1

Designs and Applications of HTS

Continuously Transposed Cable







N.J. Long
January 2005




















2

Table of Contents
Executive summary.................................................................................................... 3
Introduction................................................................................................................ 4
CTC design ................................................................................................................ 5
Conventional CTCs .............................................................................................. 5
Superconductor CTCs........................................................................................... 5
Four proposed CTC architectures .......................................................................... 7
CTC piece length and strand splicing .................................................................... 8
Defect tolerance and thermal stability ................................................................... 9
CTC splicing and current injection........................................................................ 9
Mechanical properties of CTC............................................................................... 9
AC loss of high temperature superconductors ......................................................... 10
Upper limit to acceptable AC loss ....................................................................... 10
AC loss in HTS materials .................................................................................... 10
Specific Applications ............................................................................................... 11
Transmission cables ............................................................................................. 11
Utility Transformers............................................................................................. 11
Traction transformers........................................................................................... 12
Induction heaters.................................................................................................. 12
Magnets................................................................................................................ 12
Generators and motors ......................................................................................... 13
Fault current limiters............................................................................................ 13
SMES................................................................................................................... 13
Summary.................................................................................................................. 13
References................................................................................................................ 14

3




Executive summary
Rapid progress is being made in the commercialisation of second generation (2G)
HTS wire for DC applications. Although some concepts to lower AC loss for AC
applications such as filamentization have been suggested many of the proposed
architectures are difficult to implement, require further twisting when incorporated in
a device, or do not fully address issues such as stabilisation. We propose the
development of a continuously transposed cable (CTC) using 2G wire which looks
very similar to conventional copper CTCs. Such a cable has been developed using
the 1G conductor but there has been little consideration among researchers of using
2G HTS.

We propose several different architectures for a 2G CTC which have benefits both for
AC and DC applications. Calculating the exact AC loss in a CTC is beyond this
review as it requires detailed modelling and greater knowledge of specific application
designs than we have at present. It also depends on the final filament width
achievable, at present we believe this will be in the range of 1-2mm. However, using
general considerations from previous AC loss studies, we argue that the cable can
reduce AC loss in applications such as transformers and induction heaters into the
range where they become seriously viable technologies.

4

W
d
P
Figure 1: A ten strand CTC
Introduction
Continuously transposed cables or conductors (CTC) were invented by Ludwig
Roebel in 1911 to minimise circulating current losses in the windings of electrical
machines such as generators and transformers. The conductors are commonly called
Roebel bars when used in generator stators. Innovations in Roebel conductors,
particularly in the material components, continue to be made almost 100 years after
Roebels breakthrough.

Figure 1 illustrates a CTC consisting of a
group of rectangular wires, arranged in two
stacks. In this group each wire is transposed
in turn without twist along its length. It can
be seen that each wire successively and
repeatedly takes on every possible position
inside the whole conductor cross-section.
This arrangement cleverly reduces eddy
currents in the conductor arising from
external or self generated AC fields. The
individual conductor strands have a width W,
and a thickness d. The length P, is the
transposing pitch over which the cable repeats its
configuration. The length over which an individual
strand returns to its original position is equal to the
number of conductors times the transposing pitch and is called the stranding pitch
or transposing lay.

Conventional Roebel bars usually contain an even number of strands, but may also be
made from odd numbers, with the added benefit of increasing the conductor cross
section by one strip within the same stack size. Numbers of strands typically range
from 5 to about 60. Paper tapes are wrapped around the assembly as external
insulation and this permits a large variation in the thickness and choice of insulation.

The use of CTC brings a number of advantages to electrical machine design which are
directly applicable to superconducting machines.

Reduction of AC losses
Increase in mechanical strength of windings, particularly their ability to
withstand the electrodynamic stresses caused by short circuiting
Increase in the space factor of the winding due to the reduced insulation
thickness
Improved heat dissipation
Reduced winding time for large coils
Flexibility in architecture, e.g. transposing pitch and current capacity
The strongest driver for using a CTC in superconductor applications is reduced AC
loss. This factor will be examined in detail in the following sections. However, other
5

drivers such as the above and the flexibility to create a conductor with very high Ic are
of potential interest.

CTC design
Conventional CTCs
Copper CTCs are manufactured by a four step process which involves bending in the
plane of the rectangular tape.

Step 1 Conductor #1 is bent across from the top of the right hand stack to the
top of the left hand
Step 2 The right hand stack is moved up one conductor thickness in relation to
the left hand stack.
Step 3 Conductor #4 is bent across from the bottom of the left hand stack to
the bottom of the right hand stack.
Step 4 The left hand stack is moved down one conductor thickness in relation
to the right hand stack.


Superconductor CTCs
This same process has been applied to BSCCO 1G wire
by groups such as Siemens [Leghissa et al] who made a
13 strand CTC for use in the low voltage coils of a
traction transformer. In the case of BSCCO great care has
to be taken to control the radius of curvature of the bend
or else the superconducting properties of the tape are
greatly degraded.

The second generation YBCO conductor is a thin film
conductor with a large aspect ratio. For this reason
edgewise bending of the tape has generally not been
considered feasible for producing a CTC. Other AC loss reduction strategies such as
filamentization have received much more attention [C.B Cobb et al]. An exception
Figure 2: Assembly procedure for a copper CTC. The strand is anchored at some point
and then bent into the new position.
Figure 3: An illustration of the
Siemens 1G cable
6

has been the work of Suzuki et al [Suzuki et al] who have experimented with bending
2G wire and accepting the slop in the cable due to the inability to produce plastic
deformation of the wire. In comparison with the solution described here this produces
a conductor with reduced critical current density and difficult mechanical properties.
The uneven profile of the conductor may also make it susceptible to damage.

We have invented an alternative method to produce a CTC using 2G wire. This
involves manufacturing pre-cut strands in the general shape shown in Figure 4. The
process to form the wire is then different from the process used for copper wire in that
no deformation is necessary but rather the wire must be twisted into place. A manual
assembly procedure for a 2G wire CTC is given in appendix A. Using mechanical
cutting it is expected that strands with 2mm diameter can be made. We have received
advice that 1mm strands should be possible using laser cutting, provided the laser
power and frequency are well matched to the material, this has yet to be
demonstrated.


The primary design for the CTC requires that the individual strands are essentially
isolated with only a small degree of current sharing to stabilise the conductor.
However we also consider designs with greater degrees of current sharing which may
be useful in DC or low frequency applications which require high currents and low
inductance. For illustrative purposes we consider a five strand conductor but the
principles generalise to any number of strands. In all cases it is expected that the
individual strands would be soldered together in the final package. Insulation of the
package would comprise wrapped layers of paper as for copper CTC's.

There are a number of possible architectures for the HTS wire itself which impact on
the design of the CTC. The basic architecture is shown in figure 5. In addition the
HTS wire requires extra stabilising conductor, usually copper, to provide a parallel
current path bypassing any small faults in the HTS and providing good heat
conduction to the cryo-system. By matching the copper thickness to the substrate
thickness so called neutral axis wire is created in which the HTS layer is placed in
the centre of the tape where the tensile and compressive bending strains are zero.




substrate
copper
Figure 5: a) General architecture of 2G HTS wire before packaging. b) Micrograph of
neutral axis wire. The substrate thickness is 50-75 microns.
Figure 4: The general shape of a pre cut strand used to form a CTC
7

In addition to neutral axis wire there are other possibilities for the wire architecture.
One is to place two wires face to face with a stabilising layer between. The difficulty
with this arrangement is that current is not easily injected into the composite wire.
Another proposal has been to coat both sides of a substrate with superconductor. This
has not been greatly pursued by researchers because of the difficulties of handling the
tape during production and because the superconducting layers are not easily
connected to improve stabilisation.
Four proposed CTC architectures
(1) Non-current sharing design
The asymmetry in the HTS packaging can be exploited in the CTC. For example if
the desire is to minimise current sharing between the strands of the CTC then the
tapes can be arranged as shown in Figure 6 (a). In this case there is never any
face-to-face contact of conducting elements and any current sharing can only
occur through edge contacts or through the substrate and (non-conducting) buffer
layers. This design produces a CTC which is asymmetric with respect to current
injection or removal and the CTC has a top and a bottom surface.




(2) Current sharing design
In configuration (b) half the conductors have been turned upside down (actually
for a conductor with 2n+1 strands, either n or n+1 are reversed). Now two pairs of
conductor have a high conductivity copper to copper connection. Further, as the
strands are transposed all vertically adjacent conductors will at some point have a
copper to copper contact, so a high conductivity path is available between all
strands. For purposes of current injection and splicing the conductor is essentially
symmetric. The degree of current sharing in this configuration can be controlled
by adding a further partially conductive layer to the copper surface before
assembly or by choosing a stabilising layer other than copper.

(3) High current density I
When the 2G wire has passed through the YBCO manufacturing process it
generally has an oxide scale present on the substrate surface. One option for
packaging the HTS is to remove the oxide scale and then electroplate the surface
with copper. If this is possible then a high current density CTC which partially
uses the Ni substrate as stabilisation is possible. The HTS wire as illustrated in
Figure 7(a) can be arranged as in (b) to form the CTC. A further copper
lamination could be added to the top and bottom of the CTC to provide extra
stabilisation if needed. The orientation of the HTS strands and the thickness of the
substrate
copper
copper
substrate
(a) (b)
Figure 6: A cross section of a five strand CTC illustrating (a) non current sharing and (b) current
sharing configurations.
8

copper electroplated layer could also be varied to optimise current density and
stability.





(4) High current density II
A further possibility is to use strands with superconducting layers on both faces.
Because the strands are naturally in a current sharing configuration this means that
injecting and removing current is straightforward and can be done from either face
of the cable. The use of stabilisation is also optimised in this configuration.
CTC piece length and strand splicing
Due to the complexities of the HTS wire manufacturing process discontinuities in Ic
have proved inevitable and are likely to remain for the foreseeable future. This
necessitates removing low I
c
segments of wire and limits the continuous piece lengths
available for application builders. Producing long piece lengths simplifies application
manufacture and commands a price premium. The introduction of splices in the wire
creates mechanical weaknesses which can cause problems, for example in rotating
machinery.

The proposed CTC designs offer the possibility of making conductors with
predetermined lengths in which splices can be incorporated without significant impact
on either the electrical or mechanical properties. For example if the current sharing
design is used then a new strand can be butted to the end of a broken strand. The
unaffected strands within a distance of a stranding pitch will be required to carry extra
current but only of the order of 1/N, where N is the number of strands. For the non
Electroplated copper
HTS
Ni substrate
(a)
(b)
HTS
Ni substrate
(c)
Electroplated copper
(d)
Figure 7: (a) HTS wire with electroplated copper packaging, (b) high current density I
CTC (c) HTS wire with double sided coating, (d) high current density II CTC
9

current sharing design a splice in the strand would need to be introduced, however the
impact on the total current capacity of the wire would be minimal. Likewise the
change in the mechanical properties of the CTC due to a splice in 1/N strands is
minimal.

We would foresee application developers being able to order CTCs to a
predetermined length so that no splicing by the application builder is necessary. The
incorporation of paper insulation also simplifies application building and the
insulation can easily be removed for current lead attachment.
Defect tolerance and thermal stability
The arguments around strand splicing in the current sharing design are also applicable
to the tolerance of small Ic dropouts in the conductor. A multiple strand conductor
will be more tolerant of defects due to the low probability of having a coincidence of
defects in different strands. This will make CTCs less likely to quench.
CTC splicing and current injection
Some proposed architectures for HTS packaging have foundered because splicing
and/or current injection become problematic. Because we believe CTCs can be
manufactured to a predetermined length splicing of CTCs themselves should rarely
be necessary. We discuss current injection/removal and splicing for the three
architectures we have proposed.
(1) Non-current sharing design: current injection/removal is not difficult as over
the stranding pitch all HTS layers are accessible from the top surface. As long
as current leads are soldered with an overlap of one stranding pitch then there
should not be a problem. Two cables can be spliced in the same manner by
overlapping the top surfaces.
(2) Current sharing design: Only half the HTS layers are directly accessible over a
stranding pitch. However, as there is current sharing within the cable then
provided a long enough overlap with the current leads is provided, current
injection/removal shouldnt be problematic. Also having a split current lead
contacting both top and bottom surfaces would solve the problem. Similarly
with splicing, as the CTC is essentially symmetric, either top or bottom
surfaces can be overlapped and joined. What constitutes a long enough
overlap needs to be determined for a particular cable and application.
(3) The high current density CTC is essentially the same as the non-current
sharing design and can be contacted and spliced in the same way.
Mechanical properties of CTC
An advantage of copper CTC that applies equally to superconductors is their
improved mechanical strength. The compressive and tensile strength of HTS wire has
been an issue for developers of rotating machinery and the c-axis strength of 2G wire
remains a potential stumbling block. Because in a superconductor CTC a number of
strands will be soldered together the compressive and tensile strength will greatly
exceed anything achievable by a single wire. The greater robustness of a CTC should
reduce the likelihood of introducing wire defects during winding and improve yield in
application manufacture.

The minimum bending radius of the CTC will depend on the number of strands and
the packaging of the individual strands. All strands experience the same mechanical
10

stresses if the wire is bent uniformly and the strands will alternately be in compression
and tension depending on where they are positioned within the cable.

Using a CTC may also help in applications requiring high c-axis strength. Partly this
can be achieved by using narrower strands. The use of a high J
e
design also allows
greatly flexibility in the packaging without fatally compromising the overall current
density.
AC loss of high temperature superconductors
Upper limit to acceptable AC loss
The units of AC loss which are most applicable to high current devices is W/Am the
energy loss per second, per ampere of current transported, per metre length of the
windings. To calculate the power loss for an HTS device a cooling penalty factor,
c
is introduced to account for the extra energy required to extract heat from low
temperature. The length of the conductor in the device and the total current in the
device are assumed to be fixed parameters. Therefore the relevant design parameter is
the product I P
c
/ , which should be minimised, where P is the power loss per meter
in the conductor and I is the current in the conductor.
General upper limits on AC loss for HTS can be derived from comparison with the
losses in copper based devices and allowing for the cooling penalty factor. For copper
operated at a typical value of 2A/mm
2
without forced cooling the power loss I P
c
/ is
15mW/Am for a sine-wave current [M. Oomen]. Assuming a cooling penalty of 15
for operation at liquid nitrogen temperature the total power consumption p
dev
in the
HTS device is equal to that in a copper device if P/ I
c
is 1.0mW/Am.

To justify the added complexity of a superconducting device normally the loss will
need to be significantly lower than 1.0mW/Am. Target values for transformers for
example, have been quoted as 0.2-0.8mW/Am [Marsh and Wolsky, Hornfelt,
Leghissa et al, Nagasawa et al]. This may not be the case if size and weight of the
device are the primary drivers for adopting HTS technology, e.g. for transportation
applications. In this case a loss closer to 1.0mW/Am may be justifiable. For
transformers there are also environmental advantages in the use of nitrogen as a
coolant rather than oil. Transformers and fault current limiters may also bring network
advantages which justify the added complexity.

AC loss in HTS materials
A detailed model to calculate the AC loss in a device is well beyond the scope of this
review. Neither will we attempt a comprehensive literature review. The goal is to use
some very rough approximations to estimate the usefulness of the CTC in meeting
targets for AC loss.

There are a number of loss mechanisms which are active in the conductor depending
on the local magnetic field, frequency and current conditions. The ferromagnetic
substrate will contribute a small hysteresis loss although this has shown to be
negligible under most conditions and we will ignore it [Ijaduola et al]. There are also
eddy current losses in metal components although at the frequencies of the devices we
consider these will not be significant. The loss which can be considered to be
11

dominant in the devices we consider is the magnetisation loss without transport
current. For an applied magnetic field B
a
and frequency f, the loss is
x a c
fd B I P = /
where d
x
is the largest dimension perpendicular to the field, P is the power loss per
meter and I
c
is the critical current. We assume that the field is large enough to fully
penetrate the superconductor.

Following Oomen [M.P. Oomen] we can use this to estimate the AC loss for 1G and
2G wires in various applications (see table 1).

Table 1: Approximate AC loss in power devices
Property [unit] Transmission
cable
Transformer Motor or
generator, AC
winding
Frequency of the field f, [Hz] 50 50 20
Field amplitude B
a,
parallel to the tape[T]

0.01 0.1 1
Field amplitude B
a,
perpendicular to the tape [T] - 0.01 1
P/I
c
due to B
a,
[mW/Am]

0.1 1.0 4 1G wire
4*0.2mm
2
P/I
c
due to B
z,
[mW/Am] - 2.0 80
P/I
c
due to B
a,
[mW/Am]

0.001 0.01 0.04 2G wire,
4*0.002mm
2


P/I
c
due to B
z,
[mW/Am] - 2.0 80


Specific Applications
Transmission cables
The magnetic field in transmission cables is weak and therefore the AC loss is well
within budget even for 1G wire. The use of 2G wire will improve things further. The
advantages of using a CTC therefore are likely to be in aspects other than AC loss.
Possible advantages are:
Simplified winding of the cable using a high I
c
CTC.
CTCs available to cable manufacturers in predetermined piece lengths no
splicing necessary.
Improved current sharing using the current sharing CTC design
Easier current injection/coupling to ends of cable
Improved thermal stability
Reduced space of the winding using a high current density design
Possibility of using a straight stack rather than a wound cable.
Utility Transformers
In the low voltage windings of transformers the currents are in the kA range. To use
HTS it is therefore necessary to use either a CTC winding or multiple in hand
windings of the coils. A CTC has the advantage of reducing the complexity of the
winding process and ensures good current sharing between the strands.

At the nominal goal of 0.25mW/Am for an HTS conductor the coil losses in a 10-
MVA transformer would be 5% of todays losses in copper [i.e. (15x0.25
mW/Am)/(80mW/Am)] [Marsh and Wolsky]. The AC losses in a 1G transformer
12

prototype have been reported as 3mW/Am [Hornfelt] and correspond well to
predicted values. Assuming the same design using a 4mm wide 2G wire would give
losses of 2mW/Am. The use of a CTC with 2mm strands would bring the loss down
to the magical number of 1mW/Am where HTS becomes attractive. If the loss could
be reduced further either through improved overall design or reducing the strand
width below 2mm then commercial HTS utility transformers may become a reality.
Traction transformers
Siemens have an active programme developing transformers for trains. They have
developed a 13 strand CTC using 2mm diameter 1G tape. In transportation the
reduction in size and weight is a primary driver so that slightly higher losses can be
tolerated compared to other applications. It would be expected that a 2G CTC with a
strand width of 2mm would have about 50% lower losses than the 1G cable. The 2G
CTC would clearly supplant the 1G product if this market develops.
Induction heaters
Induction heaters are used to heat billets of metal before extrusion or other forming
processes, for example aluminium billets are heated to around 500C pre extrusion.
The heaters work by inducing eddy currents in the metal. Conventional induction
heaters are only about 55-60 per cent efficient when heating high conductivity
materials such as copper and aluminium. The losses in the copper windings are
around 400mW/Am.

A prototype heater using 1G wire has been made by SINTEF energy research. The
current 1G wires give losses in the 15 to 20mW/Am range (x15 or 225-400mW/Am
including the cooling penalty at 77K) if used in a 1MW class induction heater. The
system would be viable if losses could be reduced to 10% or 100kW for a 1MW
system. The target for HTS wire therefore is about 5mW/Am to make this technology
viable.

The benefit of a 2G CTC can be estimated if we make the assumption that the ratio of
parallel to perpendicular field losses is similar to those for a transformer, i.e. 1:2 for
1G wire. For parallel field components the use of 2G wire will likely reduce the losses
to near zero. If we use a 2mm strand width rather than the 4mm then the losses in
perpendicular field are reduced by half. Therefore a CTC with a 2mm strand width
will reduce the loss from 15mW/Am to 5mW/Am making this a viable HTS
technology.

Magnets
The generic benefits of using CTC such as increased effective J
e
values due to the
smaller fill factor of insulation and higher current density design may be valuable for
applications where size and weight are important.

For NMR insert magnets the benefits would be particularly attractive:
High strength
Low inductance
High current density
Lower AC loss
13

Generators and motors
The mechanical properties of a CTC may make it useful within generators. The cable
is better able to withstand the very high electrodynamic forces which may arise during
a short circuit. Simplifying the winding process and decreasing the space of the
windings may also be major advantages. AC losses are not a problem where HTS is
used in the DC field coils of a rotating machine. The use of HTS in armature coils
would require orders of magnitude reduction in AC loss and is not considered feasible
by the use of a CTC as we have described.
Fault current limiters
Using the CTC concept allows a coil to be constructed with an Ic (A/cm-W) much
larger than that for a single HTS wire. This also allows the inductance of the coil to be
easily engineered. This flexibility may be of value in designing practical fault current
limiters. The CTC cable also has greatly enhanced mechanical strength which will be
of value in maintaining the integrity of the device under fault conditions.
SMES
The development of CTC will allow greater flexibility in SMES design to
accommodate higher currents and lower inductance coils. Time dependent fields are
present as the SMES charges and discharges however it is not clear that these will be
the dominant losses in the system.

Summary
There are possible benefits to almost all devices in using CTC. The most attractive
application is transformers as CTC has a natural use in the high current low voltage
coils. CTC is also a feasible strategy for reducing the AC loss in a transformer into the
range where the technology is viable. Transformers are a large market using
substantial quantities of conductor. The potential gains should easily justify the
development costs for a CTC.

14

References

M. P. Oomen, AC Loss in Superconducting Tapes and Cables, PhD thesis, 2000,
University of Twente
S.P. Hornfelt, HTS in Electric Power Applications, Transformers, Physica C 341-348
(2000) 2531-2533
Leghissa M, Grommoll B, Rieger J, Oomen M, Neumuller H-W, Shlosser R, Schmidt
H, Knorr W, Meinert M and Henning U, Development and application of
superconducting transformers, Physica C 372-376 (2002) 1688-1693
C.B. Cobb, P.N. Barnes, T.J. Haugan, J. Tolliver, E. Lee, M. Sumption, E. Collings,
C.E. Oberly, Hysteretic loss reduction in striated YBCO, Physica C 382 (2002) 52-56
C. Suzuki, K. Goto, T. Saitoh, T. Nakatsuka, Strain properties of transposed segment
conductors for a transmission cable, Physica C 392-396 (2002) 1186-1191
T. Nagasawa, M. Yamaguchi, S. Fukui, M. Yamamoto, Design requirements of a high
temperature superconducting transformer, Physica C 372-376 (2002) 1715-1718
A. O. Ijaduola , J. R. Thompson , A. Goyal , C. L. H. Thieme and K. Marken,
Magnetism and ferromagnetic loss in NiW textured substrates for coated conductors,
Physica C, 403, ( 2004) 163-171
G.E. March and A.M. Wolsky, AC losses in high temperature superconductors and
the importance of these losses to the future use of HTS in the power sector, IEA, May
2000
15

Appendix A: Manual Assembly Process for a CTC
For illustration we will describe forming a 5 strand CTC using 2G wire. Imagine that
the starting point is the cross section of the CTC as shown below in (a). The first
transition is for strand A to move from left to right. The next transition will be for
E to move from right to left. The strands must therefore be staggered before
stacking as shown in (b). They are then stacked in the vertical order DEABC as
shown in (c). The first transposition is then to twist tape A such that the cross over
takes the tape above tape D. Tape E is then twisted so that its first cross over takes the
tape under tape C. This process is then continued and can be summarised for an odd
numbered stack by the 2 step algorithm: 1. Twist the middle tape of the stack to the
left and to the top of the stack, 2. Twist the middle tape of the stack to the right and
the bottom of the stack.




A
B
C E
D
A B C D E
D
E
A
B
C
(a)
(c)
(b)