Académique Documents
Professionnel Documents
Culture Documents
Electrical Engineering
Archiv fr Elektrotechnik
ISSN 0948-7921
Volume 96
Number 3
Electr Eng (2014) 96:255-265
DOI 10.1007/s00202-013-0291-9
1 23
1 23
ORIGINAL PAPER
Received: 13 July 2011 / Accepted: 10 November 2013 / Published online: 30 November 2013
Springer-Verlag Berlin Heidelberg 2013
strips forming two-dimensional grids. The grounding resistance, which represents one of the most important parameters
of the grid, considerably affects the earth potential rise during earth faults. If the grounding resistance is high, personnel may be killed or injured and equipment damaged. Therefore, the grounding resistance should be estimated in an early
design phase to conceive the main geometrical parameters of
the grid. This justifies the efforts made in the past to develop
relatively simple formulas for assessing the grounding resistance of typical grounding systems including grids. Simple
methods and the designer-oriented formulas for the calculation of the grid grounding resistance are given in standards
[1] and [2]. In addition, many papers proposed simple methods and formulas for the calculation of the grounding resistance of complex grounding systems buried in uniform or
non-uniform soil. For example, the calculation of the grounding resistance of a rectangular grid installed in uniform soil
was presented in [3]. Grids and rods buried in two-layer soil
were analysed in [46] and [7], respectively. The foundation
grounding systems, which represent grounding grids encapsulated in concrete, were considered assuming uniform soil
in [810]. More complex geometries were also analysed (e.g.
foundation grounding systems surrounded by two-layer soil
[11,12], foundation grounding systems with external loops
and rods [13] and foundation grounding systems with external grids [14]). All of those methods and the derived formulas
are based on the Laplace solution of the problem, assuming
perfect contact surface between the grounding grid electrodes
and the surrounding soil. However, it was noticed in practice [15] that such formulas could not satisfactorily be used
for the calculation of the grid grounding resistance in soils
where the contact surface is significantly smaller than the surface of the electrodes (e.g. karst and sandy terrains). In such
cases, calculated grounding resistances several times lower
than the actual ones are frequently obtained. Such deviations
123
can cause dangerous situations during earth faults. It was suggested in [15] that the big difference between the measured
and computed grounding resistances is caused by a bad contact between the grounding electrodes and the surrounding
soil. However, this physical phenomenon was not theoretically investigated in [15], because it was impossible to model
a bad contact with mathematical techniques and calculation
tools available at that time.
The development of the finite-element method (FEM)
and the corresponding software, as well as improvements
of the performance of PCs, enabled the application of threedimensional (3D) finite-element models based on small size
elements surrounding the grounding electrodes. FEM is a
numerical technique for finding approximate solutions to
boundary condition problems. The method is based on the
division of the problem domain into a set of small subdomains, named finite elements, to accurately represent
its complex geometry and dissimilar material properties.
Each finite element is represented by a set of equations
related to the problem. The element equations are simple
(for steady-state problems they represent algebraic equations), locally approximating the original complex equations (often partial differential equations). All sets of element equations are systematically recombined into an overall system of equations, which can be solved using standard
techniques. A mathematical basis for the FEM was given in
[16], while the explanations for its use in electromagnetics
and electrical engineering were presented in [17] and [18],
respectively.
During the past decade, FEM has been used for the calculation of the grid grounding resistance in a number of cases
[1922]. The advantages of FEM, compared to the conventional methods for the calculation of the grid grounding resistance, are a simple representation of the total solution and
capture of local effects. It also possesses an advantage of not
having any limitations regarding the shape and size of the
grid, as well as the soil structure. For the calculation of the
electrical properties of grounding grids using FEM a reasonable level of segmentation is sufficient for practical purposes,
and an increased number of elements is needed only if highly
accurate results are requested [23]. Consequently, considerably smaller computational effort is required for practical
purposes.
In order to develop a method for precise predictions of the
grounding resistance of loops (grids) laid in soils that prevent a good contact between the grounding electrodes and
the surrounding soil, it is essential to theoretically investigate and gain understanding of the electrical behaviour of
the grounding loops installed in such soils. The aim of the
research presented in this paper was estimation of the influence of imperfect contact between the grounding electrodes
and the surrounding soil on the loop grounding resistance
and potential distribution in the soil during earth fault. To the
123
257
Fig. 2 The final finite-element model for the grounding loop under
consideration
Fig. 1 A grounding loop buried in a two-layer soil
where
(2)
J = E/
(3)
= N e ,
(4)
(1)
is the potential of a point within the finite element, including the points on the lateral faces,
e is the column vector of potentials of finite element representative nodes,
N is the correlation matrix depending on the type of a finite
element,
E is the electrical field vector,
J is the current density vector, and
is the electrical resistivity within the finite element.
It is assumed that the grounding resistance only depends
on the soil structure and the loop geometry. Accordingly,
the grid potential can be arbitrary. The following boundary
conditions are adopted for the considered case:
e0 = 200 V
(5)
e inf = 0 V
(6)
E 1t = E 2t
(7)
E 1n /1 = E 2n /2
(8)
where
e0 is the potential of the grounding loop,
einf is the potential of the external boundaries of outer
subdomains made of the infinite elements,
123
Fig. 3 The finite-element mesh close to the strip of the grounding loop
(close-up of a part of the model, containing one eighth of the loop)
E 1t , E 2t and E 1n , E 2n are the tangential and perpendicular components of the electrical field vector at the
boundary surface between the adjacent finite elements, respectively, and
1 and 2 are electrical resistivities of the material
enclosed by the adjacent finite elements.
Using Eqs. (1)(8), the potentials of all finite-element
nodes, as well as the current density at arbitrary location
of the model volume, can be calculated. This makes it possible to determine the total current, I , dissipating from the
grounding loop:
J d S
(9)
I =
e0 e inf
.
I
(10)
123
259
a(103 m)
14.85
Yes
50
21.07
Yes
50
19.96
C1
Yes
12
50
19.29
M4
C1
Yes
80
50
16.92
M5
C1
No
50
20.97
M6
C1
No
50
19.48
M7
C1
No
12
50
18.97
M8
C1
No
12
50
19.01
M9
C1
No
16
12
50
19.06
Model
Concept
of placing
air gaps
M0
C1
M1
C1
M2
C1
M3
R()
M10
C1
Yes
16
80
50
17.38
C1
Yes
16
80
80
23.92
M12
C1
Yes
16
80
92
37.25
M13
C1
Yes
16
80
95
49.12
M14
C1
Yes
16
80
96
56.64
M15
C2
16
27.20
60
23.68
M16
C2
16
13.60
80
30.32
M17
C2
16
6.80
90
42.16
M18
C2
16
4.76
93
52.06
M19
C2
16
3.40
95
64.50
where T is determined by
P
n
F(%)
M11
T =
d(103 m)
(12)
a
100,
p
(13)
123
123
(14)
where
R0 is the basic (theoretical) value of the loop grounding
resistance, representing the soil resistance between
the whole loop surface and the remote earth,
Riv is the resistance of the soil in an immediate vicinity of the loop electrodes (up to a distance d from
the electrodes), assuming perfect contact between the
electrodes and the surrounding soil; it can be approximated as
Riv =
upper d
,
pP
(15)
Rbc is the resistance of the soil in an immediate vicinity of the loop electrodes (up to a distance d from
the electrodes), assuming a bad contact between the
electrodes and the surrounding soil; it can be approximated as
Rbc =
upper d
, and
p P (100 F)/100
(16)
(17)
p P (100 F)
(18)
261
on the development of a method which will solve this problem. Nevertheless, the presented results show that the contact
resistance can have major influence on the loop grounding
resistance. In such cases the use of backfill materials, such as
bentonite [15], is recommended, because they significantly
reduce the contact resistance in karst and sandy soils.
4.2 Potential distribution in the soil
As stated in Sect. 1, using FEM not only the grounding
resistance of complex grounding systems can be calculated,
but also the potential at any point of the model volume can
be determined. This offers an opportunity to perform further analysis to obtain deeper understanding of the electrical
behaviour of grounding loops. Figure 6, serving as an illustration, displays the potential distribution in the soil obtained
for model M10.
Once the potentials at all points of the model volume are
determined, potential distribution over any plane or along
arbitrary line within the model volume can be obtained. For
the considered case, the potential distribution over x y plane
located at the depth of 0.5 m gives the best insight into the
electrical behaviour of a grounding loop buried in the soil
characterised by a bad contact with the electrodes (x and y
coordinates correspond to the coordinate system defined in
Fig. 1). Figure 7 shows the comparison of the two potential
distributions over the considered x y plane: the first is related
to the case of perfect contact (Fig. 7a, model M0), and the
second to the case of an imperfect contact characterised by
F = 50 % (Fig. 7b, model M10).
According to the diagrams shown in Fig. 7, it appears
that imperfect contact does not significantly influence the
123
123
analysed for various values of the parameter F. The potential distributions along the straight line between the points
(x, y) = (2.535 m, 0 m) and (x, y) = (2.535 m, 1 m),
placed in the considered xy plane, are presented in Fig. 9 for
models M0 and M10M14 (x and y coordinates correspond
to the coordinate system adopted in Fig. 1).
According to the diagram shown in Fig. 9, the parameter F significantly influences the fall of potential in the
soil in an immediate vicinity of the grounding electrodes,
which affects the potential distribution at the ground surface, relevant for the calculation of touch and step voltages.
The diagram shown in Fig. 10 contains curves representing
the potential distributions along the line between the points
(x, y) = (0 m, 0 m) and (x, y) = (8 m, 0 m) belonging to
the ground surface (z = 0 m), related to models M0 and
M10M14. Figure 11 presents the dependence of the maximum potential along the considered line at the ground surface
(max ) from the parameter F. The maximum potential varies
from 140.8 V (valid for F = 0 %) to 37.2 V (F = 96 %).
The diagrams shown in Figs. 10 and 11 illustrate how
the effective contact surface between the grounding elec-
263
that the touch voltage, which represents the potential difference between the ground potential rise and the ground
surface potential at a point where a person is standing while
at the same time having a hand in contact with a grounded
structure, considerably depends on the effective contact surface. The step voltages, representing the difference in ground
surface potentials experienced by a person bridging a distance of 1 m, show similar dependence. Consequently, in
soils in which the contact resistance represents the dominant
component of the total grounding resistance, the touch and
step voltagesthe parameters that indicate the quality of the
grounding systemcannot accurately be predicted using the
standard methods (derived assuming a perfect contact surface). Therefore, a new method is needed for such cases to
enable precise predictions of the grid grounding resistance
and touch and step voltages. The method should provide the
designer a possibility to take into account the variations of
the effective contact surface, as well as the effects of the use
of backfill materials. The research presented in this paper,
along with the research presented in [30], represents a theoretical basis for the development of such a method, and
our present attempts are focused on the achievement of that
goal.
5 Conclusions
Fig. 8 Potential distributions in the soil over x y plane at a depth of
0.5 m and in the vicinity of the grounding electrodes [models M0 (a)
and M10 (b)]
trodes and the surrounding soil affects the potential distribution on the ground surface during earth fault. It is obvious
123
123
cannot satisfactorily be predicted using the standard engineering methods, based on the Laplace solution of the problem, assuming perfect contact surface.
The method presented in this paper can successfully be
used for the analysis and explanation of the deviation of the
measured values of the grounding resistance from those calculated using the standard engineering methods. However,
it cannot be used for precise predictions of either the loop
(grid) grounding resistance or the potential distribution at
the ground surface for soils causing imperfect contact with
electrodes. Our present attempts are focused on the development of a method intended for the practitioners, which will
enable to take into account the influence of imperfect contacts, as well as the use of backfill materials. The research
presented in this paper, along with the research presented in
[30], represents a theoretical basis for the development of
such a method.
Acknowledgments This research was partially supported by the Ministry of Education and Science of Serbia (project TR 36018).
References
1. (1986) Guide for safety in AC substation grounding. ANSI/IEEE
Std 80
2. (2000) Guide for safety in AC substation grounding. ANSI/IEEE
Std 80
3. Nahman J, Skuletic S (1979) Resistances to ground and mesh voltages of ground grids. Proc IEE 126:5761
4. Salama MMA, Elsherbiny MM, Chow YL, Kim KC (1995) Calculation and interpretation of a grounding grid in two-layer earth with
the synthetic-asymptote approach. Electr Power Syst Res 35:157
165
265
17. Jin J (2002) The finite element method in electromagnetics. Wiley,
New York
18. Silvester PP, Ferrari RL (1996) Finite elements for electrical engineers. Cambridge University Press, Cambridge
19. Colominas I, Gomez-Calvino J, Navarrina F, Casteleiro M (2002)
A general numerical model for grounding analysis in layered soils.
Adv Eng Softw 33:641649
20. Nahman J, Paunovic I (2006) Resistance to earth of earthing grids
buried in multi-layer soil. Electr Eng 88:281287
21. Gemes JA, Hernando FE (2004) Method for calculating the
ground resistance of grounding grids using FEM. IEEE Trans
Power Deliv 19:595600
22. Gemes-Alonso JA, Hernando-Fernndez FE, Rodrguez-Bona F,
Ruiz-Moll JM (2006) A practical approach for determining the
ground resistance of grounding grids. IEEE Trans Power Deliv
21:12611266
23. Colominas I, Navarrina F, Casteleiro M (1999) A boundary element numerical approach for grounding grid computation. Comput
Method Appl M 174:7390
24. Wenner F (1916) A method of measuring earth resistances. Bull
Natl Bureau Stand 12:469482 (Report no. 258)
25. Dawalibi F, Blattner CJ (1984) Earth resistivity measurement interpretation techniques. IEEE Trans Power Appl Syst 103:374382
26. (1983) Guide for measuring earth resistivity, ground impedance,
and earth surface potentials of a ground system. ANSI/IEEE Std
81
27. Nahman J, Paunovic I (2007) Effects of the local soil nonuniformity upon performances of ground grids. IEEE Trans Power Deliv
22:21802184
28. Nahman J, Paunovic I (2010) Mesh voltages at earthing grids buried
in multi-layer soil. Electr Power Syst Res 80:556561
29. Zienkiewicz OC, Emson C, Bettess P (1983) A novel boundary
infinite element. Int J Numer Method Eng 19:393404
30. Trifunovic J (2012) The algorithm for determination of necessary
characteristics of backfill materials used for grounding resistances
of grounding loops reduction. J Electr Eng 63:373379
123