See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/3354714

Current distribution in single-core cables connected in parallel

Article  in  IET Proceedings - Generation Transmission and Distribution · October 2001

DOI: 10.1049/ip-gtd:20010430 · Source: IEEE Xplore

CITATIONS READS

24 5,943

2 authors:

Ya Ping Du Jocelyn Burnett

The Hong Kong Polytechnic University City University of New York - Hunter College
203 PUBLICATIONS   1,589 CITATIONS    72 PUBLICATIONS   1,810 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

flux modulated motor View project

Wireless Power Transfer View project

All content following this page was uploaded by Ya Ping Du on 25 March 2020.

The user has requested enhancement of the downloaded file.

 Current Distribution in Single-Core Cables Connected in Parallel

Y. Du and J. Burnett
Department of Building Services Engineering
The Hong Kong Polytechnic University

1. Introduction

In Hong Kong, single-core cables are often used for connecting transformers to main low-voltage
(LV) switchboards. With transformers being normally rated at 500 kVA per phase, installation
practice includes the use of multiple single-core cables connected in parallel (e.g., four 630mm2
XLPE copper cables per phase). As power consumption in buildings grows, overheating problems
associated with the parallel cable installations have been experienced. The problems arise from
unequal current distribution among the parallel cables in any one phase, causing unequal and
excessive temperatures. It has been estimated that an increase of 8oC to 10oC in insulation
temperature can lead to a 50% reduction of insulation life expectancy [1].

The current distribution at power frequency in a balanced three-phase circuit has been addressed in
recent decades. A number of theoretical studies have been carried out [2-4]. In [2] an iterative
procedure was derived to determine the current distribution in a set of parallel conductors, and was
demonstrated for several cable configurations. A matrix algebra procedure was proposed in [3,4],
which can overcome some difficulties encountered in the iterative procedure. A simple test was
conducted for validation, and reasonable accuracy was observed. These studies provided useful
results for estimation of temperature rise, and subsequently cable current carrying capacity. The
methods presented in [2-4], however, are difficult to apply at harmonic frequencies, as both supply
voltage and load impedance at harmonic frequencies are unknown. The current distributions in a
three-phase four-wire system or three-phase six-wire system are not discussed. The impact of
metallic tray on current distribution is not adequately addressed either.

This paper presents both theoretical and experimental investigations into current distribution
associated with parallel cable installations. Theoretical calculation of cable resistance and reactance
are discussed first. A generic current-distribution prediction method is proposed for multi-phase
distribution systems, with or without a neutral. This method is validated experimentally. The method
is applicable under non-zero neutral current conditions, as well as at power frequency and at
harmonic frequencies. The impact of cable, source and installation parameters, as well as steel tray is
disclosed. Finally, the current distribution associated with the Hong Kong practice of parallel cable
installations is evaluated, and recommendations for the design and installation are presented.

2. Cable Impedance

2.1 Cable Resistance

Various types of large size single-core cables are used in Hong Kong. PVC or XLPE insulated cables
with stranded copper conductors to BS/IEC standards are widely used for LV distribution in
buildings. Normally single-core cables do not have a magnetic armouring, but they may be installed
on galvanised iron (GI) tray.

Cable resistance is one of the parameters affecting current distribution in parallel connected single-
core cables. It is determined by conductor size, and increases with the frequency of the current
carried. The increase of cable resistance is due to both skin and proximity effects. Given that skin

1
and proximity effects can be considered separately, a general formula for cable resistance at 50 Hz is
recommended in IEC 287 [5]. This formula is extended to calculate cable resistance Rh at harmonic
order h, as follows:
Rh  RDC 1  y s ,h  y p ,h  (1)
where y s ,h and y p ,h are factors of skin and proximity effects, respectively. RDC is the cable resistance
for direct current. The value of maximum d.c. resistance for individual cable sizes is specified in IEC
228 [6].

Formulae of y s ,h and y p ,h at 50/60Hz are given in [5], which were derived originally for solid
circular conductors. To calculate harmonic resistance of stranded-conductor cables, more general
formulas are adopted [7,8]. These formulas are shown in Appendix.

2.2 Cable Inductance

Cable inductance is a major parameter affecting current distribution among parallel connected single-
core cables. For simplicity of discussion, each cable in a parallel cable installation is treated as a
current loop together with a neutral cable (reference cable). This reference cable is shared by the
cables in a circuit as the returning current path of their corresponding loops. With this assumption,
self-inductance and mutual inductance of a loop can be calculated via simple formulae.

Usually cable length is much larger than cable spacing. The self-inductance Lii of cable loop i
consisting of cables li and ni illustrated in Figure 1 can, then, be expressed by:
 2 Dli ni 
Lii  2 K  0.4 ln 
 ( H/m) (2)
 dc 
where Dli ni is the conductor axis spacing between these two cables, and dc is the external diameter of
the cable conductor. Constant K is internal inductance per conductor. It is equal to 0.0514
approximately for large single-core cables [9].

The formula of mutual inductance between two current loops can be obtained by considering a
generic case, as illustrated in Figure 1. In the figure, these two loops are labelled by i and j,
respectively. The mutual inductance Lij is given by:
 Dli n j  Dnil j 
Lij  0.2 ln
 Dl l  Dn n 
 H/m  (3)
 i j i j 
where Dxy is the distance between cables x and y.

If two loops share the same reference cable labelled by n for their returning currents, the internal
inductance of the reference cable should be taken into account. The formula, then, is modified, as
follows:
 Dli n  Dl j n 
Lij  K  0.2 ln 
 Dl l  d c / 2 
H/m  (4)
 ij 

2.3 Loop Impedance

2
A three-phase LV distribution system may be classified into one of two groups according to the
configuration of its neutral conductors. It is a three-phase four-wire system if all phase conductors
share the same neutral conductors. It is a three-phase six-wire system if separated neutral conductors
are arranged for the phase conductors of each phase. A single-phase system is considered as a special
case of the three-phase six-wire system where there is no mutual coupling among three phases. In a
three-phase three-wire system, one phase C conductor may be viewed as a virtual neutral, which is
shared by other conductors. The system is a special case of the three-phase four-wire system.

Figures 2 and 3 illustrate parallel cable installations associated with both systems, respectively. In the
analysis of current distribution, these cables are modelled by multiple current loops. For a cable
system shown in Figure 2, one neutral cable is selected to form individual current loops with other
cables. While in Figure 3, three neutral cables are selected to form current loops in phases A, B and
C, respectively. The self impedance Z ii ,h of current loop i made of cables li and n (reference cable)
is expressed by:
  2 Dli n  
Z ii ,h  Rli ,h  Rn ,h  j 2f 1 h 2 K  0.4 ln    10 6
 (5)

  dc 
and the mutual impedance Z ij ,h between loops i and j is given by:
  Dl n  Dl j n  
Z ij ,h  Rn ,h  j 2f 1 h K  0.2 ln i    10 6 (6)
  Dl l  d c / 2  
  ij 
in the case of the common reference cable, or:
 Dli n j  Dl j ni 
Z ij ,h  j 2f1 h  0.2 ln   10 6 (7)
 Dl l  Dn n 
 ij i j 

in the case of separate reference cables, where ni and n j are indexes for reference cables of loops i
and j, respectively. Both f1 and h in EQs (5-7) denote fundamental frequency and harmonic order,
respectively, and R x ,h is the resistance of cable x at order h. For simplicity, subscript h is suppressed
in the following discussion.

3. Current Distribution Analysis

As impedance of distribution cables in a LV system is much less than the load impedance, the system
loads determine phase current at power frequency. Similarly, phase current at harmonic frequency is
also determined by the loads. The system, then, can be modelled as a set of cable loops excited by
independent sources of current, as shown in Figures 2 and 3. The supply voltage sources in these
figures are suppressed, as they are connected in serial with their corresponding current sources. This
model is applicable at both power frequency and at harmonic frequency, as long as total phase
currents are given or current sequence in the system is known.

Based on this circuit model, the following equations are derived by applying Kerchief’s laws:
E a  Z a1a1 I a1   Z a1b1 I b1   Z a1c1 I c1   Z a1n1 I n1   (8)
for the 1st loop in phase A, or:
0  Z n1a1 I a1   Z n1b1 I b1   Z n1c1 I c1   Z n1n1 I n1   (9)
for the 1st loop in the neutral. By repeating the procedure for all other loops, the matrix equation is
obtained, as follows:

3
  Z AA Z AB Z AC Z AN   I A   E A 
Z Z BB Z BC Z BN   I B   E B 
ZI  E or  BA   (10)
 Z CA Z CB Z CC Z CN   I C   E C 
     
Z NA Z NB Z NC Z NN  I N  E N 

EQ (12) can be rewritten, as follows:

 I A   YAA YAB YAC YAN   E A 
I  Y YBB YBC YBN   E B 
I  YE or  B    BA  (11)
 I C   YCA YCB YCC YCN   E C 
     
I N  YNA YNB YNC YNN  E N 
where Y  Z 1 , and Z ij or Yij is mi  m j sub-matrix between phase i and phase j, and mi and m j are
loop numbers for these two phases, respectively. In EQs (10-11), E A , E B , E C and E N are phase
voltage vectors, which contain voltages of cable loops in their corresponding phases. It should be
noted that voltages are identical for all cable loops in one phase, and equal to zero for all cable loops
in the neutral. I A , I B , I C and I N are defined in a similar way.

In the analysis the total phase currents are given, or the current sequence is known. Total phase
currents I a , s , I b , s and I c , s can be expressed in relation to loop currents I ai , I bi and I ci in their
corresponding phases by the following
i a1 I ai  I a,s
m


mb
I  I b,s
i 1 bi
(12)


mc
I  I c,s
i 1 ci

If the positive sequence current, for example, is applied, I a , s , I b , s and I c , s are equal to
1, e  j120 and e j120 , respectively
o o

In order to solve EQ (11) for cable loop currents, E needs to be determined first. Substitution of EQ
(11) into (12) yields the following equation:
 I a , s  Yaa , s Yab , s Yac , s   E a 
   
I s  Ys E s or  I b , s   Yba , s Ybb , s Ybc , s    Eb  (13)
 I c , s  Yca ,s Ycb, s Ycc , s   E c 
where Yij ,s ( i, j  a, b and c ) is the summation of all elements in its corresponding sub-matrix Yij
identified in EQ (11) (e.g., Yaa , s  i a1  j a1 YAA i, j  ), and E a , Eb , and E c are voltages of cable
m m

loops in their corresponding phases. Es can be solved directly from EQ (13). Then, I is solved by
substituting Es in EQ (11).

4. Laboratory Measurements

4.1 Measurement Setup

Laboratory measurements were performed to investigate current distribution phenomena as well as to

validate the proposed method for calculating current distribution among parallel connected cables. In

4
the experiments, 400mm2 XLPE cables having stranded copper conductors were selected to form a
single-phase testing circuit. The cables were 7.5 meters long, and laid straight on wooden plates, or
on 300mm-wide (GI) tray, at the height of 1.0 m above the ground. These cables were terminated by
a 6 mm x 76 mm copper bar at one end, and connected to a current injection set at the other, as
shown in Figure 4. The magnitude of the injected current was adjusted using an autotransformer,
while the harmonic content of the injected current was controlled using a harmonic generator.

Both voltage and current in each cable were measured using a Dranetz PP1 Power analyser. Voltage
signals were recorded on Channel D of the analyser, and cable currents were recorded via 1000A
current transformers (CTs). The sensitivity of Channel D was claimed to be less than 12mV, and that
of the CT less than 0.1A. The recorded data were processed to calculate voltage and current as well
as power at harmonic orders.

4.2 Cable Resistance

Cable resistance in a single-phase circuit was investigated first. The circuit consisted of two single-
core cables, which were installed either on wooden plates or on GI tray. The measurement data were
collected when cable spacing s was set to 2d or 4d (cable diameter d = 34 mm). Given by measured
power Pk and current Ih at harmonic order h cable resistance Rh was obtained from:
P
Rh  h2 (14)
Ih
Figures 5 and 6 show the variation of cable resistance ratios (to the base of DC resistance), against
harmonic order. It is noted that both measured and calculated results match well if the cables are
installed on a non-magnetic surface. The average error for resistance ratios is generally less than 5%.
It is noted also that the cable spacing does not have significant influence on cable resistance if it is
large (e.g., s=4d). However, its influence cannot be ignored if cables are installed on the tray, as the
eddy-current effect of the tray is influential. It is observed that tray losses become dominant at
harmonic frequencies when s=4d. It would be wise to avoid using large cable spacing in cable tray
installations when the harmonic current carried by the cables is substantial.

4.3 Current Distribution

Four cables (two for the line and two for the neutral) were employed to form a single-phase circuit
for testing. To ensure that all cable currents flow through the simulated load (the termination copper
bar), parallel cables were bonded together at both ends before the connection to the load and to the
supply. In the measurements, the cables were arranged according to Figure 7, and cable currents
were measured at the frequency up to the 5th harmonic order. Tables 1 and 2 show the measured
results together with calculated results for the ratio of maxI l1, I l 2  minI l1, I l 2 . It is noted that both
measured data and theoretical data agree well, even at harmonic frequencies. The relative error is
generally less than 3%.

Both tables also show that the current distribution is worse or less even among the parallel cables if
cable spacing is smaller. It is worse still at harmonic frequencies but the ratio saturates quickly with
increasing frequency. This is because cable resistance at harmonic frequencies is comparably small.
In order words, cable resistance plays a positive role in determining the evenness of the current
distribution. It is noted also that the configuration shown in Figure 7(a) exhibits the least evenness in
distribution if cable spacing is fixed. The current is evenly distributed among the cables if they are
placed symmetrically, as shown in Figure 7(c).

5
Similar tests were performed when these cables were placed on GI tray. The results are shown
Tables 3 and 4. Because of the GI tray, the current distribution is less even in general. The difference
caused by the tray, however, is not significant, as the relative error is less than 10%, except in the
configuration shown in Figure 7(a) with zero cable spacing. The minor variation means that this
calculation method is applicable also for parallel connected cables on GI tray. The worst scenario, as
indicated in Tables 3 and 4, is the configuration shown in Figure 7(a) with zero cable spacing.

5. Current Distribution Evaluation

Cables for transformer secondary connections can be arranged in different ways. Both trefoil and flat
cable arrangements are the common practices in Hong Kong. There are many variations of the
arrangements. The typical scenarios of these arrangements are illustrated in Figures 8 and 9 for 3-
phase 4-wire systems, and in Figure 10 for 3-phase 6-wire systems. The selection of cable
installation methods for a building depends up on site conditions, electrical sub-contractor's
experience as well as preference. Normally, 640mm2 XLPE cables with stranded copper conductors
are selected for the cable installations (parameters regarding cable material and geometry can be
obtained in [10]). These cables are grouped, and either installed in concrete trench or on galvanised
iron tray.

Current distribution among parallel connected cables in these installations was evaluated using the
method discussed previously. For each cable installation, cable currents were computed at 50Hz
(positive sequence), 150Hz (zero sequence) and 250Hz (negative sequence). The results are shown in
Tables 5-12, where the values are given in terms of a current ratio (on the base of the average cable
current expected). A ratio with the value more than one indicates the cable carries the current more
than expected. It should be noted that in ideal conditions one neutral cable carries three times as
much as the phase cable current at 150Hz (zero sequence) in a 3-phase 4-wire system.

It is noted form Tables 5 and 6 that a trefoil cable arrangement yields more evenness in current
distribution than a flat cable arrangement. To minimise the unevenness in current distribution, it is
always desirable to group all three phase and neutral cables together, as shown in Figure 8(b) or (c),
and in Figure 9(b) or (c). The arrangements shown in Figure 8(c) and Figure 9(c) lead to the most
evenness in distribution at 150Hz (zero sequence), as they have a symmetric structure. Cable spacing
is another crucial parameter affecting current distribution. The currents were also computed when
cable spacing was equal to one cable diameter ( s  45 mm), as shown in Tables 7 and 8. It is found
that the current distribution tends to be even if cable spacing is increased.

In a 3-phase 6-wire system, the arrangement shown in Figure 10(a) is the most common in Hong
Kong, as cables is easy to install. This arrangement yields extreme unevenness in current
distribution. The current ratio is up to 3.31 at 150Hz (zero sequence current), as indicated in Table 9.
The currents in the arrangement shown in Figure 10(c), however, are almost evenly distributed at
150Hz due to its symmetric structure. Increasing cable spacing leads to less mutual coupling among
the three phases. The current distribution tends to be dominated by the cable arrangement of
individual phases only. The current distribution becomes more even, as shown in Tables 10.

7. Conclusions

This paper discussed current distribution associated with parallel single-core cable installations in
large buildings. A generic analytical method for solving current distribution among parallel
connected cables was derived, and validated experimentally. This method is applicable in a multi-
phase distribution system, as long as total phase currents are given or current sequence is known.

6
Current distribution in practical cable installations was evaluated also. It is concluded that the current
distribution is affected by cable characteristics and installation methods. The current distribution
tends to be even if these cables are installed symmetrically. Cable resistance plays a positive role in
evenness of the current distribution. The distribution becomes less even if the resistance is smaller,
or when the reactance is comparably larger, such as, at a harmonic frequency.

Single-core cables may be installed on GI tray. The GI tray increases both cable resistance and
reactance. It is noted that, in general, the current distribution remains the same as that associated with
the non-tray installation. In general, the calculation method is applicable to estimate current
distribution when GI tray is present. Although the GI tray does not cause a large change of current
distribution, the temperature increase could be very serious due to the significant increase of cable
resistance and consequent power loss.

7. Acknowledgements

The work leading to this paper was supported by grants from the Research Committee of the Hong
Kong Polytechnic University, and the Research Grants Council of the Hong Kong Special
Administrative Region.

Appendix

Assuming x s2,h  8f 1 h10 7 RDC and x 2p ,h  8f1 h10 7 RDC , the factor of skin effect y s ,h under the
symmetrical conditions is expressed as follows:
 xs4,h
 0.0  xs ,h  2.8
 192  0 . 8 x 4
s , h
ys ,h  0.0563 xs2,h  0.0177 xs ,h  0.136 2.8  xs ,h  3.8 (15)
 xs ,h 11
  3.8  xs ,h
 2 2 15
and factor of proximity effect y p ,h is given by
 y 2 G p ,h
 two 1/c cables
1  y A p ,h  y B p ,h
2 4
y p ,h   (16)
 3 y 2 G p ,h
three 1/c cables
 2  5 y 2 H p ,h 12

where y is the spacing ratio defined by y  d c s . s is the spacing between conductor axes, and dc is
the diameter of an equivalent circular conductor having the same cross section.
Coefficients A p ,h , B p ,h , G p ,h and H p ,h in EQ (15) are given by:
 0.042  0.012 x 4p ,h
 0.0  x p ,h  2.8
 1  0.0236 x p ,h
4


A p ,h   0.223  0.237 x p ,h  0.0154 x 2p ,h 2.8  x p ,h  3.8 (17)
0.75  1.128 x 1 3.8  x p ,h
 p ,h



7
 0 0.0  x p ,h  2.8

B p ,h  0 2.8  x p ,h  3.8 (18)
0.094  0.376 x 1 3.8  x p ,h
 p ,h

 11x 4p ,h
 0.0  x p ,h  2.8
 704  20 x p ,h
4


G p ,h   0.08 x 2p ,h  0.72 x p ,h  1.04 2.8  x p ,h  3.8 (19)
 x p ,h 1
  3.8  x p ,h
 4 2 8
 1 1  0.0283x 4p ,h
  0.0  x p ,h  2.8
 3 1  0 .0042 x 4


p ,h

H p ,h  0.0384 x 2p ,h  0.119 x p ,h  0.095 2.8  x p ,h  3.8 (20)

 2 x p ,h  4.69
 3.8  x p ,h
 x p ,h  1.16

Table 1 Current Distribution in Spaced Cable Installations in Air for a Single-Phase Circuit
( s1  d and s 2  2d )
Config.

h Mea. Cal. Mea. Cal. Mea. Cal.

1 1.49 1.48 1.25 1.25 1.00 1.00
3 1.58 1.57 1.27 1.26 1.00 1.00
5 1.60 1.58 1.27 1.26 1.00 1.00

Table 2 Current Distribution in Touching Cable Installations in Air for a Single-Phase Circuit
( s1  0 and s 2  0 )
Config.

h Mea. Cal. Mea. Cal. Mea. Cal.

1 1.85 1.88 1.46 1.49 1.00 1.00
3 2.41 2.53 1.52 1.56 1.00 1.00
5 2.55 2.68 1.54 1.56 1.00 1.00

Table 3 Current Distribution in Spaced Cable Installations on Tray for a Single-Phase Circuit
( s1  d and s 2  2d )

Config.

h Mea. Cal. Mea. Cal. Mea. Cal.

1 1.54 1.48 1.23 1.25 1.00 1.00
3 1.55 1.57 1.22 1.26 1.00 1.00
5 1.51 1.58 1.21 1.26 1.00 1.00

8
 Table 4 Current Distribution in Touching Cable Installations on Tray for a Single-Phase
Circuit ( s1  0 and s 2  0 )
Config.

h Mea. Cal. Mea. Cal. Mea. Cal.

1 2.39 1.88 1.53 1.49 1.00 1.00
3 3.08 2.53 1.57 1.55 1.00 1.00
5 3.15 2.68 1.57 1.56 1.00 1.00

Table 5 Current Distribution in Trefoil Cable Installations for a 3-Phase 3-Wire Circuit (s=0mm)
50Hz (pos.) 150Hz (zero) 250Hz (neg.)
A B C A B C N A B C
(a) AAABBBCCCNNN
I1 0.79 1.73 1.50 1.10 0.37 1.81 4.51 0.52 1.39 1.59
I2 0.94 0.88 0.89 1.01 0.98 0.78 2.68 0.86 0.83 0.85
I3 1.46 1.07 0.61 0.89 2.16 3.96 1.88 1.62 1.68 0.66
(b) ABCNABCNABCN
I1 0.90 1.01 1.04 0.23 0.83 0.98 3.38 0.92 1.01 1.00
I2 1.05 0.98 1.01 1.41 1.01 0.98 3.34 1.04 0.98 1.00
I3 1.05 1.02 0.96 1.43 1.16 1.05 2.28 1.05 1.01 1.01
(c) ABCNNCBAABCN
I1 1.09 1.05 1.00 1.20 1.04 0.96 2.82 1.12 1.04 1.01
I2 0.95 0.97 1.00 0.90 0.96 1.00 2.81 0.94 0.97 0.99
I3 0.95 0.98 1.01 0.90 1.01 1.04 3.38 0.95 0.99 0.99

Table 6 Current Distribution in Flat Cable Installations for a 3-Phase 3-Wire Circuit (s=0mm)

50Hz (pos.) 150Hz (zero) 250Hz (neg.)

A B C A B C N A B C
(a) AAABBBCCCNNN
I1 0.79 1.82 1.72 1.25 0.23 1.68 5.28 0.73 1.66 1.84
I2 0.70 0.62 0.60 0.66 0.63 0.38 1.52 0.43 0.41 0.47
I3 1.74 1.42 0.74 1.12 2.42 4.34 2.34 1.87 1.81 0.80
(b) ABCNABCNABCN
I1 0.82 0.98 1.04 0.17 0.93 0.92 3.40 0.86 1.01 1.01
I2 1.09 1.01 1.02 1.44 1.00 0.98 3.35 1.07 0.99 0.99
I3 1.10 1.02 0.94 1.49 1.07 1.10 2.25 1.09 1.00 0.99
(c) ABCNNCBAABCN
I1 1.13 1.03 1.00 1.25 0.98 0.98 2.80 1.12 1.00 1.01
I2 0.94 0.99 1.00 0.87 1.00 0.98 2.79 0.94 1.00 1.00
I3 0.94 0.99 1.01 0.88 1.02 1.04 3.41 0.94 1.00 1.00

Table 7 Current Distribution in Trefoil Cable Installations for a 3-Phase 3-Wire Circuit (s=45mm)
50Hz (Pos.) 150Hz (zero) 250Hz (neg.)
A B C A B C N A B C
(a) AAABBBCCCNNN
I1 0.86 1.41 1.32 1.13 0.41 0.80 3.80 0.69 1.12 1.31
I2 0.98 0.96 0.96 1.03 1.02 0.92 2.91 0.96 0.94 0.96
I3 1.23 0.92 0.72 0.85 1.62 2.79 2.30 1.36 1.34 0.77
(b) ABCNABCNABCN
I1 0.95 1.01 1.02 0.60 0.84 0.95 3.23 0.95 1.01 1.01
I2 1.03 0.99 1.00 1.18 1.02 0.99 3.17 1.02 0.99 1.00
I3 1.03 1.01 0.98 1.23 1.14 1.06 2.61 1.03 1.01 1.00
(c) ABCNNCBAABCN
I1 1.04 1.02 1.00 1.12 1.04 1.00 2.91 1.05 1.02 1.01
I2 0.98 0.99 1.00 0.95 0.98 1.01 2.90 0.98 0.99 1.00
I3 0.98 0.99 1.01 0.94 0.98 1.00 3.19 0.98 1.00 1.00

9
Table 8 Current Distribution in Flat Cable Installations for a 3-Phase 3-Wire Circuit (s=45mm)
50Hz (pos.) 150Hz (zero) 250Hz (neg.)
A B C A B C N A B C
(a) AAABBBCCCNNN
I1 0.92 1.52 1.52 1.31 0.49 0.68 4.54 0.89 1.40 1.57
I2 0.71 0.67 0.66 0.65 0.65 0.50 1.72 0.53 0.52 0.55
I3 1.49 1.23 0.86 1.05 1.89 3.19 2.81 1.60 1.50 0.93
(b) ABCNABCNABCN
I1 0.89 0.99 1.03 0.50 0.90 0.92 3.25 0.91 1.01 1.02
I2 1.05 1.00 1.01 1.22 1.01 0.98 3.18 1.04 1.00 1.00
I3 1.06 1.01 0.96 1.29 1.08 1.10 2.57 1.06 1.00 0.98
(c) ABCNNCBAABCN
I1 1.07 1.01 1.00 1.15 1.01 1.01 2.89 1.06 1.00 1.00
I2 0.97 1.00 1.00 0.92 1.00 1.00 2.89 0.97 1.00 1.00
I3 0.97 1.00 1.01 0.93 1.00 0.99 3.22 0.97 1.00 1.00

Table 9 Current Distribution in Cable Installations for a 3-Phase 6-Wire Circuit (s=0mm)
50Hz (pos.) 150Hz (zero) 250Hz (neg.)
A B C A B C A B C
(a) AAAANNNN / BBBBNNNN / CCCCNNNN
I1 0.96 1.02 0.98 0.43 0.58 0.43 0.98 1.02 0.96
I2 0.70 0.87 0.80 0.16 0.21 0.16 0.72 0.83 0.69
I3 0.74 0.87 0.91 0.31 0.30 0.31 0.77 0.81 0.67
I4 1.60 1.26 1.49 3.31 3.27 3.31 1.65 1.34 1.69
(b) AAAANNNN / NNNNBBBB / CCCCNNNN
I1 0.90 2.55 0.79 0.82 0.67 0.82 0.80 2.63 0.86
I2 0.66 0.53 0.51 0.67 1.11 0.67 0.53 0.28 0.56
I3 0.95 0.39 0.40 0.69 1.04 0.69 0.40 0.43 0.71
I4 2.41 0.72 2.44 1.84 1.17 1.84 2.57 0.73 2.57
(c) ANNAANNA / BNNBBNNB / CNNCCNNC
I1 1.07 1.05 1.07 1.04 1.07 1.04 1.08 1.07 1.08
I2 0.93 0.95 0.93 0.96 0.93 0.96 0.92 0.93 0.92
I3 0.93 0.95 0.93 0.96 0.93 0.96 0.92 0.93 0.92
I4 1.07 1.05 1.07 1.04 1.07 1.04 1.08 1.07 1.08

Table 10 Current Distribution in Cable Installations for a 3-Phase 6-Wire Circuit (s=300mm)
50Hz (pos.) 150Hz (zero) 250Hz (neg.)
A B C A B C A B C
(a) AAAANNNN / BBBBNNNN / CCCCNNNN
I1 0.95 1.46 0.94 0.94 0.77 0.94 0.94 1.46 0.94
I2 0.93 0.91 0.91 0.91 1.04 0.91 0.92 0.91 0.92
I3 0.97 0.78 0.96 0.96 1.11 0.96 0.96 0.78 0.97
I4 1.22 0.85 1.26 1.19 1.08 1.19 1.25 0.85 1.23
(b) AAAANNNN / NNNNBBBB / CCCCNNNN
I1 1.00 1.00 1.00 0.81 0.78 0.81 1.00 1.00 1.00
I2 1.00 1.00 1.00 0.72 0.67 0.72 1.00 1.00 1.00
I3 1.00 1.00 1.00 0.89 0.87 0.89 1.00 1.00 1.00
I4 1.00 1.00 1.00 1.58 1.69 1.58 1.00 1.00 1.00
(c) ANNAANNA / BNNBBNNB / CNNCCNNC
I1 1.00 1.00 1.00 1.03 1.04 1.03 1.00 1.00 1.00
I2 1.00 1.00 1.00 0.97 0.96 0.97 1.00 1.00 1.00
I3 1.00 1.00 1.00 0.97 0.96 0.97 1.00 1.00 1.00
I4 1.00 1.00 1.00 1.03 1.04 1.03 1.00 1.00 1.00

10
 li
ni

Dli n j
Dli l j Dni n j
Dni l j

lj nj

Figure 1 Geometry of Current Loops i and j

Figure 2 Circuit Model for 3-Phase 4-Wire Systems

Figure 3 Circuit Model of 3-Phase 6-Wire Systems

11
Figure 4 Measurement Setup

2.5

2
Rh /RDC

1.5

1
1 3 5 7 9
h

Figure 5 Resistance Ratios for non-Tray Cable Installations (20oC)

 Calculation (s=2d)
------ Calculation (s=4d)
o Measurement (s=2d)
 Measurement (s=4d)

7
Rh /RDC

1
1 3 5 7 9
h

Figure 6 Resistance Ratios for Cable Tray Installations (20oC)

O Measurement (s=2d)
 Measurement (s=4d)

12
 Figure 7 Cable Arrangements under Test

Figure 8 Trefoil Cable Arrangements

Figure 9 Flat Cable Arrangements

Figure 10 Layered Flat Cable Arrangements

13
View publication stats