Paper No.

561

A REVIEW OF AC POWER LINE COUPLING UNTO BURIED PIPELINES

John Dabkowski, Ph.D.

Electra Sciences, Inc.
P.O. Box 1438
Crystal Lake, IL 60039- 143 8

ABSTRACT

An overview of the magnetic field coupling mechanisms from overhead electric power lines col-
located with buried pipelines is presented. Such coupling results in ac voltages and currents being in-
duced unto the pipeline. Inductive coupling during normal (steady state) operation of the electric power
system and during fault periods is reviewed. Conductive coupling to the pipeline which additionally oc-
curs during fault periods is also discussed. Pipeline induced voltage levels are dependent upon the soil
resistivity along the right-of-way. For complex rights-of-way (ROW), because of computational diff-
culty, computer programs are generally utilized to model the ROW and obtain numerical solutions.
Available computer programs are based either upon a single or two layer earth resistivity model. The ad-
vantages and consequences of using one or the other models for the prediction of induced voltage levels
are discussed.

Keywords: ac induction, induced ac voltages, soil resistivity measurements, two layer soil resistivity
modeling
INTRODUCTION

Because of the lack of available ROW, the incidence of overhead ac power line and buried pipe-
line collocations is increasing. Thus the possibility of ac interference is greater as time goes on. Induced
pipeline voltages may constitute a personnel hazard, can diminish the integrity of the pipeline coating or
the line pipe, and under certain conditions lead to corrosion of the pipe steel. Hence, consideration must
be given to the prediction and subsequent mitigation design for the reduction of induced voltages that
exceed acceptable criteria.

Copyright
01998 by NACE International. Requests for permission to publish this manuscript in any form, in part or in whole must be made in writing to NACE
International, Conferences Division, P.O. Box 218340, Houston, Texas 77218-8340. The material presented and the vwvs expressed in this
paper are solely those of the author(s) and are not necessarily endorsed by the Association. Printed in the U.S.A.
 Voltage levels induced unto the pipeline are a function of electric current carried in the overhead
circuit(s) and their physical configuration, the characteristics of the pipeline(s) such as their diameter and
coating resistance, and the ROW geometry. A graphical approach for determining steady state induced
voltage levels for simple geometries is given by Dabkowski ‘.

Electromagnetic field coupling to the pipeline is a consequence of the currents carried by the
overhead conductors both during steady state and fault condition operation of the electrical circuit. These
currents generate a longitudinal electric field whose direction is parallel to the current carrying conduc-
tors. If the pipeline lies parallel to the conductors full coupling occurs. For perpendicular crossings of the
power line the coupling effect is reduced to zero. The voltage induced unto the pipeline is proportional
to the strength of the longitudinal electric field which, in turn, is proportional to the power line(s) load
current(s).

Although the longitudinal electric field may exist over the complete length of the pipeline expo-
sure, voltages on the pipeline are generally induced at locations where the physical geometry between
the power line(s) and the pipeline changes. For example, if a pipeline lies parallel and at the same dis-
tance from a power line for several tens of miles, the ac voltage level on the pipeline may be close to
zero volts over a large distance. However, if the geometry changes, e.g., the pipeline enters or leaves the
power line ROW, changes distance relative to the power line, or crosses under it, a voltage peak is in-
duced at that location. For a given level of excitation, i.e., strength of the longitudinal electric field, the
induced voltage level, in turn, depends upon parameters such as the pipe diameter and its coating
resistance.

The longitudinal electric field existing along the length of the pipeline exposure is a consequence
of inductive, i.e., magnetic field coupling. This coupling exists both for normal operation of the power
system, i.e., steady state (normal) operation, and also during fault periods such as those precipitated by a
lightning strike to the power line. However, when an electric power circuit is faulted, a portion of the
power system substation(s) fault current maybe injected into the earth at the faulted and nearby towers.
This conducted current causes a coupling to the pipeline by means of a voltage gradient developed in the
soil. As for the longitudinal electric field, voltage is induced onto the pipeline at the location of the
faulted tower.

Using a computer program for predicting the voltage induced on the pipeline, requires the physi-
cal ROW to be modeled. That is, the relative and angular positions of the pipeline, power lines, etc. are
numerically input into the program. Other pertinent parameters such as power line currents, the pipeline
diameter and coating resistance and the soil resistivity are also input. Depending upon the sophistication
of the computer program used for predictive analysis, one soil resistivity value may be input over the to-
tal length of the ROW or variations in the soil resistivity along the ROW may be accommodated. In ad-
dition, single or two layer earth modeling may be employed.

User available computer programs allow treatment of complex situations in varying degrees,
e.g., multiple pipelines, power lines, varying ROW geometry, etc, but the calculations are transparent to
the user. The primary purpose of this paper is to guide the reader through a set of induced voltage pre-
diction calculations for a simple ROW to provide an insight into the procedures used within a typical
computer program. A secondary purpose is to make the predictive calculations using both single and two
layer soil resistivity models in order to assess the suitability of each representation for the accurate pre-
diction of induced voltage levels.

56112
 SOIL RESISTIVITY MEASUREMENTS

One of the input parameters to every computer program is the soil resistivity on or along the
ROW. Generally, the resistivity exhibits variations which can be characterized statisticallf. These are
very important when considering mitigation design. However, since the intent of this paper is to provide
an insight into the differences obtained between one and two layer soil modeling for the prediction of in-
duced voltage levels, a deterministic rather than a statistical characterization will be used for simplicity.
That is, constant resistivity values will be assumed in the examples following.

Single Layer Resistivity Model

Soil resistivity measurements are usually made along the ROW using the Wenner four pin
method. Average soil resistivity values are obtained down to a depth which is approximately equal to the
pin spacing. If the soil resistivity is constant with increasing depth, the same value of measured apparent
resistivity would be obtained independent of the pin spacing used for the measurement.

One of the practical difficulties in obtaining soil resistivity measurements is that it is generally
difficult to obtain physical space on the right-of-way where large pin spacing measurements are feasible.
Hence, resistivity probing is usually confined to the more shallow depths. Another practical difficulty is
that the soil resistivity will exhibit variability along the ROW. Hence, with large pin spacing it is diffi-
cult to disassociate the lateral and the vertical variations in the soil resistivity. The single layer model
requires fewer measurements along the ROW, and therefore, is more expedient. Whether its use is more
cost effective, however, depends upon its adequacy in a particular situation.

Two Layer Resistivity Model

If resistivity measurements vary with pin spacing a single layer model may not yield an accurate
portrayal. The next step in resistivity characterization is to use a two layer model as diagrammed in Fig-
ure 1. Although more complex models may be used on occasion, e.g., N-layer, the two layer model
serves most situations. For a two layer soil resistivity, the apparent resistivity measured at a given pin
spacing will be a function of the upper and lower layer resistivities and the depth of the upper layer. This
is shown in Table 1 where typical resistivity measurement data sets are given for the cases of upper layer
depths of one, two, and five meters, respectively. For each case the upper layer resistivity, pz is as-
sumed to be 1,000 ohm-m and the lower layer resistivity, pi equal to 100 ohm-m. The Table entries
show that measured apparent resistivities are larger than the lower layer resistivity. The presence of the
upper layer adds a confounding factor to the interpretation of the measurements. If a single layer model
should be used, based upon such measurement sets, the soil resistivity would be overestimated. For
situations where the upper layer resistivity is less than the lower layer resistivity, the converse would be
true. Given a set of resistivity measurements such as shown in Table 1, methods are available for deter-
mining the individual layer resistivities and the depth of the upper layer. Usually curve matching tech-
niques are used such as discussed by Sunde’. It should be noted that to resolve these soil parameters with
a reasonable accuracy the largest pin spacing should be at least three to four times the depth of the first
layer. Hence, if the upper layer is quite deep, practical considerations such as time and cost, lack of suit-
able open terrain for obtaining the measurements, and horizontal resistivity variations usually preclude
the development of an accurate two layer model.

56113
 PREDICTION OF PIPE INDUCED VOLTAGES - EXAMPLE CALCULATIONS

Steady State Operation

During normal operation an electric power transmission circuit will deliver three phase currents
to a distant load. That is, the circuit consists of three physical conductors, wherein, the currents in each
conductor are approximately equal, e.g. 1,000, 1,000, and 1,000 amperes. However, the angular phase of
the currents differs by 120 degrees in each conductor. For example, the power company will label one
conductor current as the reference, for example, I, phase = zero degrees, and the other currents as I,
phase = -120 degrees and Ic phase = +120 degrees. Hence, each conductor current may be considered as
a vector t’phasor) having both an amplitude and an angle. The longitudinal electric field developed along
the pipe from a single conductor current is,

Ei = I, x (7 + In(%)) volts/m
I

where I, is the ith conductor current, amps

o is the radian frequency = 2*x *f
p is the permeability constant = 4 * rt * 1O-’ by/m
D, is the equivalent depth of the return current, m, and
di is the radial distance between the current carrying
conductor and the pipeline, m

The equivalent earth return depth for the current is defined as,

where p is the (single layer) soil resistivity, ohm-m, and

f is the power system frequency, Hz

For a three phase system, the all three current carrying conductors must be considered. Hence, the total
longitudinal electric field at the pipe is equal to,

E total= Ea + Eb + Ec volts/m

If overhead shield wires are present, the computer program will calculate the currents induced in
these conductors and add their contribution to the field in Equation 3. Because of the factor, D, the lon-
gitudinal electric field from any one conductor is a weak function of the soil resistivity. However, it may
be shown4, that for a balanced three phase circuit, i.e., all three conductor currents equal, the total field
as given by Equation 3 is essentially independent of the soil resistivity. This is true for all distances from
the power line if overhead shield wires are not present, and for distances close to the power line if
grounded shield wires exist.

56114
Steady State Example

Longitudinal Electric Field. Figure 2 portrays a simple ROW collocation of a pipeline with an
overhead power line. In the Figure it is assumed that coming from the left the power line and pipeline are
collocated for a long distance, e.g. >15-20 kilometers. As shown in the Figure the pipeline then leaves
the power line ROW at right angles. At this location an induced voltage peak occurs.

The following parameters will be assumed:

Configuration : Horizontal
Height of Conductors: 40 feet
Conductor Spacing: 20 feet
Conductors’ Current: 1,000 amp
Shield Wires: none
Power Line I Pipeline Separation: 50 feet
Soil Resistivity: Single layer @ 100 ohm-m or 1,000 ohm-m

From Equations (1) through (3) the longitudinal electric field at the location of the pipeline is cal-
culated as 3 1.2 mV/m for either of the above two soil resistivities. Hence, there would be no advantage
to using a two layer soil resistivity model for making this calculation.

Pine Voltage Calculation. Calculation of the pipe voltage requires that the propagation constant,
F of the pipeline be known. To calculate the constant, the following pipeline parameters will be
assumed:

Pipe Diameter: 24 inches

Pipe Coating Resistance: 800 Kohm-ft2
Coating Thickness: 15 mils
Pipe Burial Depth: 5 feet
Soil Resistivity: Single layer @ 100 ohm-m or 1,000 ohm-m

Due to the high coating resistance, the pipeline propagation constant is essentially independent of
the assumed soil resistivity, and is calculated to be approximately equal to 2.41* 1Oe4m-’ for either of the
soil resistivities.

For a simple ROW such as that considered here, where only a single physical discontinuity is en-
countered, i.e., the pipeline recedes from the power line, a simple formula may be used to determine the
peak voltage at this location, The peak voltage on the pipeline at this location, for either soil resistivity,
may be shown to be equal to,

E 31.2 x 1O-3
= 64.7 volts (4)
‘= = = 2 x 2.41 x 1O-4

This sample calculation shows that a complex model, e.g., a two layer earth, for the soil resistiv-
ity is not necessary for calculation of the steady state pipe induced voltage from a balanced three phase
power system. A simpler computer program utilizing a single layer earth model is adequate.

561/5
Fault Period Operation

Longitudinal Electric Field. When a power system circuit is faulted, e.g., by a lightning strike, the
faulted phase conductor carries a short-circuit current which is much larger than the currents in the other
phase conductors. Hence, the power circuit is unbalanced and the longitudinal electric field driving the
pipeline is determined by Equation (1) rather than from Equation (3). Because of the D, factor in Equa-
tion (l), the longitudinal electric field for this scenario will be a function of the soil resistivity and a dis-
cussion of single and two layer soil resistivity modeling is necessary.

The strength of the longitudinal electric field at the pipeline will lie between the values calculated
using the upper layer resistivity value and the lower layer resistivity value. If the upper layer is shallow,
the field will be primarily determined by the lower layer. If the upper layer is very deep, the field will be
primarily determined by the upper layer resistivity. The variation of the longitudinal electric field as a
function of the upper layer depth is plotted in Figure 3 for the ROW example discussed above. The plot
assumes a 50 foot power line pipeline separation and is based upon the assumption of a fault current
level of 15,000 amperes. It may be seen that the upper layer resistivity does not fully dominate until the
depth of the layer approaches the value of the return current depth, D, which may have values ranging
from several hundred to several thousand meters. Since soil resistivity measurements are usually made
only to a depth of 10 meters or less, information is not available as to whether an additional soil layer,
which can significantly alter the field, may present at greater depths. Hence, although the field strength is
dependent upon the deep earth resistivity, information is generally available for only relatively shallow
depths. Hence, the usefulness of a two layer soil model for determining the field accurately is severely
restricted.

For the ROW example, it will be assumed that for the two layer model, the upper layer resistiv-
ity is equal to 1000 ohm-m and the lower layer equal to 100 ohm-m. Assuming also a relatively shallow
depth for the upper layer, the longitudinal electric field strength is determined primarily by the lower
layer. For a value of p = 100 ohm-m in Equation (1) a calculated field of E = 4.1 V/m is obtained.

Induced Pine Voltage. If a single layer earth model were used, assuming a layer depth of one me-
ter, Table 1 indicates that a resistivity of 530 ohm-m would be measured at pipe depth. Using this value
in Equation (1) would yield a field of 5.0 V/m, thus leading to an overestimate of 22 percent. Con-
versely, if the upper and lower layer resistivities were reversed, the field would be underestimated by ap-
proximately the same amount.

If the power line tower is faulted at a location within the long parallel exposure, the magnetically
induced voltage across the pipe coating is equal to,

V,&= E = 4.1 = 17,000 volts (5)

r 2.41 x lOA

A fault on a tower located at the right angle bend, i.e., where the pipeline leaves the right-of-way would
result in an induced voltage equal to one-half the above value or 8,500 volts.

56116
 Conductivelv Induced Voltage. In addition to the inductively coupled field, the current flowing
into the soil at the faulted tower will cause a potential gradient along the pipeline which adds to the mag-
netically induced longitudinal electric field. Using a single layer soil resistivity model, at the location of
closest approach to the faulted tower, the pipe voltage across the coating caused by soil conduction may
be approximated by,

PXI
V
'Ond= 2 x 71x r

where I = fault current entering the soil, (15,000 amp),

r = radial distance between the tower ground
and the pipeline (15.2 m), and
p = single layer soil resistivity approximation (530 ohm-m)

The value for the soil resistivity has been arbitrarily chosen as the apparent resistivity at pipe depth as-
suming a one meter depth for the upper soil layer. For the above assumed values, Vcondis equal to,

530x15,000
V
cond= 2~7~~15.2
= 83,242 volts W

Vcondwill add phasor-wise (vectorially) to Vind, thus increasing the total fault current voltage across the
pipe coating. One of the factors leading to the large value above for the conducted voltage is the assump-
tion of no shield wires. If an electric power line has continuously grounded shield wires, much of the
fault current will return to the substation(s) via the wires. In typical installations this may reduce the soil
injected current at the faulted towers to ten to fifteen percent of the full fault current, thus reducing Vcond
proportionately.
The voltage calculated in Equation (6a) was based upon a single layer approximation to the as-
sumed two layer soil resistivity model, i.e., an upper layer resistivity of 1,000 ohm-m and a lower layer
resistivity of 100-m. For these layer resistivities, a set of computer program calculations yields the plot
of Figure 3 which gives the value of VcD,,,,at the pipeline as a function of the depth of the upper layer.
For an upper layer depth of one meter as assumed in the example, the plot indicates that the value of
Vcondwould actually be only 15,690 volts. The plot shows that for the voltage to reach the value calcu-
lated in Equation (6a), an upper layer depth of approximately 16 meters would be required. Hence, the
use of the single layer model results in an overestimation of the conduction caused pipe coating voltage
by a factor of 5.3. Conversely, if the upper and lower layer resistivities were reversed, an underestimate
of the voltage would be obtained. Thus, to obtain an accurate characterization of the induced voltage, a
two layer soil resistivity model is necessary.

In other situations where shield wires may carry the bulk of the fault current back to the substa-
tion(s) and the soil resistivities are relatively low, Vcon,,may be a small fraction of Vmd. In this case a sin-
gle layer soil resistivity model may be adequate. However, in general, caution should be exercised and
the two layer resistivity model considered when determining conductively induced voltages.

56117
 SUMMARY AND CONCLUSIONS

The mechanisms by which voltages are coupled unto buried pipelines from collocated overhead
electric power transmission circuits have been reviewed. These include magnetic field inductive cou-
pling during normal operation (steady state) of the power line and during fault current periods, and soil
injected current conductive coupling during faults. For a simple ROW scenario, equations have been
presented for the prediction of the pipe induced voltage caused by each type of coupling. For each type a
comparison has been made of the predicted voltage levels when single layer and two layer soil resistivity
models are assumed. In general it is found that for magnetic field inductive coupling, the single layer
model suffices. However, a two layer soil resistivity model is usually required to accurately predict the
pipe voltage induced by conduction current through the soil.

REFERENCES

1. J. Dabkowski, “The Calculation and Mitigation of Induced Voltages on Buried Pipelines, I’

Materials Performance 23, l(l981): p. 9.

2. J.Dabkowski, ‘A Statistical Approach to Designing Mitigation for Induced AC Voltages on

Pipelines, ” Materials Performance 35, 8( 1996): p. 9.

3. E. D. Sunde, Earth Conduction Effects in Transmission Systems (New York, NY: Dover
Publications, 1968) p. 48.

4. J. Dabkowski, “The Calculation of Electromagnetic Coupling from Overhead Transmission

Lines,“IEEE TransactionPAS-100,8(1981): p. 3850.

TABLE 1
APPARENT SOIL RESISTIVITIES (OHM-M) FOR TWO LAYER SOIL MODEL
Upper layer = 1,000 ohm-m, Lower layer = 100 ohm-m

56118
 air-soil interface
III 4

d=layer depth upper layer: resistivily = P2

.

lower layer: resistivity = Pl

Figure 1 Two Layer Soil Model

Power Line

.ong Parallel Collocation

, Induced Voltage Peak

Pipeline

Exit from ROW

Not to Scale:

Figure 2 Right-of-Way Example

56119
 Upper Layer Soil ResisMly = 1000 ohm-m:Lower Layer = 100 ohm-m

Figure 3 Fault Current Coupled Longitudinal Electric Field

Figure 4 Conductively Coupled Pipe Voltage

561110