Buschow K.H.J. (Ed.) Concise Encyclopedia of Magnetic and Superconducting Materials

Подождите немного. Документ загружается.

Superconducting Wires and Cables:

High-ﬁeld Applications

The performance of superconducting wires is funda-

mentally limited by the magnetic ﬂux density or ﬁeld,

the current density, and the operating temperature,

while the magnet applications are limited by current,

voltage, and stored energy. Thus the term ‘‘high-ﬁeld

applications’’ is really a generalization for high-per-

formance applications, involving all of these conduc-

tor and magnet constraints. As a rule, high ﬁeld is

easiest to obtain in small bores and at low temper-

atures. This article reviews progress in strands and

cables that improves the performance capabilities of

large-scale superconducting applications.

All high-current and many high-ﬁeld supercon-

ducting applications use conductors that are made up

of cables with multiple strands. A large variety of

cable topologies have been used, but two are the most

common. High-energy physics multipole (dipole,

quadrupole, etc.) magnets generally use Rutherford

cables, a high aspect ratio topology with two layers,

and a highly regular cabling and crossover pattern.

Large energy storage, fusion research, and detector

magnets more frequently use cable-in-conduit con-

ductors (CICC), round cables that are built up from

smaller round subcables, then compacted into a

square or low aspect ratio conduit. The two cable

types are illustrated in Fig. 1.

Rutherford cables are favored for multipole mag-

nets, because they can achieve tighter bends with

lower conductor strain, while multipole magnets are

characterized by modest bore sizes and high length-

to-minimum-bend-radius ratios. Cable-in-conduit

conductors are favored for the larger bore fusion,

detector, and superconducting magnetic energy stor-

age (SMES) magnets, because of their ability to be

built up to nearly unlimited current-carrying capa-

bilities, making them most appropriate for very large,

high-energy applications. CICC conductors have

been built and tested up to 200 kA for large SMES

applications (Peck and Michels 1991) and large mag-

nets have been tested to 80 kA (ITER 2001).

1. Strand

Commercial and scientific applications of supercon-

ducting cables are dominated by niobium titanium

(NbTi). NbTi, the ﬁrst practical superconducting

material for large-scale applications, is strong, duc-

tile, and relatively inexpensive. The production of

NbTi strand is approximately 200 t per year. The

largest commercial application is medical magnetic

resonance imaging (MRI) magnets, while scientific

applications are currently dominated by production

of the Large Hadron Collider at CERN (Perin 1991).

NbTi magnets have been built with peak ﬂux densi-

ties ranging from under 1 T to peaks of 8–9 T at 4.2 K

(the boiling temperature of liquid helium) and 9–11 T

at 1.8 K (Perin 1991) (near the lambda point of su-

perﬂuid helium). NbTi is a mature technology, which

has slowly been improving, so that it is now possible

to purchase large quantities with critical current den-

sities of 3000 A mm

2

at 5 T and 4.2 K.

All other magnet applications use Nb

Sn with an

annual production of approximately 2.5 tonnes per

year. Nb

Sn is used in nuclear magnetic resonance

and open MRI magnets. Scientific applications in-

clude next-generation magnets for high-energy phys-

ics and fusion, including 20 t of Nb

Sn in the KSTAR

tokamak being built in South Korea (Lee et al. 1999)

and an equivalent amount for the recently tested

ITER CS Model Coil in Japan (Martovetsky et al.

2001). Nb

Sn cables have been used in large pulsed

cables up to 13 T (Tsuji et al. 2001), dipole magnets

up to 14.7 T (Gourlay et al. 2001), small solenoids up

to 17 T (Takeo et al. 1987), and inserts up to 21 T.

Sn has superior critical temperature and ﬁeld, but

it is an order of magnitude more expensive than

NbTi, requires a long heat treatment at 600–700 1Cto

become superconducting, and is brittle, requiring

great care in handling during fabrication and cool-

down to the cryogenic operating temperature.

However, the conductor development program

for a high-ﬁeld Very Large Hadron Collider (VLHC)

has achieved spectacular progress in critical current

density, which has risen over just a few years from

700 A mm

2

to 2900 A mm

2

at 4.2 K and 12 T. Sig-

nificant progress in the programme’s goal of reducing

the cost of Nb

Sn in large quantities from $1000 to

$200 kg

1

would make several new high-ﬁeld appli-

cations possible.

New materials have been proposed, but have

not yet had practical impact on high-ﬁeld magnet

Figure 1

Cable types: (a) simple Rutherford cable, (b) Rutherford

cable with kapton/glass insulating wrap, (c) multistage

circular cable, (d) cable-in-conduit.

1190

Superconducting Wires and Cables: High-ﬁeld Applications

applications. The two most widely studied high-tem-

perature superconductor materials, BSSCO-2223

(bismuth strontium calcium copper oxide) and

YBCO (yttrium barium copper oxide), are not avail-

able as strands. Artiﬁcial pinning center strands have

been proposed as an alternative to NbTi, because of

superior properties at low ﬁeld and projected low

manufacturing costs (Rudziak et al. 1995). Nb

Al,

which has been proposed as an alternative to Nb

Sn,

has similar critical properties, but much lower strain

sensitivities.

A recently tested Nb

Al cable has shown perform-

ance at 13 T and 46 kA that equals that of previous

generation Nb

Sn and with less degradation over

time (Takeuchi 2001). BSSCO-2212 strands have

been proposed as high-temperature superconductors,

paradoxically because of their superior performance

at low temperature and high ﬁeld. Cables using

BSSCO-2212 have achieved 10 kA. None of these

alternative strands have demonstrated the existence

of more than niche markets and, despite impressive

progress, they remain single vendor prototypes. A

new strand material that may have a better oppor-

tunity to ‘‘break out’’ into commercialization is the

recently discovered medium-temperature MgB

which beneﬁts from being potentially inexpensive to

manufacture, lightweight, and ductile, although its

critical temperature is lower than BSSCO or YBCO,

while its critical ﬁeld is lower than Nb

Sn.

State-of-the-art performance of the different

strands is listed in Table 1.

2. Rutherford Cable

The Rutherford cable is a two-layer cable of round or

square strands, whose high aspect ratio allows wind-

ing around the relatively tight radii of constricted end

turns in magnets with long straight sections. The

simplest form of the Rutherford cable is a two-layer

cable with uniform superconducting strand place-

ment and uniform crossover of strands. Several vari-

ations of these two forms have been built into a

number of historic high-ﬁeld magnets. Some of these

historic cables are illustrated in Fig. 2. Rutherford

cables can be compacted for higher strength and

packing factor, and ‘‘keystoned’’ so that a winding

can ﬁt without gaps on a circumference. As shown

in Fig. 2, they can be reinforced by a central core—

either steel to add strength and reduce losses, or cop-

per to improve stability and quench protection

(Beard et al. 1988).

Figure 2 also shows how a Rutherford cable can be

inserted horizontally or vertically into a copper chan-

nel, improving strength, stability, and protection

(Beard et al. 1988, Schultz et al. 1999). It also shows

cables that functionally resemble a CICC, by being

inserted into a sealed can with forced-ﬂow helium

circulating directly through the conductor or through

side-channels (Beard et al. 1988, Chernoplekov and

Monoszon 1982). The two-layer conﬁguration can

Figure 2

Historic Rutherford cable design options, drawn to

scale: (a) large coil task, IGC/General Electric, cable

with a copper core; (b) levitated dipole experiment

ﬂoating coil, cable in a copper C-channel; (c) large coil

task, Hitachi, cable in a copper U-channel; (d) large coil

task, EURATOM/Vacuumschmelze, cable-in-conduit

with an insulated steel core; (e) large helical device, cable

with an aluminum stabilizer in a copper housing;

(f) T-15 tokamak, cable with core in a housing and

external cooling channels.

Table 1

State-of-the-art strands.

Parameter Units Value Name/reference

cnonCu

(4.2 K, 5 T, NbTi) A mm

2

3800 SSC, OST 14 mm, Furukawa 2.5 mm

cnonCu

(4.2 K, 12 T, Nb

Sn) A mm

2

2900 Scanlan

cnonCu

(4.2 K, 20 T, Nb

Al) A mm

2

650 Sumption, DOE HEP (Sumption et al. 2002)

cnonAg

(4.2 K, 0 T, BSSCO-2212) A mm

2

2000 Showa, manufacturer’s literature

Nstrands/cable (CICC) 1440 ITER Russian PF insert coil (ITER 2000)

Nstrands/cable (Rutherford) 47 LBNL D20

60 LBNL/Fermilab 2002 (not used in magnet)

1191

Superconducting Wires and Cables: High-ﬁeld Applications

also beneﬁt from the ability to achieve a highly reg-

ular cabling pattern and near-uniform contact resist-

ances, permitting analysis of cable supercurrents.

Cables with as many as 47 strands have been used in

high-energy physics dipole magnets (McInturff et al.

1997).

Since Rutherford cables are typically used in quasi-

steady-state applications, they are usually designed

for low contact resistance between strands. The spec-

iﬁcation for the Large Hadron Collider (Perin 1991)

is 10 mO per contact. Such a low contact resistance

tends to be dominated by the oxide layer on the

strand surface. In Tevatron (Cooper and Fisk 1983)

and HERA (Meinke 1991), the strands are covered

by a thin layer of SnAg (5%) to reduce surface re-

sistance. For applications requiring a significant

ramp rate, such as injector magnets, a compromise

has to be made between good current transfer for

stability and good resistance for low losses. This can

usually be accomplished by a thin steel core in the

center of the cable, allowing strand–strand current

transfer, while eliminating large circulating currents.

3. CICC

Although it has many variants, the cable-in-conduit

conductor (CICC) concept has changed little since its

invention in the 1960s. Strands are cabled into suc-

cessive stages, called subcables, in which typically 3–6

subcables are bundled into the next larger cable stage.

Cables with as many as 1440 strands have been built

up in this manner (ITER 2001). The cable is then

jacketed, i.e., inserted into a helium-tight conduit that

provides support against winding and Lorentz loads,

as well as being a pressure vessel for helium coolant.

The most common CICC is a simple circular cable,

inserted into a conduit that is squared off with Turks-

head rollers, then welded together, either with a con-

tinuous seam weld or by orbital welding of discrete

sections of pipe. The aspect ratio can be higher than

one, when this improves the overall winding packing

factor.

A circular conduit has the dual advantage of being

indifferent to angular position due to any twist dur-

ing winding or transitions and it also has the lowest

stress multiplier for holding internal helium pressure.

However, it is incapable of sustaining much com-

pression from radial loads in the winding pack, and

has a low packing factor, so that it must be placed in

a scalloped structural plate. The cable may have a

central coolant hole (Tsuji et al. 2001), or it can be

placed in a thin inner conduit, leaving four unob-

structed cusps in the corners (Steeves et al. 1989), so

that the hydraulic path will have a larger hydraulic

diameter and less friction, allowing the removal of

higher heat loads. This philosophy can be extended to

where the central hole is much larger than the cable

space for very long channels or when needed to allow

conduction cooling by superﬂuid helium (Peck and

Michels 1991). Historic examples of these design op-

tions are illustrated in Fig. 3.

CICC has several structural, electrical, and con-

ductor cooling advantages. The conduit can be a

high-strength superalloy and the winding pack can be

completely potted, giving both mechanical rigidity

and high electrical integrity. The cable is in direct

contact with helium coolant with a high wetted pe-

rimeter and can utilize the much higher enthalpy of

the helium to give it an energy margin against dis-

turbances that is orders of magnitude higher than a

dry magnet. The disadvantages of CICC are the ad-

ditional cost of the conduit, closure welds, and cable

insertion, local stress concentrations in the conduit

walls, the need for a forced-ﬂow refrigerator, and the

need for electrical isolation from the helium headers.

Because of the nearly unlimited ability to build up the

cable to high currents and the demonstration of elec-

trical isolation to 25 kV, CICC is favored in large-

bore, high-energy storage magnets.

Progress in conduit design and manufacture has

included (i) the development of reliable closure welds,

(ii) conduit materials and welds with high strength

Figure 3

Historic cable-in-conduit design options, drawn to scale:

(a) large coil task, Westinghouse/AIRCO, Nb

Sn;

(b) dual purpose coil, Japan Atomic Energy Research

Institute, cable-in-conduit with through cable and

external cooling channels; (c) US dual purpose coil,

central CICC in thick walled conduit and corner

channels; (d) POLO, insulated subcable, dual channel

CICC; (e) SSC GEM detector, cable in thin steel tube in

thick aluminum conduit; (f) superconducting magnetic

energy storage-engineering test magnet, large central

channel; (g) International Thermonuclear Experimental

Reactor (ITER) central solenoid, thick conduit, central

cooling channel; (h) ITER toroidal ﬁeld magnets, thin

conduit in plate, central cooling channel.

1192

Superconducting Wires and Cables: High-ﬁeld Applications

and toughness, even after being subjected to long su-

perconductor heat treatments, and (iii) conduit ma-

terials with a good thermal contraction match to the

superconductor. Reliable seam welds were demon-

strated in thick JBK-75 (Beard et al. 1988) and Inc-

oloy (Steeves et al. 1989) conduits in the 1980s. The

International Thermonuclear Experimental Reactor

(ITER), which would have required hundreds of kilo-

meters of high current cable, developed orbital weld-

ing techniques with two advantages. The total length

of welds in the system was greatly reduced and the

welds were fully inspectable, before the insertion of

cable.

ITER has also demonstrated excellent performance

in high-current cables using Incoloy 908 and titanium

alloy (Martovetsky et al. 2001). Incoloy 908 has the

dual advantages of a good thermal contraction match

with Nb

Sn and higher strength and toughness than

stainless steel conduits, after a long heat treatment.

Titanium puts the lowest strain in the superconductor

of any practical material. Both Incoloy 908 and ti-

tanium alloy require tight control of the atmosphere

during heat treatment to avoid severe embrittlement

during heat treatment. Less-expensive 304 and

316LN stainless steels have been used with NbTi

and a low-carbon 316 has been developed to survive

the Nb

Sn heat treatment (Miller et al. 1994).

4. State-of-the-art Magnets and Future Progress

Several metrics are available for measuring state-of-

the-art performance from magnets using strands and

cables. These include magnets of solenoidal, toroidal,

and multipole topologies, achieving the highest ﬁeld,

the highest current density for the current and volt-

age, the highest superconducting pinning force (cur-

rent density-ﬁeld product), and the highest average

stress (e.g., the current density-ﬁeld-radius product in

a solenoid). International experts were surveyed for

this article, and the highest historic values of these

metrics that were identiﬁed are listed in Table 2.

Further improvements in high-ﬁeld superconduct-

ing strands and cables will probably be driven by

steady increases in the critical current density, tem-

perature, and ﬁeld of the strands. It should also be

possible to significantly lower the manufacturing

costs of all of the advanced strand materials, when

they are produced in commercial quantities. The

same potential for reducing costs with quantity exists

for the advanced conduit materials desired for CICCs

with heat-treated strands. The copper needed for the

protection of CICC magnets should be reduced by

the development of internal heaters and sensors with

high signal/noise ratios. Advances in the understand-

ing and performance of cables are expected if fully

Table 2

State-of-the-art performance: cable and magnet.

Parameter Units Value Name/reference

max

(solenoid) T 23.4 NRIM NMR (Kiyoshi et al. 2000)

max

(cryocooled solenoid) T 15.0 INRC, Tohoko U. (Watanabe and Motokawa 2000)

max

(NbTi dipole) T 10.15 SSC/LBNL D-19

max

(NbTi solenoid) T 10.1 MIT Hybrid III (Iwasa et al. 1993)

max

(NbTi/CICC, 1.8 K) T 11.1 Euratom LCT (Darweschsad et al. 1999)

max

(Nb

Sn/CICC, 1.8 K) T 14.1 NHMFL 45 T Outsert (Miller et al. 1994)

max

(pulsed coil, 4 1Ts

1

) T 13 ITER CS model coil

max

(pulsed coil, 4 5Ts

1

) T 6.7 DPC-EX (Ando et al. 1994)

max

(dipole) T 14.7 LBNL RD-3 (Gourlay et al. 2001)

max

(quadrupole) T 9.63 LHC KEK low-beta quad

grad

(quadrupole) T m

1

260 LHC MQY Oxford/Cern IR (Strait 2001)

cold

kJ kg

1

11.8 CMS solenoid detector (Herve 2000)

cond

(CICC) kA 200 SMES/ETM (Walker et al. 1990)

cond

(dipole) kA 19.8 U. Twente 11 T (den Ouden et al. 1997)

term

kV 23 POLO (Darweschad et al. 1997)

(dipole/quadrupole) A mm

2

560 LLNL APC HCX prototype (Martovetsky et al. 2002)

(solenoid) A mm

2

415 9.5 T Cryocool outsert (Shen et al . 2001)

nonCu

(dipole/quadrupole) A mm

2

2349 LHC low I insertion quad (Lucas et al. 2000)

(solenoid) A mm

2

1343 9.5 T Cryocool outsert

(solenoid/CICC) A mm

2

550 US-DPC (Painter et al. 1992)

(dipole/quadrupole)) A mm

2

1550 LLNL APC HCX prototype

(dipole) A mm

2

1627 LBNL D20, outer

max

(solenoid) MPa 220 NHMFL 45 T outsert (Miller et al. 1994)

max

(solenoid) MPa 335 NHMFL 45 T outsert

bore

(quadrupole) m

0.73 TJNAF Hall C Q2

1193

Superconducting Wires and Cables: High-ﬁeld Applications

predictive codes become available for calculating su-

percurrent evolution in each individual strand of a

cable, providing reliable predictions of cable losses

and individual strand recovery performance.

See also: Superconducting Materials, Types of;

Superconducting Wires and Cables: Low-ﬁeld Appli-

cations; Superconducting Wires and Cables: Materi-

als and Processing; Superconducting Permanent

Magnets: Potential Applications

Bibliography

Ando T, Nakajima H, Sasaki T-Y, Hiyama T, Takahashi Y,

Nishi M, Yoshida K, Okuno K, Kato T, Sugimoto M, Isono

T, Kawano K, Koizumi N, Hamada K, Oshikiri M, Sasaki T,

Wadayma Y, Miyake A, Iwamoto S, Takahashi T, Ebisu H,

Hosono T, Yamamoto K, Yasukawa Y, Tsukamoto H, Ha-

nawa H, Seki S, Wakabayashi H, Tsuji H, Takayasu M, Turck

B, Wuchuner F 1994 The second test results on the Nb

demo poloidal coil (DPC-EX). Adv. Cryogenic Eng. 39, 335–41

Beard D S, Klose W, Shimamoto S, Vecsey G, Haubenreich P

N, Komarek P, Shimamoto S, Dresner L, Fietz W A, Kato T,

Lubell M S, Lue J W, Luton J N, Maurer W, Okuno K,

Schwenterly S W, Takahashi Y, Ulbricht A R, Zichy J A

1988 The IEA Large Coil Task: development of supercon-

ducting toroidal ﬁeld magnets for fusion power. Fusion Eng.

Des. 7, 1–229 (special issue)

Chernoplekov N A, Monoszon N A 1982 T-15 facility and

tests. IEEE Trans. Magn. 23, 826–30

Cooper W E, Fisk H E 1983 Fermilab Tevatron quadrupoles’’.

IEEE Trans. Magn. 19, 1372–5

Darweschsad M, Fink S, Friesinger G, Gruenhagen A, Heller

R, Hofmann A, Herz W, Komarek P, Maurer W, Noether G,

Rietzschel K, Schleinkofer G, Suesser M, Ulbricht A,

Wuechner F, Zahn G 1999 Test of the EURATOM LCT

coil (NbTi conductor) with forced ﬂow He II. Fusion Eng.

Des. 45, 361–75

Darweschsad M, Friesinger G, Heller R, Irmisch M, Kathol H,

Komarek P, Maurer W, No

ther G, Schleinkofer G, Schmidt

C, Sihler Ch, Su

er M, Ulbricht A, Wu

chner F 1997 Devel-

opment and test of the poloidal ﬁeld prototype coil POLO at

Forschungszentrum Karlsruhe. Fusion Eng. Des. 36, 227–50

Gourlay S A, Benjegerdes B, Bish P, Byfford D, Caspi S, Chow

K, Dietderich D, Gupta R, Hafalia R, Hannaford R, Harn-

den W, Higley H, Jackson A, Lietzke A, Liggins N, McInturff

A, Millos G, O’Neill J, Palmerston E, Sabbi G, Scanlan R,

Swanson J 2001 Fabrication and test of Nb

Sn racetrack coils

at high ﬁeld. IEEE Trans. Appl. Supercond. 11 (1), 2164–7

Herve A 2000 The CMS detector magnet. IEEE Trans. Appl.

Supercond. 10 (1), 389–94

ITER Design Group 2000 Technical Basis for the ITER FEAT

Outline Design. ITER EDA Documentation Series No. 19.

IAEA, Vienna

ITER TFMC Operation and Testing Group 2001 First results

of test of the ITER toroidal ﬁeld model coil in the TOSKA

facility of Forschungszentrum Karlsruhe. (in press)

Iwasa Y, Leupold M J, Weggel R J, Williams J E C, Itoh S 1993

Hybrid III: the system, test results, the next step. IEEE Trans.

Appl. Supercond. 3 (1), 58–65

Kiyoshi T, Kosuge M, Yuyama M, Nagai H, Wada H, Kit-

aguchi H, Okada M, Tanaka K, Wakuda T, Ohata K, Sato J

2000 Generation of 23.4 T using two Bi-2212 insert coils.

IEEE Trans. Appl. Supercond. 10 (1), 472–7

Lee G S, Kim J, Hwang S M, Chang C S, Chang H Y, Cho M

H 1999 The design of the KSTAR tokamak. Fusion Eng. Des.

46, 405–24

Lucas J, Kirby G A, Ostojic R, Rodrigeuz-Mateos F, Krischel

D, Schillo M 2000 Design and construction of a 1 m model of

the Low Current Superconducting Quadrupole for the LHC

insertions. IEEE Trans. Appl. Supercond. 10 (1), 150–3

Martovetsky N, Manahan R, Lietzke A 2002 Development of

superconducting focusing quadrupole for heavy ion driver.

Magnet Technology Conf. MT-17, Geneva, Sept 24–28, 2001

Martovetsky N, Michael P, Minervini J, Radovinsky A, Ta-

kayasu M, Thome R, Ando T, Isono T, Kato T, Nakajima

H, Nishijima G, Nunoya Y, Sugimoto M, Takahashi Y, Tsuji

H, Bessette D, Okuno K, Ricci M 2001 ITER CS model coil

and CS insert test results. IEEE Trans. Appl. Supercond. 11

(1), 2030–3

McInturff A D, Benjegerdes R, Bish P, Caspi S, Chow K,

Dell’Orco D, Dietderich D, Hannaford R, Harnden W, Hi-

gley H, Lietzke A, Morrison L, Morrison M, Scanlan R,

Smithwick J, Taylor C, van Oort J 1997 Test results for a high

ﬁeld (13 T) Nb

Sn dipole. Proc. Particle Accelerator Confer-

ence, Vancouver, Canada, May 12–16, 1997, pp. 3212– 14

Meinke R 1991 Superconducting magnet system for HERA.

IEEE Trans. Magn. 27, 1728–31

Miller J R, Bird M D, Bole S, Bonito-Oliva A, Eyssa Y, Kenney

W J, Painter T A, Schneider-Muntau H-J, Summers L T, Van

Sciver S W, Welton S, Wood R J, Williams J E C, Bobrov E,

Iwasa Y, Leupold M, Stejskal V, Weggel R 1994 An overview

of the 45-T hybrid magnet system for the new national

high magnetic ﬁeld laboratory. IEEE Trans. Magn. 30 (4),

1563–70

den Ouden A, Wessel S, Krooshoop E, ten Kate H 1997 Ap-

plication of Nb

Sn superconductors in high-ﬁeld accelerator

magnets. IEEE Trans. Appl. Supercond. (June), 733–8

Painter T A, et al. 1992 Test data from the US Demonstration

Poloidal Coil Experiment. MIT Plasma Fusion Center Re-

port, PFC/RR-91-1, pp. 1–281

Peck S D, Michels P H 1991 Test results from the 200 kA

SMES/ETM conductor. IEEE Trans. Magn. 27 (2), 1724–7

Perin R 1991 The superconducting magnet system for the LHC.

IEEE Trans. Magn. 27, 1735–40

Rudziak M K, Seuntjens J M, Renaud C V, Wong T, Wong J

1995 Development of APC Nb-Ti composite conductors at

Supercon, Inc. IEEE Trans. Appl. Supercond. 5 (2), 404–7

Schultz J H, Kesner J, Minervini J V, Radovisnky A, Pourra-

himi S, Smith B, Thomas P, Wang P W, Zhukovsky A, Myatt

R O, Kochan S, Mauel M 1999 The levitated dipole exper-

iment (LDX) magnet system. IEEE Trans. Appl. Supercond. 9

(2), 378–81

Shen W, Coffey M, McGhee W 2001 Development of 9.5 T

NbTi cryogen-free magnet’’. IEEE Trans. Appl. Supercond.

11 (1), 2619–22

Steeves M M, Hoenig M O, Takayasu M, Randall R N, Tracey

J E, Hale J R, Morra M M, Hwang I, Marti P 1989 Progress

in the manufacture of the US-DPC test coil. IEEE Trans.

Magn. 25 (2), 2089–92

Strait J 2001 Very high gradient quadrupoles. Proc. 2001 Par-

ticle Accelerator Conf., Chicago, IL, pp. 176–80

Sumption M D, Buta F, Collings E W, Tomsic M, Austen A,

Gregory E, Parrell J, Hong S, Zhang Y, Scanlan R M 2002

Al strand processing transport properties and cabling.

Proc. Applied Superconductivity Conf., Houston, TX, Aug

2002, to be published

1194

Superconducting Wires and Cables: High-ﬁeld Applications

Takeo M, Funaki K, Sato S, Yamafuji K, Iwakuma M, Hira-

matsu M, Matsumoto K, Hamada M 1987 A 17 tesla super-

conducting magnet with multiﬁlamentary superconductors.

IEEE Trans. Magn. 23 (2), 565–8

Takeuchi T 2001 Nb

Al conductors. Magnet Technology Conf.

MT-17, Geneva, Sept 24–28, 2001

Tsuji H, Okuno K, Thome R, Salpietro E, Egorov S, Marto-

vetsky N, Ricci M, Zanino R, Zahn G, Martinez A, Vecsey

G, Arai K, Ishigooka T, Kato T, Ando T, Takahashi Y,

Nakajima H, Hiyama T, Sugimoto M, Hosogane N, Mats-

ukawa M, Miura Y, Terakado T, Okano J, Shimada K,

Yamashita M, Isono T, Koizumi N, Kawano K, Oshikiri M,

Nunoya Y, Matsui K, Tsuchiya Y, Nishijima G, Kubo H,

Shimba T, Hara E, Imahashi K, Uno Y, Ohuchi T, Ohtsu K,

Okayama J, Kawasaki T, Kawabe M, Seki S, Takano K,

Takaya Y, Tajiri F, Tsutsumi F, Nakamura T, Hanawa H,

Wakabayashi H, Shimizu T, Kuramochi K, Omine T, Ta-

miya T, Harada J, Nishii K, Huguet M, Mitchell N, Bessette

D, Minervini J, Vieira R, Michael P, Takayasu M, Be-

vilacqua G, Maix R, Manahan R, Jayakumar R, Savoldi L,

Herz W, Ninomiya A 2000 Progress of the ITER Central

Solenoid Model Coil Program. IAEA-CN-77, 18th IAEA

Energy Conf., Sorrento, Italy, 4-10, IAEA 2001, pp. 354–61

Walker D L, et al. 1990 Design of a 200 kA conductor for

superconducting magnetic energy storage (SMES). Adv. Cryo-

genic Eng. 35, 573–9

Watanabe K, Motokawa M 2000 Cryogen-free high ﬁeld su-

perconducting magnets. IEEE Trans. Appl. Supercond. 10 (1),

489–94

J. H. Schultz

Massachusetts Institute of Technology, Cambridge

MA, USA

Superconducting Wires and Cables:

Low-ﬁeld Applications

The discovery in 1986 of high-temperature supercon-

ductors (HTSs) with the prospect for operation at

liquid nitrogen temperature (77 K) or above provided

new impetus for superconducting electric power ap-

plications. HTS materials also exhibit higher heat

capacity than low-temperature superconductor (LTS)

materials that markedly improves their electrical sta-

bility, i.e., ‘‘normalization’’ of the superconducting

state, during thermal transients. The higher transition

temperatures also ease the refrigeration requirements

and facilitate cryogenic operation (Blaugher 1996).

1. Cables

Type II superconductors available in the shape of wires

and tapes appear as almost ‘‘natural’’ candidates for

such electrotechnical applications as power transmis-

sion cables. Traditionally, network engineers have re-

lied on increasing voltage to transmit a larger amount

of power. Superconducting cables have the capability

of carrying much higher currents, so providing the

opportunity for reducing the number of circuits, or

using lower voltage rating.

In the 1980s a significant research effort, including

design studies and experimental activity, was devoted

to the assessment of power cables based on low T

liquid-helium-cooled materials. Such research led to

the conclusion that, although technically attractive,

such cables were not economically viable, unless for

very high transmitted power, above 2–5 GVA. The

main cause for this was identiﬁed as the projected

cost of refrigeration, which resulted from the high

cost and low efﬁciency of liquid helium, combined

with the intrinsically unfavorable geometry (very high

surface to volume ratio) of cables. Since 1986 the

availability of the new class of high T

superconduct-

ing materials, potentially suitable for (far more ef-

fective) cooling with liquid nitrogen, has therefore

given a renewed impetus to the development of su-

perconducting power cables.

A range of applications is envisaged. A few instanc-

es worth mentioning are the upgrading of existing un-

derground circuits, the matching of the power rating

of an overhead line for sections to be relocated un-

derground, and the elimination of intermediate volt-

age transformations in urban distribution networks.

The expected beneﬁts include, for example, in-

creased transmission efﬁciency (energy saving, im-

proved stability of high-power circuits), drastic

reduction of electromagnetic pollution, an effective

solution to the expanded energy demand in congested

metropolitan areas, etc.

Development activities are underway in several

countries, and include both feasibility studies and

experimental work with prototype fabrication and

testing. In a number of such experiments, prototype

terminations and joints are also being developed. The

lengths of such prototypes range from a few meters to

more than a hundred meters (Norman et al. 1999,

Kelley et al. 1999).

The ﬁrst experimental installations of a three-phase

superconducting connection in a ‘‘real’’ environment

are being realized in the USA. Actually, one link

made with cold dielectric (CD) cables (see below)

fabricated by SouthWire has already been installed at

an industrial plant, and another link, made by Pirelli

based on a warm dielectric (WD) design (see below),

was planned to be installed in a part of a utility

(Detroit Edison) network by the end of 2001.

As type II superconductors show true zero resist-

ance only under d.c. conditions, d.c. cables would

appear as the ‘‘natural’’ way of best exploiting HTS

materials, by achieving a virtually zero dissipation in

the conductor, in addition to negligible dielectric

losses. In fact, most power lines transmit a.c. power,

and so a.c. designs are also typically assumed for su-

perconducting cables. These are in turn generally

classiﬁed, according to two main choices that can be

made regarding electrical insulation, into ‘‘warm’’ or

‘‘cold’’ dielectric types.

1195

Superconducting Wires and Cables: Low-ﬁeld Applica tions

In the WD cable design (Fig. 1), the HTS, made up

of one or more layers of current-carrying HTS tapes,

is enclosed in a cryogenic environment (the so-called

‘‘cryostat’’), which in turn is covered by conventional

dielectric materials.

This design has several advantages, such as:

 it has similar performance and requires similar

installation procedures as a conventional cable.

 it can carry, at the same level of losses, more

than twice the power of a conventional cable.

 it uses already developed dielectric materials.

 it needs accessories, both termination and joint,

that can be derived in a relatively easy way from

conventional ones.

 it requires fewer HTS tapes than the CD coaxial

design.

In the CD design, two coaxial HTS conductors,

separated by a dielectric at liquid nitrogen tempera-

ture, are enclosed in the cryostat (Fig. 2).

This approach, which is characterized by the pres-

ence of a superconducting ‘‘sheath’’ that acts as a re-

turn path for the current and cancels out the external

magnetic ﬁeld, represents the best technical solution

for very high-power transmission. It allows one to:

transmit a large amount of power with reduced

losses in ‘‘compact’’ cable systems, eliminating a.c.

losses due to external magnetic ﬁeld of the other

phases; and

eliminate eddy current losses in metallic com-

ponents, allowing installation close to other cables in

service.

However, it requires about twice the amount

of superconducting material compared to the WD

design, and requires a suitable cryogenic insu-

lation in addition to the development of specific

accessories.

Generally, the design of the conductor of a super-

conducting cable includes a multilayer stranded

Figure 1

Example of a three-phase WD cable in a pipe.

Figure 2

Example of a three-phase CD cable in a pipe.

1196

Superconducting Wires and Cables: Low-ﬁeld Applications

structure around a suitable former of adequate in-

ternal dimension.

Optimization of the conductor involves several pa-

rameters such as the tape laying angle, the laying

tension, the use of binding and/or insulating materials

between superconducting layers, the number of tapes

and layers, and the former material and dimensions to

comply with both electrical and mechanical con-

straints during fabrication, handling, and service.

Cable manufacturing requires that the supercon-

ducting materials are available in the shape of long,

ﬂexible wires. Most of the cable prototypes fabricated

have been made with BSCCO 2223 tapes typically a

few millimeters wide, with a thickness of a few tenths

of a millimeter.

For the optimization of the overall electrical and

mechanical performance, superconducting tapes are

made with a ‘‘multiﬁlamentary’’ structure, with the

number of ﬁlaments typically ranging from a few tens

to over a hundred.

It is worth remarking that the long length required

for a wire to be used in cable technology also imposes

obvious constraints in terms of performance uni-

formity. This in turn requires a very good control of

the processing parameters (mechanical, ‘‘chemical,’’

and thermal) throughout the whole wire fabrication

process.

The development of ﬂexible tapes with biaxially

textured YBCO, being pursued in the USA, Europe,

and Japan with promising results, has not yet reached

a stage as to allow the manufacturing of long proto-

types. The possibility of replacing BSCCO sintered

tapes with YBCO thick ﬁlms deposited on a ﬂexible

substrate is, however, considered extremely interesting

for cable applications. In fact, provided that the per-

formances already demonstrated on shorter lengths

can be maintained to a significant extent over kilo-

meter-long tapes, the use of YBCO is expected to

provide an important improvement in terms of cost-

effectiveness, and will therefore possibly give further

impetus to the development of superconducting cables.

2. Fault Current Limiter

A fault current limiter (FCL) is, in principle, a device

that is capable of protecting a power grid from very

high surge currents that can arise when a fault (short

circuit) occurs anywhere in the grid. As such fault

currents are expected to increase with increasing

grid interconnection, which is presently a worldwide

trend, the need for an effective protection will become

more necessary. The problem has been solved par-

tially with the use of nonsuperconducting devices as

circuit breakers, fuses, high-impedance transformers,

etc, which in general have some drawbacks

such as increased operating losses or at least nonide-

al efﬁciency. In fact, there is no direct ‘‘equivalent’’

for a superconducting FCL among conventional,

nonsuperconducting power components.

A superconducting device can take advantage of

the unique property of superconducting materials to

undergo a rapid transition to normal state upon a

quench, and therefore can in principle offer ideal

performance in terms of fast response, fast recovery,

and zero impedance throughout normal operation.

As it did for cables, and substantially for the same

reasons (insufﬁcient cost-efﬁciency with low T

ma-

terials), the discovery of HTSs has given new impetus

to research activities on FCL development in Europe,

the USA, Israel, and Japan. Such activities include

design studies as well as demonstrations of small- to

full-scale prototypes (Leung 1997).

Various design approaches are considered, namely:

‘‘Resistive,’’ i.e., the superconducting device is

connected directly in series with the power line. The

sudden increase of current due to a fault would in-

duce the superconductor-to-normal transition of the

material, hence increasing the impedance of the line.

‘‘Inductive,’’ or ‘‘screened core,’’ when the super-

conductor is magnetically (and not galvanically) con-

nected to the line. Such a device is substantially based

on a transformer, which has a superconducting shield-

ing tube as the secondary. Under fault conditions the

HTS becomes unable to screen the magnetic ﬁeld gen-

erated by the primary coil and the ﬂux penetrates into

the iron core thus increasing suddenly the impedance

of the device. Different from the resistive type, the

HTS component will experience higher current but at

lower voltage and does not need current contacts.

‘‘Hybrid,’’ which is also based on a transformer

conﬁguration that includes a resistive superconduct-

ing element combined with resistive or inductive

shunts.

The different designs are typically based on differ-

ent superconducting materials. Being studied for

FCL applications are BSCCO 2212, BSCCO 2223,

and YBCO, in various shapes, namely wires, bulk,

and ﬁlms.

Long tapes with good uniformity are available in

the case of a composite tape with BSCCO 2223,

which would therefore be a candidate for resistive

FCLs. Unfortunately, the composite includes a high-

conductivity matrix, which poses severe limits on the

efﬁciency of a type of device based on such tapes.

Tapes made with YBCO ﬁlms deposited on a high-

resistivity metallic substrate are available only in

short lengths, but, assuming that kilometer-long

tapes can be produced, they could represent a good

candidate for that purpose. An alternative approach

has already been demonstrated with prototypes of

resistive FCLs made with YBCO thick ﬁlms depos-

ited on nonmetallic substrates in a suitable pattern.

YBCO ﬁlms have also been used as the resistive

element in a prototype of a hybrid (with inductive

shunt) type of FCL. BSCCO 2212 is more suitable for

the manufacture of large pieces, and has been used in

prototypes of inductive FCLs in the shape of bulk

cylinders.

1197

Superconducting Wires and Cables: Low-ﬁeld Applica tions

In general, key material features of superconduct-

ing elements for both resistive- and hybrid-type FCLs

are essentially their critical current density and normal

state resistivity, which determine the required length

of the element and the amount of power that will be

generated during quench transition, and the sharpness

of the superconducting-to-normal transition, which

determines the time response of the device.

Other extremely important requirements concern

the stability of the superconducting element against

the creation of ‘‘hot spots’’ and subsequent thermal

instability. From this viewpoint, thin ﬁlms on non-

metallic substrates imply a different design approach

compared to tapes with a metallic matrix or substrate.

As a general consideration, however, it must be

noted that there will never be a universal supercon-

ducting FCL, since depending on its location inside

the network the requirements in terms of recovery,

limitation current, and protection coordination are

very different. The requirements for the material are

also very different. There will therefore probably be

opportunities for different materials to be imple-

mented in different types of FCLs in the future.

See also: Superconducting Wires and Cables: High-

ﬁeld Applications; Superconducting Wires and

Cables: Materials and Processing

Bibliography

Blaugher R D 1996 Superconducting electric power applica-

tions. Adv. Cryogenic Eng. 42, 883–98

Kelley N, Cinquemani P, Belcher P, Wen Y, Nassi M, Corsaro

P, Ladie

P, Norman S R, Rahman M M 1999 Applications of

HTS cables to power transmission: state of the art and op-

portunities. In: Proc. IEEE Transmission and Distribution

Conf. IEEEE, Piscataway, NJ, Vol. 1, pp. 13–19

Leung E 1997 Surge protection for power grids. IEEE Spectrum

July, 26–30

Norman S R, Ladie

P, Nassi M, Schroth R 1999 Latest devel-

opment of high temperature superconducting cable systems.

In: Proc. Jicable ’99. Jicable, Versailles, France, paper A3.5,

pp. 97–102

L. Gherardi

Pirelli Cavi e Sistemi Spa, Milan, Italy

Superconducting Wires and Cables:

Materials and Processing

1. Introduction

The phenomenon of superconductivity was ﬁrst ob-

served in 1912 in the laboratory of Heike Kamerlingh

Onnes, at the University of Leiden (Holland). At low

temperatures certain materials (B 800 superconducting

compounds have been recently tabulated by Poole and

Farach 2000) suddenly lose their resistance to the ﬂow

of electricity. The phenomenon remained a laboratory

curiosity until 1954 when G. B. Yntema at the Uni-

versity of Illinois made the ﬁrst successful supercon-

ducting magnet. At that point, magnet performance

fell well below that predicted by the properties of the

superconductor. The slow progress can be partially at-

tributed to (i) the difﬁculty in obtaining the extremely

low temperatures required and (ii) the lack of under-

standing of how to create electrically and thermally

stable wires from superconductors. The primary lim-

itation, however, was that the superconducting phe-

nomenon was not only limited to low temperatures but

also to a restricted range of electrical current density

and magnetic ﬁeld and that the range of these prop-

erties was particularly limited in early superconductors.

The maximum values of temperature, electrical cur-

rent, and magnetic ﬁeld are interdependent and, when

plotted in three axes, form a ‘‘critical surface’’ (Fig. 1).

Those early superconductors, such as Pb, In, and Hg

are classiﬁed as type I superconductors. These are su-

perconductors in which magnetic ﬂux is excluded from

their bulk and the critical current density is limited to a

surface layer of approximately one-tenth of a microm-

eter. The maximum ﬁelds in which these superconduc-

tors can operate are usually less that 0.1 T (similar to

the ﬂux density between the poles of a horseshoe per-

manent magnet). These limitations make type-I super-

conductors impractical for wire and cable applications.

All ‘‘technical superconductors,’’ i.e., superconductors

that can be readily fabricated into wires and cables for

high-current applications, are type-II superconductors.

In type-II superconductors, magnetic ﬂux penetrates

the bulk of the superconductor to form individual ﬂux

Figure 1

Schematic illustration of the critical surface for a type-II

superconductor. The applied ﬁeld (B), the temperature

(T), and the current density (J) must be maintained

below the critical surface in order to retain

superconductivity.

1198

Superconducting Wires and Cables: Materials and Processing

quanta (see Electrodynamics of Superconductors: Flux

Properties). In the most widely used technical super-

conductors, Nb–Ti and Nb

Sn, the upper critical ﬁelds

are 13 T and 27 T, respectively, and current densities

410

2

can be carried (compared to B10

2

for domestic Cu wire). In Fig. 2, the critical current

densities at 4.2 K are compared for strands or tapes

that are fabricated in lengths of more than 100 m. The

temperature range is still limited (the critical temper-

atures for Nb–Ti and Nb

Sn are 10 K and 18 K, re-

spectively), but for many applications the cost of

insulation and refrigeration is more than offset by the

reduced energy costs resulting from nonresistive cur-

rent ﬂow. Because the amount of energy lost in current

ﬂow through a superconducting magnet is extremely

low, it can be operated without a power supply once

the magnet has been charged. This mode of operation,

termed ‘‘persistent’’ is very desirable for MRI and

NMR applications because the current ﬂow (and con-

sequently the ﬁeld) is very stable. Liquid He is now

readily available and provides cooling to 4.2 K at at-

mospheric pressure and as low as 1.8 K at reduced

pressure. Recent advances in insulation and refrigera-

tion technology have also made the application of su-

perconductors more widespread. The discovery of

high-temperature superconductiv ity (HTS) supercon-

ductors (ceramic superconductors with critical temper-

atures above the boiling point of liquid nitrogen (77 K)

in 1986 (see High-temperature Superconductors: Thin

Films and Multilayers)) dramatically increased interest

in the application of superconductivity. The discovery

of superconductivity at 39 K in MgB

in 2001 further

extended the temperature range for intermetallic su-

perconductors.

Superconductors must be formed into composite

wires in order for them to be used in high-current

applications and these wires may also be cabled. The

wide variety of mechanical properties of supercon-

ductors: the range from strong ductile alloys like Nb–

Ti to brittle ceramics such as HTS superconductors

means that the wire processing routes vary greatly. Of

the HTS superconductors only one, Bi-2212, is man-

ufactured in round-wire form and HTS supercon-

ductors are more commonly manufactured as

multiﬁlamentary tapes. The brittle superconductors

cannot undergo wire drawing and are consequently

mechanically processed in the form of ductile pre-

cursors or ﬁne powders that are reacted to form the

ﬁnal superconductor. For Nb–Ti, the alloy can be

cold-worked to a remarkable extent but heat treat-

ments are also applied during processing in order to

introduce a second phase into the microstructure that

will ultimately pin the ﬂuxoids in the ﬁnal wire. Be-

cause of these variations, the types of processing will

be introduced separately and by superconducting

material. There are, however, key components that

are common to all superconducting strands.

2. General Strand Requirements

2.1 Matrix and Filaments

Superconducting wires are almost always of a multi-

ﬁlamentary design in which individual continuous

ﬁlaments (o30 mm in diameter) are embedded in a

high-conductivity matrix. When conditions in a lo-

calized region of a superconductor exceed the critical

surface conditions, that area will start to conduct re-

sistively and produce local heating. The resulting heat

and current transfer will start a cascade effect and

result in a rapid transition of the whole superconduc-

tor to the normal state. The ability to transfer heat

and current away from localized regions by sub-di-

viding the ﬁlaments into a high-conductivity matrix

was a key step towards producing practical supercon-

ductors. The primary components of cryogenic stab-

ilization were not formulated until the mid-1960s

(Stekley and Zar 1965). The most desirable matrices

are Cu or Al, but Ag is required for the processing

of HTS superconductors where oxygen diffusion

through the matrix is required during heat treatment.

External stabilizer is often added after heat treatment;

this is particularly the case for Al, which has thermal

and density advantages over Cu, but is difﬁcult to

Figure 2

Comparison of critical current densities (at 4.2 K unless

otherwise stated) available in strands and tapes of more

than 100 m.

1199

Superconducting Wires and Cables: Materials and Processing