Electric Power Systems Research 153 (2017) 111–118

Contents lists available at ScienceDirect

Electric Power Systems Research

journal homepage: www.elsevier.com/locate/epsr

Development of a new methodology for measurements of earth

resistance, touch and step voltages within urban substations
Adroaldo Raizer a,∗ , Wilson Valente Jr. b , Vilson Luiz Coelho c
a
Federal University of Santa Catarina, UFSC, Florianópolis, SC 88040-970, Brazil
b
Federal Institute of Santa Catarina, IFSC, Itajaí, SC 88307-303, Brazil
c
Faculty SATC (FASATC), Criciúma, SC 88805-380, Brazil

a r t i c l e i n f o a b s t r a c t

Article history: This paper discusses the development of a new methodology for measurements of earth resistance and
Received 29 April 2016 touch and step voltages in ground mesh of urban power substations. The main challenge of the work is
Received in revised form to find a feasible solution for performing earth resistance measurements from short distances, given that
26 December 2016
this is the condition actually found in the majority of urban substations. A test field with four ground
Accepted 17 January 2017
Available online 22 January 2017
mesh units of different geometric configurations was implemented and used for performing numerous
measurement tests. Based on all the data collected and analysis of the measurement results, a mathemat-
ical model was developed to estimate and predict the correct distance in order to obtain earth resistance
Keywords:
Grounding measurements from short distances, the PRED Method – Polynomial Regression from Database Method.
Earth resistance A discussion of the proposed method, as well as analysis of its accuracy and susceptibility to external
Touch and step voltage interference is performed, in addition to its validation in real substations, so as to prove the efficiency
Power system measurements and applicability of the proposed method.
Substation protection © 2017 Elsevier B.V. All rights reserved.

1. Introduction
Obtaining grounding resistance values is a major factor in the
analysis of electrical safety in electrical substations. This informa-
tion is considered essential, not only to maintain a low impedance
path for the protection of the electrical system in any outbreak,
but also to ensure that the potential increase on the ground does
not reach levels above the limits set for touch and step voltages in
substations.
In view of its fundamental importance, it is possible to histor-
ically observe a major concern of the scientific community and
engineers in obtaining measurements that are able to set the cor-
rect value of the grounding system resistance with a high degree of
accuracy [1–9]. These methods are often the subject of work for the
establishment of technical safety standards, with well-established
protocols and measurement methods [11–13]. The ANSI/IEEE Std
81 [11] is the main Guide for Measuring Earth Resistivity, Ground
Impedance, and Earth Surface Potentials of a Ground System.
There are several methods for measuring resistance of the
ground electrode system. Among them, as verified in many field
∗ Corresponding author.
E-mail addresses: raizer@eel.ufsc.br (A. Raizer), wilson.valente@ifsc.edu.br
(W. Valente Jr.), vilson.coelho@vlc.eng.br (V.L. Coelho).
tests, the fall-of-potential method is widely applied for almost all
types of grounding systems [1–6]. However, the fall-of-potential
method and its regulatory procedures require certain conditions
that are often difficult to obtain for substations in operation in the
electric power system, especially in urban environments. As a rule,
it is necessary a distance of approximately 4–10 time the greater
diagonal (D) of the mesh for the proper positioning of the current
electrode (dA) and 2–6 time D for the placement of the potential
electrode (dP) (sometimes referred as 62% rule). This measure-
ment procedure is often infeasible, either for operational reasons
or for practical reasons in urban substations, given the extent of
cables and physical obstacles at measuring points. In an attempt to
overcome the above-mentioned aspects, some alternative methods
have been adapted, using other electric power system structures as
auxiliary electrode, for example, grounded transmission line tow-
ers and other substations grounding meshes. In these cases, the
distances can be very long and measurements may suffer from the
influence of external factors (e.g. electromagnetic interference due
to coupling with transmission lines). So either the assessments are
made providing incorrect values, which contradicts safety, or the
measurements end up not being performed.
In this paper, we propose an effective method for the evalua-
tion of earth resistance, and touch and step voltages within urban
substations in order to contribute with this area of research and
development (R&D). It employs a mathematical model based on

http://dx.doi.org/10.1016/j.epsr.2017.01.025
0378-7796/© 2017 Elsevier B.V. All rights reserved.
112 A. Raizer et al. / Electric Power Systems Research 153 (2017) 111–118

linear regression of a parametric analysis to compute the error esti-

mate and evaluate the best electrode position to perform the test. In
this sense, we addressed the aforementioned problem employing a
method that can be easily applied to short distances and not as sus-
ceptible to outside interference, when compared to the traditional
methods [11].

2. Literature review

2.1. General fall-of-potential method (FoP)

Several papers presented in the technical literature contribute

to solving some of the aforementioned problems. In Ref. [5] an
extended analysis is presented for the ground impedance mea-
surement using the fall-of-potential method. It introduces a set of
curves to represent the exact placement locations for the potential
probe when the potential and current probes are in different direc-
tions. Also, curves representing measurement error are presented
in the case when the potential probe is placed in locations where
ground impedance cannot be measured correctly. A similar analysis
Fig. 1. Tagg Method (solid line) and Extended Tagg Method (ETM) (dashed line)[2].
for ground impedance measurements in multilayer soils is also dis-
cussed. In Ref. [10] is described a method and instrumentation for
measuring touch and step voltages near the grounding system. The d Selection of the dPT /dA value corresponding to the “” value in a
method injects the transient electric current between the ground correction table proposed by Tagg [11] (for  ranging from 0.4 to
under testing and an auxiliary ground electrode, allowing measure- 1.59).
ments of ground potential differences at different locations. The e Measurement of the true grounding resistance by placing the
touch or step voltages are related with the system short circuit potential probe at the distance dPT established by the Tagg’s table.
capability. The paper Ref. [9] proposes a technique for measuring
ground resistance without using auxiliary electrodes. The mea-
2.3. Extended Tagg Method
surement technique and a measurement tool are introduced and
discussed. In Ref. [4], an alternative fall-of-potential method is pre-
In a number of measurement tests performed in practical situa-
sented. The method is a voltmeter and ammeter method, and does
tions, it was noticed that Tagg’s method showed a certain amount
not require any auxiliary electrodes. Using the ground electrodes
of points in which the  coefficient could not be obtained in the
of a nearby substation, the method allows determining ground
interval of the correction table proposed by Tagg [1]. Taggs’s table
electrode resistance of one of the premises and of the effective
presented in the IEEE guide shows values for the relation dPT /dA due
ground electrode resistance of the substation. The method yields
to the tilt variation coefficient () in a range from 0.4 to 1.59 and
accurate results and is practically contributive for electrode resis-
the rate dPT /dA from 0.643 to 0.341 (referred as PPT /CP at Ref. [11]).
tance measurements in sites where the use of auxiliary electrodes
When the tilt coefficient obtained by the 3 consecutive resistance
is difficult. Also, numerical simulation analysis was presented in
measurements of Tagg’s method (i.e. 20%, 40%, 60%) is outside this
[14] in order to evaluate the fall-of-potential method and to assess
range, the result of this method could not be defined.
the influence of auxiliary electrodes at near distances, providing an
Thus, in order to establish boundary conditions for measur-
overview about the voltage distribution phenomena and grounding
ing points with undefined values by Tagg’s original method, Ref.
resistance.
[2] proposes the use of the Extended-Tagg Method (ETM), which
extrapolates the table of the original method, from a mathematical
2.2. Tagg Method approximation presented in Eq. (1):

Tagg’s Method, also recommended by IEEE Std. 81-2012 [11],
is based on the work developed by Tagg [1] since 1970, and also
known as the “slope” method. In this method, uniform resistivity
assumptions of the soil and representation of the ground elec-
trode system as an equivalent hemispheric electrode are used.
Additionally, this method allows performing the measurement at
near distances, by introducing the correction factor to the fall-of-
potential procedure.
In general, Dr. Tagg’s slope method [1] can be applied from the
following protocol measures [11]:
a Choosing of a convenient starting point for linear measurements
and the selection of a suitable distance for the positioning of the
current electrode (dA ).
b Measurement of three resistance values R1, R2, R3 inserting the
potential electrodes (dP ) on the distance of 0.2 dA , 0.4 dA , and 0.6
dA , respectively.
c Calculation of the inclination variation coefficient () using the
equation:  = (R3 − R2)/(R2 − R1).
PPT /CP = −0.1242u3 + 0.2339u2 − 0.3049u + 0.738 (1)
From this mathematical approximation, it is possible to perform
the extrapolation of Tagg’s method for applications in near distance
situations, and frequently encountered in practical evaluations. In
Fig. 1, original Tagg’s method [1] is shown by the solid line, while
the Extended-Tagg Method is shown by the dashed red line. This
allows the use of Tagg’s method for analysis of more near distance
measurements, which is the main goal of this paper.
3. Formulation of the proposed method (PRED)
The proposed method, denominated PRED method (Polynomial
Regression from Database), was developed from a series of exper-
imental results considering the parametric assessment of dA and
dP and subsequent analysis of the measurement database. This
method presented a different approach from the other works found
in the technical literature, but can also be employed in order to pre-
dict the correct probe placement for resistance measurement at
 A. Raizer et al. / Electric Power Systems Research 153 (2017) 111–118 113

Fig. 4. Fall-of-potential test arrangement.

10 dA/D:
9 Mesh 1 6
Fig. 2. CT UFSC test site – Lat. 27◦ 40 52 , Long. 48◦ 32 11 . 5
8 4
7 3
6 2
1
5

R (p.u.)
0.9
4 0.8
3 0.7
0.6
2 0.5
1 0.4
0.3
0 0.2
0.2 0.4 dP/dA 0.6 0.1

Fig. 5. Typical resistance versus potential probe spacing – mesh 1.

3.5 dA/D:
Fig. 3. Grounding mesh 1: 10 m × 10 m. Mesh 2 6
3 5
4
near distances. The basic formulation is presented in this section, 2.5 3
and all validation tests will be presented latter. 2
2 1
R (p.u.)

3.1. Survey of experimental data
A test site was built for data collection (CT UFSC) with 4 ground-
ing meshes of different geometries (10 × 10 m, 5 × 10 m, 2.5 × 10 m,
5 × 5 m,) in a lot within the Agricultural Sciences Center of Fed-
eral University of Santa Catarina – UFSC, far enough away and free
from external interference, as shown in Fig. 2. The resistivity and
soil stratification parameters in this test site were systematically
monitored so as to control their properties conditions during the
tests.
In this piece of land, the 4 grounding meshes were built with
bare copper wire, of medium hardness, 95 mm2 , and compression
type connectors. Fig. 3 shows the geometric configuration of the
10 × 10 mesh, buried at 0.5 m from the surface. The other meshes
follow a similar configuration and were built spaced apart on the
available lot.
After the implementation of the test site, a series of resistance
measurement tests were performed, so as to form a database with
characteristic information of the ground meshes. Such tests follow
the configuration of the fall-of-potential method (IEEE Std 81) [11]
presented in Fig. 4. The systematic variation of the dP and dA dis-
tances, using parametric assessment, allows the evaluation of the
grounding resistance influence of reference electrodes, which is the
main focus of this work.
3.2. Analytical data treatment
It is understood that, as performed by Tagg in his work [1,2,11],
the main challenge in obtaining grounding values for short dis-
0.9
1.5 0.8
0.7
1 0.6
0.5
0.5 0.4
0.3
0 0.2
0.2 0.4 dP/dA 0.6 0.1
Fig. 6. Typical resistance versus potential probe spacing – mesh 2.
tances is the definition of equations or analytical models in order
to correct the error due to the use of electrodes in nearby regions.
With the PRED method, this is performed analytically, through lin-
ear regression of all parametric measurement results, as shown
below.
The graphs in Figs. 5–8 show the behavior of the measured val-
ues, in p.u. of the normative theoretical value (Tagg Method [11])
for dP values equal to 20, 40 and 60% of dA . The dA values were
used between 6D and 0.1D, where D is the greatest diagonal value
of the analyzed mesh. By observing the graphs, one can see that
for the dP interval between 0.2dA and 0.6dA , the typical resistance
curve is approximately a straight line. By using the linear regres-
sion techniques, it is possible to determine the linear and angular
coefficients of the lines for the different values of the dA /D ratio.
For each set of measurements performed for the same dA value,
the dP for R equal to 1 p.u. was determined. The graph in Fig. 9
shows the behavior of the dP , values for R = 1 p.u., due to the differ-
ent spacing of the current electrode. A strong correlation is noticed
between the values toward a line with a logarithmic tendency.
114 A. Raizer et al. / Electric Power Systems Research 153 (2017) 111–118

3.5 dA/D:
Mesh 3 6
3 5
4
2.5 3
2
2 1
R (p.u.)

0.9
1.5 0.8
0.7
1 0.6
0.5
0.5 0.4
0.3
0 0.2
0.2 0.4 dP/dA 0.6 0.1

Fig. 7. Typical resistance versus potential probe spacing – mesh 3.

Fig. 10. Step and touch voltages on ground grid area.

3 dA/D:
Mesh 4 6 4. Measurement of earth surface potentials
2.5 5
4 Traditionally, the analysis of touch and step voltages is
2 3
2 performed from measurement procedures and equipment in accor-
1 dance with well-established standards [11,13].
1.5
R (p.u.)

0.9 In these procedures, the current injection circuit is similar to

0.8 the one shown in Fig. 4 with the current injection performed at a
1 0.7
0.6 central point of the mesh. For data acquisition, in the measurement
0.5 0.5 of the touch voltage, a voltmeter is connected between an exposed
0.4 point connected to the mesh and a point to be measured, 1 m apart,
0.3
0 0.2
while in the step voltage measurements, potential measurements
0.2 0.4 dP/dA 0.6 0.1 are taken between rods located in predetermined points 1 m apart
(see Fig. 10).
Fig. 8. Typical resistance versus potential probe spacing – mesh 4. The measured values of the touch and step voltages refer to the
test current and should be corrected to the maximum fault current
in accordance with Eq. (3).
90 V 
e
80 VR = IM (3)
Ie
70
where: Ve is the voltage value obtained in the test (step or touch
60
voltages), Ie is the current used in the test, IM is the maximum fault
dP(R) (%dA)

50 current of the mesh and VR the real voltage (step and touch).
40 Therefore, the main result of this test is the relation between the
30 touch or step voltage values and the test current, which represent
such voltages referred to a unitary current, or to a touch and step
20
“resistance”.
10
0 4.1. Measuring earth surface potentials by PRED method
0 2 4 6 dA/D 8
In the mesh potential measurement test, the ground terminal of
Fig. 9. Behavior of voltage probe spacing (dP ) as a function of the ratio of the current
probe spacing and the major diagonal of the ground grid (D). the voltmeter is connected to the potential electrode (PP ) located
at a dP distance from the mesh, and the current electrode (CP ), posi-
tioned at a dA reference distance, it is connected to the generator
and center point of the grounding mesh, in a procedure similar to
3.3. Analytical expression of PRED method the configuration of the fall-of-potential method (Fig. 4).
In this specific test, measurements were performed for various
In view of the analysis performed in the previous section (Fig. 9), dA distances, and the dP distances were determined according to
it is seen that the logarithmic Eq. (2) can determine with a certain the PRED method.
degree of accuracy the relative distance between the potential elec- By means of this method, the maximum touch and step voltages
trode and the current electrode for a determined distance between can be estimated indirectly, using the mesh potential curve and the
the current electrode and the mesh. application of Kirchoff’s Law in the measurement circuit. Just like
the traditional method, the results refer to the test current. Further
d 
A on Eq. (2) is also applied.
dP (R) = 0.14 ln + 0.41 (2)
D
5. Validation tests of the proposed method
This equation can be useful for measurements with small val-
ues of dA , where it is not possible to obtain a typical behavior in This section presents the validation results of the Extended Tagg
determining the resistance curve. Method [2] and PRED Method, when compared to the standard
 A. Raizer et al. / Electric Power Systems Research 153 (2017) 111–118
Fig. 11. Typical resistance curve at hemispherical condition.
fall-of-potential method prescribed by IEEE Std. [11,14] for obtain-
ing the grounding resistance. A large set of measurements taken
from different grounding meshes was used, in order to compare
the results and estimate the average error of the aforementioned
methods. It is divided in grounding resistance measurements at
hemispherical condition (ideal scenario) and grounding resistance
measurements at near distances, as follows.
5.1. Grounding resistance measurement at hemispherical
condition (standard reference)
According to the standards [11,13], the correct value of ground-
ing resistance can be estimated by employing fall-of-potential
method (Fig. 4.) when ideal measurement condition is observed,
where distances are long enough to meet all the compliance crite-
ria for hemispherical geometry of potential distribution. At uniform
soils, the hemispherical potential distribution is usually observed in
fall-of-potential method when the current electrode dA is placed at
least at 5 times the mesh diagonal size (5D) and potential electrode
dP complies with the 62% rule position [11]. IEEE Std 81-2012 also
suggests the use of Tagg’s method [1] in the case of non-uniform
soils [11].
Fig. 11 shows an example of a set of measurements that perfectly
meets the hemispheric condition recommended by the norm [11].
It is performed on a 10 × 10 m mesh, where dA is set at 5 times
the diagonal (5D), with resistance values measured for dP ranging
from 10 to 90% of dA . In this curve it is possible to clearly observe
the resistance level where the influence of current and reference
electrodes is minimized (it is almost constant about 62% dP /dA ).
This standard condition of fall-of-potential method was employed
in this paper as reference result, according to IEEE Std 81-2012 [11],
in order to compare the performance of the proposed methods and
error estimate in this paper.
5.2. Grounding resistance measurement at hemispherical
condition (proposed methods)
Exploratory measurements considering hemispherical condi-
tion were performed by Extend Tag Method and PRED Method. It
will be briefly presented here only as example, since the main goal
of this paper is to explore near distances measurements. Consider-
ing hemispherical conditions, measurement results for all methods
are also marked in Fig. 11 (fall-of-potential 62% standard method
[11], Tagg Method [1], and PRED Method). The grounding resis-
tance value of the mesh considering all methods employed were
around 16.75 . Additionally, the inflection point of the curve is
also presented to evaluate the adequate spot for the application of
the potential electrode. Similar results were also achieved in many
115
other trials, and considering all different mesh sizes in this survey.
Therefore, it is observed that at distant placements of dA the per-
centage of error between the presented methods is usually less than
1%, and quite similar to the reference standard at hemispherical
condition [11].
5.3. Grounding resistance measurement at near distances
(proposed methods)
An extensive number of measurement tests were developed
in order to evaluate the performance of Extended Tagg and PRED
Methods at near distances conditions and compare it to the stan-
dard fall-of-potential method result at hemispherical condition
(considered here as the correct resistance value). Part of the val-
idation results is presented in Table 1, which presents the Standard
Reference Method Resistance (using fall-of-potential method [11]),
and the resistance results of proposed methods, both Extended
Tagg and PRED methods.
It is important to mention that many different distances of the
current electrode (dA ) were evaluated in relation to the size of the
mesh diagonal (ranging from 5 to 0.25 times the diagonal of the
mesh D), as defined by the parametric analysis. For each set of dA
distance, 10 equidistant variations of the potential electrode (dP )
were obtained according to the dP /dA (ranging from 1% to 100% of
the dA value). This parametric survey leads to many graphs similar
to the one presented in Fig. 11. However, as dA values decreases, the
typical constant resistance level becomes distorted because hemi-
spherical condition is not complete satisfied. Simulation results can
easily show this situation as published at Ref. [14]. However, even at
this non-hemispherical situation, Extended Tagg and PRED meth-
ods suggest a dP /dA rate to estimate grounding resistances. At this
dP /dA rate, resistances were measured and error evaluated for each
test configuration. All results can be observed in Table 1.
It is possible to note that the percentage errors of both methods
are equivalent for measurement conditions where dA is larger than
5 m (average error of 6% for the Extended Tagg Method, against 4%
for the PRED method). For extremely close measuring conditions
(less than 5 m), further analysis is still required due to the strong
mutual coupling between the electrodes. Taking into account these
situations (identified by the gray lines in Table 1) the error of the
Modified Tagg and PRED methods drastically increases to respec-
tively 26% and 17%. Since mutual coupling at extreme conditions is
not within the scope of this work, all results where dA are small than
5 m were neglected, in order to keep error estimates under reason-
able values. So, both methods are not suitable to be employed for
cases where dA is smaller than 5 m, but can provide good results
when this situation is solved.
6. Results
6.1. Application of proposed methods at urban substations
The ETM and PRED methods presented in this paper were
applied to urban power substations meshes, connected to the
electric power system, in order to analyze its applicability and per-
formance in real and uncontrolled cases. The reference value (RT )
was provided by the power company and obtained according to
Ref. [11]. Two urban substation were evaluated as study-case dur-
ing this analysis, named here as Substation A and Substation B, both
chosen by the State Company of Electric Power (CEEE). Substation
A is an urban substation where the grounding mesh presents an
80 m diagonal size D. The reference resistance estimated at this
substation was about 0.75 . The second urban substation, named
Substation B, was built with a grounding mesh with 77.4 m of diag-
onal size. The reference resistance estimated at Substation B was
116 A. Raizer et al. / Electric Power Systems Research 153 (2017) 111–118

Table 1 Table 2
Measurement of resistance values and error estimate (STANDARD Method, Extended Measurement of resistance at urban substations (PRED Method).
Tagg Method, and PRED Method).
Test dA /D dA R Reference PRAD Method
Test dA/D dA Standard Method Extended PRAD Method dP /dA R Abs. Error Error
(FoP) Tagg Method (m) () (m) () () (%)
R Error R Error 1 5 400 0.75 256 0.56 0.19 25
(m) (Ω) (Ω) % (Ω) % 2 4 320 0.75 192 0.57 0.18 24
3 3 240 0.75 134.4 0.71 0.04 5
1 9.9 140 18.2 18.5 2 18.7 3 4 2 160 0.75 81.6 1 −0.25 33
2 8.5 120 18.2 18.5 2 18.7 3 5 1 80 0.75 32.8 0.78 −0.03 4
3 7.1 100 18.2 18.4 1 18.5 2 6 0.75 60 0.75 22.2 0.72 0.03 5
4 5.6 80 18.2 17.2 5 18.9 4 7 0.5 40 0.75 12.4 0.83 −0.08 10
5 4.2 60 18.2 18.2 0 18.2 0 8 5 387 2.36 247.7 2.28 0.08 3
6 2.8 40 18.2 17.8 2 18.7 3 9 4 310 2.36 185.8 2.28 0.08 3
7 1.4 20 18.2 18.7 3 18.7 3 10 3 232 2.36 130.7 2.26 0.1 4
8 0.99 14 13.6 13.9 2 13.9 2
11 2 155 2.36 78.9 2.2 0.16 7
9 0.91 13 13.6 14.3 5 14.3 5
12 1 77 2.36 31.7 2.28 0.08 3
10 0.85 12 13.6 10 26 14 3
11 0.77 11 13.6 11 19 14 3 13 0.75 58 2.36 21.5 2.34 0.02 1
12 0.70 10 13.6 16 18 15 10 14 0.5 39 2.36 12.0 2.49 −0.13 6
13 0.64 9 13.6 15 10 14 3 Total Average Error 10
14 0.56 8 13.6 13 4 13 4
15 0.49 7 13.6 14 3 14 3
16 0.42 6 13.6 12 12 16 18
Table 3
17 0.35 5 13.6 19 40 19 40
Touch and step voltages measured by the STANDARD Method.
18 0.28 4 13.6 16 18 18 32
19 0.21 3 13.6 35 157 25 84 dA /D Generator Voltage (V) Ve/Ie
20 0.14 2 13.6 51 275 41 201
21 0.07 1 13.6 110 709 68 400 V (V) I (A) Touch Step Touch Step
22 5 70.7 13.9 13.9 0 13.9 0
23 4 56.56 13.9 14 1 14.2 2 5 221.7 0.279 0.64 0.07 2.28 0.26
24 3 42.42 13.9 14 1 14.2 2 4 220.9 0.246 0.56 0.06 2.28 0.25
25 2 28.28 13.9 14.6 5 14.5 4 3 221.6 0.376 0.85 0.09 2.27 0.25
26 1 14.14 13.9 13.3 4 14 1 2 221.6 0.821 1.82 0.2 2.22 0.24
27 0.75 10.6 13.9 14.91 7 14.1 1 1 221.8 0.949 2.14 0.23 2.25 0.24
28 0.5 7.07 13.9 13.9 0 13.9 0 0.75 221.5 0.714 1.66 0.17 2.32 0.23
29 0.25 3.53 13.9 21 51 19 37 0.5 221.7 0.583 1.43 0.12 2.45 0.2
30 5 35.35 62.8 63.5 1 64 2 0.25 221.6 0.396 1.18 0.06 2.98 0.15
31 4 28.28 62.8 64.4 3 65.5 4
32 3 21.21 62.8 64.7 3 64.5 3
33 2 14.14 62.8 66 5 64 2
34 1 7.07 62.8 68.1 8 59 6
35 0.75 5.3 62.8 70 11 56 11 about 2.36 . Table 2 presents the results achieved by PRED method
36 0.5 3.53 62.8 77.8 24 46.8 25 during the evaluation of both urban substations. In this table, 14 sets
37 0.25 1.76 62.8 80 27 43 32 of measurement tests can be observed. Grounding resistance and
38 5 55.9 31.2 32.5 4 31.8 2
39 4 44.72 31.2 32.1 3 32 3 error estimate is also provided at this table. The total average error
40 3 33.54 31.2 33 6 32.1 3 was about 10%, and it was considered a reasonable accuracy for the
41 2 22.36 31.2 33.9 9 31.9 2 proposed method, especially because the resistance values in this
42 1 11.18 31.2 32 3 29 7
43 0.75 8.38 31.2 33.8 8 27 13 situation are very small. It is important to note that at Substation A,
44 0.5 5.59 31.2 37 19 25.4 19 the largest error estimate was about 33% and it was presented dur-
45 0.25 2.79 31.2 57 83 23 26 ing test number 4. The average error for this substation was 15%, but
46 5 51.53 49.7 50 1 50.5 2
47 4 41.23 49.7 50.5 2 50.5 2 it is also considered a good result because the absolute error was
48 3 30.92 49.7 51.2 3 50.8 2 less than 1 . At Substation B, the maximum error was 7% obtained
49 2 20.61 49.7 52.5 6 50.7 2 during test 11, and the average error for this substation was about
50 1 10.3 49.7 55 11 50 1
51 0.75 7.73 49.7 54 9 48 3 4%. It can be also considered a reasonable grounding estimate by
52 0.5 5.15 49.7 61 23 44 11 PRED method.
53 0.25 2.57 49.7 74 49 43 13
54 5 70.7 16.74 16.8 0 16.8 0
55 4 56.56 16.74 17.03 2 17.04 2 6.2. Application of proposed methods for touch and step potential
56 3 42.42 16.74 16.8 0 17.2 3
57 2 28.28 16.74 17.3 3 17.2 3 estimate
58 1 14.14 16.74 18 8 17.2 3
59 0.75 10.6 16.74 15.9 5 17.2 3 An important goal of this work is to evaluate the application
60 0.5 7.07 16.74 15 10 16.41 2
61 0.25 3.53 16.74 24 43 22 31 of the proposed method in order to obtain the touch and step
62 5 35.35 81.3 81.5 0 81.6 0 potential, since this information is very important for safety rea-
63 4 28.28 81.3 82.4 1 83.1 2 sons. During the preview section, it was demonstrated that PRED
64 3 21.21 81.3 81 0 81 0
65 2 14.14 81.3 83 2 79.2 3
Method was capable to obtain earth resistance measurements with
66 1 7.07 81.3 85.9 6 75 8 reasonable accuracy, since average error remains about 5% when
67 0.75 5.3 81.3 78 4 75 8 compared to Standard references [11]. Therefore, in this section the
68 0.5 3.53 81.3 92.5 14 70 14
indirect measurements proposed by the PRED method are again
69 0.25 1.76 81.3 101.4 25 70 14
compared with the traditional method set by IEEE Std. 81-2012
Total Average Error (considering all results) 26 17
[11], but now considering touch and step potential results.
Total Average Error (neglecting dA < 5m) 6 4
Tables 3 and 4 present results of touch and step potential
measurements performed by the standard and PRED methods
respectively. Here, both tests were performed on experimental
mesh 1.
 A. Raizer et al. / Electric Power Systems Research 153 (2017) 111–118 117

Table 4 In regard to the earth resistance measurement methods

Touch and step voltages measured by the PRED Method.
explored in this work, ETM and PRED, both methodologies allow
dA /D dP /dA Generator Voltage (V) Ve/Ie the taking of earth resistance measurements, and the estimate of
touch and step potential, in a ground mesh, with current and volt-
V (V) I (A) Touch Step Touch Step
age electrodes positioned at a short distance from the ground mesh
5 0.64 221.7 0.279 0.64 0.07 2.28 0.25
of a substation. It is noteworthy that both methods are able to lead
4 0.6 220.9 0.246 0.56 0.07 2.28 0.28
3 0.56 221.6 0.376 0.85 0.09 2.26 0.24 to satisfactory results and are entitled to define analytical equations
2 0.51 221.6 0.821 1.81 0.19 2.2 0.23 or approaches that correct the error by using electrodes inserted
1 0.41 221.8 0.949 2.16 0.22 2.28 0.23 in nearby regions. However, it is important to note that despite
0.75 0.37 221.5 0.714 1.67 0.17 2.34 0.24 similar results, mathematical approaches employed for each of the
0.5 0.31 221.7 0.583 1.45 0.11 2.49 0.19
0.25 0.22 221.6 0.396 1.2 0.05 3.03 0.13
methods are completely different. While the Tagg Method is based
on calculating the slope coefficient (), the PRED method takes
into account a linear regression and a logarithmic approximation
3.5 that correlates the pattern of variation of the potential and current
Touch (S) electrodes when brought closer to the mesh.
3.0
In practical terms, the Tagg Method requires taking 4 measure-
2.5 ments in order to estimate the resistance value of the mesh. It
Step (S)
Ve/Ie (V)

2.0 takes three exploratory resistance measurements taken at 20%, 40%

Touch (R)
1.5
1.0 step (R)
0.5
0.0 dA/D
0 2 4 6
Fig. 12. Touch and step voltages – standard (S) and PRED (R) methods.
Columns Ve/Ie represent the touch and step voltage values in
relation to a unitary current (p.u.).
Fig. 12 shows in a comparative way, the touch and step voltage
values, measured according to the traditional method and the PRED
method.
It can be observed that the values obtained for the two methods
are very similar, leading to the conclusion that the concern regard-
ing the position of the voltage electrode is not crucial when the
objective is to determine the touch and step potential. Regarding
the position of the current electrode (dA ), it is seen that for distances
less than 1D there is a systematic change of the measured values. In
other words, the measurements are valid until dA = 1D and below
it there is a deviation of values which increases as dA is reduced. It
presents an important contribution verified by this work, since it
suggests that voltage electrode can be now placed in near distances
area when employing PRED to evaluate touch and step potentials.
7. Discussion and conclusions
The PRED method developed and presented in this paper aims to
contribute to the discussion of grounding resistance measurement
within urban substations. The reduction of measurement distances
has always been a technical challenge to be overcome in these
analyses, considering that the minimum distances required by the
technical norms are increasingly unfeasible in urban substations.
Besides that, long distances imply the possibility of measurement
errors, since the effects of the couplings and electromagnetic inter-
ferences significantly increase in a larger test area. As an example, in
the case of urban substations, a distance equal to 5D can be greater
than 380 m, such as the two examples presented in this paper.
Efforts at reducing the distances published in the technical liter-
ature, until now, systematically required measurement procedures
of difficult practical implementation and a number of limitations,
which eventually marginalized the use of these methods compared
to the non-conventional alternatives. On the other hand, such non-
conventional alternatives, as the use of grounded transmission lines
or closer substations as the reference electrode, are also not con-
trollable procedures.
and 60% of the dA distance (current electrode), which are used to
estimate the slope coefficient, and finally another additional mea-
surement to obtain the true resistance value. At this aspect, the
PRED method has a considerable advantage as it allows one to pre-
dict, with good degree of precision, the relative distance between
the potential electrode and the current electrode, not requiring
any additional exploratory measurements. So, this method requires
only a single measuring point in order to establish the correct value
of ground resistance. Besides that, the correct estimate point can
be calculated in advance, needing only previous knowledge of the
greater dimension of the D mesh, which in practical terms makes
the application of this method very versatile.
It is recommended to apply these methods respecting the min-
imum distance, which already consolidates a major advance when
compared to the dimensions recommended by the current stan-
dard (4–5 times the longest dimension of the mesh). According to
the results obtained in this study, one can establish the distance
of 0.75 times the longest dimension of the mesh as the minimum
distance for implementing the method without includes a great
amount of error in measurements.
8. Final considerations
In this paper was presented a number of grounding resistance
measurement results and error estimate for Extended Tagg Method
(ETM) and Polynomial Regression from Database (PRED) Method at
short distances, taking the standard fall-of-potential (FoP) method
at hemispherical condition as reference [11]. An extensive set of
measurement tests, with parametric analysis at different meshes
configurations, were developed in order to fully validate the pro-
posed method. Part of this validation data was presented in this
paper. Eventually, more work can be required for fine adjustment
and definitive consolidation of the developed method. However, at
this point the proposed method can present the following contri-
butions:
• The possibility of applying measurement techniques at close dis-
tances in order to estimate the true grounding resistance values
and error estimate.
• In view of the mathematical treatment given to the problem,
the method is able to achieve good results in cases where the
traditional methods cannot be applied.
• The possibility of reducing the number of exploratory measure-
ments in order to perform a reliable estimate at reduced distances
(Taggs Method or ETM requires at least 4 measurements – three
exploratory measurements followed by calculation, and one
additional measurement to estimate the grounding resistance.
118 A. Raizer et al. / Electric Power Systems Research 153 (2017) 111–118
PRED Method can predict the measurement point in advance,
based on mesh dimension D, and requires only one direct mea-
surement).
However, as in the case of any scientific development process, it
is important to note some possible limitations that may be related
to the following aspects:
• The experimental developments of this project were based on 4
meshes of different dimensions, located in the same region, in
which the soil has similar characteristics.
• Since only two real urban substations meshes were evaluated
using PRED Method, it is important to provide the use of this
procedure in large scales analysis and compare the results with
standard methods to confirm the effectiveness of the methods
under general conditions.
To summarize, after a series of technical analyses, which
involved, soil resistivity measurement processes, computer sim-
ulation, parametric evaluation, mathematical analysis and data
correlation, it was possible to gather relevant information for the
assessment of earth resistance measurement methods at reduced
distances. The PRED method underwent a results validation pro-
cess and it is currently regarded as a functional tool that can be
applied by design engineers of grounding systems, as well as for
monitoring the electrical performance at CEEE-D utility company.
Acknowledgement
This work was developed under the State Company of Electric
Power Distribution (CEEE-D) R&D program which is regulated by
the Brazilian Electricity Regulatory Agency (ANEEL).
References
[1] G.F. Tagg, Measurement of the resistance of physically large earth-electrode
systems, Proc. IEE 117 (11) (1970) 2185–2190.
[2] A. Raizer, V.L. Coelho, W. Valente Jr., C.I. Cardoso, Contribution to Tagg’s
methodology in the resistance measurement of earth-electrode systems at
reduced distances, in: International Symposium on Lightning Protection (XIII
SIPDA 2015), Balneario Camboriu, 2015, pp. 109–115.
[3] C. Korasli, Ground resistance measurement with alternative fall-of-potential
method, in: Proceedings of the Transmission and a Distribution Conference
and Exhibition-IEEE-PES, 21–24 May, Dallas-TX, USA, 2006, pp. 942–946.
[4] C. Wang, T.T. Takasima, T. Sakuta, Y. Tsubota, Ground resistance measurement
using fall-of-potential method with potential probe located in opposite
direction to the current probe, IEEE Trans. Power Deliv. 13 (October) (1998)
1128–1135.
[5] J. Ma, F.P. Dawalibi, Extended analysis of ground impedance measurement
using the fall-of-potential method, IEEE Trans. Power Deliv. 17 (October)
(2002) 881–885.
[6] J. Ma, F.P. Dawalibi, Influence of inductive coupling between leads on ground
impedance measurements using the fall-of-potential method, IEEE Trans.
Power Deliv. 16 (October) (2001) 739–743.
[7] R.J. Heppe, Computation of potential at surface above and energized grid or
other electrode, allowing for non-uniform current distribution, in: IEEE
PAS-98, November/December, 1979, No. 6.
[8] G. Parise, M. Lucheroni, Measurements of touch and step voltages adopting
current auxiliary electrodes at reduced distance, in: Proceedings of the
Industrial and Commercial Power Systems Technical Conference, 8–12 May,
Saratoga Springs-NY, USA, 2005, pp. 191–198.
[9] K. Murakawa, H. Yamane, M. Hattori, Earthing resistance measurement
technique without using auxiliary electrodes, 16 May, Instabul, Turkey, in:
Proceedings of the IEEE International Symposium on Electromagnetic
Compatibility, vol. 1, 2003, pp. 213–216.
[10] A.S. Meliopoulos, Shashi Patel, G.J. Cokkinides, A new method and instrument
for touch and step voltage measurements, IEEE Trans Power Deliv. 9 (4)
(1994).
[11] IEEEStd. 81-2012, IEEE guide for measurement earth resistivity, ground
impedance, and earth surface potentials of a grounding system, in: IEEE
Power and Energy Society, New York-USA, 2012.
[12] C.-H. Lee, A.P. Sakis Meliopoulos, Comparison of touch and step voltages
between IEEE Std 80 and IEC 479-1, September, in: IEEE Procedings
Generation Transmission Distribution, vol. 146, 1999, pp. 593–601, No. 5.
[13] ANSI/IEEE Std 80-IEEE Guide for Safety in AC Substation Grounding, in: IEEE
Power and Energy Society, New York-USA, 2000.
[14] W. Valente Jr., V.L. Coelho, A. Raizer, C.I. Cardoso, Ground impedance
assessment employing earth measurements, numerical simulations, and
analytical techniques, in: International Symposium on Lightning Protection
(XIII SIPDA 2015), Balneario Camboriu, 2015.