Equivalent Circuit of Single-Core Cable Lines Suitable For Grounding Systems Analysis Under Line-to-Ground Faults
Equivalent Circuit of Single-Core Cable Lines Suitable For Grounding Systems Analysis Under Line-to-Ground Faults
Equivalent Circuit of Single-Core Cable Lines Suitable For Grounding Systems Analysis Under Line-to-Ground Faults
Abstract—This paper presents two approaches where the single- conductor is a bare conductor directly buried into the ground
core cable line and an additional conductor laid in parallel can and it is in good contact with it in its full length.
be reduced to a simple equivalent -circuit containing an active The modeling of grounding system elements was the subject
element beside the usual passive impedances. The parameters of
in many papers, and methods for complex systems analysis with
the equivalent circuit are calculated using a numerical approach
where the distributed parameter line was discretized and an ana- underground cables were proposed [1]–[3]. In [1], three-phase
lytical approach is based on the wave propagation theory. A series cable line with a common sheath is considered. The sheath is
of tests has been performed to determine error levels in both cases partitioned by introducing the number of fictitious nodes so that
showing that they are less than 1%. With numerical examples, we a set of linear equations is formed. The method in [2] is an ex-
show that the cable equivalent circuit offers a simple way to cal- tension of [1], where all mutual impedances were considered
culate voltages in complex grounding systems. The results are in as well as the influence of saturation of cable steel armor on
good agreement with the previously published calculated and/or
measured values. Both approaches offer the possibility of calcu- sheaths impedances. The method for analysis of the grounding
lating voltage and current distribution along the cable line which system of a single cable feeder supplying an arbitrary number
gives insight into the significance of each cable component in the of transformer stations, whose grounding electrodes are con-
complex grounding system. nected through the cable sheath, is presented in [3], where rel-
Index Terms—Cable sheaths, distributed parameter circuits, atively simple expressions are derived by assessing most rele-
equivalent circuits, fault currents, grounding electrodes, under- vant grounding effects of coated cables. Laying an additional
ground cable lines. conductor in parallel with the cable line as a measure to reduce
the current flow into the ground electrode was analyzed in [4].
More complex configurations comprised of three-single core
I. INTRODUCTION cables with or without additional conductors are considered in
[5]–[11]. In these papers, the main topic is the determination of
G ROUNDING systems of power distribution networks
are complex structures comprised of many elements,
including substations groundings which are interconnected
the fault current reduction factor and in [10] and [11], there are
comparisons with experimentally obtained measured data.
Following the idea of [9], in this paper, we develop two ap-
with overhead and cable lines. These systems should ensure
proaches where the single-core cable line can be reduced to a
that in cases with ground faults, there will be a safe level of
simple equivalent –circuit containing an active element be-
step-and-touch voltages at any point in the system, and that the
side the usual passive impedances. This equivalent circuit can
relay protection will function properly under any condition. To
be further utilized to solve complex grounding systems and to
check whether these requirements are met, one has to construct
calculate step-and-touch voltages, fault current reduction fac-
a circuit model of the grounding system and calculate the
tors, etc. The proposed methods are compared against previ-
voltage and current distribution among its element using basic
ously published calculated or measured data. The comparison
circuit theory. For the sake of result accuracy, proper models
is made with [11], where experimental measurements are per-
for each of the grounding system elements are necessary. In this
formed on a real cable line, which radially supplies two substa-
paper, we focus on the underground cable lines as components
tions connected in series. We also compare our results with the
of grounding systems which are practically most common ele-
calculated values of [11].
ments in the distribution networks of large cities. Specifically,
we consider single-core cables with metallic sheath insulated
from ground (coated cables) which, in many cases, are laid out II. GENERAL CONSIDERATIONS
in the same trench in parallel with an additional conductor. This
Let us consider an underground cable line comprised of three
single-core cables whose metallic sheaths are isolated from the
Manuscript received March 07, 2013; revised June 19, 2013; accepted August ground (coated cables). Very often, in the same trench with the
07, 2013. Date of publication August 28, 2013; date of current version March
20, 2014. Paper no. TPWRD-00264-2013. cables, an uncoated conductor is installed, referred to as an ad-
The authors are with the Faculty of Electrical Engineering and Information ditional conductor. In such cases, it is common practice to con-
Technologies, University Ss. Cyril and Methodius, Skopje, Republic of Mace- nect all ends of metallic sheaths, as well as the additional con-
donia (e-mail: mirko@feit.ukim.edu.mk; acko@feit.ukim.edu.mk).
Color versions of one or more of the figures in this paper are available online
ductor, to the groundings at both ends of the cable line. Three
at http://ieeexplore.ieee.org. cases of possible configurations of the underground cable line
Digital Object Identifier 10.1109/TPWRD.2013.2277887 comprising of three single-core coated cables and an additional
0885-8977 © 2013 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.
752 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 29, NO. 2, APRIL 2014
Fig. 3. Segment of the cable line with equivalent current generators. Fig. 5. Passive part of the cable line divided in a finite number of segments.
(10)
segments is the same, that is, they are pointing from the end to- where is the segment length. Note that we assume
ward the beginning of the segment. If we consider two consec- equal earth resistivity along the line. Sometimes this assumption
utive segments and , it is obvious that at their common point may be too rigorous for a cable length up to several kilometers.
, there will be a quadruplet of currents flowing out of (due to In these cases, we should split the cable in sections where earth
segment ) and a quadruplet of currents flowing into (due to resistivity is constant and apply the same approach on a section
segment ). Since the currents are equal, the net current injection base.
at point is zero. This is the case for all cable segments except For the circuit in Fig. 5, we construct the bus admittance ma-
the first (starting at ) and the last (ending at ) since there are no trix in two steps. In the first step, we consider the serial branches
segments before and after them, respectively. Consequently, we only while the shunt branches, which are not mutually coupled,
will have a single quadruplet of current generators going from will be added in the second step.
point to point . These four current generators can be repre- Throughout this section for all matrices, we use sparse storage
sented by a single equivalent generator and operations which are readily available in or
GNU Octave.
(7) First, we form the incidence matrix for the serial branches
with dimensions . For the case with three segments
(Fig. 5), the incidence matrix is
The rest of the circuit from Fig. 2, which is comprised of a pas-
sive element, can only be represented by an equivalent –circuit
so that the equivalent circuit for the four-wire cable line will ob-
tain the form given in Fig. 4, which is also used as a modeling
basis of the EMTP [16], [18]. In the following text, we give two
procedures for calculating the parameters and .
number of segments. In the case with 3, it has the following Besides the parameters of the equivalent –circuit, it is useful
structure: to calculate the distribution of voltages and currents across the
cable line. Assume that the voltages and , with refer-
ence to remote earth, which depend on the situation outside the
cable line, such as the location and intensity of fault current, are
(17) known (see Section IV-B). Then, the voltages of all inner nodes
in Fig. 5 can be calculated from the bus admittance matrix which
is known through (18)–(20).
Having the incidence and branch admittance matrices in the We start with (21) by omitting the first and the last row in the
first step, we construct the bus admittance matrix for the serial matrix equation since the first and the last voltage are known.
branches as Then, in the remaining rows, we transfer the terms from the first
and the last column on the right-hand side of the equation so that
(18) we obtain the following system of linear equations:
In the second step, we add to the bus admittance matrix the shunt (26)
branches in the following way:
where is a submatrix of comprised of rows and
for (19) columns from 2 to , is a vector of voltages for the
for and (20) inner nodes, and and are the first and last columns of ,
where the row number also goes from 2 to .
where the subscript “old” denotes the bus admittance matrix The vector of all voltages consists of the voltage fol-
obtained in the first step and the subscript “new” denotes the lowed by the vector calculated from (26) and the voltage
same matrix at the end of step two. at the end
In order to proceed, we now need to calculate the input and
mutual impedances for buses 1 and , that is, , , , (27)
and . The most obvious way to calculate them is to perform
matrix inverse on the bus admittance matrix. But this is not the Once we know the voltage vector , the branch currents are
most efficient procedure since it is both memory and time con- simply
suming. We will use a different approach instead. We will solve
the following system of equations: (28)
(23) At this stage, we will assume that all off diagonal elements of
are equal as given in (32) where we use for the impedance
of the cable sheaths, for the impedance of the additional con-
Finally, having the elements of , we calculate ductor, and for the mutual impedances
(24)
and
(32)
(25)
TODOROVSKI AND ACKOVSKI: EQUIVALENT CIRCUIT OF SINGLE-CORE CABLE LINES 755
Mutual impedance is an average value of all mutual im- From (37) for , we have
pedances, given with (2), and may be calculated as
(42)
(43)
where the terms under the root operator are the distances be-
tween cable sheaths (indices 1–3) and the additional conductor
(index 4). The influence of mutual impedances approximation where and are constants that could be determined from
on the accuracy of the results is discussed in Section IV. Diag- the boundary conditions for which are the same as for
onal elements in (32) are calculated using (1). since all conductors are connected together at both ends of the
The matrix is line.
In this case, the first boundary condition for is
(34) (44)
(46)
(37)
where we have also used the following identity
The system of (37) is usually solved using modal transforma- .
tion [19], which is used to decouple matrix equations. However, In (46), is expressed in terms of for which there is
since the last equation in (37) is completely decoupled from the a solution given by (41). The same can be done for the first
other and it is a homogeneous differential equation of the second derivative of which is
order in one unknown , we may solve it without modal trans-
formation. This equation can be easily solved using the method
known as the variation of constants. According to this method, (47)
the solution can be written as
Here, we would like to stress the fact that in this case, the
(38)
following identity holds since from (37), it is
obvious that the differential equations for these three voltages
where and are constants that should be determined.
are the same. Once we have all four voltages, that is, their first
It happens that (35) is equal to the fourth modal propagation
derivatives, we may use (30) to find the currents
constant since eigenvalues of are . Please
note that (36) is not a modal propagation constant.
These two constants can be determined from the boundary
conditions for , where and for , where
. From the boundary conditions, we obtain (48)
(39) After certain algebraic operations for the total current, which is
(40) a sum of all four currents, we obtain
and after solving for and and substituting in (38) for the
voltage , we obtain the following:
(41) (49)
756 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 29, NO. 2, APRIL 2014
For the last term of (49) keeping in mind (41), we can write Since all derivations in this section are performed on the four-
the following: wire model of the cable, where only the sheaths are represented,
currents given with (61) and (62) do not include the effect of
mutual coupling with the phase conductors. Therefore, vector
with currents in the sheaths and the additional conductor,
including this coupling is
(50)
(63)
(64)
(59)
(65)
In this case, (7) can also be expressed in an analytical form as
In the first test, we estimated the largest value of the segment
length in such a way that both errors and are less than
1%. The results are presented in Fig. 6 where the horizontal axis
(60)
is the cable length and the vertical axis is the segment length.
At the end, we use (48) to derive the expressions for current
distribution along the cable line from where we obtain Different curves are for different earth resistivity as given in the
legend on the right-hand side. From the results, we see that in
cases with low earth resistivity, shorter segments are required
(61) but in any case, the required segment length is no shorter than
100 m. Therefore, as a general rule, we recommend that seg-
ments of 100 m be used in any case so that we are certain that
the errors will be less than 1%.
In the second test, we have computed the errors in the calcula-
(62)
tions with the procedure from Section III-C. In the same manner
TODOROVSKI AND ACKOVSKI: EQUIVALENT CIRCUIT OF SINGLE-CORE CABLE LINES 757
Fig. 6. Largest value of segment length which guarantees errors less than 1%
– for different earth resistivity.
–circuit for the cable line are calculated with (7), (24), and (25)
using 1000 and they are
Fig. 7. Error in calculation with the procedure from Section III-C for dif-
ferent earth resistivity.
V. CONCLUSIONS
The paper presents two approaches in modeling underground
cable lines where it was shown that the complex systems of the
cable line comprised of four coupled wires can be reduced to a
simple equivalent –circuit. The parameters of the equivalent
circuit can be calculated using numerical approach based on a
simple circuit theory where the distributed parameter line was
discretized or on an analytical approach based on the wave prop-
agation theory where the mutual impedances were averaged.
Fig. 13. Radial currents along the additional conductor. The main advantage of the method based on circuit theory is the
possibility for exact modeling of all mutual impedances. How-
ever, usage of an average value of mutual impedance (33) in the
but if we sum all of them, that is, if we integrate this current approach based on wave propagation theory does not affect the
along the whole cable length, we get kA, which results by more than 1%. For practical application we may con-
is exactly the sum of the two currents flowing through both sider both approaches equally good and the second one is easier
components of the equivalent –circuit of Fig. 10. to apply since the elements of the equivalent circuit are obtained
2) Case 2: Using the same type of cable as in Case 1 in [11] with simple formulas (58)–(60).
three phase single-core cable line without the additional con- The tests have shown that in the first case a discretization step
ductor is considered. In this case we may apply the procedure of 100 m guaranties errors lower than 1%, while on the other
from Section III-B with only one section since there is no con- hand in the second case the errors are lower that 1% as well.
ductor in contact with the ground. This significantly simplifies Therefore both approaches are practically applicable, and they
the calculations and we are in a position to perform them sym- have been applied on real cable lines whose data were taken
bolically which gives the following formulas for parameters of from the literature.
the equivalent –circuit It was shown that they offer a simple way in calculating volt-
ages in complex grounding systems and the results are in a good
(66) agreement with the previously published calculated and mea-
sured values [11]. Both approaches offer a possibility to calcu-
(67) late voltage and current distribution along the cable line which
gives us an insight of the significance of each cable component
while . in the complex grounding system.
TODOROVSKI AND ACKOVSKI: EQUIVALENT CIRCUIT OF SINGLE-CORE CABLE LINES 759
The numerical approach offers more possibilities which [13] P. R. Karmel, G. D. Colef, and R. L. Camisa, Introduction to Electro-
should be further exploited. They include its application in magnetic and Microwave Engineering. Hoboken, NJ: Wiley, 1998,
p. 290.
problems where the additional conductor is not laid along the [14] L. M. Wedepohl and D. J. Wilcox, “Transient analysis of underground
entire cable length and/or cases with variable earth resistivity power-transmission systems. System-model and wave-propagation
along the path. In these cases, the equivalent –circuit will not characteristics,” Proc. Inst. Elect. Eng, vol. 120, no. 2, pp. 253–260,
Feb. 1973.
be symmetrical but the voltage calculation procedure in the [15] A. Ametani, “A general formulation of impedance and admittance
grounding system will remain the same. of cables,” IEEE Trans. Power App. Syst., vol. PAS-99, no. 3, pp.
902–910, May 1980.
[16] H. W. Dommel, EMTP Theory Book, 2nd ed. Vancouver, BC,
ACKNOWLEDGMENT Canada: Microtran Power Syst. Anal. Corporation, 1996.
[17] EMTDC, Transient Analysis for PSCAD Power System Simulation,
The authors are thankful to Prof. Dragoslav Rajicic for proof- User’s Guide. Winnipeg, MB, Canada: Manitoba HVDC Research
reading the manuscript. Centre, 2010.
[18] H. W. Dommel, “Digital computer solution of electromagnetic tran-
sients in single-and multiphase networks,” IEEE Trans. Power App.
REFERENCES Syst., vol. PAS-88, no. 4, pp. 388–399, Apr. 1969.
[19] L. M. Wedepohl, “Application of matrix methods to the solution of
[1] J. Nahman and D. Salamon, “Effects of the metal sheathed cables upon
travelling-wave phenomena in polyphase systems,” Proc Inst. Elect.
performances of the distribution substation grounding system,” IEEE
Eng., vol. 110, no. 12, pp. 2200–2212, Dec. 1963.
Trans. Power Del., vol. 7, no. 3, pp. 1179–1187, Jul. 1992.
[20] Alternative Transients Program (ATP) Rule Book, Bonneville Power
[2] J. M. Nahman, V. B. Djordjevic, and D. D. Salamon, “Grounding ef-
Administration (BPA) . Vancouver, WA, USA: Canadian/American
fects of HV and MV underground cables associated with urban dis-
EMTP User Group, 1992.
tribution substations,” IEEE Trans. Power Del., vol. 17, no. 1, pp.
111–116, Jan. 2002.
[3] J. Nahman, “Earthing effects of coated underground cables with Mirko Todorovski was born in Skopje, Macedonia,
metallic shields bonded to earth electrodes,” Proc. Inst. Elect. Eng., on January 21, 1972. He received the B.Sc, M.Sc.,
Gen. Transm. Distrib, vol. 144, pp. 26–30, Jan. 1997. and D.Sc. degrees in electric power engineering from
[4] L. M. Popovic, “Efficient reduction of fault current through the University Ss. Cyril and Methodius, Skopje, in 1995,
grounding grid of substation supplied by cable line,” IEEE Trans. 1998, and 2004, respectively.
Power Del., vol. 15, no. 2, pp. 556–561, Apr. 2000. From 1997 to 2005, he was with the Research
[5] I. Sarajcev, M. Majstrovic, and I. Medic, “Current reduction factor of Center for Energy, Informatics and Materials of
compensation conductors laid alongside three single-core cables in flat the Macedonian Academy of Arts and Sciences,
formation,” in Proc. IEEE Int. Symp. Electromagn. Compat., 2003, vol. Skopje. In 2006, he joined the Faculty of Electrical
2, pp. 1020–1023. Engineering and Information Technologies, Univer-
[6] I. Sarajcev, M. Majstrovic, and R. Goic, “Determining currents of cable sity Ss. Cyril and Methodius, where he is currently
sheaths by means of current load factor and current reduction factor,” an Associate Professor in the Department of Power Systems. His research
presented at the IEEE Power Tech Conf., Bologna, Italy, 2003. interests includes software design and modeling in power system analysis under
[7] L. M. Popovic, “Determination of the reduction factor for feeding cable steady-state and transient conditions, as well as the application of optimization
lines consisting of three single-core cables,” IEEE Trans. Power Del., methods.
vol. 18, no. 3, pp. 736–743, Jul. 2003.
[8] Y. X. Shi, H. Q. Niu, Y. Zhang, Z. P. Yi, X. B. Wang, and Y. G. Liu, “A
new method for evaluating fault reduction factor in HV cable system,”
in Proc. IEEE/Power Energy Soc. Transm. Distrib. Conf. Exhibit,: Asia Risto Ackovski (M’95) was born in Skopje, Mace-
Pacific, Dalian, China, 2005, pp. 1–5. donia, on February 14, 1950. He received the B.Sc.
[9] D. Rajicic and R. Ackovski, “Modeling underground lines as compo- degree in electric power engineering from University
nents of distribution network grounding systems,” in Proc. Comput. Ss. Cyril and Methodius, Skopje, in 1973, the M.Sc.
Tool, EUROCON , Belgrade, Serbia, 2005, vol. 2, pp. 1255–1258. degree in electric power engineering from University
[10] L. M. Popovic, “Improved analytical expressions for the reduction of Belgrade, Belgrade, Serbia, in 1978, and the D.Sc.
factor of feeding lines consisting of three single-core cables,” Eur. degree in electric power engineering from University
Trans. Elect. Power, vol. 18, pp. 190–203, 2008. Ss. Cyril and Methodius in 1990.
[11] L. M. Popovic, “Influence of metal installations surrounding the Since 1974, he has been with the Faculty of Elec-
feeding cable line on the ground fault current distribution in supplied trical Engineering and Information Technologies,
substations,” IEEE Trans. Power Del., vol. 23, no. 4, pp. 2583–2590, University Ss. Cyril and Methodius, where he is
Oct. 2008. currently Full Professor and Chair of the High Voltage Laboratory. His areas of
[12] J. R. Carson, “Wave propagation in overhead wires with ground re- interest include power system analysis, power system reliability, power quality,
turn,” Bell Syst. Tech. J., vol. 5, pp. 539–554, 1926. and high voltage.