
and approximate expressions yield a higher average fault currents through the CB in HVDC grids during pole-to-
of the wavefront up to the peak than the PSCAD benchmark ground faults have been proposed. These take into account the
BUCHER AND FRANCK: ANALYTIC APPROXIMATION OF FAULT CURRENT CONTRIBUTIONS 81

planar skin effect and an adjacent feeder in case of meshed [16] E. W. Kimbark, “Transient overvoltages caused by monopolar ground
HVDC networks. In general, good accuracy is achieved with fault on bipolar DC line: Theory and simulation,” IEEE Trans. Power
App. Syst., vol. PAS-89, no. 4, pp. 584–592, Apr. 1970.
the expressions based on the exact back-transformation from the [17] M. P. Bahrman and B. K. Johnson, “The ABCs of HVDC transmission
Laplace domain. The peak values and the average of the technologies,” IEEE Power Energy Mag., vol. 5, no. 2, pp. 32–44, Mar.
wavefronts stay in good agreement with the benchmark simula- /Apr. 2007.
[18] B. R. Andersen, L. Xu, P. J. Horton, and P. Cartwright, “Topologies for
tions. The formulas of the exact transformations have performed VSC transmission,” Power Eng. J., vol. 16, no. 3, pp. 142–150, 2002.
better for small dc capacitors and the scaled approximate solu- [19] CIGRE Working Group B4.52, HVDC grid feasibility study, Brochure
tions better for high values of the capacitor. The proposed equa- 533, Apr. 2013.
[20] L. Ronström, M. L. Hoffstein, R. Pajo, and M. Lahtinen, “The estlink
tions enable a simple CB specification and a highly accurate HVDC light transmission system,” in Proc. CIGRE Reg. Meeting Se-
representation of the temporal development of transient CB cur- curity Rel. Elect. Power Syst., Tallinn, Estonia, Jun. 2007.
rents without detailed simulations, even for very complex cur- [21] ABB High Voltage Cables. HVDC Light Cables- Submarine and land
power cables. Oct. 2006. [Online]. Available: http://library.abb.com
rent curves with multiple surges. These equations may serve as [22] Transmission Lines and Lumped Circuits, G. Miano and A. Maffucci,
a basis for future HVDC fault current calculation standards. Eds. London, U.K.: Academic Press, 2001.
[23] R. L. Wigington and N. S. Nahman, “Transient analysis of coaxial ca-
bles considering skin effect,” Proc. IRE, vol. 45, no. 2, pp. 166–174,
1957.
[24] P. Magnusson, “Transient wavefronts on lossy transmission lines—ef-
REFERENCES fect of source resistance,” IEEE Trans. Circuit Theory, vol. 15, no. 3,
pp. 290–292, Sep. 1968.
[1] D. v. Hertem and M. Ghandhari, “Multi-terminal VSC HVDC for the [25] Handbook of Mathematical Functions, M. Abramowitz and I. A.
European supergrid: Obstacles,” Renew. Sustain. Energy Rev., vol. 14, Stegun, Eds. New York, USA: Dover, 1964.
no. 9, pp. 3156–3163, 2010. [26] N. S. Nahman, “A discussion on the transient analysis of coaxial cables
[2] Friends of the Supergrid FOSG WG2 Roadmap to the Supergrid Tech- considering high-frequency losses,” IRE Trans. Circuit Theory, vol. 9,
nologies. Final Report. Mar. 2012. [Online]. Available: http://www. no. 2, pp. 144–152, 1962.
friendsofthesupergrid.eu [27] D. R. Holt and N. S. Nahman, “Coaxial-line pulse-response error due
[3] OffshoreGrid. Offshore Electricity Grid Infrastructure in Europe. Final to a planar skin-effect approximation,” IEEE Trans. Instrum. Meas.,
Report. Oct. 2011. [Online]. Available: http://www.offshoregrid.eu vol. IM-21, no. 4, pp. 515–519, Nov. 1972.
[4] Greenpeace and 3E. A North Sea Electricity Grid Revolution. Sep. [28] C. Xu, L. Zhou, J. Y. Zhou, and S. Boggs, “High frequency proper-
2008. [Online]. Available: http://www.greenpeace.org ties of shielded power cable—Part I: overview of mechanisms,” IEEE
[5] L. Tang and B.-T. Ooi, “Locating and isolating DC faults in multi- Elect. Insul. Mag., vol. 21, no. 6, pp. 24–27, Nov./Dec. 2005.
terminal DC systems,” IEEE Trans. Power Del., vol. 22, no. 3, pp. [29] N. S. Nahman and D. R. Holt, “Transient analysis of coaxial cables
1877–1884, Jul. 2007. using the skin effect approximaton ,” IEEE Trans. Circuit
[6] C. M. Franck, “HVDC circuit Breakers: A review identifying future Theory, vol. CT-19, no. 5, pp. 443–451, Sep. 1972.
research needs,” IEEE Trans. Power Del., vol. 26, no. 2, pp. 998–1007, [30] Operational Mathematics, R. V. Churchill, Ed. New York, USA: Mc-
Apr. 2011. Graw-Hill, 1972.
[7] M. Callavik, A. Blomberg, J. Häfner, and B. Jacobson, “The hybrid [31] Conduction of Heat in Solids, H. S. Carslaw and J. C. Jaeger, Eds.
HVDC breaker—An innovation breakthrough enabling reliable HVDC Oxford, U.K.: Clarendon, 1959.
grids,” ABB Grid Syst. Tech. Paper, Nov. 2012.
[8] J. Haefner and B. Jacobson, “Proactive hybrid HVDC breakers—A key Matthias K. Bucher (S'12) received the B.Sc. and
innovation for reliable HVDC grids,” presented at the CIGRE Symp., M.Sc. degree in electrical engineering from the ETH
Bologna, Italy, Sep. 2011. Zurich, Zurich, Switzerland, in 2009 and 2011, re-
[9] M. K. Bucher, M. M. Walter, M. Pfeiffer, and C. M. Franck, “Options spectively, where he is currently pursuing the Ph.D.
for ground fault clearance in HVDC offshore networks,” presented degree in high voltage engineering.
at the IEEE Energy Convers. Congr. Expo., Raleigh, NC, USA, Sep. He joined the High Voltage Laboratory, ETH
2012. Zurich, in 2011. His research is dedicated to tran-
[10] M. K. Bucher and C. M. Franck, “Contribution of fault current sources sients in multiterminal HVDC networks.
in multi-terminal HVDC cable networks,” IEEE Trans. Power Del.,
vol. 28, no. 3, pp. 1796–1803, Jul. 2013.
[11] M. K. Bucher and C. M. Franck, “Analysis of transient fault currents
in multi-terminal HVDC networks during pole-to-Ground faults,”
presented at the Int. Conf. Power Syst. Transients, Vancouver, BC,
Canada, Jul. 2013.
[12] V. Akhmatov, M. Callavik, C. M. Franck, S. E. Rye, T. Ahndorf, M. Christian M. Franck (M'04–SM'11) received the
K. Bucher, H. Müller, F. Schettler, and R. Wiget, “Technical guidelines diploma in physics from the University of Kiel, Kiel,
and prestandardization work for first HVDC grids,” IEEE Trans. Power Germany, in 1999 and the Ph.D. degree in physics
Del., vol. 29, no. 1, pp. 327–335, Jan. 2014. from the University of Greifswald, Greifswald,
[13] Short-Circuit Current in D.C. Auxiliary Installations in Power Plants Germany, in 2003.
and Substations , Int. Standard IEC 61660, IEC, Jun. 1997. He was Scientist and Group Leader for gas
[14] A. Wasserrab, B. Just, and G. Balzer, “Contribution of HVDC con- circuit breakers and high-voltage systems with the
verters to the DC short circuit current,” presented at the Univ. Power Swiss Corporate Research Center of ABB, Dättwil,
Eng. Conf., Dublin, Ireland, Sep. 2013. Switzerland, from 2003 to 2009. Currently, he is
[15] Submarine Power Cables: Design, Installation, Repair, Environmental Assistant Professor of High Voltage Technology,
Aspects, T. Worzyk, Ed. New York, USA: Springer, 2009. ETH Zurich, Zurich, Switzerland.