Cable Heating Effects due to Harmonic Distortion

in Electrical Installations
Kevin OConnell, Martin Barrett, Jonathan Blackledge and Anthony Sung
AbstractThe increasing use of non-linear loads in elec-
trical installations has exacerbated the problems of harmonic
distortion in industrial and commercial electrical systems. In
the UK the current practice to determine the cable size for
an electric circuit is to use BS7671. However, previously the
16th edition IEE Wiring Regulations only dealt with situations
where cables attain the conductor temperature generated by
sinusoidal currents at the fundamental power frequency. This
paper outlines the methods available to determine the minimum
size of line conductors for protection against overload currents,
taking into account the harmonic content of the load current,
and explains the harmonic rating factor Cf introduced in 2008
for cables that are under signicant harmonic inuences. Since
the effect of harmonic currents is to increase the joule losses
in a cable, the ampacity of the cable will need to be corrected
to ensure the maximum conductor operating temperature is
not exceeded. An experiment on how cable temperature can be
measured under harmonic inuence is described, and several
sets of measurements taken on a typical cable are analysed. The
paper concludes that direct usage of the BS7671 rating factor
for harmonics appears to be rather conservative and could lead
to over-sizing of the line conductors for three-phase circuits, but
is deemed benecial in the long run.
Index TermsNon-linear loads, harmonics distortion, har-
monic contents, triplen harmonics, heat emitter, ampacity,
cables, correction factors, derating factor, BS7671, IEC, NEC,
skin effect, proximity effect.
I. INTRODUCTION
H
ARMONIC distortion in low voltage electrical in-
stallations is now a common occurrence in the built
environment. It is caused by non-linear loads and historically
was only associated with industrial power systems that
used large static power converters. The increased usage of
information technology equipment and low energy devices
in buildings over the past twenty years, which result in non-
linear electrical loads has introduced a high level of harmonic
distortion into the LV electrical system. As a result, it has
become necessary to establish criteria for limiting problems
from system quality degradation. E. W. Fuchs et al [1]
reported that the present versions of IEEE519-1992 [2] and
IEC61000-3-2 [3] harmonic standards are too restrictive for
low-frequency voltage and current harmonics, as they apply
to residential power systems.
II. BS7671 METHOD FOR DETERMINING CABLE SIZE
FOR PROTECTION AGAINST OVERLOAD
In the UK the general method used to determine the size of
a line conductor is based on the method described in BS7671
(17th edition IEE Wiring Regulations) [4].
Manuscript submitted on 1 March, 2012.
K. OConnell, School of Electrical Engineering Systems, Dublin Institute
of Technology
M. Barrett, School of Electrical Engineering Systems, Dublin Institute of
Technology
J. Blackledge, Stokes Professor, Dublin Institute of Technology
A. Sung, Director of Research of N. G. Bailey Ltd
In appendix 4 of BS7671, tables of ampacity and the
associated impedance drop of common types of cables can
be found. However, the ampacity published are based on the
assumption that there is no harmonics present in the cabling
system. Clearly the three-phase four wire ampacity rating
column in the tables does not count the neutral conductor as
a current carrying conductor hence it has no heat emission.
The basic method to determine the size of a line conductor
for protection against overload is given in BS7671 Appendix
4, Section 5 and is as follows:
I
Z
= I
t
C
a
C
g
C
i
C
c
I
n
I
b
(1)
where I
Z
is the continuous service ampacity in amperes
of a cable having taken all the applicable rating factors
into account under dened installation conditions; I
t
is the
tabulated ampacity in amperes of a cable (BS7671 Table
4A2, gives a schedule of appropriate ampacity tables in-
cluded in BS7671); C
a
is the rating factor for ambient
temperature and is given in BS7671 Tables 4B1 and 4B2;
C
g
is the rating factor for conductors that are grouped in
dened installation arrangements (given in BS7671 Tables
4C1, 4C2, 4C3, 4C4 and 4C5); C
i
is the rating factor for
conductors embedded within thermal insulation, (given in
BS7671 Part 5 Regulation 523.7 and Table 52.2); C
c
is
the rating factor for the type of protective device or under
dened installation conditions (given in BS7671 appendix 4,
section 5.1 and Tables 4B3); I
n
is the nominal rated current
or current setting in amperes of the over-current device (its
value can be selected from either BS7671 Appendix 3 or
device manufacturers technical data literatures); I
b
is the
design current in amperes of the circuit under normal steady
state operating conditions and calculated using the declared
nominal voltage level.
For single phase loads and for single phase motors
I
b
=
P
U
0
cos
and I
b
=
P
m
U
0
cos
respectively, where P is the total active power of the load in
W, P
m
is the total mechanical power of the load in W, U
0
is the nominal a.c. rms line to an earthed neutral voltage in
V, cos is the displacement power factor without harmonic
contents and is the mechanical efciency of the motor.
Similarly for three phase loads and motors,
I
b
=
P

3U cos
and I
b
=
P
m

3U cos
respectively, where U is the line-to-line voltage in V, and

3U
0
= U.
Depending on the actual installation circuit arrangement,
not all rating factors C
a
, C
g
, C
i
or C
c
need to be applied.
For example, if the circuit is not buried and an approved
type of circuit breaker (BS EN 60898) is being used, which
Fig. 1. Current Waveforms of (a) linear load vs (b) non-linear load for a
sinusoidal ac supply at fundamental frequency of 50 or 60 Hz.
is usually the case, C
c
should be omitted from the equation
(1). Further, if the cable is not totally surrounded by thermal
insulations equation (1) can be reduced to
I
Z
= I
t
C
a
C
g
I
n
I
b
The ampacity tables found in BS7671 and other international
standards such as the ET 101 [5] and IEC 60364 [6] generally
assume a balanced three-phase linear load when the current
in the neutral conductor of a three-phase four-wire circuit
will be negligible.
Most commercial and industrial establishments tend to
employ balanced three-phase four-wire distribution systems
with a reduced size neutral conductor. However, with the
proliferation of non-linear loads in this type of electrical de-
sign, the triplen (multiples of 3
rd
order harmonics) harmonic
current from each phase sums in the neutral. P. Cook et. al.
[7] report that neutral currents can be higher than the phase
currents where high harmonic distortion exists. The line and
neutral currents that were taken by two separate but balanced
three-phase four-wire linear loads and non-linear loads are
shown in Figure 1(a) and Figure 1(b), respectively.
In Figure 1(a), with a linear load, all three phases draw per-
fect sinusoidal current waves. Since they are balanced they
cancel out in the neutral, thus the neutral conductor carries
negligible current. Although in practice, even where load
currents are at the standard fundamental power frequency
of 50Hz or 60Hz there is rarely zero current in the neutral.
Any value of current that exists in the neutral simply reects
an out-of-balance in the three-phase load. Assuming that no
phase is overloaded and the cable was sized using the basic
method outlined above. The effect of an out of balance three
phase load causing the neutral conductor to become a heat
emitter should not cause any overheating to the group of
three-phase four-wire cable. However, when harmonics are
present as shown in Figure 1(b), the line current in the three
phases are no longer balanced sine waves and if they are in
triplen order (i.e. 3n), they will be additive in the neutral.
Now the neutral conductor becomes a fourth and additional
Fig. 2. Schematic for an Electric Cable.
current carrying conductor. As a result, it is an additional
heat-emitting source in the group of four conductors.
In view of the fact that the neutral conductor is now a
current carrying conductor, hence a heat emitter, steps need
to be taken to take account of the extra heat that is produced
by the neutral conductor in a three phase circuit. The update
published by the IEE (now the IET) [13] states that for every
8
o
C increase above the maximum core conductor continuous
operating temperature the life of the cable will be halved (e.g.
25 years reduced to 12.5 years). A method is thus required
to size the cable accordingly to dissipate the extra heat that
is being generated within a group of three-phase four-wire
conductors to ensure that the group of four conductors does
not overheat
III. HEAT TRANSFER MECHANISMS IN ELECTRIC
CABLES
A comprehensive review and research on the heat transfer
mechanisms of electrical cables, which focused mainly on
overhead line cables is given in [8]. In general, the heat
balance of any cable can be considered by the law of
conservation of energy and on a rate basis, we have

t
q
in
+

t
q
gen
=

t
q
stored
+

t
q
out
where q/t is the rate of heat energy per unit volume, q
in
is the heat input, q
gen
is heat generated internally, q
stored
is heat stored by the medium and q
out
is heat loss to the
external environment.
In [8] it is shown that the heat transfer mechanisms
associated with an electric cable immersed in air can be
approximated using the following assumptions: (i) using
cylindrical coordinates, it is a one-dimensional radial conduc-
tion, convection and radiation system; (ii) it has uniform vol-
umetric heat generation; (iii) the thermal contact resistance
between the conductor material and electrical insulation ma-
terial is negligible; (iv) the electrical and thermal properties
of the conductor and insulation materials are constant (i.e.
homogeneous); (v)the surroundings are large compared to
the cable; (vi)the analysis is for steady state conditions.
An energy balance rate basis analytic technique can be
applied to an electrical cable (see Figure 2) to evaluate the
surface temperature of the conductor or the cable. For the
control surface (see Figure 3) placed around the inner and
outer surfaces of the insulation material:
E

in
E

out
= 0
since energy in is taken to be equal to energy out,
E

in
= E

out
= q

r
= q

r
2
1
(2)
Fig. 3. Energy exchanged at the inner and outer surfaces of the electrical
insulation of a conductor.
q

r
q

cv
q

rad
= 0 (3)
hence, from equation (2) and (3), at the outer surface
q

r
2
1
h(2r
2
)(T
s,2
T

(2r
2
)(T
4
s,2
T
4
sur
) = 0
This general equation can be used to determine the surface
temperature T
s,2
of a cable in terms of q

, r
1
, r
2
, h, T and
where h is the convection coefcient in W/m
2
K and is
the dimensionless emissivity of the cable surface.
IV. INDUSTRY METHODS AVAILABLE FOR DETERMINING
THE SIZE OF A CABLE UNDER HARMONIC DISTORTION
A. The Neher-McGrath Method
In 1957 Neher and McGrath [9] derived a set of Neher
and McGrath (NM) cable rating equations to predict the
resulting ampacity of a group of four single core cables.
They are a more complex version of the Fourier heat transfer
equations. There are many variables in the 66 equations used
to account for the number of conductors, number and size of
adjacent conduits, number and size of adjacent duct banks,
coefcient of surface emissivity, number of cables, axial
spacing between cables, extraneous heat sources, and wind
velocity. All these factors and more, effect the calculation
of ampacity. Two of the factors affecting the nal ampacity
value of a cable under harmonic conditions are the skin and
proximity effects. The NM equation is given by
I =

T
c
(T
a
T
d
)
r
dc
(1 +Y
c
)

r

ca
(4)
where I is the conductor current in kA, T
c
is the conductor
temperature in degree
o
C, T
a
is the ambient temperature
in degree
o
C, T
d
is the temperature difference due to
dielectric loss in degree
o
C, r
dc
is the direct current resis-
tance of the conductors in at the conductors operating
temperature per unit length,

r

ca
is the effective resistance
between the conductor and ambient for a conductor loss in
at the conductors operating temperature per unit length.
The parameter Y
c
is the increment of ac/dc ratio in p.u. due
to losses originating in the conductor, having components
Y
cs
(the skin effect) and Y
cp
(the proximity effect) where
Y
c
= 1 +Y
cs
+Y
cp
Y
cs
= 0.875
_
fk
s
r
dc
Y
cp
= x
p
_
D
c
S
_
2
_
_
1.18
x
p
+ 0.27
_
+ 0.312
_
D
c
S
_
2
_
and
x
p
=
6.80
_
r
dc
/k
p
at 60Hz
where f is the frequency in Hz, D
c
is the conductors outer
diameter (in inches), S is the axial spacing between cables
(in inches) and k
s
and k
p
are the skin and proximity effect
factor respectively (with recommended values of 60Hz [9]).
The NM method does not cater for the inclusion of a range
of harmonic components in the generalised equation (4). The
NM equation is very similar to the IEC60287 method [15]
as they are both based on the same principle.
B. The Meliopoulos and Martin Method
The Meliopoulos and Martin method [10] provides an
extension of the Neher-McGrath equation using power losses
in the cable under harmonic conditions to derive a derating
factor for cables given in the NEC [11]. For single phase
circuits
=

2
1
I
2
B
r
ac
(1)
P
loss
and for three phase circuits
=

2
1
I
2
B
[r
ac,A
(1) +r
ac,B
(1) +r
ac,C
(1)]
P
loss
where is the desired harmonic derating factor
r
ac,A
(1),r
ac,B
(1) and r
ac,C
(1) are the ac resistance of
phase A, B and C conductors at fundamental frequency,
P
loss
is the total ohmic losses of the cable including
harmonic effects, I
B
is the base RMS value of the design
current and
1
is the p.u. value of the fundamental with
respect to the base I
B
. For harmonics at a frequency of
h f
fundamental
, additional values can be found by the
equations for x
s
(h), x
sp
(h) and x
cp
(h) as given below:
x
s
(h) =
_
ka
2
M
0
(ka)
M
1
(ka)
sin[
1
(ka)
0
(ka) /4]
_
1
(5)
where k =

2fh, a is the conductor radius in me-
tres, f is the fundamental power frequency in Hz, is
the relative permittivity of the conductor, is the con-
ductivity (of the conductor), h is the harmonic order and
M
0
(ka), M
1
(ka),
0
(ka) and
1
(ka) are Bessel functions
obtained from [10].
x
sp
(h) = F(x
p
)
_
D
c
S
_
2

_
_
1.18
F(x
p
) + 0.27
_
+ 0.312

h
_
D
c
S
_
2
_
(6)
where x
p
= k
_
k
p
/(r
dc
) at the h
th
harmonic and k
p
is
the empirical factor at fundamental power frequency from
[9].
F(x
p
) =
_
x
p
2
M
0
(x
p
)
M
1
(x
p
)
sin[
1
(x
p
)
0
(x
p
) /4]
_
1
(7)
The contribution to the increase of conductor ac resistance
due to proximity to a steel pipe or magnetic conduit is given
by the following expressions: For a trefoil arrangement
x
cp
(h) =

n
_
0.89S 0.115D
p
r
dc
_
0.3048 10
6
(8)
and for a at cradled arrangement
x
cp
(h) =

n
_
0.89S 0.175D
p
r
dc
_
0.3048 10
6
(9)
where =1.7 for a steel pipe and 0.8 for an iron conduit, D
p
is the inside diameter of the pipe or conduit in metres and
r
ac
(h) = r
dc
[1 +x
s
(h) +x
sp
(h) +x
cp
(h)] (10)
Two examples were given by Meliopoulos and Martin [10]
to illustrate the full computation procedures to use the above
equations. It will not be repeated here and readers who wish
to consider these equations should refer to the original paper
[10].
However, if the designer nds that the Meliopoulos and
Martin approximation derates the cable leading to an inac-
curacy or when signicant zero sequence harmonic currents
are present in the neutral, then the Neher-McGrath equation
should be used to re-rate the cable.
C. The AH Generalised Ampacity Model
Hiranandani [12] develops a simple general equation
that can be used to evaluate separate harmonic derating
factors for line and neutral conductors. The AH method for
calculating a cables ampacity in the presence of harmonics
for NEC cables can be summarized as follows:
1. Determine the harmonic signature of the line and
neutral conductors by either calculation or measurement.
The Harmonic Signature (HS) is then determined by the
equation
HS =
_
I
1
,
y
=
I
y
I
1
,
y+1
=
I
y+1
I
1
,
y+2
=
I
y+2
I
1
, ...
_
(11)
where y = 2. For example, a three-phase distribution circuit
with a THD=41.9% has a phase current I
rms
= 99.12A;
I
50Hz
= I
1
= 90A; I
150Hz
= I
3
=35A; I
250Hz
= I
5
=20A;
I
350Hz
= I
7
=10A and a neutral current I
150Hz
= I
3
= 335
= 105A. Hence from equation (11), the harmonic signatures
are: Line HS=(90,
3
=0.39,
5
=0.222,
7
=0.111) and
Neutral HS= (90,
3
=1.17).
2. Determine the total ac resistance (r
ac
) of the line
and neutral conductors including skin effect and proximity
effect using equation (10).
where h is the order of the harmonic, r
dc
is the dc
resistance of the conductor in at the conductors operating
temperature per unit length, x
s
is the contribution factor to
ac resistance due to skin effect, x
sp
is the contribution factor
to ac resistance due to proximity effects of neighbouring
conductors, x
cp
is the contribution factor to ac resistance
due to proximity effect of a metallic pipe or conduit,
r
ac
(h) is the ac resistance, and x
s
(h), x
sp
(h) and x
cp
(h)
are skin effect and proximity effect factors calculated for
each harmonic order h as given in [10] from equations (5)-(9)
TABLE I
RATING FACTORS FOR TRIPLE HARMONIC CURRENTS IN FOUR- AND
FIVE-CORE CABLES (SOURCE: BS7671).

THE 3RD HARMONIC
CONTENT IS EXPRESSED IN TERMS OF THE TOTAL HARMONIC
DISTORTION.
3rd harmonic
content of line Rating Factor Rating Factor
current

(%)
Size selection based Size selection based
on the line current on the neutral current
0-15 1.0 -
15-33 0.86 -
33-45 - 0.86
>45 - 1.0
3. The derating factor for line and neutral conductors
in the presence of harmonics can be evaluated using
equation (12):
HDF =
_
1 +
n

h=2

2
h

h
_

1
2
(12)
where
h
is the harmonic distribution factor per unit har-
monic content due to each harmonic with respect to base load
current (i.e.
h
= I
h
/I
1
) and
h
is the normalized harmonic
ac resistance factor, i.e. the ratio of conductor resistance at
n
th
harmonic frequency to resistance at fundamental power
frequency -
h
= r
ac
(h)/r
ac
(1).
Similar to the approach proposed by Meliopoulos and Martin,
it is necessary to assume a certain cable size and derate it by
the factor HDF accordingly. Hiranandani presents a worked
example similar to the one given by Meliopoulos and Martin
but giving a different set of results for the derived derating
factor.
D. The BS7671 Appendix 11 Method
In 2004 the International Electrotechnical Commission
(IEC) published a set of harmonic rating factors so that
allowance can be made for 3
rd
harmonic currents in 4 and 5
core cables, where all the cores have the same conductor size.
It is now included in BS7671 Appendix 11 as informative
guidance - see Table I.
Cook and Coate [7] explained the guidance given by the
IEC such that if the third harmonic content of the current in
each phase is between 33% and 45%, i.e. the neutral current
is greater than the fundamental phase current, then selection
of the conductor size should be based on the current in the
neutral conductor divided by the given factor. If the harmonic
content is greater than 45% then the size of the line conductor
chosen is based on the neutral conductor current. In this case
the line conductors will be larger than that required to carry
the line current and this spare capacity allows the factor of
0.86 to be omitted.
Three examples are given in BS7671 Appendix 11 illus-
trating how to apply the rating factor in practice. It will not
be repeated here and readers who wish to consider the rating
factor should refer BS7671.
V. EXPERIMENTAL DETERMINATION
An experiment was set up to inject discrete harmonic
currents into a 185 sq.mm solid core aluminum cable. A
Fig. 4. Experimental setup for measuring harmonic heating effects in
cables.
schematic diagram of the equipment used is as shown in
Figure 4. A signal generator was connected to the input
of a 400W power amplier, which, in turn, was connected
to a current transformer as shown. The cable under test
was connected to the primary of a current transformer. This
arrangement allowed the full rated current of 400A to be
injected into the cable at discrete frequencies, which could be
set by the signal generator and measured by the grip amme-
ter. Thermo couples embedded in the cable at points B and C
(see Figure 4 accurately measure the conductor temperature.
Thermo couple A measures the ambient air temperature. The
cable was shaped as shown so that thermo couple B would
indicate the temperature of a single conductor suspended in
free air. Thermo couple C would indicate the temperature of
conductors in close proximity running parallel to each other
also suspended in free air. The thermo couples B and C can
thus measure the additional heating effect due to the skin
and proximity effects respectively as the cable is run at full
load at frequencies of n, (50Hz), 3n, 5n, 7n ... up to 20n.
Initially, the cable was fully loaded at 50 Hz and the
result compared with the value of 70
o
C quoted in BS 7671.
Adjustments to the readings were made to compensate for
the actual ambient temperature in the laboratory at the time
of the test. This test validated the accuracy of the measuring
system used as shown in the graph in Figure 5. Further tests
were carried out by injecting the full rated current value (400
A) at various harmonic frequencies. and the results are shown
in Figure 6. It can be seen that the conductor temperature
at 50Hz is approximately 70
o
C, which agrees with BS7671.
However as the frequency is increased whilst maintaining
the current at 400A full rated value, it can be seen that
the temperature of the conductor increases signicantly. The
temperature of the parallel conductors has increased by a
larger amount reecting the combined skin and proximity
effects.
Load currents that have signicant harmonic distortion
such as those supplying personal computers (see Figure 7)
will therefore experience additional heating due to both the
skin and proximity effects. Arising from this, cables will
operate at a higher temperature than would be the case
Fig. 5. Temperature rise of cable under harmonic inuences.
Fig. 6. Temperature Vs Harmonics for various cable congurations.
Fig. 7. Spectral analysis of current.
without harmonic distortion. If one can predict the degree
of harmonic distortion in the load current then it is possible
to determine the degree of additional heating that will occur
and apply a suitable de-rating correction factor.
The experimental data collected and analysed in the above
experiment proves that there are indeed signicant heating
effects in a conductor carrying harmonic currents as predicted
by Meliopoulos and Martin [10] and in the AH Generalised
Ampacity Model by Hiranandani [12]. Those effects must be
taken into consideration and it is likely that the oversizing
of conductors by the BS7671 Appendix 11 method can
adequately cover the excess heating caused by the harmonic
content of the currents.
VI. DISCUSSION
There are a number of ways to reduce and combat the
detrimental effects of high levels of harmonic distortion in
an electrical installation, e.g. by the application of lters
etc. Active lters are devices which actively inject opposite
harmonics into a system to cancel out the harmonics created
by the non-linear loads. Passive lters trap or resist the ow
of harmonics through them. They do this through various
capacitors or reactors. Harmonics rated transformers known
as K factor transformers are specically designed in order to
cope with the excess heating problem caused by the presence
of high level circulating harmonic currents. The thermal and
neutral connections are sometimes being sized at around
200% of the size required in order to accommodate the
harmonic loads [3]. Most of the remedial systems that are
put into place do work, but are usually quite costly especially
if they are not initially included at the design stage. Also,
lters and transformers may require maintenance or could
suffer failures if not designed and installed properly. The dire
consequence of which will render the system unprotected
and the harmonics present may cause damage in this time
period, especially if there is a failure in the equipment as
there may be a certain length of time until the fault is found
and rectied.
Reducing the temperature of the conductors is one of
the most important cabling design aspects in an electrical
installation. It is has been clearly demonstrated in this paper,
that in order to maintain the operating temperature of the
cable within the specied maximum tolerable temperature,
an increase in the cross-sectional area of the conductor is
required. With a larger cross-sectional area, even if the lter
or the transformer is faulty, the cables will be sized to cope
with the extra currents, reducing the damage that can be
caused. Another fact to consider is that these calculations
have been carried out on the assumption that the neutral
conductor is carrying 100% third harmonic load. However,
at certain times of the day, if the equipment that causes
harmonic distortion is not operating, the harmonic load will
be reduced and as a result of this, the voltage drop will reduce
further, making the circuit much more energy efcient.
There are several advantages to increasing the cable size
as a result of harmonic derating, in most cases, only up to
the next size. They include:
harmonic loads are accounted for and even if preventa-
tive measures (e.g. lters, transformers) fail, the cables
are adequately sized to carry the load;
temperature rise of the cable is reduced, reducing losses,
maintenance and running costs (increasing the life ex-
pectancy of the cable);
larger cross-sectional areas can reduce the voltage drop
along the circuit, proving more efcient by delivering
close to the declared voltage to the current using equip-
ment.
VII. CONCLUSION
Harmonic distortion in electrical installations of tomorrow
is likely to get worse as the rise in use of low energy
electrical equipment in the built environment increases. Steps
need to be taken by electrical designers and installers to
minimise its detrimental inuence on the interconnecting
cables, busbars, energy sources and neighbouring equipment.
With the newly published BS7671 (17
th
edition IEE Wiring
Regulations) in 2008, at long last designers and installers
now have a set of harmonic rating factors which can be
used in the initial design calculations to determine reasonably
accurately the size of cables to allow for conditions when
harmonic distortion is present in a system. Alternatively,
they can use the other calculation methods given in this
paper to calculate the heating effect of harmonic currents
from rst principles. Often the BS7671 method may result
in an oversized cable, however, it was discussed earlier that
this process is only benecial as it can reduce the operating
temperature of the cable and as a result the voltage drop
in the cable is also reduced, thereby decreasing losses and
increasing transmission efciency.
ACKNOWLEDGMENT
The authors would like to thank Dublin Institute of
Technology, the Science Foundation Ireland and the Dublin
Energy Laboratory for their continuing support.
REFERENCES
[1] E. F. Fuchs, D. J. Roesler, and M. A. S. Masoum, Are Harmonic
Recommendations According to IEEE and IEC Too Restrictive?,
IEEE Transactions on Power Delivery, Issue: 99 , 2004.
[2] IEEE, IEEE Recommended Practices and Requirements for Harmonic
Control, in Electric Power Systems, IEEE-519, 1992.
[3] IEC, Consolidated Edition. Electromagnetic Compatibility (EMC)-Part
3-2 Limits for Harmonic Current Emissions, IEC 61000-3-2 (2001-10).
[4] IEE/BSI, BS 7671:2008, Requirements For Electrical Installations,
IEE Wiring Regulations 17
th
Edition, 2008.
[5] ETCI, ET101: 2004 - National Rules for Electrical Installations,
Fourth Edition, 2004, Electro-Technical Council of Ireland.
[6] IEC, IEC60364: 2002-06 - Electrical installations of buildings, 2
nd
Edition, Geneva, Switzerland.
[7] P. Cook and M. Coates, Double Neutrals, The Answer To Triple
Harmonics?, International Seminar, Santiago, CHILE 26-28 April
2000.
[8] V. T. Morgan, Thermal behaviour of electrical conductors, Wiley,
(ISBN: 0863801196), 1991.
[9] J. H. Neher and M. H. McGrath, The Calculation of the Temperature
Rise and Load Capability of Cable Systems Vol. 76, pp. 752-772,
AIEE, March 1957.
[10] A. P. S. Meliopoulos, M. A. Martin, Calculation of secondary cable
losses and ampacity in the presence of harmonics, IEEE Transactions
on Power Delivery, Vol. 7, No. 2, April, 1992.
[11] NFPA, National Electrical Code 2008, NFPA, 2008.
[12] A. Hiranandam, Calculation of Conductor Temperatures and Am-
pacities of Cable Systems using a Generalised FDM Method, IEEE
Transactions on Power delivery, Vol. 6, No. 1, Jan. 1991.
[13] P. Cook, (Update published by the IEE - now the IET), Commentary on
IEE Wiring Regulations 16
th
edition BS7671:2001 Peregrinus, ISBN:
0 85296 237 1, April 2004.
[14] F. P. Incropera and D. P. Dewitt, Introduction to Heat Transfer, Wiley,
(4
th
Edition. ISBN: 0-471-38649-9
[15] IEC 60287: Electric cables - Calculation of the current rating, IEC
Central Ofce, GENEVA 20, Switzerland.