You are on page 1of 7

GROUNDELECTRODERESISTANCEMEASUREMENTSINNONUNIFORMSOILS

Abstract - Ground electrode resistance measurements are usually done by Fall


of Potential method. The potential probe position (0.618 rule) has been
calculated for homogeneous soil 1,2]. Theoretical computation and
experimental verification of the potential probe position in non homogeneous
soil is presented. The experimental work was carried out on a twin layer
laboratory model [3,4,6].

INTRODUCTION
The ground electrode resistance is measured by two methods. In the classical
fall of potential method (fig. 1-a & 1-c) the potential probe is between the
ground and current electrodes. In the alternative fall of potential method (fig. 1-
b & 1-c) the potential probe is exterior and in line to the ground and current
electrodes. The 0.618 rule for the potential probe position in homogeneous soil
has been verified in the field [1].
When earth is non uniform, it has been recommended that the true resistance
of the ground electrode is given by the zero slope region on the fall of potential
curve [ 1,51].
An analytical study for the correct position of the potential probe for normal
ground electrode is presented. The theoretical calculations are verified on a
twin layer laboratory model [3,4,6].

THEORETICAL STUDY
The theoretical study is carried out in two steps. In the first step, the influence
of electrode geometry is neglected and in the second, its influence is
considered.
1st step
Small ground electrode: In this case, the following assumptions are made:
a) Metallic resistivity of the electrode is neglected.
b) Electrode is equivalent to an hemisphere whose radius is r.
c) The distance d between the ground and current electrode is large enough so
that the ground electrode dimensions do not modify the mutual resistance
value given by the point source electrode assumption. A reasonable value is a
ratio r/d < 1/20 [ 7] .
d) Radius r is small with respect to first layer thickness h, so that point source
electrode assumption is still valid in the two layer soil. A value r/h < 1/5 is
suggested.
e) The ground electrode is buried in the first layer only.
f) Uniform current density in the electrode is assumed.
g) All the impedances have only a resistive compbnent. The reactive
component being negligible in steady state conditions and at industrial
frequencies.
Let's consider the electrode configuration of figure 2, we have [2] (see
Appendix I):

If now three electrodes are involved (fig. 1-a) the following holds (see Appendix
I):

The true resistance of electrode a will be given by equation 2, if we can achieve


the following:

Equation 4 determines the required potential electrode position x0.

The mutual resistance Rij between two electrodes i and j, depends on earth
layer structure and electrode shape. Calculations become more and more
tedious with increase in complex structures. However, with large spacing d and
relatively small ground electrodes such mutual resistances can be evaluated in
a two layer soil (fig.3). Appendix II shows how such calculations can be done.
We have then:

If now we write:

If the values of equation 5 are substituted in equation 4, the resulting equation


becomes:

Equation (8) can be written in a symbolic but convenient form:

The required potential electrode position is then:

Figures 4 and 5 give the computed values of x0 as a function of H for the


classical and alternative potential probe positions respectively.
RESULTS AND DISCUSSION
Analysis of figures 4 and 5 show clearly that the required potential probe
position x0 depends considerably on the following parameters:
- reflection factor K
- ratio H = height of first layer-distance between ground and current electrodes
- physical position of electrode b with respect to a and c.

On the basis of this theoretical study the following conclusions hold:


1. When K = 0 (uniform earth) xo = 0.618. This is confirmed by previous
analysis [1,2,7].
2. When K is positive and increases gradually to + 1 the potential electrode
must be put closer to the current electrode. This means x0 increasing from
0.618 for the classical electrode position and decreasing from 0.618 for the
alternative electrode position.
3. In the case of negative values of K potential electrode must be put further
from the current electrode.
4. In both cases the deviation from 0.618 increases with the absolute value of
K. For values of H between 0.05 and 5.0 the deviations are quite rapid.
5. With the alternative fall of potential method, the potential electrode position
may vary practically between zero to one for high values of K. This makes the
classical fall of potefitial method prefetable since such position is bounded by
approximately 0.50 and 0.90. Note that in this case for positive values of K the
deviation is greater than for negative values. This confirms that high resistivity
second layer may have very adverse effects on ground resistance
nieasurement if its presence is ignored.
6. Finally if H is greater than 5.0, which means a very thick layer of resistivity
ro1, x0 approaches 0.618. This is a logical conclusion since the earth
approaiches urtiformity. For small values of H (0.01 and less) however, the
uniform condition (earth With resistivity ro2) is reached by the classical fall of
potential tiethod only. The alternative gives constant values of xo but different
from 0.618. (Even for very thin first layer).
Although these concluisions are for relatively small electrodes, they show
nonetheless that great care must be taken when measuring ground resistance
in non uniform earth. The curves of figures 4 and 5 may be of great help in
determining an order of magnitude for xo. When earth structure is more than
two layers a reasonable equivalent two layer soil must be chosen before using
such curves [ 2 ].
With large ground electrodes d c,an not be large enough with respect to
electrode dimension arid first layer height will be often relatively small. More
than that, non uniform current density may be expected in the electrode due to
its shape and its proximity to the current electrode. Hence a more elaborate
analytical study becomes necessary; The following section shows how such an
analytical study can be carried out.

2nd step

Effect of electrode geometry:


Influence of electrode size and shape in a uniform earth has been studied by
Curdts [7]. His conclusions were that if r/d < 1/10, the potential probe position,
is still very close to 0.62. For r/d = 1/2 this position must be increased to 0.70.
In the case of buried wire or driven rod, he shows that potential electrode
position decreases slowly with increased electrode length (down to 0.55 if wire
length is 1.30d).
In the case of two layer earth, the effect of electrode configuration will be a
complicated function of d, k, n, gama. In this case only assumptions a, e & g
are necessary.
Consider a complex ground electrode buried in the first layer of a two layer
earth as shown itn figure 6. Since potential is an additive scalar quantity, the
potential due to current 'I' flowing from the electrode at any point at the earths'
surface is the sum of potentials due to each linear part of the electrode
carrying a fraction delta i of the current I.
The electrode is divided in such a way that the linear element length is small
with respect to h and d, and the factor delta i is chosen such that:
1) The closest linear eiement to the current electrode has the highest delta i
2) delta i will also depend on the electrode configuration.

It is obvious that the choice of delta i depends on the experience of the


designer if tedious calculations are to be avoided.
The final expression for equation 4 is (see Appendix III):

n and gamma are easily determined for a given electrode.


Figures 7 and 8 show the computed values for potential probe position of a
ground wire buried in the first layer for different lengths and various depths.
The study being limited to classical fall of potential method only.
The value of

was chosen such that conditions 1 and 2 are satisfied. The ground wire is
placed in the same physical direction as the electrode line abc.
Figure 7 indicates that when electrode length is more than d/10 the value of x0
increases rapidly from the theoretical point source electrode value to 1.0 in the
case of an electrode length L1 = 2d. The non uniform current assumption
curves correspond to higher values than the uniform current assumption.
The analysis of equation 11 shows that if the ground electrode wire is placed
perpendicular to the abc direction, the effect of increased length will be a
decrease of the required position x0.
Figure 8 shows that xo decreases slowly with increased depth. The deviation is
relatively much greater with lower values of K. In fact for high positive values of
K it can be assumed that xo does not vary with ground wire depth.

EXPERIMENTAL MEASUREMENTS
In order to illustrate the theoretical results by experimental measurements, a
laboratory model [3,4,6] has been used to determine earth resistance of
different driven rod electrodes. The model's first 'layer is simulated by water
while the second layer was made up of concrete. Tables I, II and figures 9 & 10,
resume the experimental results respectively for different rod length L2 with h
and d held constant and different spacing d with L2 and h held constant.
Curves of figures 11 & 12 show the potential electrode position as function of
L2/h and h/d respectively. The potential electrode position x0 has been
determined as follows:

The resistance of a driven rod is calculated from the theoretical formula [1].
The resistance value on the classical fall of potential curve which is equal to the
previous calculated value, gives the required probe position.
The experimental curve of figure 11 has been extrapolated for very small
values of L2 since in this case the rod is practically a point source electrode.
Figure 4 then gives the required position which is 0.705; Figure 11 shows that
the potential electrode position begins to increase from 0.705 to 0.84 before
decreasing for larger L2. In the case of small L2 the driven rod behaves like an
hemisphere since its diameter is no longer negligible. With this remark in mind,
it can be said that these results agree qualitatively with Curdts' [7] conclusions
in uniform earth.
In the case of various spacing d, figure 12 shows that the experimental curve
has the same shape as the theoretical curve for a point source electrode and
same reflection factor K (0.62), but is shifted up because of the rod length
which is not small enough to be considered as a point source electrode.

CONCLUSIONS
This study has shown that earth resistance measurement with the fall of
potential method must be done carefully in non uniform soils.
Measurements must take into account the electrodes spacing d, the resistivity
variation, the ground electrode dimension and shape. When the earth can be
reasonably considered as a two layer structure, the following conclusions have
been drawn:
1. The classical fall of potential method gives less variations in the required
potential electrode position than the alternative method. Hence classical fall of
potential method is recommended.
2. In uniform soil such a position xo is 0.618. However, the presence of a
second layer may modify considerably this value and the required position will
vary between 0.50 to 0.90 in the case of the classical fall of potential method.
3. The magnitude of the variations increases with the absolute value of the
reflection factor K. However, a more conductive second layer has less adverse
influence on such variations than the alternative.
4. For complex ground electrodes the value of xo will vary from those
calculated for a point source electrode. These values will depend uniquely on
the ratios h/d and h/di.

ACKNOWLEDGEMENT
The authors would like to express their appreciation to the National Research
Council of Canada and the Department of Energy, Mines & Resources of
Government of Canada for providing the necessary financial support of the
work. The authors would like to thank the management of The Shawinigan
Engineering Company for their cooperation.

APPENDIX I
Figure 2 shows three electrodes, electrode a which is the ground electrode
under test, electrodes b and c which are the auxiliary grounds. If resistance
between each pair of grounds is measured in such a way that the current is
kept constant in each measurement the following can be written [2]:

Adding (14) and (15) and subtracting (16) from the previous result gives:
Since electrode b is the potential electrode which is connected to a high input
impedance voltmeter, practically no current will flow through b and the
following holds:

APPENDIX II

Potential distribution around a point source electrode in a multilayer earth has


been studied by Stefanesco and Schlumberger [9]. Since we are involved only
by potential on soil surface in a two layer earth this potential will be written
only (fig. 3):

The mutual resistance between the point source electrode i injecting a current I
and point j at earth surface spaced x from i is:

APPENDIX III

Consider a ground electrode buried in the first layer of a two layer earth as
shown in figure 6. The potential at point b due to current I flowing from
electrode a can be written:

i being a subscript which indicate a linear small part of the ground electrode.
Each Vib can be written again according to the images method [8] or by using
Laplace's equation [6,9]:

Identical equations can be derived for Vic and V(cb)i the subscript i in Vcb
means potential in b due to current delta iI returning through electrode c. The
current flowing through b is assumed very small and so is neglected.
With the previous considerations equation (4) becomes:

You might also like