EVALUATION OF GROUNDING RESISTANCE AND

ITS EFFECT ON UNDERGROUND CABLE SYSTEMS

Andreas I. Chrysochos1*, Lysimachos A. Rozik1, Ioannis Dolianitis1,
Dimitrios Chatzipetros1, Theofilos A. Papadopoulos2, Georgios Georgallis1
1
Hellenic Cables, Athens, Greece
2
Department of Electrical and Computer Engineering, Democritus University of Thrace, Xanthi, Greece
*achrysochos@hellenic-cables.com

Keywords: ATP/EMTP SOFTWARE, CABLE SYSTEM, EARTH POTENTIAL RISE, FINITE ELEMENT
METHOD, GROUNDING RESISTANCE

Abstract
This paper presents a complete methodology for the calculation of the grounding resistance and its effect on the induced sheath
voltages and earth potential rise that a cable system may encounter under fault conditions. First, the derivation of the two-layer
soil model is performed by using a dataset of measured apparent resistivity. Then, the resistance of the grounding system in
cable applications is calculated by employing different methodologies, i.e. analytical, numerical and finite element methods.
The accuracy of these methods is further assessed, while their results are also compared with field measurements. Finally, given
the calculated or measured grounding resistance, the induced sheath voltages and earth potential rise are calculated via
electromagnetic transients software and analytical formulae. Different fault scenarios are investigated and conclusions are drawn
regarding the operation of the cable sheath bonding system.

1 Introduction calculated by employing numerical [9] and finite element

method (FEM) techniques [10], which are proved to be more
In power systems analysis and substation design, the versatile and accurate than any analytical formulae. In
evaluation of earth potential rise (EPR) in terms of step, touch addition, these methods are also compared and validated with
and mesh voltages is of paramount importance for the field measurements of the grounding resistance. Finally, given
personnel safety [1], [2]. However, the evaluation of the the calculated or measured grounding resistance, the induced
grounding resistance is also essential for the integrity and sheath voltages and EPR are calculated via electromagnetic
secure operation of the cable sheath bonding system under transients (EMT) software [11] and analytical formulae [12].
fault or transient conditions and the associated sheath voltage Different fault scenarios are investigated and conclusions are
limiters (SVLs) when subject to EPR [3], [4]. drawn regarding the impact of the grounding resistance on the
operation of the cable sheath bonding system and the
The grounding systems in cable applications commonly dimensioning of the SVLs.
consist of single rods, arrays of rods, grounding grids and
combinations of all previous types. The calculation of the 2. Soil Characterization
grounding resistance for that system requires first the
determination of soil resistivity and then the application of a The adopted methodology presented in this section consists of
robust and accurate calculation scheme [2]. Different two major steps; the parameter calculation of the equivalent
analytical methods have been proposed in the literature that two-layer soil model and the evaluation of the grounding
depend on the type of the grounding system and the soil resistance.
structure, either uniform or multilayer [5]-[7]. However, these
methods show significant differences between their results, 2.1 Derivation of the Two-Layer Soil Model
especially when used in multilayer soils. Such deviations
would eventually lead to different results in the calculation of Although the ambient soil may contain a uniform resistivity to
induced sheath voltages and EPR in cable systems. a significant depth, it is a common practice to adopt models of
soil that are stratified. Among them, the two-layer soil model
This paper presents the steps for a complete evaluation of the of Fig. 1 is frequently used, which is also suggested by various
grounding resistance as well as the overall procedure for the international standards [1], [2].
calculation of the induced sheath voltages and the EPR that a
cable system with special bonding scheme may experience. Within this context, the soil is characterized by an upper layer
Following the state of the art and the latest technical reports, of definite depth h [m] and resistivity ρ1 [Ωm], and a lower
the methodology starts with the derivation of the two-layer soil layer of infinite depth and resistivity ρ2 [Ωm]. A resistivity
model from the dataset of measured apparent resistivity using determination using the well-known Wenner method results in
the iterative least-square scheme proposed by the international a measured apparent resistivity 𝜌𝑎𝑚 [Ωm] that is a function of
standards [2], [8]. Then, the grounding resistance of the grid is the electrode separation a [m]:

1
 𝑉 [6], [7], which exhibit the highest accuracy among all
4𝜋𝑎 ( 𝑎 )
𝐼𝑎 proposed expressions:
𝜌𝑎𝑚 = (1)
2𝑎 𝑎
1+ 2 − 2
√𝑎 + 4𝑏 2 √𝑎 + 𝑏 2  The rod lies entirely in the upper soil layer, i.e., 𝑙 < ℎ:
∞
where Ia [A] the injected current, Va [V] the resulting voltage 𝜌1 2𝑙 2𝑛ℎ + 𝑙
measured by the voltmeter, and b [m] the probe length. 𝑅𝑔 = ln + ∑ (𝐾 𝑛 ln ) (5)
2𝜋𝑙 𝑟 2𝑛ℎ − 𝑙
𝑛=1
Ia
 The rod lies in both soil layers, i.e., 𝑙 ≥ ℎ:
Va

𝜌2 2𝑙𝑒
b 𝑅𝑔 = ln (6)
2𝜋𝑙𝑒 𝑟
h
𝜌2
𝑙𝑒 = 𝑙2 + 𝑙1 (7)
Upper layer (ρ1) 𝜌1

Lower layer (ρ2) where l1, l2 [m] the length of rod within the upper and the lower
soil layer, respectively.
Fig. 1: Two-layer soil model with measurement of apparent
Numerical Approach: The grounding grid is divided into a
resistivity via the Wenner method.
suitable number of n elements, where the partial element
Each measurement 𝜌𝑎𝑚 shall also satisfy the following equivalent circuit (PEEC) method is applied [9]. Assuming
expression, as computed by the two-layer model [2]: only the resistive coupling between those elements through the
soil, the self and mutual potential coefficient of and between
elements are calculated. Then, the resistive coupling is
∞
4𝜌1 𝐾 𝑛 4𝜌1 𝐾 𝑛 obtained by the following set:
𝜌𝑎𝑚 ≅ 𝜌𝑎𝑐 = 𝜌1 + ∑ − (2)
2 2
𝑛=1 √1 + (2𝑛ℎ ) √4 + (2𝑛ℎ) 𝑅𝑝 𝐼𝑙 = 𝑉𝑝 (8)
( 𝑎 𝑎 )
where Rp [Ω] the matrix of potential coefficient of order 𝑛 ×
where K the reflection coefficient:
𝑛, Il [Α] the vector of leakage currents of order 𝑛 × 1, and Vp
𝜌2 − 𝜌1 [V] the vector of potentials of order 𝑛 × 1.
𝐾= (3)
𝜌1 + 𝜌2
In the above formulation, the grounding grid may include m
The unknown variables ρ1, ρ2 and h of the two-layer soil model distinct electrodes and each of them is assumed equipotential
are found by solving (1)-(2) in a least-square sense, but in general with a different potential from the others. As a
minimizing the normalized objective function [2], [8]: result, the linear system of (8) contains n equations and 𝑛 + 𝑚
unknowns, namely n leakage currents and m potentials. In
∞ 𝑚 𝑐 2 order to set the system as determined, a second set of m
𝜌𝑎,𝑖 − 𝜌𝑎,𝑖
𝑓(𝜌1 , 𝜌2 , ℎ) = ∑ ( 𝑚 ) (4) equations is applied to each distinct electrode [9]:
𝜌𝑎,𝑖
𝑖=1
∑ 𝐼𝑒 = 𝐼𝐸 (9)
where the sum spans all the available measurements.

2.2 Calculation of the Grounding Resistance where IE [A] the total leakage current.

The grounding resistance depends not only on the adopted soil Eventually, the linear system is solved, calculating the leakage
model but also on the examined grid configuration. Several current for each element and the potential for each distinct
closed but approximate formulae have been proposed for electrode. From these, the resistance to earth of each electrode
various simplified cases [5]-[7]. For more complex grids, is calculated:
numerical and FEM methodologies are typically employed,
providing higher accuracy. 𝑉𝐸
𝑅𝐸 = (10)
𝐼𝐸
Analytical Formulae: The calculation of ground resistance
Rg [Ω] for a single rod is first presented for validation where VE [V] the associated EPR.
purposes. The rod with length l [m] and radius r [m] is buried
in a two-layer soil model. Its grounding resistance can be given The total grounding grid resistance finally results from the
approximately by the following two analytical formulae superposition of the leakage current distribution of all grid
elements [9].

2
FEM: The computational scheme of FEM is based on 3 Calculation of Grounding Resistance
Maxwell’s equations, explicitly formulated by the
electromagnetic problem that is to be solved [10]. Considering 3.1 Apparent Soil Resistivity and Two-Layer Soil Model
purely resistive coupling, FEM computes electric field,
Table 1 summarizes the measurements conducted according to
current, and potential distributions in conducting media where
the guidelines of [2], in two different sites, namely site 1 and 2.
inductive effects are assumed negligible. Specifically, the
They are located in north Germany and south Greece,
stationary problem can be described with the following
respectively. Results show that, apart from the first two
equations:
measurements on site 1, the apparent resistivity is a monotonic
∇∙𝑱=0 (11) function of distance in both cases. This indicates that a two-
layer soil model would be adequate for modeling purposes. In
𝑱 = 𝜎𝑬 + 𝑱𝑒 (12) particular, in site 1 its value decreases with increasing distance,
revealing that the resistivity of upper layer in the two-layer soil
𝑬 = −∇𝑉 (13) model shall be higher than the lower layer, i.e. 𝜌1 > 𝜌2 . On the
other hand, the apparent resistivity of site 2 increases with
where J [Am-2] the current density, σ [Sm-1] the material respect to distance, indicating that 𝜌1 < 𝜌2 [8].
conductivity, E [Vm-1] the electric field, Je [Am-2] the
externally imposed current density and V [V] the voltage Table 1 Measurements of apparent resistivity in both sites
potential which is also the dependent variable to be solved for.
Distance Site 1 Site 2
As shown in the indicative 3D example of Fig. 2, the boundary a [m] 𝜌𝑎𝑚 [Ωm] 𝜌𝑎𝑚 [Ωm]
conditions include an arbitrary injected current IF [A] into the
grounding grid (orange color) and the electric insulation of the 0.5 833.2 54.7
top soil surface (blue color), which is realized via: 1 950.3 63.8
1.5 962.3 -
𝒏∙𝑱= 0 (14) 2 928.2 90.3
2.5 887.7 -
where n the normal unit vector. 3 853.8 105.0
3.5 804.7 -
The imposition of the zero potential at the remote earth (green 4 727.0 110.8
color) is somewhat more challenging, since the electric field 4.5 665.8 -
can be considered unbounded. In order to limit the extent of 5 613.5 -
the FEM model to a manageable region of interest with 7 392.3 -
reasonable execution time, a coordinate scaling is adopted to 10 157.6 -
layers of virtual domains surrounding the physical region of 15 40.5 -
interest [11], i.e., to soil as indicatively shown in Fig. 2. These 20 18.3 -
virtual layers can be mathematically stretched out towards 30 17.3 -
infinity, where 𝑉0 = 0 is imposed. As a result, the model is
computationally efficient, while the solution inside the region The measurements are further processed based on the
of interest is not affected by the artificial geometric methodology of Section 2.1 in order to derive the two-layer
boundaries. Eventually, the grounding resistance is evaluated earth model. Table 2 presents the results for the unknown
by the injected current and the solution of voltage potential. variables of (4), characterising the soil model for both sites.
n J=0
Fig. 3 shows the measurements and the corresponding best
least-square fit according to (2). It is evident that the two-layer
IF model captures accurately the dependence of apparent
resistivity to the rod distance. This would not be the case if a
uniform soil model was considered instead.

Table 2 Characterization of two-layer earth model in both sites

Variable Site 1 Site 2

V0 = 0
ρ1 [Ωm] 985.81 51.00
Fig. 2: Indicative FEM model with the imposed boundary ρ2 [Ωm] 14.26 141.79
conditions. Zero potential is practically assigned to infinity. h [m] 4.18 0.83

3
 is highly accurate yielding a maximum error of only 1.20 %,
which is observed for rod lengths close to the soil upper layer
depth. The error is maximized in that length due to the image
reflections of the stratified earth which are dominant as the rod
length tends to become equal to the soil upper layer depth.

It can be concluded that the numerical and FEM approaches

can be used in more complex geometries, preserving their high
accuracy.

3.3 Grounding Resistance of a Complex Grid

A complex grid typically used in cable joint pits is further
assessed in terms of grounding resistance. The grid geometry
is shown in Fig. 5, consisting of horizontally laid conductors
(blue lines) and vertically buried rods (red lines). The main
grid plane xy is buried at 2.1 m below soil surface, assuming
that both sites are represented by the soil models of Table 2. In
addition, the specific grounding grid has been also measured
Fig. 3: Measured apparent soil resistivity data and least-square at site 2, based on the guidelines of [2].
fit. (a) Site 1 and (b) site 2.

3.2 Grounding Resistance of a Single Rod

For validation purposes, the grounding resistance of a simple
single rod is first examined by comparing all methodologies Ø8 mm

described in Section 2.2. The rod is assumed to be buried in

both sites, which are represented by the corresponding two- 1.5 m
0.85 m
layer soil model of Table 2. Fig. 4 presents the grounding
Ø17 mm
resistance of the single rod with respect to length l. As
expected, the value of Rg decreases with increasing length.
z
Furthermore, an abrupt decrease is observed when the rod y
x
length exceeds the height h of the two-layer model. In site 1,
this occurs between the lengths of 4 m and 5 m, while in site 2 Fig. 5: Geometry of the complex grid.
between the lengths of 0.5 m and 1 m.
Table 3 summarizes the simulation results derived by the
numerical and FEM methodologies. The quarter symmetry of
the grid with respect to planes yz and zx is illustrated with
brown arrows in Fig. 5. This is further exploited in 3D FEM,
reducing its computational burden to ¼ of the original
problem. The field measurement at site 2 is also presented in
the table. Specifically, the extrema of multiple measurements
are reported along with the most representative value. Results
reveal a very good agreement between the numerical and FEM
methodologies with a maximum error of 0.57 %. In addition,
both models are in compliance with the field measurement of
site 2. Finally, from the resulting values at both sites, it can be
concluded that the characteristics of the two-layer soil have a
significant impact on the value of grounding resistance.

Table 3 Simulations and field measurements of complex grid

grounding resistance in both sites

Grounding resistance
Site 1 Site 2
Rg [Ω]
Fig. 4: Grounding resistance of single rod calculated by all
methodologies. (a) Site 1 and (b) site 2. Numerical 5.45 1.79
3D FEM 5.43 1.78
Results of the analytical formulae generally present Field measurement - 1.98 ∈ [1.81,2.25]
differences when compared to the other two methods.
Specifically, a maximum relative error of 9.15 % is found with
respect to 2D-axisymmetric FEM, which is considered as
reference method. On the other hand, the numerical approach

4
4 Application to Cable Systems Outer semi-conductive ρ 10
εr 2000
The 150 kV underground cable system of Fig. 6 is examined μr 1
in order to assess the impact of the grounding resistance on the Sheath ρ 2.84∙10-8
induced voltages and EPR, based on the recommendations εr 1
of [3], [4]. The circuit consists of a cross-bonded (CB) part μr 1
with minor sections of 1.2 km length each and a single-point Oversheath ρ ∞
bonded (SPB) section of 0.5 km length. The cables are single- εr 2.3
core XLPE insulated with 1000 mm2 copper conductor and μr 1
aluminium sheath, arranged in flat formation with spacing
S [m] equal to 3De, where De [m] the total cable diameter. The For the sake of simplicity, only the inductive mechanism of the
SPB section is also equipped with a copper earth continuity cable system is investigated by employing ATP/EMTP
conductor (ECC) of 150 mm2, placed at its optimum position, software [11]. As a result, the three phase conductors are
i.e., at distance 0.7S to centre phase. The ECC is transposed in excited by ideal current sources, while the phase-to-earth
the middle of the SPB section in order to cancel out the induced voltages and any source or load impedances are neglected.
electromotive force (EMF) and the resulting circulating Three-phase and single-phase short-circuit scenarios at
current under normal operation. Rest geometric and electrical substation 2 are investigated, assuming a fault current of
properties of cables and ECC are given in Tables 4 and 5, 31.5 kA in both cases.
respectively.
The grounding resistances Rg1 and Rg2 at substations 1 and 2 of
CB minor 1 CB minor 2 CB minor 3 SPB the cable circuit are equal to 0.5 Ω. At the transition between
the CB and SPB sections, the ECC of the SPB part as well as
Ia
the sheaths of the CB part are grounded at the joint pit via the
grounding resistance Rg3. This grounding grid follows the
Ib
VSE,SPB design of Fig. 5 and its resistance is deliberately varied
between 0.1 Ω and 20 Ω in order to investigate its impact on
Ic the induced sheath voltages and EPR in the examined cable
VSE,CB ECC R system.
Rg1 EPR Rg3 g2

Substation 1 Joint pit Substation 2 Results of the three-phase fault scenario are given in Fig. 7.
Fig. 6: Underground cable system with CB and SPB sections. Specifically, the EPR at both substations and the transition
joint pit are shown (green arrows of Fig. 6), as well as the
Table 4 Geometric properties of cables and ECC maximum voltage VSE [V] between the cable sheath and local
earth in both CB (red arrows) and SPB (blue arrows) sections.
Outer diameter It is evident that the influence of Rg3 on the EPR is significant
Layer Material
[mm] for the three-phase short-circuit scenario, in particular for the
EPR at that point. However, the EPR at the substations is also
Conductor CU 39.5 influenced due to the current distribution between sheaths,
Inner semi-conductive Semi-PE 42 ECC and earth return path which is changing with varying Rg3.
Insulation XLPE 76 On the other hand, the calculated maximum VSE, which is a
Outer semi-conductive Semi-PE 78.5 crucial value for the dimensioning of SVLs [3], is much less
Sheath AL 80.5 dependent on Rg3 than the EPR. Results in Fig. 7b also include
Oversheath HDPE 89 the corresponding calculations based on IEEE simplified
Burial depth equal to 1.5 m on top of cables formulae for both sections [12]. It can be concluded that these
ECC insulation thickness of 8.5 mm formulae are in very good agreement with ATP/EMTP for the
SPB part but give an optimistic estimation for the CB part.
Table 5 Electrical properties of cables and ECC (if applicable)
in SI units Results for the single-phase fault scenario are given in Fig. 8.
Similar to the three-phase short-circuit scenario, the influence
Layer Variable Value of Rg3 on EPR is notable, in particular at the transition joint pit
itself. The voltage level significantly exceeds the
Conductor ρ 2.16∙10-8 corresponding of Fig. 7 due to the return path of the short-
εr 1 circuit current. The maximum VSE shows again less
μr 1 dependency with Rg3, as in the case of three-phase short-circuit
Inner semi-conductive ρ 10 scenario. In addition, the IEEE formulae for the SPB part in
εr 2000 this case gives a good conservative estimation when compared
μr 1 to the corresponding results of ATP/EMTP. For the CB part, it
Insulation ρ ∞ is noted that the IEEE standard does not provide formulae for
εr 2.5 the calculation of VSE.
μr 1

5
 geometries, presenting high accuracy in the calculation of
grounding resistance.

In the last step, the induced sheath voltages and EPR are
calculated via EMT software, based on the derived grounding
resistance in joint pit. The magnitude of EPR is affected by the
grounding resistance of the transition pit, highlighting the
importance of accurate simulation and frequent field
measurement of the latter. The assessment of EPR is important
for the secure operation of the cable system, especially in cases
of single-phase faults. The cable sheath induced voltages,
which are important for the dimensioning of the SVLs, are less
affected by the grounding resistance of the transition pit. The
maximum voltage level for the CB part is found in the case of
three-phase short-circuit fault, while the corresponding for the
SPB section is observed for the single-phase short-circuit fault.
Approximate formulae by the international standards are
generally in good agreement, although there is no standardized
Fig. 7: Three-phase short-circuit scenario. (a) EPR at both formulation for the case of CB under single-phase fault
substations and the transition joint pit. (b) Maximum voltage conditions. In addition, caution should be taken in the results
between sheath and local earth in CB and SPB sections via of the CB section, since they may lead to more optimistic
ATP/EMTP software and IEEE formulae. voltage, as the present paper demonstrates.

6 References
[1] IEEE80: 'Guide for safety in AC substation grounding',
2013.
[2] IEEE81: 'Guide for measuring earth resistivity, ground
impedance, and earth surface potentials of a grounding
system', 2012.
[3] CIGRE283: 'Special bonding of high voltage power
cables', 2005.
[4] CIGRE347: 'Earth potential rises in specially bonded
screen systems', 2008.
[5] Schwarz, S. J.: 'Resistance of grounding systems', Trans.
Amer. Inst. Elec. Engrs., 1954, 73, pp. 1010-1016.
[6] Nahman, J., and Salamon, D.: 'Analytical expression for
the resistance of rodbeds and of combined grounding systems
in nonuniform soil', IEEE Trans. Power Del., 1986, 1, (3),
pp. 90-96.
[7] Thug, G. F.: 'Earth resistances' (Georges Newnes
Limited, 1964).
Fig. 8: Single-phase short-circuit scenario. (a) EPR at both [8] L. del Alamo, J.: 'A comparison among eight different
substations and the transition joint pit. (b) Maximum voltage techniques to achieve an optimum estimation of electrical
between sheath and local earth in CB and SPB sections via grounding parameters in two-layered earth', IEEE Trans.
ATP/EMTP software and IEEE formulae. Power. Del., 1993, 8, (4), pp. 1890-1899.
[9] Ruehli, A. E., Antonini, G., and Jiang, L.: 'Circuit
5 Conclusion oriented electromagnetic modelling using the PEEC
techniques' (Wiley-IEEE Press, 2017).
This paper deals with the evaluation of the grounding
[10] Katsanou, V. N., and Papagiannis, G. K.: 'Comparison of
resistance calculation and its effect on the induced cable sheath
substation grounding system design methods using a FEM
voltages and EPR under fault scenarios.
approach'. Proc. IET MedPower, Athens, Greece, 2014,
The methodology starts with the derivation of the two-layer pp. 1-6.
soil model from the dataset of measured apparent resistivity. [11] Chrysochos, A. I., et al.: 'Capacitive and inductive
The iterative least-square scheme proposed by the coupling in cable systems – Comparative study between
international standards is found to be quite robust for the calculation methods', Proc. JICABLE, Paris, France, 2019,
calculation of the unknown variables. After the extraction of pp. 1-6.
the two-layer soil model, the grounding resistance can be [12] IEEE575: 'Guide for bonding shields and sheaths of
calculated based on numerical or FEM methodology. Both single-conductor power cables rated 5 kV through 500 kV',
methods are demonstrated in simple and complex grid 2014.