IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 28, NO.

1, JANUARY 2013 427

Effects of Segmented Shield Wires on Signal

Attenuation of Power-Line Carrier Channels
on Overhead Transmission Lines—Part I:
Modeling Method
Shanshan Yang, Student Member, IEEE, and Gregory A. Franklin, Senior Member, IEEE

Abstract—Shield wire segmentation has been used for quite lines. Single-phase-to-ground coupling methods are most fre-
some time as an effective method for reducing transmission-line quently used for lower voltage level transmission lines [1], [2].
power losses, with most segmented shield wire applications on As efficiency becomes more important for electric utilities,
extra high-voltage lines utilizing phase-to-phase power-line car-
rier (PLC) coupling. As transmission system efficiency becomes shield wire segmentation has been used as an effective way
more of a priority, some utilities may consider applying segmented to reduce power loss from shield wires, and has been applied
shield wires to lower voltage lines, such as 115-kV and 230-kV to some high-voltage (HV) transmission lines [3], [4]. It was
lines, which most often utilize single-phase-to-ground PLC cou- concluded in [5] that the segmented shield wire system is not
pling. Thus, knowledge of the effects of segmented shield wires on suitable for use with PLC communications applications, which
PLC signal attenuation for various coupling methods would be
useful to relay engineers. The purpose of this paper is to present a are critical systems in transmission-line protection schemes.
generalized method to model transmission lines for signal attenu- From [5], the unsuitability of segmented shield wires with PLC
ation calculation at PLC frequencies, with special consideration applications results from pronounced resonance that can occur
given to shield wire connections. Terminating impedance and in a band of frequencies, resulting in excessively high PLC
phase wire transposition are considered as well. The modeling signal attenuation. However, the previous work described in
method presented is validated with field measurements. In a com-
panion paper, signal attenuation results considering the effects of [5] is based on a particular type of segmentation and coupling
various parameters are analyzed. method, which determines the particular frequency band where
the resonance occurs. For other frequencies, PLC coupling
Index Terms—Power-line carrier (PLC), signal attenuation,
methods, and shield wire segmentation methods, the resonance
transmission-line modeling.
effects could be much less severe. Therefore, a generalized
signal attenuation calculation model of the transmission line,
I. INTRODUCTION including the shield wires, was developed to analyze the effects
of shield wire segmentation on PLC signal attenuation.
Several papers have been written on transmission-line mod-
eling and modal analysis. References [6]–[10] provide gener-

P OWER-LINE carrier (PLC) is a crucial communication

system for power system protection. The high-frequency
carrier signals are transmitted over the existing transmission
alized information on this topic along with calculation results
compared with field tests. These papers assume that the shield
wires are continuously grounded (i.e., grounded at each tower).
lines. The frequency range of the PLC signals is 30 to 500 kHz, When one-fourth of the wavelength of the carrier signal is less
which is classified as low radio frequency. Line traps, coupling than a tower-to-tower span of the transmission line, the shield
capacitors, line-tuning equipment and communication equip- wires can be eliminated since it can be assumed that the shield
ment are the essential components of the PLC system, along wires have zero potential along their entire length. However,
with the transmission line. The descriptions of these compo- for higher frequencies, especially in the vicinity of frequencies
nents can be found in [1]. To propagate PLC signals along the that make the electrical length between towers odd multiples
line, different coupling methods can be used to feed one or more of one-half wavelength, the aforementioned assumption of zero
conductors of a multiconductor line. The high importance of potential along the shield wire cannot be made, and it is shown
extra high-voltage transmission lines typically warrants the use in [11], [12] that the standing-wave effects on the shield wires
of phase-to-phase coupling methods for PLC systems on these can cause drastic changes in attenuation, impedance, and modal
vector distributions.
Manuscript received April 30, 2012; accepted August 06, 2012. Date of pub- Since field measurements of signal levels are usually made in
lication October 16, 2012; date of current version December 19, 2012. Paper
phase domain. Section III considers modeling in the phase do-
no. TPWRD-00444-2012.
The authors are with the Electrical and Computer Engineering Department, main. Information on the modeling of transmission lines with
University of Alabama at Birmingham, Birmingham, AL 35205 USA (e-mail: segmented shield wires needed for actual calculations of signal
[email protected]; [email protected]).
attenuation in the phase domain is provided. The chain-param-
Color versions of one or more of the figures in this paper are available online
at http://ieeexplore.ieee.org. eter matrix method is used to model the effects of the nonuni-
Digital Object Identifier 10.1109/TPWRD.2012.2217759 formly distributed towers and different shield wire connections.

0885-8977/$31.00 © 2012 IEEE

428 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 28, NO. 1, JANUARY 2013

Since the phase domain signal attenuation of the entire PLC

channel also depends on the particular coupling method, the
total line length, and the terminal impedances, all of these pa-
rameters are considered in the generalized model.
To illustrate the steps to calculate phase domain parameters
at the receiving end of transmission lines, a sample calculation
is provided in Section IV for a 500-kV transmission line with
segmented shield wires where the towers are nonuniformly dis-
tributed. Furthermore, field tests were performed on the line to
validate the calculations using the described methods. Results
from these tests in the form of a comparison of calculated versus
measured voltage levels are provided.
In addition to the effect of the coupling method on PLC signal
attenuation, other problems, such as resonance conditions re-
sulting in high signal attenuation, may also occur depending on
the PLC signal frequency and the length of the shield wire seg-
ments. The considerations when applying PLC systems on over-
head transmission lines with segmented shield wires are pro-
vided in a companion paper.

II. SHIELD WIRE CONNECTIONS

Shield wires (also called ground wires or earth wires) are in-
stalled on overhead transmission lines to minimize the likeli-
hood of direct lightning strikes to the phase conductors. On most
lines, the shield wires are grounded at each tower, forming a
continuously grounded system [as shown in Fig. 1(a)]. How-
ever, the closed loop between shield wires and ground make it
possible for induced currents to flow, resulting in power loss.
As efficiency becomes more of a priority, utilities want to re-
duce the power loss from shield wires. Shield wire segmenta-
tion is implemented by opening the shield wires at some points
along the line to break the closed loop. As a result, shield wire
segmentation is an effective way to reduce the power loss and
Fig. 1. Shield wire connections. (a) Continuous shield wires. (b) Type A seg-
improve efficiency. mentation. (c) Type B segmentation.
Type A and Type B segmented shield wires shown in
Fig. 1(b) and (c) are different in the opening points. Type A
segmentation is opened in the middle, and grounded at each procedure much more complex for multiconductor transmis-
tower. Type B segmentation is grounded at the beginning and sion lines. Therefore, a more generalized and simplified phase
opened at the end for each section. It should be noted that the domain PLC signal attenuation calculation model in which the
shield wire sections in Fig. 1 can span several towers. various shield wires connections, coupling methods, and line
terminations are considered is provided in this paper.
The signal attenuation in decibels can be calculated by the
III. MODELING METHOD
logarithmic equation through power, voltage, or current. The
[13] described the modeling process for PLC signal at- voltage equation is used in this paper. Reference
tenuation calculation on overhead transmission lines with [10] provides an equation relating sending-end voltage to
segmented shield wires. The model in [13] assumes that the receiving-end voltage as
system is terminated in its characteristic impedance and uses
outer-to-outer phase coupling. This assumption is useful to
check the signal attenuation of the transmission lines without (1)
disturbance from other parameters using the particular coupling where
method. However, the actual line terminations are always far • is a unit matrix;
from being that of the characteristic impedance of the line, • is the phase domain propagation constant matrix;
and other coupling methods are often used, especially for • is a multiphase reflection coefficient matrix at the re-
lower voltage levels. Refeences [14]–[16] provide a method ceiving end of the line.
to calculate the signal attenuation for PLC signals accounting The main task for the modeling process in this paper is to
for different terminating impedances. However, this method is find out the exact expression for and in (1). For a typ-
based on the concept of input impedances, which makes the ical transmission system consisting of phase conductors and
YANG AND FRANKLIN: EFFECTS OF SEGMENTED SHIELD WIRES ON SIGNAL ATTENUATION OF PLC CHANNELS—PART I 429

TABLE I
BOUNDARY CONDITIONS FOR DIFFERENT SHIELD WIRE CONNECTIONS

Fig. 2. Transmission system including its terminating impedance.

shield wires with a number of sections cascaded together as

shown in Fig. 2, the end-to-end system response can be ob-
tained by multiplication of the transmission matrix of each sec-
tion shown in Fig. 1.

A. Basic Chain-Parameter Matrix

all along the entire line. As a result, it is possible to eliminate the
The modeling process starts from two towers in which the shield wire components from the transmission matrix and make
shield wire discontinuities occur by relating the voltages and use of the phase conductor continuity to combine all of the trans-
currents at both ends. Alternative port representations of the line mission matrices together. For example, consider a transmission
can be used. The chain-parameter representation is one of the system consisting of 3 phase conductors and 2 shield wires. Re-
most useful expressions of two-port parameters for a two-con- ferring to Fig. 1, the boundary conditions for the shield wires
ductor system, and can be extended to a multiconductor trans- are provided in Table I.
mission system. The chain-parameter representation relates the By applying the boundary conditions shown in Table I, the 10
voltage and current at the termination of the line to the sending 10 transmission matrix of (2) for each section or half section
end voltage and current according to (2) and (3) [17] can be reduced to a 6 6 matrix with only phase conductors.
The reduction process for different shield wire connections can
be found in [13]. As a result, the chain parameter matrix in (2)
(2) can be reduced to

(5)

(3) where for the continuous shield wire section as shown in

Fig. 1(a)

In (3), , , , and are by matrices; ,

, , and are by matrices, , , , and (6)
are by matrices and , , , and are For the type A segmentation section shown in Fig. 1(b)
by matrices. The , , , matrices elements can be
found from (7)

• is the chain parameter matrix of the phase conductors

(4) for the first half

where (8)
• is the phase domain voltage propaga- • is the chain parameter matrix of the phase conductors
tion constant matrix in per-unit length, and the series for the second half
impedance per-unit length matrix and shunt admit-
tance per-unit length matrix can be calculated by
following the steps stated in [10]; (9)
• is the phase domain characteristic
(surge) impedance of the line; C. Overall Chain Parameter Matrix
• is the line length between two towers or discontinuities. For transmission systems without termination as shown in
Fig. 4, all discontinuities are considered from the shield wire
B. Shield Wire Elimination discontinuities. After applying the boundary conditions to re-
The shield wires are either grounded or opened at the end of duce the shield wires from the transmission matrix, the overall
each section, or open at the middle point for the type A segmen- chain parameter matrix for the phase conductors that relates the
tation. Phase conductors are continuous without discontinuities sending end to the receiving end can be obtained as the product
430 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 28, NO. 1, JANUARY 2013

The equivalent phase domain propagation constant matrix

can be calculated from

(17)

D. Including Terminating Impedance

Fig. 3. Transmission-line diagram between two towers. The unknown parameter left in (1) is . The termination is
shown in Fig. 2. The multiphase reflection coefficient method
in [9] proves to be suitable for analyzing PLC signal propaga-
tion. This method is concise in that the terminating impedances
do not have to be converted to the modal domain. The verified
phase domain multiphase reflection coefficient matrix at the re-
ceiving end of the line is derived in [10] as

(18)
Fig. 4. Transmission system with sections.
where is the transmission system terminating impedance.
For the special case where the line is terminated in its char-
of the chain parameter matrices of each section in the proper acteristic impedance, the reflection coefficient in (18) is calcu-
order. The process is described as lated to be zero matrix. Thus, the overall equation between the
sending end and the receiving end voltages in (1) can be reduced
to
(10)
(19)
(11)
E. Including Transpositions in Phase Conductors
Phase conductor transposition is an effective way to equalize
(12) the phase impedances and reduce the induction from normal
operating currents and voltages [18]. However, transposition
(13) can cause additional attenuation of a PLC signal. As a result,
accounting for phase conductor transposition is necessary to
where model the PLC channel performance accurately. The defined
• is the chain parameter matrix of the phase conductors model can easily be extended to include the transpositions and
for the entire line; applied to the analysis of the effects.
• consists the elements of relating the re- For the two different phase conductor transposition types as
ceiving-end voltages to sending-end voltages; shown in Fig. 5, the relationship of the voltages after (at the
• consists of the elements of relating the receiving points ’ and ’) and before (at the points and ) transposi-
end voltages to sending end currents; tion can be expressed as
• consists of the elements of relating the receiving
end currents to sending end voltages; (20)
• consists of the elements of relating the receiving
end currents to sending end currents.
Due to the discontinuities from the shield wires and the
(21)
nonuniform distribution of section lengths, the phase domain
propagation constant for the entire line should be recalculated
from the transmission matrix generated in (12). By comparing So the phase transposition matrices of the two different types of
(12) and (4), the following equations can be used to derive the transpositions are
equivalent phase domain propagation parameters:

(14) (22)
(15)
and
where the equivalent phase domain characteristic impedance
matrix can be calculated through
(23)
(16)
YANG AND FRANKLIN: EFFECTS OF SEGMENTED SHIELD WIRES ON SIGNAL ATTENUATION OF PLC CHANNELS—PART I 431

Fig. 7. Configuration of the 500-kV transmission line for tests.

Fig. 5. Different phase conductor transposition types.

• shield wires are 3/8-in steel.

It should be noted that the accurate data for the parameters,
such as the height of the conductors and the ground resistivity,
normally vary from span to span. Therefore, they were roughly
approximated.

B. Measurement Description
Fig. 6. Section distribution of the transmission line used for tests.
Two tests were performed on the test line with the frequency
at 100 kHz and 300 kHz with the signal applied to the center
phase. Phase-to-ground coupling equipment is installed on the
The or matrices can be inserted into the original trans- center phase, which made it possible to measure voltages at the
mission chain matrices to generate the modified chain matrix for receiving-end. The 500-kV test line has 1000- wideband line
the PLC channel traps that present almost purely resistive impedance to frequen-
cies within their stopband. The frequencies of the test signal,
which were 100 kHz and 300 kHz, were within the stopband of
(24) the line trap.
For the case in which several transpositions are present, the The tests consisted of applying a signal at one end of the line
chain matrix can be generated by inserting each corresponding and measuring the magnitude of the signal at the other end of
transposition matrix into (24). the line using a frequency-selective voltmeter. At the sending
end of the line, the signals were applied after the IMT, which is
direct to the line-tuning unit. As a result, when the series LC unit
IV. FIELD MEASUREMENT
was tuned to resonate at the test frequency, the phase-to-ground
voltage level applied to the transmission line should approxi-
A. Test Line Description
mately be the voltage level measured at the terminals of the
The field tests were performed on a nontransposed 500-kV signal generator. The applied voltage to the center phase was
transmission line with segmented shield wires as shown in 5 V for both tests. At the receiving end of the line, the voltage
Fig. 6. The shield wires have six segmentations along the total was measured across the drain coil, which was approximately
line length. The segmentation here is recognized as type A the phase-to-ground voltage level on the line. However, the only
segmentation. So the shield wires are opened at the towers parameters known at the receiving end are the line trap imped-
which are at the ends of each segmentation, including the very ances, which are 1000 on the center phase and one outer
beginning and very end of the line. The total line length is phase. The bus impedance can only be approximated.
around 49.6 km with the actual value of each section shown
in Fig. 6. The shield wires are only grounded at the towers C. Calculation Procedure
which are in the middle of each segmentation and insulated at
all remaining towers. The receiving end voltages can be calculated using the mod-
The test line is a delta configuration as shown in Fig. 7. Re- eling method described in Section III. Using 100 kHz as an ex-
ferring to Fig. 7: ample, the calculation steps are as follows:
• phase conductors are Ortolan 1033.5 kcm 45/7 ACSR; the 1) Calculate the series impedance and shunt admittance
geometric mean radius (GMR) of the phase conductors is of the line at 100 kHz using the steps stated in [10]. The
137 mm; ground resistivity in this example calculation is considered
432 IEEE TRANSACTIONS ON POWER DELIVERY, VOL. 28, NO. 1, JANUARY 2013

to be 100 -m. The calculated and are shown in the D. Comparison of Measured and Calculated Voltages
equation at the bottom of the page. The phase-to-ground line terminating impedance consists of
2) Generate the chain parameter matrix for each section with the line trap impedance in series with a capacitive reactance
phase conductors and shield wires included. All of the representing the phase-to-ground bus capacitance. The line
chain parameter matrices are 10 10 matrices. trap impedance value is the only parameter known. As recom-
3) Apply boundary conditions to reduce the shield wires. The mended in [19], the bus impedance used at PLC frequencies
reduced chain parameter matrices are 6 6 matrices. below 100 kHz is approximately , and for frequencies
4) Combine the chain parameter matrix of each section to- above 100 kHz, a value of to should be used.
gether to obtain the phase conductors’ chain parameter ma- Four different sets of terminating impedance networks were
trix for the entire line. compared as follows: is the characteristic impedance of
5) Calculate the phase domain propagation constant and char- the line, is the recommended [19] phase-to-ground bus
acteristic impedance from elements of the chain parameter impedance in series with the line trap impedance with short cir-
matrix using (12) and (13). The calculated results are as cuits between phases, is the recommended phase-to-ground
follows. bus impedance in series with the line trap impedance with open
1) Characteristic Impedance: circuits between phases, and is the line trap impedance
only with short circuits elsewhere

2) Propagation Constant:

1) Specify the terminating impedance and use (1) to calcu-

late the receiving end voltage. Here, take the characteristic
impedance termination as an example. Since the center
phase-to-ground coupling is used, the sending end voltage
is known as .

As a result, with the characteristic impedance as the termi-

nating impedance, the calculated receiving end voltage for the
center phase-to-ground coupling is 3.68 V.
YANG AND FRANKLIN: EFFECTS OF SEGMENTED SHIELD WIRES ON SIGNAL ATTENUATION OF PLC CHANNELS—PART I 433

TABLE II [4] B. Foehner and P. Thomas, “Sectionalizing static wire prevents 500-kV
MEASURED AND CALCULATED VOLTAGE LEVELS AT THE RECEIVING END outages,” Transm. Distrib., pp. 56–57, Sep. 1978.
OF LINE [5] J. A. B. Faria, “Broadband resonance effects in power lines with
sectionalized shield wires,” in Proc. IEEE Southeast Conf., 1990, pp.
654–658.
[6] M. C. Perz, “Natural modes of power line carrier on horizontal three
phase lines,” IEEE Trans. Power App. Syst., vol. PAS-83, no. 7, pp.
679–686, Jul. 1964.
[7] M. C. Perz, “A method of analysis of power line carrier problems on
three-phase lines,” IEEE Trans. Power App. Syst., vol. PAS-83, no. 7,
pp. 686–691, Jul. 1964.
[8] L. M. Wedepoh and S. E. T. Mohamed, “Multiconductor transmission
lines: Theory of natural modes and fourier integral applied to transient
analysis,” Proc. Inst. Elect. Eng., vol. 116, no. 9, Sep. 1969.
[9] L. M. Wedepohl, “Electrical characteristics of polyphase transmission
systems, with special reference to boundary value calculations at power
line carrier frequencies,” Proc. Inst. Elect. Eng., vol. 112, no. 11, pp.
2103–2112, Nov. 1965.
[10] G. A. Franklin, “Using modal analysis to estimate received signal
levels for a power-line carrier channel on a 500-kV transmission line,”
IEEE Trans. Power Del., vol. 24, no. 4, pp. 2446–2454, Oct. 2009.
[11] L. M. Wedepohl and R. G. Wasley, “Wave propagation in polyphase
transmission systems: Resonance effects due to discretely bonded earth
wires,” Proc. Inst. Elect. Eng., vol. 112, no. 11, Nov. 1965.
[12] J. A. B. Faria and J. F. B. da Silva, “The effect of randomly earthed
ground wires on PLC transmission: A simulation experiment,” IEEE
Trans. Power Del., vol. 5, no. 4, pp. 1669–1677, Oct. 1990.
[13] S. Yang and G. A. Franklin, “Modeling of overhead transmission lines
with segmented shield wires for signal attenuation calculation in power
line carrier systems,” in Proc. IEEE Southeast Conf., 2011, pp. 12–16.
[14] S. Cristina and M. D’Amore, “Propagation on polyphase lossy power
Table II shows a comparison of calculated versus measured lines: A new parameter sensitivity model,” IEEE Trans. Power App.
received signal levels for the aforementioned terminating im- Syst., vol. PAS-98, no. 1, pp. 35–44, Jan. 1979.
pedances, with three different values of ground resistivity of [15] S. Cristina and M. D’Admore, “Performance sensitivity of a power line
carrier channel,” IEEE Trans. Power App. Syst., vol. PAS-98, no. 4, pp.
100, 200, and 300 -m. As shown in Table II, the characteristic 1337–1344, Jul. 1979.
impedance, which is the reflectionless terminating impedance, [16] S. Cristina and M. D’Amore, “A new efficient method to evaluate
produced the least accurate results. It can be expected since the carrier channel performances on long EHV power lines with trans-
positions,” IEEE Trans. Power App. Syst., vol. PAS-101, no. 9, pp.
actual line termination was far from being that of the charac- 3053–3060, Sep. 1982.
teristic impedance of the line. The other three termination im- [17] C. R. Paul, Analysis of Multiconductor Transmission Lines, 2nd ed.
pedances resulted in much better estimates of the receiving end Hoboken, NJ: Wiley/IEEE.
[18] P. D. Morgan and S. Whitehead, “The impedance and power losses of
signal levels. three-phase overhead lines,” J. Inst. Elect. Eng., vol. 68, no. 399, pp.
367–408, Mar. 1930.
V. CONCLUSION [19] Relaying Communications Channels Application Guide. Markham,
ON, Canada: Pub. No. GET-8034, GE Power Management, 2001.
A generalized phase domain signal attenuation calculation [20] J. A. B. Faria, “On the resonance effects due to ground wires in trans-
method for a PLC channel on overhead transmission lines with mission lines with non-uniform soil conductivity and non-uniform
tower resistances,” IEEE Trans. Power Del., vol. 7, no. 1, pp. 29–38,
segmented shield wires was described in this paper. A group Jan. 1992.
of field measurement data along with the model calculation re-
sults is provided to verify the modeling method. The method Shanshan Yang (S’10) received the B.S. and M.S. degrees in electrical power
presented herein can be used by relay engineers or PLC appli- system and power system protection from Shandong University, Jinan, China,
in 2005 and 2008, respectively, and the Ph.D. degree in computer engineering
cation engineers to help select appropriate PLC frequencies and from the University of Alabama at Birmingham in 2012, with research in the
coupling methods for PLC applications on overhead transmis- area of computer-based modeling and analysis of power system protection and
sion lines with continuous or segmented shield wires. power-line carrier applications on overhead transmission lines.
Currently, she is an Applications and Development Engineer with Electrocon
Future work will be on the analysis of other effects that seg- International, Inc. Ann Arbor, MI. Her responsibilities include the development
mented shield wires may bring into the transmission system, of detailed models of protective relays.
such as influences to transient overvoltage studies.

REFERENCES Gregory A. Franklin (M’03–SM’11) received the B.S.E.E., M.S.E.E., and

[1] IEEE Guide for Power-Line Carrier Applications, IEEE Standard 643- Ph.D. degrees in electrical and computer engineering from the University of
2004, 2005. Alabama at Birmingham in 1996, 1997, and 2005, respectively.
[2] M. P. Sanders and R. E. Ray, “Power line carrier channel & application From 1997 through 2006, he was with Southern Company, Birmingham, AL,
considerations for transmission line relaying,” Pulsar Doc. No.C045- working in the area of substation protection and control. In 2007, he resigned
P0597. from Southern Company as a Senior Engineer to join the Faculty of the Elec-
[3] A. J. F. Keri, A. Nourai, and J. M. Schneider, “The open loop scheme: trical and Computer Engineering, Department at the University of Alabama at
an effective method of ground wire loss reduction,” IEEE Trans. Power Birmingham. Dr. Franklin is a registered Professional Engineer in the State of
App. Syst., vol. PAS-103, no. 12, pp. 3615–3624, Dec. 1984. Alabama.