Escolar Documentos
Profissional Documentos
Cultura Documentos
A.M. Wolsky, E.J. Daniels, R.F. Giese, J.B.L. Harkness, L.R. Johnson, D.M. Rote, S.A. Zwick
Argonne National Laboratory Illinois with Argonne,
in collaboration
R.A. Thomas,
Brookhaven Upton,
E.B. Forsyth
Laboratory
National
New York
J.D. Rogers
Los Alamos National Laboratory Los Alamos, New Mexico
J.B. Kirtley
Massachusetts Institute of Technology Cambridge, Massachusetts
B.W. McConnell
Oak Ridge National Laboratory Oak Ridge, Tennessee
J.G. DeSteese, J.A. Dirks, M. K. Drost, S.B. Merrick, R.M. Smith, T.A. Williams
Pacific Northwest Richland, Laboratory Washington
T.A.
University
Lipo
of Wisconsin
Washington,
NOYES
DATA
CORPORATION
New Jersey, U.S.A.
Park Ridge,
Copyright 01989 by Noyes Data Corporation Library of Congress Catalog Card Number: 88-38251 ISBN: O-8155-1 191-4 Printed in the United States Published in the United States of America Noyes Data Corporation Mill Road, Park Ridge, New Jersey 07656 10987654321 by
Data
Applied superconductivity / by A.M. Wolsky . . . [et al.1 in collaboration with R.A. Thomas. . . [et al.1 ; Department of Defense. cm. P. Bibliography: p. Includes index. ISBN 0-8155-l 191-4 : 1. Superconductors. 2. Superconductors--Industrial applications. I. Wolsky, A.M. II. United States. Dept. of Defense. TK7872.S8A67 1989 537.623--dcl9 88-38251 CIP
Acknowledgments
The thank
(ANL)
staff
who
organized
this
information Los Alamos (MIT), W. Laboratory Kenneth Conservation saw the need His John (ORNL), (BNL) final G.
National and
(BNL), Northwest
Laboratory
of Technology
Director
Energy
Storage of that
and Renewable a preliminary (PNL), leadership DeSteese John provided comments draft The nett, pletion
promoted
cooperation Richard A.
possible
L. Kirtley,
D. Rogers important of
considerations. contributed
The thoughtful
four
peer
reviewers,
very
B. Poeppel Lou
Donald were
W. Capone
for
valuable C. BencomKaatz
discussions
technical document,
J. Schmidt,
Bluth
vii
NOTICE
The ment agency implied, curacy, product, infringe materials by of in this book the were prepared States as accounts and the U.S. Government of work Departnor any express or for the acapparatus, not any specific or imply States state or
sponsored
Argonne
National
Laboratory United
Defense.
Neither
thereof,
legal liability
or represents Reference
commercial manufacturer,
recommendation, expressed
or the Publisher.
States Government
or the Publisher. Final determination of the suitability of any information for use purbe exbe
by any user, and the manner The reader is warned materials such as found that
caution
hazardous,
industry;
VIII
Foreword
impact
of
recent indicate
research
is discussed and,
in this
and desired
as actual appear
achievements, realistic
diverse
commercial
applications only
of the new materials. circumstances, temperature) which are unusual expense differ and
circumstances
have been expensive although medical rial become further lower some magnetic
practical The
resonance recent
sudden there
superconducting
at temperatures Now,
77 K raises the
is at hand. of applications
of the technology
by utilities
and industry. The book is organized and applied such research as to make The which Electricity its contents overview follow accessible presents and the the to readers principal with chaltopic:
varying benefits a
l l l
interests available.
lenges facing
on superconductivity
potential
economic
The eight
Sources
Generation
AC Transmission Superconducting Motors Industrial Magnetic Separations Levitation and Material for Transportation
V
0 0 0
l
Magnetic
Energy
Storage
Handling
vi
Foreword
covers
military program.
applications Both
of superconDepart-
and development
in a menu
for a 5-year
of Defense
and development
are considered. in the book in Applied and Impacts, L.R. is from the following documents: A Preliminary E.J. Rote, Daniels, and with S.A. R.A. J.D. Ridge S.B. 1988. and Developfor Exploitaissued by the Evaluation R.F. Giese, of Thomas Rogers National Merrick, Zwick
Laboratory; J.A.
of Massachusetts
Smith,
Williams
Pacific
Laboratory;
prepared
Department
of Energy,
Defense Options:
Superconductivity of Defense,
of contents
is organized
index
and provides
contained
Advanced composition and production methods developed by Noyes Data Corporation are employed to bring this durably bound book to you in a minimum of time. Special techniques are used to close the gap between manuscript and completed book. In order to keep the price of the book to a reasonable level, it has been partially reproduced by photo-offset directly from the original reports and the cost saving passed on to the reader. Due to this method of publishing, certain portions of the book may be less legible than desired.
and Scope.
.I
.2 .2
2. OVERVIEW...........................................3 A.M. Wolsky, E. J. Daniels, SOURCES and T.A. Impacts and R. F. Giese ELECTRICITY GENERATION
FOR Williams
......
.I2
S. B. Merrick, 3 14
Technologies
...
Background
Introduction.
.I4
14
.I5
5 15
System
impact.
...................... ............................
Storage Capability. Potential. ............... ................. Energy Technologies. ... Integration.
.I5
16 16 16
....................
Energy
................................
on Renewable
.I6
17 17 18 .............. 19
...........................
........................ Concepts. Receiver
Ponds.
Contents
and Subject
index
Wind
Energy
............
Conversion
......
..........
Conversion.
Magnetohydrodynamic Generation
Energy
............
.......................
. . . . . . . . . . .. . . . . . . . . . ... . . . . .. . . . . . . . . . . . . .. . . . . . . . . .
25 26 28 29 30 31 .33 34 37 37 38 39 40 41 48 48
E .J. Daniels
........................... and J. L. Kktley, Jr. Summary. .......................... Impact of HTSCs on Generators. Introduction. ......................
Superconductors Other Applications Case Study: Development Conclusions Reference Applied 300-MVA Effortsand
...................
....................... ........................
.................................
49
50 Transformers Transformers
.....................................
Application of HTSCs to Power
....... .....
51 51 52 52 54 56 56 57 .58 .59
Introduction. Application Method Results.. Transformer Conclusions References. 6. AC TRANSMISSION. R. F. Giese, R.A. Summary. Preliminary
.................................
of Superconductors of Analysis. ................................. to Power
..........................
...................................... Economic
........................ ..................................
Analysis the Power for Transmission System
.61
.61 2
Method.......................................6 About
........
.62
.63
Assumptions Costs
...........................
Losses and Refrigeration.
Cost of Energy
...........
................................
Properties Losses. Losses. and Current-Dependent Losses
.63 .64
4
Losses........................................6 Superconductor Voltage-Dependent Cryogenic Refrigerator Total Capital Losses. Enclosure Efficiency.
....................... ......................
.........................
.67 .67
Contents
and Subject
Index
xi
..................................
......... ....... Assumptions Regarding Properties of Cable Materials. l,OOO-MVA Transmission Systems .................... Conclusions ................................... Comparison of Electrical Losses and Costs .............
Cost of the Aerial/Underground Comparison Future of High-T, Superconducting ....................... Cable System with NbaSn Cable System Systems Studies
......................... ..................... Enclosures and Optimization. References. ................................... Supplement: Levelized Annual Cost Method ............... Introduction. ..................................
.81 .81 .82 .83 .83 .... .84 230-kV Superconducting AC Power Transmission System. 500-kV HPOPT Cellulose-Insulated Naturally Cooled System. .. .85 .86 500-kV Aerial/Underground System ................... .87 Conclusions, ..................................
MAGNETIC ENERGY STORAGE.
7. SUPERCONDUCTING R. F. Giese
.........
.88 9
and J. D. Rogers
Load-Leveling Superconducting Magnetic .90 .90 .90 .91 .92 6 and T.A. Lipo 7 ................ .98 .98 .99 102 103 Super105 105 Motors
Summary........................................8 HTSCs in Diurnal Energy Storage Introduction. Discussion Conclusions References. .................................. .................................. .................................... ................................... ...................................
8.MOTORS............................................g E .J. Daniels, Potential B. In/: McConnell Application Summary........................................9 of HTSCs to Motors. ........................... Introduction. Applications Conclusions References. Supplement: conducting Motivation Application HTSC HTSC HTSC Introduction. .................................. to Motors.
................................... ...................................
The Potential for High-Temperature AC and DC Motors. for Development Considerations
....................... ..................................
of HTSC for HTSC Electric Machines.
....... ...........
DC Motors Induction
HTSC Synchronous
Induction/Synchronous
.................
xii
Contents
and Subject
Index
HTSC HTSC
Reluctance Homopolar
Motor. Inductor
...................................
for Supplement, ........................ MATERIAL S.A. Zwick,
SEPARATIONSAND 6. W. McConnell,
......
118
J. 6. L. Harkness, 119 121 121 121 126 127 ....... 129 129 130 13 1 132 ........ 133 133 133 135 135 Separation 138 138 139 140 141 142 143 144
...................... Introduction. .................................. Materials Separation. ............................. ................... Materials Handling and Fabrication. References. ...................................
Application of HTSCs to Magnetic Separations. Introduction. Discussion ..................................
Potential
.................................. ................
Systems ............................ ................................... for High-Gradient Magnetic
Magnetic
............................ Diffusion Relations. ............................. Limiting Magnetic Effects. ......................... Estimation of Diffusion Rates. ...................... References. ...................................
Valuesfor02 10. MAGNETIC Larry LEVITATION FOR TRANSPORTATION
............
146 147
R. Johnson
......................................
of HTSCs to Magnetically .................................. Transportation Trains. Technology of Magnetically of High-T, Technology Needs Options. ..................... Levitated Vehicles. .............. ........................... Ground Levitated Trains ........
Background. Conventional
.......
151 152
Superconductors
.......................
Technology Development
of Magnetic-Levitation
.......................... .......
155
Contents
and Subject
Index
xiii
Bibliography.
. ..
. .
. .
. . .. . .
. ..
. . . . . . . . . . . . 156
7
APPENDICES..........................................15 . J. Daniels, Appendix R. F. Giese and A.M. A: Economic Wolsky Assumptions. for Preliminary
...................
Economic Evaluation
158 158
Summary.
....................................
for Superconductivity. ................
158 158 162 165 5 for 165 165 166 167 ...... 169 170 175 175 176 177 177 177 178 178 178 of High179 ..I7 9 179 180 184 185 191 191 193 193 193 194
Introduction.
....................................
of High-Temperature ........................
.............................. ........................
of High-Temperature
.............................
........................ ........................
of High-Temperature
DOD Superconductivity
......
.....
Processing Accomplishments
for Program Scope of DSRD ................... ...................... Program Work Statements. of and Search for High Temperature Materials. ........................ T,. ...................... Depth, h. ................ Temperature, Field Penetration
Energy Gap, 2A
..............................
(JJ) Tunneling with and Weak-Link Electromagnetic Fields. ..
194
xiv
Contents
and Subject
Index
Magnetic
Approaches Vortices.
to Controlled
Inhomogeneities
.................................. Determination of the Magnetic-Field/Current-Density/ Temperature Mitigation Mechanical Thermal Atomic Effects Critical Surface. Flux .................... Creep, Aspects and Jumps. ............ ..... of Magnetic Flow, Effects. ........................ ......................... ..................... Ginzburg-LandauTheory ........ Macroscopic with Structure Microscopic and Other
................
Electromigration
Abrikosov-Gorkov Electronic-Energy-Band
Theories.
.....
Normal-State
.............
206
Deposition.
............................. ........................
of Films. ............... Processing.
.206
208 209 210 217 220 222 227 231
................. Bulk Superconductors .......................... Single Crystals ............................... ............ Small Scale Applications and Demonstrations. Magnetometers and Gradiometers. .................. ........ Hybrid Semiconductor-Superconductor Systems. mm Wave Receivers. ..........................
Infrared Digital Digital Systems Large-Scale Shields Sources (Mid (Mid Sensors Systems Systems ..............................
.236
239 242 246 249 252
Three-Terminal Refrigeration
................... ............
Wave
...............................
and Demonstrations.
.254
258 262 265 Systems 267 (SMES) 270
...........................
and Millimeter
..........................
Ship Propulsion Energy Storage
...........................
................................
Contents
and Subject
Index
xv
Term).
.......... ......
. 273
Term)
.................
. 276 . 278
280
Term) ......... .............. ELF Communication (Far Term) ............... ........................ Other Applications .................. DSRD Budget Recommendations. Sweeping Supermagnets Pulsed Power Systems (Far Term). ADDENDUM II: MILITARY Summary. SYSTEM APPLICATIONS.
......
..............................
Superconductors
....... ...... High Temperature Superconductors (HTS). .............. Status of Supporting Technologies ....................... Cryogenic Cooling. .................... High Strength Materials .......... Military Applications of Superconductors ............................... Introduction. ............................ Electronics. Overview. ........................... IR Sensors. ........................... Microwave and MMW Sensors ............... ..................... DC to UHF Sensors. Magnetic Sensors. ....................... Signal Processing. ....................... ...................... AID Converters .............. Delay Line Signal Processor .......... Digital Signal and Data Processing ....................... High Power Applications. ...................... Magnets-Applications ....................... Electrical Machinery Launchers ..............................
Low Temperature U.S. and Foreign Superconductivity Conclusions Appendix Terms Research Expenditures in High Temperature
,300 ,302 ,304 ,304 .304 .306 .306 .306 .306 .306 ,307 ,310 ,310
.310
........................... ................................
to the DSB Task Force and Development on Military of Superconductors
..................
Into
of Research
Superconductors.
1. Introduction.
................... ............................
.332 .332
xvi
Contents
2. General
Issues.
.........................
for Electronics.
. . . 332
333
. .
............................. ..................... Ground-Based Systems. Large Systems. ......................... Small Systems. ......................... .................... Space System Cryogenics ....................... High Strength Materials ................ A Josephson 4-Bit Microprocessor
Back-Up R&D Data on Japanese ................................... Funding for Superconductivity
Government. Corporate
. . . 356
356
....
..........................
. . . 361 362 .
1 Introduction
Applied
Superconductivity
PURPOSE
MD
SCOPE
This document is meant to aid the U.S. Department of Energy (DOE), Assistant Secretary for Conservation and Renewable Energy, Office of Energy Storage and Distribution, by discussing the likely impacts of recent results from research on superconducting materials. Substantial discussion is also given to the impacts of hoped for, but not yet achieved, advances. These discussions indicate research goals that appear realistic and, if reached, would enable commercial application of the new materials. The discussions that follow concern terrestrial applications that would substantially affect the production and use of electricity. Such applications occur on the utility side of the meter (e.g., transmission of electricity) and on the customer side (e.g., magnetically levitated trains). The prospects for such applications cannot now be described precisely. The engineering properties of the recently discovered superconductors have not yet been explored, and forecasts of energy prices and supplies are difficult at best. However, promising applications and needed property improvements can be discerned, and these are emphasized. One area of application, utility system process monitoring and control, was not discussed but does deserve future consideration. Two other areas of application, digital computation and measurement of very weak magnetic fields, are not discussed because their direct impact on energy production and use appears negligible. Nonetheless, products made for these applications may have significant impact on the economy (e.g., Moreover, the measurement of weak fields may enhance geologic exploration). manufacture of these products will increase the number of persons familiar with superconductivity, thus increasing the likelihood that superconducting solutions will be found to problems that now appear remote from superconductivity.
ORGANIZATION
OF THIS REPORT
This report is organized to make its contents accessible to various readers, each The Overview (Sec. 2) introduces the with his or her own interest and background. principal challenges facing applied research on superconductivity and the economic benefits that may result from success. The sections that follow the Overview, prepared by different teams of experts, each address a particular topic. Because these sections vary in length and technical detail, a summary of each has been prepared to serve as an introduction. Finally, the base economic assumptions used by the authors and information about the properties of one bulk sample of YBa2Cu307_x (one of the new superconductors) are presented in Apps. A and B, respectively.
2 Overview
A.M. Wolsky, E.J. Daniels, and RF. Giese Argonne National Laboratory
Applied Superconductivity
The recent and sudden discovery of a family of materials that become superconducting at temperatures above 77 K raises the likelihood that further advances are at hand and that these advances will lead to commercial applications that conserve energy. Materials in their superconducting state offer a means to circulate direct electric currents (DC) with no resistive loss. Materials in their superconducting state also offer a means to convey low-frequency alternating currents (i.e., AC at 60 Hz) with unusually small losses. The absence or significant reduction of losses prompts universal interest in superconductors as energy savers. Materials become superconducting only in certain circumstances, which differ for each material. These circumstances (e.g., low temperature) are unusual and have been expensive to arrange and maintain. In the past, that expense has been too great to permit widespread commercial applications of superconductivity, although commercial applications have been made in high-energy physics, medical magnetic resonance imaging (MRI), and -- most recently -- industrial materials separation. Now, there is hope for further advances that will lower the cost of applications and enable adoption of the technology by utilities and industry. The most well-known characteristic affecting superconductivity is the temperature of the material. Niobium-tin, Nb$n, becomes superconducting when its temperature is less than 16.05 K; the corresponding transition temperature for niobiumtitanium, NbTi, is 9.8 K. (On this scale, the Kelvin scale, room temperature is generally considered as 298 K.) The total cost of refrigeration to cool these materials to 1.8-4 K and maintain their operating temperatures is formidable. This cost includes capital and Capital is required to purchase thermal insulation, which slows operating components. the rate at which ambient heat reaches the superconductor and, in some cases, to purchase equipment to refrigerate the coolant. Operating costs pay for the coolant (i.e., helium) makeup and, in some cases, for the energy required to remove the heat that penetrates the thermal insulation. As noted above, the new materials (e.g., YBa2Cu307_x) become superconducting at temperatures in the range 77-100 K. This range of temperatures is above a significant threshold -- it provides the opportunity to use liquid nitrogen instead of liquid helium as Furthermore, operating in this temperature regime would the superconductor coolant. reduce the total cost of refrigeration for two reasons: (1) for the same insulation, the rate of heat transfer from ambient temperature to cold superconductor declines as the cold temperature increases (alternatively, the same heat-transfer rate may be obtained with less costly thermal insulation) and (2) the cost of removing the heat that penetrates the thermal insulation declines as the cold temperature increases. The cost savings for heat removal depend on the type of refrigeration and insulation system used in a particular superconductor application. Under idealized conditions, the energy required to remove one unit of heat at 77 K is less than 5% of the energy required to remove the same amount at 4 K, and the amount of heat that can be removed by the vaporization of 1 L of liquid nitrogen is 60 times that of 1 L of liquid helium. Because the cost of liquid nitrogen (per liter) is less than 10% of the cost of a significant potential for cost reduction in the liquid helium, this represents As a practical example, a typical MRI solenoid, refrigeration of superconductors.
Overview
maintained below 4.2 K, provides a magnetic field of 1.5-2.0 T in a l-m bore. The capital cost of the thermal insulation (also known as a cryostat) is about $100,000, and the annual cost of liquid helium makeup is about $30,000. Were the solenoid maintained at 77 K, the capital cost of the needed thermal insulation would be $50,000, and the annual cost of liquid nitrogen makeup would be $3,000 - a very substantial reduction in the total cost of refrigeration. The second circumstance affecting superconductivity is the strength of the magnetic field around the material. if this field strength is too great, superconductivity cannot be achieved. The new materials are expected to maintain superconductivity at field strengths greater than those that would prevent superconductivity in the This property could enable the production of lighter-weight commercial materials. magnets with strong fields induced by currents circulating within the superconductor itself. Present practice is to insert iron, with a density of 7.9 g/cm, within the core of an electromagnet, where the field it contributes is at most 2.2 T. However, magnetic field strength is also limited by the ability to accept the mechanical stress that the magnetic field exerts on the currents that produce it. For example, the outward stress or pressure on the interior walls of a long, air-filled solenoid producing the magnetic field B is given by 3.9 atm x (B/l T)2 - thus, a 5-T field exerts a stress of 97.5 atm -and the concomitant tension (tangent to the solenoids wall and perpendicular to its radius) is given by the product of that pressure and the solenoids radius. The third circumstance affecting superconductivity is the electric current density, usually described in amperes per square centimeter (A/cm2), within the The maximum or critical current density depends on the material, its material. temperature, and the magnetic field around it. Although the popular press has given much more attention to the critical temperature than to the critical current density, the latter is now equally important, or more so, for the following reasons: 1. Weight. The weight of material (the density of YBa2Cu307_x is about 6.3 g/cm3) required to convey a given total current for a given distance is inversely proportional to the current density within the material. Reduced weight means reduced cost for supporting structures or increased payload for levitation (cranes or trains). This is a reason for avoiding the use of iron. The volume of material required to convey a given total Size. current for a given distance is inversely proportional to the current density within the material. Reduced size means increased opportunity to replace equipment for which floor space has already been allotted. Flexibility. Over equal lengths, material with a large cross section is less flexible than material with a small cross section, and thus less easily wound in the form of wire or tape. The needed cross section is inversely proportional to the current density within the superconducting filaments embodied in the wire. Increased flexibility means increased ease of handling and increased reliability in the face of mechanical perturbations.
2.
3.
Applied Superconductivity
4.
Cost of Raw Materials. The cost of raw materials is likely to be proportional to the weight or volume of the final product superconductor, which (as noted above) is inversely proportional to the current density within it.
In addition to current density, three other classes of engineering properties deserve attention. The first is the ability of new superconductors to join with or be coated by other materials. Present practice often requires that superconductors form composites with other materials. For example, the tape used in Brookhaven National Laboratorys transmission line is a sandwich of stainless steel (for strength), Nb3Sn, and copper (to shunt current during a fault). The usefulness of new superconductors will almost certainly increase when they, too, can be part of such composites. The second class of properties involves chemical stability. The new materials show a propensity to lose oxygen and, with it, their superconducting properties. It may be important to know if the composites required for electrical systems also act to preserve the chemical stability of the superconductor. The third class of properties affects the AC losses in the new superconductors. As already noted, superconductors circulate only direct currents without loss. However, many applications in the electric power system require superconductors to experience time-dependent magnetic fields, or AC currents. Hysteresis loss deserves attention, as does the effect on losses of the condition of the superconductors surface. When superconductors with favorable properties ;Ipe fabricated, they are likely to find profitable applications on both sides of the meter. Below, we describe our essential findings, including the essential findings of the topical sections that follow. Some of these findings are also presented in Table 2.1. 1. The critical current densities that have been observed in bulk samples of the new superconductors are too small to permit their terrestrial commercial application. Research should be devoted to increasing these critical current densities. Because the chemical stability of the new superconductors in the presence of oxygen (e.g., air) and water is unknown, their potential for terrestrial application cannot be evaluated without speculation. Research should be devoted to measuring the chemical stability of these materials and, if needed, to developing suitable protective coatings. These coatings might also serve to add mechanical strength (e.g., stainless steel) or provide a heat sink and electrical shunt (e.g., copper and aluminum). The AC properties of the new superconductors are unknown. Thus, their potential for application in generators, transformers, AC power lines, and motors cannot be evaluated without Research should be devoted to measuring the AC speculation. properties of new superconductors.
2.
3.
Overview
TABLE 2.1 Design Goala and Economic Benefits for Selected &@icatiOns
Application
Life-Cycle Dollar Savings of High-T System (%I Design Operating Field (T) Compared with Liquid Helium Systema Compared with Conventional Systemb
300 MU
3d 10
2 0.30g
27e 36
63e 60
Transmission lines, 113,000 HVA, 230 kV SMES systems, 5,000 MWh Motors
Haglev systems
23h
no.1
23
43i
60i
1.6-5
5-a
Note
0.1-0.251 1 3p
2-3 3 2-5
llrn NA 15
21m NA 20
Magnetic
separators
aSavings bSavings
T,)/LHe] - High
x 100. T,)/Conventional]
x 100.
Generators, which account for l-22 of the capital cost of conventional power plants, convert shaft power to electrical power. The rest of the plant produces shaft power and is unaffected by superconductivity. Superconductivity may substantially affect future power plants using MHD or fusion. dDesired bulk critical current density = 4.5 x IO4 A/cm2; operating current density in wire (including copper cladding) = lo4 A/cm2. eBased on materials and operating costs, with refrigeration costs proportional to refrigeration power. fl MVA = 1 MU, if there is go.30 T maximum in the coil no phase windings difference and 1.75 between T in the current and voltage. transformer core.
hBulk critical current density = 230 x lo4 A/cm2; bulk operating current den,sity = 23 x lo4 A/cm2, or equivalent operating surface current = 500 A/cm. tconventional underground transmission. JDesired bulk critical current density = 70 x lo4 A/cm2. kDepends on utility characteristics (e.g., load shape and capacity mix). Based on copper windings with a iron core. mAssuming a 20% capital cost reduction for coolant refrigeration. Based on both U.S. and Japanese research during the 1970s. Not available. pBased on a small prototype.
Applied Superconductivity
4.
Although the foregoing research and the for below may increase the efficiency and transmission from all sources, the different. In particular, the choice selected conventional sources may be following examples:
specific advances called of electricity production impact on each may be between renewable and We offer the affected.
- Peaking power is now supplied by units fueled by natural gas. In the future, such units may compete with superconducting magnetic energy storage (SMES) for the peak market. Thus, SMES may provide the means for solar energy (e.g., wind power or photovoltaic cells) to displace natural gas. Solar energy will continue to compete with coal and nuclear fuel, burned in otherwise idle capacity (if any), for the SMES charging market. - Because of their ability to charge and discharge rapidly, SMES units may also play a role in absorbing transient and discharging level power. This conversion of transient to level output may ease the burden of incorporating generation into the grid. - The cost of electricity ocean thermal energy using superconducting, lines under water. 5. delivered to the shoreline conversion facilities might rather than conventional, small power input wind
If current densities of about IO4 A/cm2 can be achieved in wire (including copper cladding) at about 77 K and 2 T, and if the superconductor otherwise behaves as Nb3Sn or NbTi, then large (300-MWe) generators using the new superconductors will be more economical than either conventional generators or low-T generators. In particular, a high-T, 300-MWe generator mig& have an efficiency of 99.7% (compared with efficiencies of 99.5% for a low-T, generator and 98.6% for a conventional generator). Increased efficiency would reduce the quantity of air pollution from combustion or reduce the cost of air-pollution control. Engineering research and economic evaluation should be devoted to smaller generators (e.g., 60 MWe), for which there is now a greater demand than for 300-MWe generators. If current densities of about 10 x lo4 A/cm2 can be achieved at 77 K, and if the material otherwise behaves as Nb3Sn or NbTi, then the cost of service of a l,OOO-MWe, high-T, superconducting transformer would be 64% of the cost of service of a low-T, transformer and 40% of the cost of service of a conventional transformer. These cost comparisons reflect the higher efficiency of the high-T, transformer (99.92%) compared with the
6.
Overview
the
low-T,
(99.85%)
and
conventional
If current densities of about 100 x lo* A/cm2 can be achieved in wire at 77 K and 1.8-5.0 T, and if the material otherwise behaves as NbTi, then the capital cost of large (l,OOO-MWe, 5,000-MWh) SMES facilities might be reduced by 3-8%. The low end of this range accounts for savings in thermal insulation and refrigeration, whereas the high end includes savings from inexpensive (2.2-e/g) Under reasonable assumptions, these superconductor materials. savings might make SMES competitive with gas-fired peaking plants. Lower current densities (e.g., 60 x lo* A/cm21 might be sufficient to make SMES economical. Research should be devoted to determining the effect of increased specific heat, concomitant with the increase in operating temperature from 1.8 to 77 K, on SMES reliability. If operating current densities of about 23 x IO* A/cm2, with critical current densities of about 230 x lo* A/cm2, can be achieved in tapes at 77 K and less than 1 T, and if the material otherwise behaves as Nb3Sn, then the cost of service for a 66-mi, lO,OOO-MWe, AC superconducting transmission line appears to be roughly 60% of the cost of service of conventional underground, This cost advantage reflects lower oil-filled-pipe transmission. transmission loss (0.73%) in the superconducting line than in the conventional underground line (3.60%). Both lines are more expensive than a conventional aerial transmission line. However, concern about the health and environmental effects from aerial transmission and the ability to obtain aerial rights of way may result in future mandates to construct underground lines. Research and economic evaluation should be devoted to lowercapacity (e.g., 300-1,000 MWe) transmission lines, for which there is a greater demand than for lO,OOO-MWe lines. If current densities of about 0.1-0.25 x lo* A/cm2 could be achieved in wire at 77 K in the range of 2-3 T, and if the material otherwise behaves as Nb3Sn or NbTi, then a conservative estimate indicates that a large (e.g., 1,500-hp) high-T, superconducting motor, with an iron alloy core, might provide shaft power for 90% of the cost of service of a conventional motor. This saving reflects the assumed high efficiency (97%) of the high-T, superconducting AC motor and the lower efficiency (95%) of a conventional AC motor. If the capital cost of the system were reduced by about 20% by redesign of the refrigeration system, the high-T, superconducting motors cost of service would be about 80% that of a conventional motor.
8.
9.
10
Applied Superconductivity
10.
Most recently, low-T, superconductors have been commercially applied to high-gradient magnetic separation (HGMS) of magnetic contaminants in kaolin processing. Superconductors offer a number of advantages in industrial processing (e.g., reduced weight, increased throughput, and reduced floor space), in addition to their 80% reduction in power consumption (including refrigeration power) compared with conventional HGMS. The primary advantage of high-T, superconductors for industrial applications, compared with low-T, systems, would be a capital cost reduction of lo-1596 due to elimination of the helium refrigeration/reliquefaction system. Thus, compared with low-T, or conventional HGMS, the cost savings of a high-T, superconducting HGMS system would be about 15 or 2096, respectively. In addition to competing with conventional HGMS systems in industry, high-T, superconducting magnets may be applicable to other industrial processes, including (1) gas/gas separation, (2) materials handling, and (3) materials fabrication (e.g., press fitting of components). High-speed rail is being actively considered for at least a dozen corridors in the United States. Like other systems, magnetic levitation (maglev) is unlikely to be economical without indirect benefits being added. Advances in superconductivity are unlikely to change this situation, because present designs allocate only about 1% of the system capital cost to the levitating magnets on the train. However, if the new superconductors can operate at 77 K as well as NbTi operates at 4.2 K, these superconductors may offer an ease of operation and promise an increase in system reliability that will make high-T, maglev systems the preferred choice among high-speed rail technologies.
11.
Many of the superconductor applications discussed above and illustrated in Table 2.1 exhibit large economic savings, even for the liquid-helium-cooled versions. .Moreover, several have been developed through the prototype stage. Why have none of them been commercialized? First, most technologies employing superconductivity have large economies of scale that require large capital investments and the associated financial risks. Second, many of the technologies (generators, transformers, transmission lines, and SMES) are in the electric utility sector. This sector has curtailed investments in recent years because of (1) recent completion of a large capacity-expansion program, (2) slow growth in electricity demand, (3) an existing capacity that consists of equipment with long lifetimes, and (4) an uncertain regulatory environment. Also, the electric utility industry places a very high premium on system reliability. These factors have combined to delay the adoption of any new, superconducting technologies. The overall impact of successful development of the new Not only is there a superconductors cannot be gauged precisely. but both conventional and concerning the emerging technologies, However, if even superconductor technologies continue to improve. potential improvement in energy efficiency is realized, the associated high-temperature large uncertainty low-temperaturea fraction of the economic benefits
Overview
11
may be important. In 1983, about 7% of the electricity generated in the United States was lost before it reached the customers meters. Superconductivity may enable increased profitability and improved electrical system efficiency, with concomitant reductions in environmental impact. For example, a 3.6% loss is expected from a 66-mi conventional, underground AC transmission line, while the loss from the competing superconducting, underground AC transmission line is expected to be only 0.7%. Improvements are also likely to extend to the customers side of the meter. About 64% of the electricity sold is transformed to shaft power by motors with efficiencies that now range from 72% for small motors to 95% for large industrial motors (e.g., 1,500 hp or The efficiency of large motors might well be raised to 97%. Further, 1,119 kWe). materials-separation processing in industry is now energy-intensive (about 3 quads per year*). As familiarity with new superconductors increases, magnetic separation may replace present practice in several applications (e.g., cleaning boiler feedwater). At present, no one knows if or when needed advances will be made. Enthusiasm among researchers is very high, and progress is reported each week. If confirmed, a recent announcement that critical current densities of about lo3 A/cm2, at 1 T, have been observed in a bulk superconductor marks an important step toward commercially useful material.
12
Renewable
Generation
13
II
Summary
calls attention to the fact that more efficient electrical generation and will lower the cost of electricity derived from all forms of primary energy. advances In superconductivity wlll have a bC3iYYti irn~~& 'th bn Vsk tit
respect,
cwev+&2&&
However, attention is also called to the great importance of storage in conjunction with generation from wind and solar energy. Advances in superconductivity promise to lower the cost of superconducting magnetic energy storage (SMES) and make it the lowest-cost form of storage. Thus, the total cost of a system including wind energy and SMES, or solar energy and SMES, would be lowered. Such combined systems should be compared with combined systems using SMES and conventional sources of electricity (e.g., off-peak power from coal or nuclear units). The result of this comparison will depend on the load profile, existing stock of generation capacity, fuel prices, and environmental regulation. (Section 7 describes SMES.) Advances in superconductivity may also reduce the cost of electricity delivered to the shoreline from ocean thermal energy conversion systems, if underwater superconducting transmission lines are less costly than underwater conventional transmission lines. Section 6 compares the cost of superconducting and conventional transmission underground.
14
Applied Superconductivity
3.1
BACKGROUND
This section summarizes a preliminary Pacific Northwest Laboratory assessment of the potential for superconducting materials and devices to change or enhance the future value of renewable energy technologies. The approach taken was to identify the possible interfaces between these technologies and high-temperature (greater than 70 K) superconducting subsystems and project the nature of resulting changes in overall system performance. Four impact categories were considered: neutral impact, enhanced energy storage capability, improved system integration, and new energy conversion potential. The value of incorporating superconductors appears to range from a neutral impact for the renewable technologies that are operated as base-load systems to the facilitation of advanced energy conversion opportunities that are impractical with normal conductors. The smallest impacts are in technologies such as geothermal energy conversion, where superconductors add no value to the intrinsic power or availability of the resource. In such a system, superconductors might replace conventional electric-power generating, transmission, protection, and control components: this is also possible with other thermal or hydroelectric power systems. Superconductors do not appear capable of improving the energy conversion process of intermittent resources, such as solar energy systems, but benefits are likely to be realized from superconductor-enhanced energy storage, stand-alone capability, and/or utility system integration. The highest value expected from superconductor applications was found to be in technologies where a new intrinsic capability might be provided in the energy conversion An example is the possible facilitation of magnetohydrodynamic conversion process. from resources such as biomass, where previously the magnet power required with lowconductivity, low-temperature working fluids would have made these concepts Finally, assuming that fusion power will become possible and can be impractical. the plasma containment and energy conversion considered a renewable resource, processes would be impractical without the incorporation of superconducting subsystems.
3.2 INTRODUCTION Recent indications that a new class of metallic oxide superconductors exhibit superconductivity at liquid nitrogen temperatures and above have fueled speculation on their practical value. Participants in the April 1987 meeting on superconductors held by
15
the U.S. Department of Energy (DOE), Office of Energy Storage and Distribution, were assigned to assess the benefits and changes that may be sssociated with hightemperature superconductors in a number of possible applications. Pacific Northwest Laboratory (PNL) was assigned the task of assessing the potential for the new superconductors to change or enhance the future performance and value of renewable energy technologies. This section presents the results of the PNL effort in this activity up to the July 1, 1987, deadline for the delivery of preliminary results.
3.2.1 Approach The approach taken was to review systems from a top-down perspective to identify possible opportunities for inserting or substituting high-temperature (greater than 70-K) superconducting subsystems in the place of normally conducting components. The nature of the resulting changes in overall system performance and other significant characteristics was projected at least qualitatively for all cases considered, and quantitatively when appropriate data were at hand.
3.2.2 Scope For the purposes of this assessment, the term renewable technologies was taken to mean electric power production from inexhaustible energy resources. These include hydropower, solar, wind, geothermal, ocean-thermal, ocean-mechanical, and biomass resources. Fusion power is also classified as a renewable technology in this assessment, because the deuterium fuel can be derived from seawater, an inexhaustible resource.
3.2.3 System Impact Classification From the overall system-level perspective, superconductors may have an impact on an electric-power-producing technology in four principal ways, according to the nature of the resource. The four impact categories considered were:
l l l l
Neutral impact, Enhanced energy storage capability, Improved system integration, and New energy conversion potential.
Neutral Impact
Some systems, such as the larger-scale geothermal power plants, produce an intrinsic base-load, grid-synchronized AC output. In renewable resource systems of this type, superconductors may replace conventional electric-power generating, transmission, protection, and control components in the same manner as is possible with regular fossilor nuclear-fueled stations. The resource side of the system is constrained by the geographical location, thermal power, and diversity of the source, which neither gains
16
Applied Superconductivity
nor loses value from superconductors being incorporated in the balance of the system. In situations of this type, the value added by superconductors is identical to that achieved in systems powered by nonrenewable fuels. Inasmuch as superconductors make no difference to the performance of the energy conversion technology, impacts of this type are considered to be neutral.
Enhanced Energy Storage Capability Systems that convert solar energy to electric power typically produce an intermittent output, requiring energy storage if a constant output is desired. Energy storage in this type of renewable technology is a significant advantage that can increase system flexibility and, in some cases, reduce the delivered energy cost. Superconducting magnetic energy storage could enhance the storage potential of such systems and could, therefore, be a means of extending their overall performance and value.
Improved System Integration Intermittent power systems typically require energy storage and power conditioning to be compatible with end-use needs and/or utility integration. Superconducting generation, storage, and transmission can influence and enhance intermittent power technologies by introducing new system integration options. As an example, the ability of SMES to rapidly switch from charge to discharge or vice versa makes it attractive for use in controlling unstable systems. This may be needed at the grid interface to inhibit any plant output variations that might cause undesirable voltage fluctuations on the transmission system. The high round-trip cycle efficiency potential of superconducting systems could provide energy-management and cost advantages.
New Energy Conversion Potential The final category of impact is the potential for superconducting devices, such as high-field-strength magnets, to facilitate previously impractical energy conversion For example, low-temperature, low-conductivity, magnetohydrodynamic, and options. magnetofluid mechanical concepts would fall into this category.
3.2.4
Organization
Section 3.3 contains a portfolio of technology-specific reviews summarizing the potential influence of high-temperature superconductors on each of the renewable technologies considered. The technologies are reviewed in the following order:
l l l l l
Hydroelectric energy Solar salt gradient ponds Solar thermal central receiver concepts Solar thermal dish concept Solar photovoltaic cells
Renewable
Generation
17
l l l l l l
Geothermal energy conversion Wind energy conversion Ocean thermal energy conversion Biomass energy conversion Magnetohydrodynamic energy conversion Fusion power generation
The results of this assessment are summarized by the matrix in Sec. 3.4, which shows the overall impact of superconductors on the above systems according to impact category.
3.3
SUPERCONDUCTOR
IMPACTS ON RENEWABLE
ENERGY
TECHNOLOGIES
3.3.1
Hydroelectric
Energy
Technology
Description
Hydropower constitutes about 12% of the nations electric energy generation. Hydroelectric energy is converted from the fluid-mechanical energy of rivers, streams, and sometimes ocean water, by causing this water to flow through turbines located in a dam. The dam typically provides storage for large volumes of water, generally sufficient contain multiple generators and are for base-load or load-following use. Dams typically capable of delivering bulk power (loo-6,600 MW) to the transmission grid. A major exception to this generalization is the TVA system, which collects water from a number of sources through an elaborate system of penstocks. Smaller hydropower projects are often built without dammed storage on rivers, streams, and irrigation canals with variable water flow rates. Generators of this type generally have capacities of between 1 and 80 MW.
Current
System
Integration
Approach
Most hydroelectric energy is generated at the synchronous power frequency of the electric grid and is transformed to high voltages on site to supply bulk power transmission lines. Some hydropower projects also supply bulk power directly to large The smaller generators (more than 80 MW) are typically industrial users located nearby. base-loaded to the capacity provided by the stream flow, which may vary or even be interrupted during the year. Most of the generators in this class are connected to the utility grid and, in many cases, supply power that the utility is obliged to purchase under the requirements of the Public Utility Regulatory Policy Act (PURPA).
18
Applied Superconductivity
Performance
Characteristics
Hydropower is the most flexible resource and can be brought on line within minutes of the start-up command. The water storage provided by dams and the ability to respond to rapid fluctuations in demand allow hydroelectric systems to be dispatched as load-following units. Hydroelectric systems are typically the lowest-cost base-load or load-following capacity available.
Impact of Superconductors The use of superconductors in hydropower systems can be expected mainly to improve generator efficiency, reduce electric losses, and increase the capacity in the power buses connecting generator outputs to the primary side of the transmission transformer. In large systems, generator efficiency may be improved by 24%. The size, efficiency, cost, and routing of bus ducts may be similarly improved. When practical materials and components are developed, superconductors should find relatively early application in the larger hydroelectric plants. This is because the existing facilities and caliber of personnel should be well suited to accommodate the complexity of whatever cryogenic refrigeration system may be required. There is less application potential at the lower end of the power range, where operations and There is probably a negligible prospect for personnel are typically less sophisticated. Dammed systems have inherent storage, and SMES applications in any hydroplants. storage systems are generally not cost-effective to implement on small systems. If storage were desirable where it did not already exist, comparisons between options in the 5,000-7,000 MWh, 660-1,000 MW energy storage and power ranges show that underground pumped hydropower storage would be much cheaper ($740/kW) than a SMES system ($1,90O/kW). However, in a few particular cases, a SMES system could possibly be viable despite its cost, if environmental restrictions were to eliminate pumped hydropower or other cheaper storage options. The transmission systems that deliver hydropower to load centers are potentially amenable to the use of superconductors, based on the operational and economic criteria that would apply to conventional systems.
Technology Description The solar salt gradient pond is a device for trapping and storing solar energy. The salt gradient pond achieves this by means of a concentration gradient, where salinity increases from a low value at the pond surface to high salinity a meter or two below the surface. Hence, deeper waters are heavier than the water above them. This eliminates buoyancy-induced convection, impeding the upward movement of the warmed water. Buoyancy-induced convection is the major heat-loss mechanism in a solar pond.
Generation
19
In practice, the salt gradient pond has three layers: a thin surface convecting zone, the salt gradient zone, and the storage zone located under the salt gradient zone. Useful energy is absorbed in the storage zone, resulting in the storage zone having a temperature substantially above ambient. In operating ponds, temperatures above 100C By extracting storage zone water from the have been achieved in the storage zone. pond, a heat source is made available for a low-temperature organic Rankine-cycle heat engine, which can be used to drive an electric alternator. Studies from the early 1980s project a levelized energy cost of about lOQ/kWh for large solar salt ponds.
Current System Integration Approach A solar salt gradient pond system is normally designed to deliver synchronized AC power directly to the grid without on-site electric energy storage.
Performance
Characteristics
The solar salt gradient pond concept is a low-temperature solar energy conversion concept that results in low efficiency. In many cases, however, the pond can be constructed at a very low cost, offsetting the efficiency penalty. The major feature of this concept, compared with other solar thermal concepts, is that the storage zone A solar salt gradient pond may take provides very large amounts of thermal storage. several years to warm up, but once it is warm, the plants output will not be affected by diurnal or short-term weather-induced variations in incident solar radiation. There will be a significant seasonal variation in output, but overall, a solar salt gradient pond power plant can be considered as a base-load power generator. For economical operation, solar power plants must be located in areas with high incident solar radiation, such as the southwestern United States.
Impact of Superconductors The impact of superconductors would be essentially neutral in this technology. Application potential would exist in the generator and transmission system according to the same criteria that apply to these components in nonrenewable base-load systems. Storage is an intrinsic feature of solar salt gradient ponds. Therefore, superconducting storage would add little value.
Concepts
Technology Description The central receiver concept calls for a field of mirrors or heliostats that completely or partially surround a tower-mounted receiver. The heliostats can move about two axes and track the sun as it moves through the sky, concentrating the incident solar radiation on the tower-mounted receiver. The reflected solar radiant energy is
20
Applied Superconductivity
absorbed on the receiver, converted to thermal energy, and transferred to a heattransfer fluid. Molten salt, liquid sodium, and water/steam have been proposed as heattransfer fluids. The heat-transfer fluid is transported to ground level, where it is used to generate steam for a Rankine-cycle heat engine. The heat engine provides shaft power to a conventional electric generator. To extend the amount of time that a central receiver can provide energy, a storage subsystem is included. During operation, a fraction of the thermal energy in the heat-transfer fluid is used to charge a thermal storage unit. The stored energy is then available to generate steam for the heat engine during periods when solar radiation is not available. A similar system uses parabolic trough collectors to focus solar radiation on a linear receiver tube, where it is used to heat a heat-transfer fluid (such as oil). With the exception of the collector, the trough system is similar to the central receiver system. Statements regarding performance characteristics, current approach for end use integration, and impact of superconduction on the technology will apply equally to both trough and central receiver systems. Recent evaluations of solar thermal technology project a levelized energy cost of for central receiver plants. Trough systems are projected to be much more with levelized costs approaching 15@/kWh.
S-?C/kWh expensive,
Current
System
Integration
Approach
difficult system.
Only one prototype central receiver plant is in operation. Therefore, it is to predict the preferred approach for integrating this technology into a utility There are at least three options:
l
Quasi Base-Load. A central receiver power plant can be designed to operate with capacity factors equal to conventional base-load plants. This is accomplished by adding substantial thermal storage. Reasonable designs have been proposed with annual capacity factors up to 0.8. Unlike conventional plants, the solar plant (even with storage) is still vulnerable to unusual weather conditions, such as a long period of cloudy weather. Peak Load. A central receiver amount of storage and operated plant can be designed as a peaking plant. with a small
Hybrid. A central receiver plant can use fossil fuels to either increase the temperature of the Rankine-cycle heat source or replace solar energy during periods of low insolation. In the second case, storage is not required.
Performance
Characteristics
to have
Depending on the design of the plant, a central receiver facility can be designed a capacity factor between 0.25 (no storage) and 0.8 (maximum reasonable
Renewable
Generation
21
storage), but in all cases, the plant is vulnerable to unusual weather. The use of thermal because this type of storage is essentially is an advantage, energy storage commercialized and inexpensive ($12/kWh thermal). For economical operation, a solar power plant must be located in areas with high incident solar radiation, such as the southwestern United States.
Impact of Superconductors Superconductor application potential would exist in the generator and transmission system according to the same criteria that apply to these components in Because the solar central receiver can reach high nonrenewable base-load systems. capacity factors with inexpensive thermal storage, there appears to be little prospect Superconducting that electric energy storage, of any form, would be competitive. magnetic energy storage might be cost-effective in particular cases, such as the integration of a stand-alone power system for a remote community or industry not connected to the grid. Liquid-metal magnetohydrodynamic (MHD) systems have been considered for fluid. application with central receiver systems , using sodium as a heat-transfer Superconducting magnets can provide many times the magnetic field strength of The application of conventional magnets and can be more cost-effective. superconducting magnets could facilitate the use of liquid-metal MHD energy conversion in the central receiver concept, thereby greatly increasing its relative importance as a renewable resource.
Technology
Description
This concept involves an array of parabolic dish-shaped collectors that track the sun in two axes, redirecting the incident radiation onto individual receivers located at The cavity-type receiver absorbs the solar the focal point of each concentrator. radiation on the heater tubes of a small (about 2%kWe) heat engine. Stirling-cycle, and supercritical organic Rankine-cycle heat engines have been Brayton-cycle, proposed. In all cases, the heat engine drives a generator to produce electricity, which is transported to either a storage unit or the utility power grid. When storage is included, battery storage has typically been selected for use with solar dish concepts. An alternative solar thermal dish concept uses an array of dish receivers to produce thermal energy, which is then transported to a central heat engine. Thermal energy can be transported by a heat-transfer fluid or in a thermochemical transport system. Either way, the thermal energy generated by a large number of receivers is transported to a central Rankine-cycle heat engine, which generates electricity. To extend the amount of time that the dish concept can provide energy, a thermal storage system can be included. During operation, a fraction of the thermal energy from the dish
22
Applied Superconductivity
array is used to charge storage. The energy is then available to generate heat engine during periods when solar radiation is not available.
steam in the
The dish systems using dish-mounted heat engines and electric storage have been extensively investigated, but electric storage using batteries has proved to be very expensive. The dish concept with a central heat engine is being considered as an alternative, but the thermal energy transport system has also proved to be expensive. The dish concept with a central heat engine has performance characteristics similar to those of a central receiver system, and the impact of superconductivity on this concept should be about the same: therefore, this discussion will concentrate on the dish system using small dish-mounted heat engines.
Current System
integration Approach
Two approaches for integrating the solar thermal dish concept into a utility system have been proposed. First, the dish system can be installed without any battery In this ease, the dish systems storage and feed power into the grid as it is produced. The second approach is to include a small amount of depend on system-wide storage. storage to allow the power to be delivered during peak demand periods, but the economics of battery storage are so unattractive that adding more storage will rapidly make the solar thermal dish concept uneconomical.
Performance
Characteristics
The solar thermal dish concept is characterized by high efficiency but is strongly Compared with other solar thermal concepts, penalized by expensive battery storage. the dish concept normally shows the lowest cost up to a capacity factor of 0.27. Above this capacity factor, storage must be included, and the cost of the concept soon exceeds Due to the modular nature of this concept, that of the other solar thermal technologies. it is particularly suitable for remote installations, but the lack of cost-effective electric storage again offsets this advantage. For economical operation, solar thermal dish power plants must be located in areas with high incident solar radiation, such as the southwestern United States.
Impact of Superconductors Superconductors could replace generation and transmission system components of the solar thermal dish system, with potential advantages similar to those achievable with Costsuperconductor applications to these components in nonrenewable systems. effective SMES would have a major impact, because the application potential of the solar thermal dish concept is currently limited by battery storage. Large-scale superconducting storage could improve grid-wide storage, eliminating the need for onsite battery storage. Small-scale superconducting storage could replace on-site batteries and improve the prospects for remote applications.
Renewable
Sources
for Electricitv
Generation
23
3.3.5
Solar
Photovoltaic
Cells
Technology
Description
Solar photovoltaic cells convert solar power directly into DC electric power. The solar cell is a semiconductor (typically silicon or gallium arsenide) that exhibits a photoelectric effect when illuminated by sunlight. The freed electrons are gathered and transported from the cells by metallic contacts on each cells surface. The cells are arranged in modules, which can either be fixed (typically pointing due south with a tilt equal to the latitude of the installation) or mounted in arrays of modules that track the sun in one or two axes. Sun-tracking arrays can also employ devices to concentrate the solar radiation. Photovoltaic cells represent the least site-restricted solar technology. However, due to the periodic nature of the solar resource, a typical photovoltaic system operates at a capacity factor of 20-30%.
Current
System Integration
Approach
The use of photovoltaic cells in large-scale commercial power applications has been prevented, to date, by their high cost. Worldwide, in 1986, over 80% of the photovoltaic panels purchased for electric power production were used in stand-alone Photovoltaic cells are often the lowest-cost power alternative in remote applications. applications. Small stand-alone applications with constant or variable power demand (e.g., telerepeaters or railroad signals) are cost-effective and work very well with photovoltaic cells and battery storage. Larger stand-alone photovoltaic systems that could be used to power entire communities not connected to the electric power grid would require energy storage. However, the cost and efficiency of current electric energy storage technology (batteries) precludes their use in this application. Thus, photovoltaic systems of this type are currently considered as fuel-saver systems for more conventional generating technologies. Grid-connected photovoltaic systems would require DC-to-AC inverters and power conditioning. Currently envisioned plants would supply energy to the grid as it is produced (sun-following mode).
Performance
Characteristics
Fixed flat plate photovoltaic arrays have lower efficiency and capacity factors than tracking arrays; however, fixed arrays do not require complex tracking mechanisms. Most stand-alone (non-grid-connected) applications use this technology. One-axis tracking improves the performance of photovoltaic arrays over fixed arrays by allowing higher input (and hence output) during the day. The capacity factor is increased by more than 10% over that of a fixed array.
24
Applied Superconductivity
Two-axis tracking maximizes the output of a photovoltaic array by always keeping the array normal to the suns rays. The capacity factor is more than 25% higher than that of a fixed array. Concentrating photovoltaic arrays use either mirrors or Fresnel lenses to concentrate the solar radiation on the cells. Concentrators are used with more efficient and more costly cells to decrease the number of cells required. However, the use of concentrators precludes using the diffuse portion of the solar radiation.
Impact of Superconductors The primary potential value of superconductors could be to provide costeffective energy storage for photovoltaic systems of most sizes (except very small standalone units, in which SMES would probably not replace batteries cost-effectively). The inherent DC output of photovoltaic systems is highly compatible with superconducting storage. Energy storage might be provided as either utility system-level storage or dedicated plant storage. Cost-effective SMES could vastly increase the potential market for power from large photovoltaic systems and would allow the output of the plant to be stored and dispatched by the utility at the time of its maximum need. Thus, instead of having the output be completely sun-following, a photovoltaic plant could be designed to operate at any capacity factor from the base-load to peaking mode. Superconducting devices may also have the potential to improve both the efficiency and economics of solar cell production (for example, as a component of the doping process).
3.3.6 Geothermal
Energy Convemion
Technology
Description
Geothermal technologies employ conventional steam-turbine generation to produce electricity from the naturally occurring heat sources below the Heat can be extracted from any part of the earth and used to raise earths surface. steam; however, extraction costs can be prohibitive if the resource is not close to the surface. Four geothermal resources currently under consideration are hydrothermal deposits (steam and hot water), geopressure, hot dry rock, and magma. The most economical method is to use the naturally occurring hydrothermal deposits found in areas near volcanic zones.
technology
Current System Integration Approach Currently, geothermal electric plants operate as grid-connected, base-load The average geothermal unit is available on-line more than 95% of the time.
capacity.
Renewable
Sources
for Electricity
Generation
25
Performance
Characteristics
Dry-steam (superheated steam with little or no liquid) geothermal resources can be directly coupled to a steam turbine. Wet-steam or hot-water resources that are under high pressure currently use flash systems to generate steam suitable for input to the turbine. Dual-flash, binary-cycle, and flow systems are also being developed to make lower-temperature sources economical. Geopressured geothermal technology is based on high-pressure geothermal resources (3,000-10,000 lb/in.2 above hydrostatic pressure). This technology is still in the research phase, but it would employ conventional electrical generation technologies. Hot-dry-rock geothermal technology is still in the research phase and would also employ conventional electrical generation technologies. Magma technology is still only a hypothetical concept, but it would most likely use conventional steam-turbine technology to generate electricity.
Impact of Superconductors The impact of superconductors on large-scale, grid-connected geothermal energy conversion appears, in general, to be neutral. The resource side of the system is constrained by the geographical location, thermal power, and diversity of the source, which neither gains nor loses value if superconductors are incorporated in the balance of Superconducting components may replace conventional electric power the system. generating, transmission, protection, and control components of geothermal plants in the same manner as is possible with regular fossil- or nuclear-fueled stations. The use of SMES could possibly allow geothermal plants to operate in a load-following mode, which could offer the potential of new stand-alone (non-grid-connected) development of the Superconducting magnets could possibly find application in devices smaller resources. that remove materials in geothermal fluids that cause corrosion and fouling.
Technology
Description
Wind energy conversion systems extract power from the wind by the use of a wind turbine-alternator set. The kinetic energy of the wind is first converted into mechanical energy, and then electrical energy. The amount of power that can be obtained from the wind increases with the square of the blade diameter and the cube of the wind speed.
Current System Integration Approach The majority of wind turbines installed today have a rated output of between 50 and 100 kW and are deployed in arrays known as wind farms delivering AC power
26
Applied Superconductivity
directly to the grid. Power-conditioning equipment relatively poor intrinsic quality of the power produced applications are limited to water pumping, because unpredictable. If wind power is used in a stand-alone high-cost battery storage is often required.
is often required to improve the by wind farms. Most stand-alone the resource is intermittent and electrical application, relatively
Performance
Characteristics
There are two basic types of wind turbines: vertical axis and horizontal axis. Horizontal axis wind turbines (HAWTs) are by far the most common type currently in use. They operate with the blade either up- or downwind and produce higher torques than the vertical axis wind turbines, which are suitable for providing mechanical energy or producing electricity. Vertical axis wind turbines (VAWTs) are the most efficient variety, extracting the most power from the wind at any given speed. They also operate at higher speeds than other wind turbines, require no equipment to point them into the wind, and enable the generator to be located on the ground. Gusts of wind traveling across hundreds of generators in the wind farm pose special control problems for the facility operator to ensure system stability and maintain adequate power quality. Though fewer, larger wind turbines are expected both to enable lower-cost power production and to simplify the control of wind farms, multiple small turbines are currently preferred, because mass production techniques can be used in their manufacture. Turbines with outputs of up to several megawatts have been built, but technical problems have delayed their development towards commercialization.
Impact
of Superconductors
Due to the variability, seasonality, and unpredictability of wind resources, wind turbines are generally not considered for stand-alone applications. Thus, in this type of system, SMES would probably be of minimal benefit. However, in grid-connected applications, SMES could be used to interface wind turbines with the grid, permitting dispatch by the utility at the time of its maximum need. Power conditioning using superconducting components may become an efficient and cost-effective means of removing voltage and frequency transients induced by variations in wind velocity.
3.3-a
Ocean
Thermal
Energy
Conversion
Technology
Description conversion (OTEC) exploits the small thermal differential water and water at depths of about 1,000 m. There are currently under consideration: openand closed-cycle plants evaporate the warm surface water and use the drive a turbogenerator. Closed-cycle OTEC systems use
Ocean thermal energy (about 2OC) between surface two types of OTEC plants Open-cycle OTEC systems. resulting low-density steam to
27
the warm surface water to boil a different working fluorocarbons), which then drives the turbogenerator.
fluid
(ammonia,
propane,
or
Current System Integration Approach OTEC facilities produce energy continuously with very little diurnal or seasonal variation in energy output. Shore-based OTEC generating stations often require prohibitively long lengths of large-diameter inlet pipe to transport cold water from suitable depths. Offshore plants may dispense with the need for an extremely long coldwater inlet, but they still require some method of transporting generated power to shore. With a suitable means of transmission, AC power may be generated and delivered directly to the grid. The best resource areas in the Gulf of Mexico are often more than 100 mi offshore, so the ability to transmit electricity to the grid remains one of the main limitations to current application of this technology. For this reason, the offshore production of many electric-energy-intensive products (e.g., aluminum, ammonia, chlorine, hydrogen, magnesium, methanol, etc.) has been proposed as a more costeffective use of OTEC power.
Performance
Characteristics
It has been estimated that more than 14 x lo6 MW may be generated through the conversion of less than 0.1% of the heat energy stored in tropical surface waters. This represents more than 20 times the current generating capacity of the United States. Desalinated water produced in open-cycle plants may be used for drinking water or irrigation in arid regions. In addition, cold-water return from an OTEC plant may be used to cool nearby buildings. Several serious drawbacks remain prior to commercialization of OTEC technology. Extremely large low-pressure turbine sizes are required to gather energy from the low-density steam generated in open-cycle systems, seawater is extremely corrosive to the heat-exchanger elements, and the removal of dissolved gases from the seawater through the OTEC process may alter the ecological balance of the surrounding region. The main drawbacks of the closed-cycle design have been fouling and corrosion of the large and very costly heat exchangers. Low-cost, easily deployable deep-water pipe and suitable means for transporting or storing the electricity generated offshore need to be developed before this resource can be fully exploited.
Impact of Superconductors There are several areas where advances in superconducting technology may contribute to improvements in the performance and operating efficiency of OTEC generating stations. Extremely large quantities of both warm and cold water must be pumped through an OTEC plant. Superconducting motors may improve the efficiency of the inlet water pumps. Significant frictional losses in the extremely large low-density steam turbines may be substantially eliminated through the application of
28
Applied Superconductivity
superconducting magnetically levitated bearings. In addition, the cold inlet water may provide an excellent heat-sink for the superconductor refrigeration equipment. The lack of a suitable means of transportation of power generated in OTEC generating stations is presently a significant limiting factor to this technology. Excitation current losses in conventional AC transmission cables limit their usefulness to distances of less than 30 mi. The development of an undersea superconducting transmission cable may open new markets for OTEC-generated power. As the critical temperatures of the new generation of superconductor materials continue to rise, it may one day be possible to deploy superconducting power transmission cables cooled simply by the surrounding seawater. It may also be feasible to employ tankers equipped with giant SMES coils to store offshore OTEC-generated power for later transportation to shore.
3.3.9
Technology
Description
Biomass energy is derived from plant or animal matter. There are numerous methods of extracting this energy currently in practice, and others have been proposed. Biomass conversion can be used to produce heat or electricity near the feedstock source, or the feedstock can be converted to other, more-transportable fuels. The most common methods considered are direct burning; producing synthesis gas under oxygenated, aerobic, or anaerobic conditions; producing alcohols by hydrolysis and fermentation; or producing methane by anaerobic digestion.
Current System Integration Approach Due to the extremely high cost of transporting the biomass feedstocks and their generally low energy density, biomass facilities are typically located close to the feedstock source. Direct burning applications (e.g., wood-fired boilers) are typically used to raise steam for use in a process or for generating electricity. Other biomass processes that produce liquid or gaseous fuels typically have their output transported in fuel form to other end users. However, if biomass processes can be made economical, electricity may be made during the process, or the fuel may be used to generate electricity at the site.
Performance
Characteristics
Direct burning of biomass waste is often used to provide heat, steam, and electricity. Typical feedstocks are wood wastes and peat. When electricity or steam is being produced, conventional steam-boiler technology similar to that used in coal plants is used.
Generation
29
Ethanol is typically produced by fermentation from feedstocks containing starch, The main differences between the fermentation processes occur sugar, or cellulose. because of the differences in the pretreatment the various feedstocks require. Methanol production from wood biomass is generally accomplished by wood gasification, modification and cleaning of the resulting gas, and then liquefaction. Methane from biomass is produced by anaerobic digestion by various types of animal manures, aquatic plants, sewage bacteria. Typical feedstocks are wet biomass: sludge, or food processing wastes.
Impact
of Superconductors will probably have a neutral impact on most biomass conversion systems that produce electric power directly are large enough, find beneficial application in associated generation and as is the case with other renewable and nonrenewable energy
3.3.10
Magnetohydrodynamic
Energy
Conversion
Technology
Description
In a conventional electric power generator, electric current and voltage are induced in conductors that are caused to move orthogonally to the direction of a magnetic field. In a magnetohydrodynamic (MHD) generator, a conducting fluid replaces the solid conductor windings in the conventional machine. Many device configurations and fluid systems have been proposed, including DC and AC machines energized by nuclear-heated inert gas plasmas, combustion products, liquid metals, and two-phase working fluids. In most concepts, the heated working fluid expands into a duct containing a magnetic field perpendicular to the flow direction. Electric current is generated in the fluid in a direction mutually perpendicular to both the magnetic field and the flow. Insulated electrodes built into the side of the duct collect the current and connect with power-conditioning equipment for delivery of energy to the grid.
Current
System
Integration
Approach
While MHD research and development has been going on for nearly 30 yr in the United States, no machine has been developed past the pilot plant size or connected to the grid on a continuous basis. The DOE-funded MHD experiment conducted by Mountain States Energy Corp. in Butte, Montana, has operated for about 150 h. The DC output (about 1.5 MW) of this coal-fired machine has been inverted to AC and supplied to the grid for short durations. Most MHD plant designs are large base-load systems, connected to the grid through inverters in the case of DC machines. Large DC machines could also supply DC bulk power transmission lines directly. MHD plants may be developed to
30
Applied
Superconductivity
generate AC power, possibly with a continuing need for power conditioning the synchronization, waveform, and harmonic quality of the output.
to improve
Performance
Characteristics
MHD operation in the United States to date has been less than that needed to encourage the expectation of near-term commercialization. The Soviet Union has, however, forged ahead in this area and has several experimental units. One is a 250-MW MHD system topping a gas-turbine/combined-cycle plant of equal size that is planned for on-line operation following an experimental evaluation period.
Impact of Superconductors The power of an MHD generator is proportional to the fluid conductivity and the square of fluid velocity and magnetic field strength. The one-to-two order of magnitude increase in magnetic field strength provided by superconductors over conventional magnets could improve the performance of essentially all MHD devices. However, of particular value, the availability of superconducting magnets opens up the potential for operation at lower MHD temperatures and with fluids of lower conductivity. The renewable energy technologies that may become attractive in combination with MHD conversion include biomass and waste combustion, ocean energy, and some solar thermal concepts. The use of superconductors in MHD systems would also be applicable to the power buses and the balance of the transmission system, according to criteria that would apply to conventional nonrenewable energy systems. There is probably a negligible prospect for SMES applications plants, because almost all concepts appear to be base-load systems. possible and economical to drive a low-temperature MHD generator and store the output in a shipboard SMES. in land-based MHD However, it may be by an OTEC source
3.3.11
Technology
Description
Nuclear fusion is the joining together, or fusing, of nuclei from light elements, such as the deuterium or tritium isotopes of hydrogen, to form a new atom (helium) with The mass difference between the fusion less mass than the sum of the reactants. products and the reactants is converted to heat, which can then be used to generate electrical energy. While nuclear fusion is not usually classified as a renewable resource, the abundant supply of deuterium fuel that can be easily extracted from sea water may enable fusion to become a virtually inexhaustible energy resource. Three primary methods are available These are high-energy laser-induced fusion, While great catalyzed cold nuclear fusion. for achieving controlled nuclear fusion. magnetic-confinement fusion, and muonstrides have been made in each of these
Generation
31
technologies, output.
none
has
as yet
approached
the break-even
point
to yield
a net power
Approach
One approach to nuclear fusion generation involves the use of high-energy lasers to compress a fuel pellet to an equivalent of more than lo6 atm pressure and over The development of higher-efficiency lasers is necessary prior to 100 x 106 Oc. achieving the break-even point where this technology approaches viability. Another option involves the generation of intense magnetic fields to confine the high-temperature plasma fuel until fusion occurs. Superconducting magnets offer the most plausible means of efficiently generating the intense magnetic fields necessary to enable this technology to become viable. A third approach, which is beginning to show great promise, is known as muon-catalyzed cold nuclear fusion. This method involves the generation of negatively charged muons with a particle accelerator. These particles bind to the hydrogen atoms and facilitate the fusion process. Cold fusion technology is presently limited by the efficiency with which the muons may be generated in the particle accelerator. Advances in accelerator design employing high-intensity superconducting magnets may soon enable commercial cold fusion generation to become a reality.
Performance Characteristics When fusion power generation becomes a reality, heat generated in the reaction process will most likely be used to generate steam to run a conventional turbogenerator. The ionized plasma produced may also be used to generate electricity magnetohydrodynamically.
Impact of Superconductors While laser-induced fusion is likely to experience little initial impact from advances in superconductor technology, both magnetically confined and muon-catalyzed fusion are likely to make significant use of advanced superconducting magnets. As fusion power generation becomes a reality, superconducting alternators and MHD generators should further enhance its performance and efficiency.
3.4
CONCLUSIONS
The principal impacts of superconductors on the 11 renewable energy technologies considered in this section are summarized in Table 3.1. The impact categories are (1) neutral impact, (2) enhanced energy storage capability, (3) improved system integration, and (4) new energy conversion potential, as discussed in Sec. 3.2.3.
32
Applied Superconductivity
in Renewable
Energy Technologiesa
Technology
Neutralb
Hydroelectric energy Solar salt gradient ponds Solar central receivers Solar thermal dishes Solar photovoltaic cells Geothermal energy Wind energy systems OTEC systems Biomass conversion MHD conversion Fusion power generation
X X X X X X
(X)
(X) (Xl
(Xl
(X) (X)
X X
X
(Xl
X X
aThe
dominant impact category is indicated by an X for each technolSecondary, conditional, or hypothetical impacts are indicated by ogy. If the primary impact is neutral, secondary (X) where appropriate. impacts generally reflect special-case considerations.
bAll technologies with impacts in Category 1 (neutral) can benefit from superconductors replacing conventional electric-power generation, and control components in the same manner as transmission, protection, is possible with conventional fossilor nuclear-fueled systems.
4 Generators
Summary
E.J. Daniels Argonne National Laboratory
33
34
Applied Superconductivity
Summary
Section
4 indicates
the following:
1.
Compared to a conventional 300-MW generator, a liquid-heliumcooled generator is projected to be more cost-effective (due to its higher efficiency); this would be the case even if the capital cost of the conventional generator were zero. The gain in efficiency due to reduction of refrigeration power requirements for a high-temperature superconductor is a modest 0.2%. However, the capital cost of the higher-temperature superconducting generator would be reduced significantly by the reduction in refrigeration system costs. If the refrigeration cost is proportional to refrigeration power, the cost of the $500,000 refrigeration system could be virtually eliminated. Unless the current density of wire, including both copper stabilizer and high-temperature superconductor, is on the order of lo4 A/cm2 (i.e., equivalent to that of the low-temperature conductor), the cost savings due to elimination of the helium refrigeration system will be offset by increased superconductor materials costs to achieve the 300-MW rated power. For example, at lo3 A/cm2, the higher-temperature superconductor machine would have a cost of $800,000 in excess of that of a machine based on the lower-temperature superconductors.
2.
3.
Therefore, the conclusions regarding economics are that superconducting machines are more cost-effective than conventional machines and higher-temperature more cost-effective than superconducting machines are lower-temperature superconducting machines at current densities of lo4 A/cm2. The analyses leading to these conclusions are summarized in Table 4.1, which presents five 300-MW generators: a conventional system, a helium-cooled system, and three nitrogen-cooled systems at different current densities. The costs and losses are those estimated by Kirtley. The value of losses is based on a 65% capacity factor, rather than the 80% used in the Kirtley analysis, and is presented on an annual basis. As shown, the total annual value of the conventional machine losses is $1.4 million. The total annual costs of the helium-cooled superconducting machine are more than $0.9 million less, including the annual capital costs at an 18.7% fixed charge rate. Even if the cost of the conventional machine were zero, the value of the annual losses of the conventional materials cost and losses of the helium-cooled machine would exceed the At a current density of 0.075 x lo4 A/cm2, superconducting machine by $0.75 million.
Generators
35
Nitrogen-Cooled Conventional System 298.23 98.6 a84= 4.17 1,411 165 HeliumCooled Systema 298.23 99.57 1,238 1.28 433 231 0 075 It 10 A/cm2 (Ml) 296.70 99.67 3,200 0.98 331 598
Systemsb
Parameter Rating (MU) Efficiency (I) Materials cost ($103) Losses (NW) Levelized val e of losses ($10 Y/yr) Annual capitalized materials cost ($103/y=)
1,576
664
929
454
421
aCurrent
density
= 0.80
described a helium-
in
on the basis
of
1.4 for
superconducting
relative
a conventional
machine.
superconductors is somewhat lower than that for the helium-cooled generator. However, this gain is offset by increased At a current density of 0.75 x lo4 A/cm2, the reduction in materials materials costs. costs makes the higher-temperature superconducting system more cost-effective than the lower-temperature system. If the current densities of the two superconductors are the higher-temperature system will have a cost advantage of about equivalent, $25O,OOO/yr. Section 4 also points out that the advantage of superconductors with higher transition temperatures would also be enhanced by expected improvements in reliability and availability when operating at liquid nitrogen temperatures relative to liquid helium temperatures. Higher-temperature superconducting machines may be cost-effective compared with conventional machines at smaller capacities. This is apparently attributable to the consideration that the capital cost of higher-temperature superconducting machines is significantly lower than that of the machines based on Therefore, a smaller absolute difference in losses helium-cooled superconductors.
36
Applied Superconductivity
(associated with the use of smaller machines) would be adequate to justify the capital cost premium of a higher-temperature superconducting machine relative to a conventional generator. To date, have been built. almost completed Electric built a Hitachi built a United Kingdom, several small generators incorporating low-temperature superconductors The Massachusetts Institute of Technology (MIT), with DOE support, has construction of a IO-MW generator. In the early 199Os, General ZO-MW machine and, in Japan, Mitsubishi built a 30-MW machine and SO-MW machine. In addition, other work is underway in Japan, the Europe, and the USSR.
In the early and mid-1970s, smaller machines were built: 5-MVA Westinghouse, 3-MVA MIT, 60-kVA MIT, and 45-kVA MIT. These machines were tested by using them as synchronous condensers -- i.e., they were spun with no input of shaft power; instead of generating electricity, they shifted the phase between current and voltage.
Generators
37
4.1
INTRODUCTION
This section discusses the possible impact on generators of superconductors capable of operating at temperatures achievable with liquid nitrogen coolant. Such conductors can be very useful in electric machinery if they can be fabricated into windings with appropriate cross sections carrying sufficiently high current densities in reasonable magnetic fields. Superconductors with transition temperatures may be exploited in at least three ways: 1. as high as that of boiling liquid
nitrogen
For the same type of thermal isolation systems used in prototype superconducting machines, higher operating temperatures will reduce the power required to provide refrigeration; Thermal expensive, Thermal between that isolation systems can be more reliable and efficient stability operating simplified, leading designs; and to less-
2.
3.
limits may be increased by increasing the margin temperature and transition temperature. modes of exploitation will be used in some combination.
It is likely
all three
of these
Cryogenic cooling exhibits a very strong economy of scale: although it takes a lot of work to refrigerate any space to liquid helium temperature, it does not take much more work to refrigerate a large space than a small space. For this reason, it has been generally accepted that superconducting technology is applicable only to the largest machines or to those with the most stringent power density requirements. Because the effort required for refrigeration to liquid nitrogen temperature is only on the order of l-2% of the effort required for refrigeration to liquid helium temperature, it is reasonable to expect that a much broader range of machines will be candidates for the application of superconductivity. The improvement in efficiency afforded by the reduction in cooling energy may be reduced if the superconductor is more limited in current density or magnetic flux density than its liquid-helium-temperature counterpart. A numerical study has been done to establish quantitatively, in one class of machines, the benefits of a number of postulated superconductors. The machine chosen is a turbogenerator rated at 300 MVA and 3,600 rpm.
38
Applied Superconductivity
4.2
SUPERCONDUCTORS
APPLIED TO GENRRATORS
Because superconductors carry current with no dissipation, their application to a large electric machine should increase its efficiency, but the need to cool the field However, even with superconductors that winding will tend to counter this advantage. must be cooled to liquid helium temperatures, turbine generators will be more efficient Estimates are that a superconducting than their normally conducting counterparts. machine rated at 300 MW would have a net efficiency, including refrigeration power, of 99.596, as opposed to 98.6% for a conventional machine. The reduction in refrigeration afforded by higher temperatures would improve efficiency still further, perhaps by another 0.2%. The usefulness of superconductors in electric machinery does not arise solely because of lossless conduction. It comes in part from their ability to carry very large current densities and from the high flux densities that can be produced. High current density allows a field winding in a generator, for instance, to produce both magnetization and reaction in a space containing no magnetic iron. This, in turn, allows the armature to be located in a low-permeance space (no iron), so that it can carry large reaction currents with little reactive voltage drop. Absence of iron has other useful attributes, such as an increase in the armature space factor and a reduction in core losses. An increase in useful flux density helps still further by shortening the armature path around each unit of flux, thus reducing the ratio of armature loss to power produced. The usefulness of superconductors in machines, therefore, depends on their having reasonably high useful current and flux density limits, as well as lossless conduction. These limits have an impact on the efficiency and on the first cost of the generator. It appears that a wire (made of copper cladding and superconductor) capable of carrying a current density of lo4 A/cm2 in a flux density of 2 T could be very useful. Two thirds of the wires cross-section might be copper, with the remaining third being superconductor. In this case, the current density in the superconducting material would be 3 x lo4 A/cm2, a value that could be comfortably achieved in superconducting material having a critical current density of 4.5 x IO4 A/cm. Hereafter, for ease of will be used to mean superconducting wire like that expression, superconductor described above, and current density will mean the ratio of the current to the crosssectional area of the wire. A major effect of the increase in operating temperature will be to make superconductors attractive for smaller machines, because the effort required to cool small spaces to liquid helium temperature is still substantial. Thus, for small machines the efficiency impact of cooling is larger than the gain in efficiency from the use of superconductors. Since cooling to liquid nitrogen temperatures takes only about 1% of the energy required to cool to liquid helium temperatures, it is reasonable to expect the boundary between the domain of superconducting machines and normally conducting machines will move downward with respect to size. There is some disagreement over just where that boundary is located for machines cooled by liquid helium, but it is certainly in the range of hundreds of megawatts. It is conceivable, therefore, that if liquid-nitrogen-temperature superconductors become practical, the boundary may be as small as one or only a few megawatts. This is small enough that all utility generators and many other applications would be within the domain of superconductivity.
Generators
39
Another major effect of the increase in operating temperatures will be a major reduction in the complexity of the thermal isolation system. In machines using liquid helium, it is necessary to use vacuum insulation, intermediate-temperature thermalradiation barriers, and vapor cooling of both structural shaft elements and current leads. Vacuum insulation, in particular, adds substantially to the cost, complexity, and potential unreliability of a machine. With the lower cost of refrigeration to 77 K, it is likely that much of this complexity can be eliminated, including the need for vacuum in difficult places. Superconductors, when applied to synchronous machines, are used in the field winding. Two attributes of the field winding are of concern to us. First, it is a winding, with a complex shape; second, the absolute current that the field must carry must not be either too large or too small. Taken together, these attributes restrict the type of superconductor that might be used in generators. The superconductor must be capable of fabrication. The complex shape of the field winding would indicate flexible wires or tapes (wires are better), but it is possible that innovative developments in fabrication might make other forms useful. In addition, the field winding of a generator is subject to very large forces arising from the magnetic field produced by the field winding, by reaction forces from the armature winding, and from rotation. The conductor used must be capable of withstanding these forces or be compatible with a scheme that can provide sufficiently strong support. This means that the support structure must provide enough restraint for the superconductor, allow for proper cooling, and accommodate thermal contraction. For operational reasons, it is necessary to control the current in this winding. Fairly large and rapid variations in this current are required. Thus, the field current cannot be too large or too small. If the field current were too large, leads required to carry current from the slip rings or rotating exciter would impose too large a thermal load. If the current were too small, control voltages would become too high. Field currents in the range of 103-lo4 A seem reasonable; these current values imply a conductor size of 3-10 mm (in conductors with lo4 A/cm2).
4.3
OTHER
APPLICATIONS
For reasons discussed in Sec. 4.2, a substantial increase in transition temperature is likely to extend the range of machines for which superconductivity will be helpful to much smaller ratings. It is possible superconducting windings will be economically justified in machines with ratings as small as a few megawatts. Superconducting machines have high power density, and so they have been investigated for such applications as high-power ship drives (tens of megawatts) and air and space applications. Refrigeration has always been something of a problem for these applications, because the refrigerator is of substantial size, complexity, and cost. Clearly, the advent of SUperCQndUCtOrS that could be cooled by liquid nitrogen would virtually eliminate this difficulty for many such applications. Further, it is reasonable to expect that other applications, not now considered feasible, will become so. One could imagine superconducting electric locomotives, pump drives, etc.
40
Applied Superconductivity
4.4
CASE STUDY:
JOO-MVA TURBOGENERATORS
In order to quantify the potential impact of liquid nitrogen temperature The case used was that of superconductors, a first-cut design study was made. 300-MVA, two-pole generators. This size was chosen because (1) it seems likely to be a common size for generators in the near future, (2) a starting point design already and (3) a comparison between conventional and liquid-helium-temperature exists, superconducting machines is available. In this study, six machines were designed. One of these, the base case machine, is assumed to be built using a liquid-helium-temperature superconductor. This machine is actually a modification of one designed b Westinghouse Corp. as part of an Electric P Four different liquid-nitrogen-temperature Power Research Institute (EPRI) program. superconductors were suggested in the call for this paper (see Table 4.2). These all have limited flux density capabilities, so a fifth, hypothetical conductor (M5) was added; this superconductor is assumed to have the same current and flux limitations as the liquidhelium-temperature superconductors. Table 4.3 lists the study results. Briefly, while it is possible to design a generator using the lower-current materials, such a machine does not appear to be economically feasible. On the other hand, the higher-rated superconductor results in an economically attractive design. The reduction in refrigeration capital cost and losses are large enough to reduce the importance of the cost of the superconducting material. Table 4.4 presents a more detailed listing of the pertinent design details, and Table 4.5 presents a complete listing of the spreadsheet that produced these results. Direct comparison with a conventional generator is difficult, because the dissimilarities in construction make our estimates of capital cost crude. However, evaluating losses at 5C/kWh, we note that the base case superconducting generator prevails over the conventional machine on the basis of losses. A 300-MVA conventional generator has an efficiency of about 98.6%. The base case superconducting generator has an efficiency of about 99.5%. The difference between these two efficiencies TABLE 4.2 Assumed Superconductors for translates to about 2.7 MW at full load. 300-MVA Generatorsa Now, a loss of 5C/kWh is, at 80% capacity factor, worth about $351/kW-yr, and that, at a 12% discount rate over 20 yr, is worth about $2,657/kW. Thus, the losses alone Cost Current Den ity give the superconducting generator a total Material (104 A/J) ($/kg) cost advantage of roughly $7 million. It is more difficult to estimate how much more expensive the superconducting machine might be, but the difference is not likely to be great. For reference, the refrigeration equipment for the liquid-helium-temperature machine will cost in the vicinity of $0.5 million. Ml M2 M3 M4 220 440 220 440 0.1 0.1 1 1
?4sss flux
density density
and
Generators
41
Component Refrigerator cap. Superconductor Armature copper Back iron Rotor steel Losses Capital cOstC Total aComponent costs
greater or less
M4
MS (499)
0 0 0
increments
base value.
(i.e., cost is this much
indicate negative values the base-case value). cost is the sum of system; they are the not
4.5 DEVELOPMENT EFFORTS AND IlUPEDIMENTS Turbogenerators with superconducting field windings have already been substantially developed, and any development using higher-temperature materials will benefit from this earlier work. Generators built with liquid-nitrogen-temperature superconductors will not differ in many respects from the types of machines already built (e.g., with respect to armature windings). The major differences will be that the liquidnitrogen-temperature machines will require less-elaborate thermal isolation systems. It is possible that so-called liquid-nitrogen-temperature superconductors may at temperatures lower than 77 K. The reason for this is that substantially higher margins for current density and magnetic flux density might be reached by cooling to a lower temperature. In essence, the advent of higher-temperature superconductors adds a degree of freedom to the design. It will be necessary to investigate the behavior This investigation, in turn, will require an of designs over this degree of freedom. understanding of many characteristics of the superconducting material at different temperatures.
be operated
42
Applied
Superconductivity
Base
Parameter
Case
Ml
I42
M3
n4
MS
Machine dimensions (m) Field inner radius Field outer radius Rotor outer radius Armature inner radius Armature outer radius Core inner radius Core outer radius Active length Curre t den ity of field (10' A/cmqja Maximum flux density (T) Efficiency (I)
5.96 99.57
1.56 99.67
1.56 99.67
1.90 99.71
1.90
5.96
99.71
99.72
'Current density in the copper windings on the armature is 1.2 x lo6 A/m2 in all cases.
steps in developing
electric
include
Development of forms of the material that can be fabricated into field windings. These could be wires, tapes, or green forms that can be sintered into place. Full characterization of the behavior of the material, including:
2.
Magnetic flux density vs. current density frontier, as a function of temperature; Dissipation resulting from time-varying magnetic fields; and Determination of sensitivities of other material properties on stress.
3.
Temperature optimization (determination of the best design temperature). Cooling will be easier at higher temperatures, but lower temperatures will provide batter material performance.
Generators:
Zeroth-Order
Design
Variable/Parameter Performance Rating (MW) Sync. reactance (X) Efficiency t%) Maximum field CT) Input Variables Machine dimensions Field inner radius (m) Field thickness (m) Rotation gap (m) Armature thickness (m) Active length (m) Rotational speed (rad/s) Pole pairs Rotor shell thickness (m) Arm to shield space (m) TT length (m) Field angle Arm angle Power factor Field space factor Arm space factor Material parameters
Symbol
Cold
Ml
H2
H3
H4
M5
MU xd eta Bmar
0.197 0.082 0.150 0.265 1.900 377 1 0.1 0.101 2.269 0.2 1.047 0.900 0.5 0.15
0.260 0.200 0.100 0.150 5.200 377 1 0.075 0.025 0.2 2.269 1.047 0.900 0.6 0.15
0.260 0.200 0.100 0.150 5.200 377 1 0.075 0.025 0.2 2.269 1.047 0.900
0.320 0.025 0.100 0.175 4.600 377 1 0.1 0.050 0.2 2.269 1.047 0.900 0.6 0.15
0.320 0.025 0.100 0.175 4.600 377 1 0.1 0.050 0.2 2.269 1.047 0.900 0.6 0.15
0.197 0.082 0.150 0.265 1.900 377 1 0.1 0.101 0.2 2.269 1.047 0.900 0.5 0.15
Field current density (i06 Aim2) Arm current density (10 A/m )
ifa
80.0 1.20
7.50 1.20
7.50 1.20
75.0 1.20
75.0 1.20
80.0 1.20
TABLE
4.5 (Contd)
% D B _7 g v,
Variable/Parameter
Symbol
Cold
Ml
M2
M3
t44
MS
Sat. flux density (T) TT conductance W;; TT strength (10 Pa Field density (kg m ) Arm density (kg/m ) Steel density (kg/m32 Arm conductivity (10 S) Free space permeability (10m6 Derived dimensions
1.4 1,790
?J
2 0 L? P
a ,400 a ,400
7,800 67.0 1.26
200
4 a _. 2
P F angle (rad) Shield radius (m) Field outer radius (m) Armature inner radius (m) Armature outer radius (m) Self length (m) Field length (m) Heuristic rules
psi
rS if: al R a0 1,
1f
trqf fvc etac eta,
Torque factor Vapor cooling factor Carnot efficiency efficien y Refrig. Windage loss (10 6 W) Field concentration Miscellaneous Armature coefficients ratio
P W
cfc
0.618
0.789
0.789
0.718
0.718
0.618
TABLE
4.5
(Contd)
Variable/Parameter
Symbol
Cold
Ml
H2
M3
H4
H5
Field ratio Field coefficient Arm coefficient Y coefficient Self coefficient Mot coefficient Flux coefficient Permeance coefficient
E kf ka ky k1 2 AP
Machine
rating xa 1. Power (T) b=f 0.398 0.752 298 5.19 0.475 0.685 297 1.36 0.475 0.685 297 1.36 0.365 0.779 296 1.69 0.365 0.779 296 1.69 0.398 0.752 298 5.19
Internal reactance Voltage ragi Rating (10 W) Fundamental field Thermal calculations
Maximum torque (1 6 Nom) TT thickne s (10 -8 m) 9) TT area (m Low-temp. heat leak (W) Back iron field (T) Back iron thickness (m) Mass calculations Shield outer radius Arm ass (kg) Field mass (kg) (m)
t=q, ttt ;E br th
a ;:
2 z
TABLE
4.5 (Contd)
b D
Variable/Parameter
Symbol
Cold
Ml
M2
M3
M4
MS
B r:
E v,
Shield mass (kg) Torque tube mass Rotor shell mass Losses (103 W)
(kg) (kg)
ii s R z-. r 2.
2
Refrig. input power Armature loss Core loss Total losses Core loss density Financial parameters
P,
pda
PC
hot
pcd
Interest rate Tax rate Lifetime (yr) Operating hours per year Power price ($/kWh) Caoitalized Dower ($/kW) Ta; kicker Costs of materials
f t
hrs
fp
dPr tax
($/kg)
220 20 4 20 220 20 4 20 440 20 4 20 220 20 4 20 440 20 4 20 220 20 4 20
Symbol
Cold
Ml
H2
H3
H4
MS
Refrigerator cap Superconductor Armature copper Back iron Rotor steel Losses Capital cost Modified Total for tax base
16 436 78
262 304 2,305 1,097 683 2,988
1,238 3,436
771 4,201
3,200 2,582
1,992 4,514
5,321 2,582
3,312 5,894
2,305 a79
541 2,052
Refrigerator cap Superconductor Armature copper Back iron Rotor steel Losses Capital cost Total Parentheses indicate negative values (i.e.,
cost
much lower
48
Applied Superconductivity
4.
Design and fabrication of a prototype. Because of prior experience with liquid-helium-temperature machines, this prototype could be fairly large. Conceivably, a prototype machine could be made as a modification to an existing experiment.
The major impediment to adoption of this technology is the current shrunken market for large electric machinery in this country. Because there have been virtually no orders for new power plants, the result of a sharp reduction in load growth rates, manufacturers of turbine generators have been retrenching. If and when demand growth catches up with installed capacity, the turbogenerator market wilI improve.
4.6
A cursory look at the possible advantages cooled) superconductors indicates substantial superconductors:
l
Higher temperatures would result in a sharp reduction in the cost of equipment to refrigerate the field winding and a hundred-fold reduction in the power required to keep the field winding cool. Operation at liquid nitrogen substantially the thermal isolation increase thermal stability margins capacity of the materials. temperatures would simplify scheme of the rotor and would because of the increased heat
These advantages may result and the power level at which smaller. These conclusions
l
attractive becoming
are contingent
on the following
requirements:
The new class of superconductors can be made so as to be capable of fabrication into complex shapes, such as field windings. The new superconducting wire (including superconductor stabilizer) can be made with reasonably high current density (lo4 A/cm2) in reasonably high magnetic fields (at least 2 T). The new superconductors such as unusual sensitivity do not turn out to have to strain or alternating and limits
4.7 1.
REFERENCE Westinghouse Electric Corp., Superconducting Generation Design, Electric Research Institute Report EPRI-EL-577 (Research Project 429-l) (Nov. 1977). Power
5 Transformers
50
Applied Superconductivity
Summary
One-sixth of the annual losses associated with transmitting electricity over the national grid occur in power transformers. Losses in power transformers are equal in magnitude to the output of five large-scale, base-load power plants. Installation of superconducting power transformers could reduce these losses.
Section 5 considers a design for a l,OOO-MVA generation step-up transformer with superconducting windings. incorporating Nb3Sn developed by Westinghouse Electric Corp. under contract to DOE in 1981. This design, together with the cost assumptions, formed the basis of B.W. McConnells following evaluation of the potential impact of the new high-temperature superconductors (HTSCs) on power transformers. Since almost nothing is known concerning the AC properties of the new HTSCs, Nb3Sn properties were assumed (except for the high critical temperature). The results of this analysis indicate that use of the new HTSCs will result in total life-cycle costs that are 35% lower than for Nb3Sn and 60% lower than for conventional power transformers of this size. To date, no full-scale superconducting power transformer has been built or tested. This is probably due in large part to the high value electric utilities assign to reliability; failure of the power transformer could result in a shutdown of the entire generating plant. Furthermore, although the cost savings associated with the superconducting power transformer appear to be substantial, the power transformer itself represents only a small part of the entire generating plant.
Transformers
51
5.1 INTRODUCTION at temperatures above the The recent discoveries of materials that are superconducting boiling point (77 K) of liquid nitrogen (LN2) may allow the development of power apparatus with significantly higher operating efficiencies and, hence, greatly reduced These materials also might have the advantage of remaining in the operating costs. superconducting state at significantly higher magnetic fields than previously seen in Type I and II superconductors. (However, the high field region has not yet been studied in At present, these high-temperature superconductors (HTSCs) appear to be detail.) extremely brittle and have a low current density (nominally 100 A/cm). However, reports of wires and ribbons fabricated from the materials offer hope that potential fabrication problems can be solved. In addition, IBMs announced increase of the current density in thin films by a factor of 100 is encouraging. The use of LN2 as a coolant implies immediate economic advantages over the previously required liquid helium (LHe). LN2 is considerably less expensive, because the basic raw material is free and the production process is considerably more efficient. In fact, the process is so inexpensive that the operation of HTSC apparatus at LN2 temperatures may well be considered for other technical reasons, even if highertemperature superconductors are found. This section presents a first evaluation of power transformers ss one technological application of the new HTSCs. This evaluation is based on the following general assumptions: 1. Extension of previous HTSC operating region These materials will working configurations Adequate bulk current designs using is possible. LHe superconductors to the
2.
into
3. 4.
of the
materials
will be favorable
In addition, the best technological estimates of realistic improvements in efficiencies consistent with other engineering constraints are applied where No credit is taken for the higher heat capacities or the greater thermal These latter credits may well further range present at LN2 temperatures.
52
Applied Superconductivity
improve the HTSC economic advantage and may provide for technical solutions to some perplexing problems seen in LHe designs. Also, no credit is taken for the elimination of any iron or the subsequent reduction in losses that may be possible with these materials. Transformer technology is evaluated using the set of baseline economic assumptions presented in App. A. The total life-cycle costs (TLCC) are compared for conventional and HTSC applications, and a time to break-even is estimated. Potential problems and research areas for the technology are summarized.
TO POWER TRANSFORMERS
5.2.1
Method of Analysis
The application of high-magnetic-field, high-current-density Type II superconductors has presented a challenge to power engineers for the last 30 years. However, the design of a power transformer using Type II superconductors has proven to be an extremely difficult engineering problem. First attempts at designing a superconducting transformer began in 1961 and continued through 1981. Over a ZO-yr period, a truly viable design was not found. However, near the end of this period, a joint did succeed in achieving a transformer design DOE/Westinghouse (DOE/WH) project that showed favorable economic results and appeared capable of prolonged steady-state Prior to this 1981 design, designs were unsuccessful due to a lack of operation. knowledge of AC losses in Type II superconductors, the excessive volumes of the configurations, and high AC losses due to large AC magnetic fields or large superconductor volumes. The 1981 DOE/WH study produced a design for a l,OOO-MVA generation step-up transformer, which had superconducting windings and operated at LHe temperatures. The study included an economic comparison of the new design with a conventional design of the same rating. The superconducting design was seen to have an economic advantage as a result of (1) a careful design of the conductors and windings, which substantially reduced AC losses, and (2) the inclusion of all costs associated with ownership over the transformers useful lifetimes (i.e., TLCC). This evaluation of HTSC application to transformers is based on (1) a careful extension of the results of the 1981 study using the ANL guidelines for TLCC analysis in the economic evaluation, (2) the inclusion of common costs that were not previously considered, and (3) a conservative replacement of the HTSC design during the design life of the system. This last change in the economic evaluation is based upon the present trend of replacing or overhauling large power transformers at the midpoint of the 30-yr book life. In this evaluation, the conventional transformer is not replaced during its lifetime; however, the HTSC transformer is replaced at the 10th and 20th years. The design parameters of the generator step-up transformer Power Voltage Basic impulse level 1,000 MVA 22-500 kV 1,300 kV under study are:
Transformers
53
Impedance Construction
12% Three-phase,
core-form
The original economic study considered only the components of the two designs that would differ (i.e., core, windings, refrigeration, and losses). Other items, such as the tank, manufacturing, instrumentation, and bushings, were not included because their costs were judged to be the same for both designs. Relative costs were computed, with 100 being the total cost of the items considered for a conventional unit. The present study includes these latter costs to obtain a more realistic economic evaluation. The economic parameters used in the present study are those provided in App. A. The losses include (1) for the conventional design, conductor IR losses, iron (hysteresis) and stray (or unknown) losses, and dielectric losses and (2) for the superconducting design, conductor AC losses, iron and stray losses, dielectric losses, heat leakage through the leads and dewar, and input power to the refrigerator. The procedure for adapting the results of the previous study at LHe temperature to a design at LN2 temperature was to identify the most significant items that would be changed and to estimate the impact of these changes on the cost. The items that were identified are:
1.
2.
requirements
3. 4.
windings, around
The refrigeration plant is required to remove about 2,000 W from the lowtemperature area. From Fig. 10 of Ref. 1, the efficiency of such a refrigerator is about 18% of the Carnot efficiency. Combined with a Carnot efficiency of 77/(300 - 77) and expressed as a reciprocal of efficiency, the coefficient of performance for an LN2 refrigerator is calculated to be 16.1. For this study, a more conservative value of 20 is assumed. Using the two coefficients-of-performance (COP) values, the cost of an LN2 refrigerator was determined (from Fig. 11 of Ref. 1) to be about one-eighth that of an LHe refrigerator. The second item to be altered was the cost of powering the was accounted for by multiplying the portion of the cost of losses refrigeration by the ratio (20/400) of the COP of the two systems. refrigeration plant itself and the cost of refrigeration power, the sufficiently low that it no longer represents a significant portion Therefore, the result is not sensitive to the exact value of the applied refrigeration. This attributable to the In both cases, the resultant value is of the total cost. correction factors.
Finally, the superconducting windings and the thermal insulation were assumed to be equal to LHe values. These represent a small portion of the total costs. The HTSC materials are undefined at this time, although there are indications that their brittleness and difficult handling characteristics will be quite similar to those of Nb3Sn. Since
54
Applied Superconductivity
fabrication costs will be a large part of the total cost of this material and fabrication processes may be quite similar, it is reasonable to assume that the cost of the new material will be close to that of the old. Perturbations on the costs of these materials were evaluated, and an extreme case, which demonstrates the effect on the final result, is included in Fig. 5.1.
5.2.2
Reaulta
The results of the comparison are shown in Table 5.1, with the base data for the conventional and LHe-cooled units taken from Table 6 of Ref. 2. The assumption is made and that the previously ignored costs of the tank, manufacturing, instrumentation, bushings account for about 94% of the capital cost of a conventional transformer. This value is added to the cost of all three designs, and the other component costs are adjusted so that the TLCC of the conventional transformer continues to be expressed as 100, as in the original study. The result of this adjustment 60% for the complete transformer. is to show a present-value, life-cycle savings of
On the basis of the data from Table 5.1, the effect of significant changes in the Factors of up to 10 times were cost of the superconducting materials was explored.
go-
Conventional
10
00 0
12
20
24
28
Time i;r)
FIGURE 5.1 Relative Costs of 1,006-MVA Power Transformers (costs are normalized to the cumulative costs of the eomrentional system in year 30)
Transformers
55
TABLE 5.1 Relative Costs of 1,000~MVA Transformers (costs are normalized to the cumulative costs of the conventional system in yeer 30)
Conventional Cost Item Conventional materials Superconducting materials Refrigeration plant Miscellaneous costs Efficiency Cost of lossesb Total life-cycle Percent savingsd aIncludes tank, costsC Case 1 0.47 0 0 7.28 0.997 92.25 100.00 Case 2 0.63 0 0 9.83 0.997 92.25 102.71 -3
Superconducting LHe 0.38 0.50 1.76 12.67 0.9985 46.14 62.20 38 HTSC 0.38 0.50 0.22 12.67 0.9992 25.53 39.61 60
manufacturing,
instrumentation,
bPl-esent value based on 11.55% discount rate, 30-yr and 4% inflation; the capacity factor is 80%.
The conventional Case 1 unit is assumed to have a full operating life of 30 yt. The conventional Case 2 unit is replaced at 15 yr, and the superconducting units are replaced in the 10th and 20th years. The present values of the capital costs are adjusted to reflect these assumptions. dCompared with the conventional Case 1 unit.
applied to this cost, and the effect on present-value cost of the LN3 superconducting transformer was computed. The results of this comparison indicate that significant variations in the value of these materials do not greatly change the final result. The fabricated materials cost used in the LHe study was $150/lb. Figure 5.1 shows the relative costs of the three designs as a function of time and includes a case where the cost of the LN3 superconducting materials exceeds the cost of LHe superconducting materials ten-fold. The payback time for the LN2 design, about three years, is still less than five years if the superconducting material is ten times as costly. Also, the conventional transformer has several distinct capital advantages in this analysis. If the superconducting transformers are assumed to have an effective life of 30 yr, the break-even time is about six months for the HTSC base case. The incremental capital costs were calculated to be $4,835/MVA, which means that an HTSC transformer in the l,OOO-MVA size range can have about 250% greater equivalent capital costs than a conventional transformer.
56
Applied Superconductivity
The LHe transformer design is based on using Nb$n as the superconductor, with a current density of about lo5 A/cm2. Present superconducting materials at higher temperatures may not be able to sustain currents of this level within the near future. If the current density cannot be increased to at least this level, the HTSC transformer size would become excessive. The LHe transformer design had an unresolved technical problem If a quench occurred as a result concerning short-circuit conditions. of overcurrent, the LHe refrigeration could not provide adequate cooling to return the windings to the superconducting state. Because of the substantially lower cost of LN2 refrigeration, sufficient cooling capacity could feasibly be included to overcome this problem. concern when a complicated Reliability becomes a primary apparatus, such as a cryogenic refrigerator, is installed in a system. However, replacing an LHe refrigerator with an LN2 unit greatly simplifies the system, and the cost of the LN2 unit is sufficiently low that redundancy can be built into the system with little economic penalty. A central concept in the transformer design used in this study is a configuration of four windings, with a main and an auxiliary winding At a selected overcurrent level, the main at each voltage level. windings switch to the normal conduction state in response to the magnetic leakage field strength, with the auxiliary windings then carrying the current and limiting it to a small multiple of the fullload current while remaining superconducting. The LHe or LNg superconducting designs will be physically the same volume as a conventional unit, but they should have a moderate weight advantage. Hence, transportation costs will be comparable.
5.4
CONCLUSIONS
The technology of power transformers, which represents a potential application for the new HTSCs, has been evaluated. This evaluation was predominantly an economic scoping study developed from previous work on a similar device using earlier, LHe-based technology. Power transformers show a strong potential for significant cost reductions using HTSCs when evaluated on a life-cycle basis. Break-even occurs at between six months and three years, and the analysis is considered to be conservative (i.e., favorable to conventional technologies).
Transformers
57
These evaluations assume that the new HTSCs can be made to perform at least as well as LHe superconducting materials in their magnetic, current density, and material properties. Specifically, the AC properties of the HTSCs have not yet been determined, but they are expected to be similar to the earlier Type II superconductors. If this is indeed the case, AC power applications may not be so easily achievable. However, the knowledge gained in applying LHe materials to both AC and DC power devices should reduce the amount of time required to achieve useful applications. For example, the 20-yr period required to produce a reasonable power transformer may be cut in half for the HTSC application. Several key areas of research appear to have been uncovered by this evaluation. The obvious need for higher current densities and bulk current capability has been previously stated by many researchers. A better understanding of HTSC physics and material properties is also needed. In particular, experimental and theoretical research on HTSC properties under time-varying magnetic fields must be conducted as soon as possible. If the HTSCs reported to exist above 150 K are consistently reproducible, some severe thermal difficulties encountered in earlier designs for transformers may be essentially solved by operating these HTSC materials at LN2 temperatures. A more detailed study of the application of HTSCs to transformers could also identify certain needed properties that may be producible by materials researchers.
5.5 REFERENCES
1.
Westinghouse Electric Corp., Application of Low Temperature Technology Transformers, U.S. Dept. of Energy Report DOE-ET-29324-l (Feb. 1982).
to Power
2.
Riemersma, H., et al., Application of Superconducting Technology to Power Transformers, IEEE Trans. on Power Apparatus and Systems, PAS-100(7):3398-3407 (July 1991).
6 AC Transmission
Supplement: Levelized Annual Cost Method R.A. Thomas and E.B. Forsyth
Brookhaven National Laboratory
58
AC Transmission
59
Summary
The annual losses associated with transmitting electricity over the national grid are equal to about one quad.* One-third of these losses (equal in magnitude to the output of ten large-scale, base-load power plants) occur in the transmission system. Installation of a superconducting transmission grid could greatly reduce these losses. Both AC and DC superconducting systems have been proposed. The DC system requires the use of AC-to-DC and DC-to-AC converters. The combined losses of these converters may exceed the losses of a conventional transmission line shorter than several hundred miles. Therefore, current research is concentrating on development of a superconducting AC transmission line. Recently, Brookhaven National Laboratory (BNL) has designed, built, and tested a helically wound, superconducting, coaxial cable 115 m in length and made of Nb3Sn. The cable system exhibits three types of losses: (1) current-induced losses in the superconducting cable, (2) voltage-induced losses in the dielectric, and (3) losses associated with refrigeration. In order to keep the current-induced losses at an acceptable level, current must be a factor of 8-10 lower than the critical current. Section 6 presents an analysis of the potential impacts of high-temperature lines that is based upon a superconductors on superconducting transmission DOE-sponsored analysis performed by the Philadelphia Electric Company (PECO) in 1977. Almost nothing is known concerning the AC properties of the new hightemperature superconductors, so this analysis assumed AC properties identical to those of Nb3Sn. Because the refrigeration losses exhibit strong economies of scale (due to surface-to-volume effects), the PECO analysis is based on a very large transmission line (10,000 MVA). The required critical current for this system is 200 x lo* A/cm2. Modifying the PECO analysis to account for the higher critical temperature (77 K) results in a reduction of the system cost, including capitalized energy costs, of about 30%. Table 6.1 summarizes the losses and transmission costs of service for the new superconducting material and two conventional systems: (1) an underground high-pressure, oil-filled-pipe transmission (HPOPT) line and (2) a combined aerial/underground system. Each system is 66 mi in length, haa a capacity of The superconducting system has the 10,000 MVA, and has a substation at each end. lowest losses and a cost of service that is higher than that of the aerial/underground system, but lower than that of the HPOPT system. Recent studies have high-intensity electromagnetic indicated fields. possible adverse health effects associated If this finding should lead to the requirement with that
60
Applied Superconductivity
TABLE 6.1 Comparison of Losses and Costs of Service for Superconducting and Conventional Transmission Systems
System
Loss (Xl
0.73 3.60
3.46 6.05
1.68
2.08
all future transmission lines be placed underground, superconducting could turn out to be the lowest-cost alternative.
transmission
lines
The losses associated with a transmission line of more typical size (1,000 MVA, LOO km) were also analyzed. In this case, several systems designed for a variety of voltage-current conditions were all found to have refrigeration losses of less than 0.2%. The cost-effectiveness of these systems was not analyzed.
AC Transmission
61
6.1
INTRODUCTION
The economic evaluation of power transmission cables has a long history and is fraught with uncertainties. This is especially true of force-cooled cables, since the force-cooling components and their energy costs add another axis to the optimization of all the other components of the system. Moreover, because power transmission cables generally have a physical life of more than 40 yr, the evaluation of the cost of losses (or of energy for refrigeration) re uires knowledge of the cost of energy 40 yr in the future. As.H.D. Short comments: 4 [Ejngineers, when discussing costs, should tread cautiously -- for there are many pitfalls in the imponderable paths of accountancy . . . . The design of any cable transmission circuit is more a matter of sound engineering judgment, having due regard to circuit security, standby and overload capacities, and its intrinsic commercial features, rather than academic formulae which attempt to convert the future into the present. Neither mans organic computer, nor any man-made computer, is clever enough to foretell the future, and let us pray they never will be, for then life itself would be intolerable. The calculations by Short.* presented below should be viewed with the skeptical attitude expressed
The costs associated with this section have been extrapolated from an earlier, comprehensive study on underground power transmission systems. That study was produced under a DOE contract by the Philadelphia Electric Company (PECO) in 1977. The superconducting system considered there was for the transmission of a very large block of power (10,000 MVA) and used three 230-kV circuits. In the near term, there appears to be no need to transmit such a large amount of power underground over a long distance, but the use of the cost figures from the PECO study allows different transmission systems to be compared rapidly, and it also highlights where attention should be focused to produce an economically competitive system. It is recommended that a careful systems study be performed on a system with circuits rated at 500 MVA
*Shorts
comments
are part
of the Discussion
of Walldorf
(Ref.
2).
62
Applied Superconductivity
(as compared with the 3,330-MVA circuits of the PECO study) to obtain appraisal of the potential of the new superconducting materials.*
a more realistic
A subsection on 500-MVA-per-circuit power transmission systems appears near the end of this section; Experimental data are available on superconducting cables of of the engineering knowledge about such systems is much this sixe,4 and the extent greater. If it is assumed that the characteristics of the superconducting material should be similar to those of materials used in present cables, it is possible to derive specifications for the new superconductors. Such specifications are given in Sec. 6.9.
6.2
METHOD Two methods are widely used in the economic evaluation of power transmission
cables: 1. The cost is given as the sum of all the capital costs, capitalization of the energy for losses and refrigeration present worth over the physical life of the installation. plus the to their
2.
The cost is expressed as a levelized annual cost obtained by converting the capital cost to an annual payment and adding the annual cost of losses. In this method, the conversion of the capital cost to an annual payment should be based on the economic or book life of the system, while the annual cost of losses should reflect the cost of the energy for losses and refrigeration over the physical life of the cable. as a per-unit cost, where per kilometer (or mile) the unit is either cost or cost per megavolt-
These costs are then usually expressed per megawatt-hour (or megawatt-year) amp per kilometer (or mile).
In the baseline assumptions given in second method. Studies done in the past articles have all used the first method. The values used in the PECO study, so the first assumptions will be those suggested in App. supplement to this section.
App. A, Daniels et al. recommend the and explained in textbooks and journal numbers presented below are scaled from method will be used, but the economic A. The second method is applied in the
6.3
ASSUMPTIONS
The power transmission system (106 km). The system cost is to include
is to move 10,000 MVA over a distance the cost of substations and their losses,
of the meaning of these circuit power-level maximum continuous thermal rating of the
AC Transmission
63
cost varies with the type of transmission cable used. (Direct-current lines, for example, have a high substation cost.) The cost for compensation is also included. The system Therefore, the superconducting must be able to operate after a single contingency. system consists of three 3,500-MVA (230-kV) circuits in three separate cryogenic enclosures, but it is capable of carrying 5,100 MVA on the two circuits remaining if one of the circuits should fail. In addition, if two of the circuits should fail, it is possible to carry 7,500 MVA on the remaining circuit for a period of up to four hours. Finally, the cables are each capable of carrying a fault current of 122.5 kA for the clearing time of 3.25 cycles. (The current carried by each cable is 12.55 kA at the 5,000-MVA maximum steady-state power level.)
6.4.1
The economic assumptions are generally those given in App. A, but they are modified according to the recommendations given in EPRIs Technical Assessment Guide. According to the EPRI guide,* the cost of energy for losses and refrigeration should include an energy cost and a demand cost. The energy cost should be evaluated at the average incremental cost generation ... . The average incremental cost of generation is close to the average cost of fuel per kWh of generator output. Also, the demand cost should be evaluated at the incremental cost of increasing the size of new a kilowatt of incremental loss would be facilities. For transmission facilities, evaluated at two-thirds the cost required to supply a kilowatt of new load. Therefore, the fuel cost of electricity was taken as 1.7elkWh and the demand cost as $1,20O/kW. (The PECO study used 1.76e/kWh and $460/kW.) Next, in order to calculate the capitalized energy cost over the life of the system, it is necessary to compute the 40-yr annual . ..qrrying charge rate for the cost of energy for losses and refrigeration. The 40-yr carrying charge rate is the inverse of the following sum: 40 (1 + e) ci=l (1 + r1*
where e is the inflation rate, 4%, and r is the discount rate, 11.55%. This gives a carrying charge rate of 7.73%. (The rate for the PECO study was 16.3%.) This means that the energy component (as opposed to the demand component) of the capitalized cost of energy for losses and refrigeration will be 2.1 times more important in this evaluation than in the ten-year-old PECO study. It is not clear why the 40-yr carrying charge rate for energy was such a high value for the PECO study.
*A copy of this guide could not be obtained in time to use for this report. are taken from the direct quotations in Walldorf and Eich (Ref. 2).
The quotations
64
Applied
Superconductivity
The demand component of the cost of energy is also much greater in this study than it was in the PECO study. The total capitalized cost of one watt of energy for the 40-yr operating life of the cable was $1.41 in 1976 dollars. In this study, it is $3.13 in 1987 dollars (2.22 times greater). Capital costs in general increased by a factor of 1.80 Therefore, the capitalized energy costs, as over this same time period (see below). compared with capital costs, will be 24% more important in this study than they were in the PECO study.
6.4.2
Capital
Coats
In order to properly scale the capital costs from 1976 dollars as given in the PECO study to 1987 dollars, the Producer Price Index was used. The latest edition of the Statistical Abstract of the United States: 19878 has values only for the years through 1985, so the 1987 value was obtained by using the change in the index from 1984 to 1985 This approach gives a multiplication factor of 1.80 for the and extrapolating to 1987. 1976 prices if the Producer Price Index for either capital equipment or electrical machinery and equipment is used. The costs for excavation, backfilling, and clearing and roads will be scaled by using the index for pipeline construction, which is about the same as the index for the cost of construction of dams and reclamation projects. This factor is 1.73. The right-of-way cost is more difficult to scale, because the route includes not farmland, but also suburban and urban areas. (A high percentage of the route is farmland.) Farmland prices peaked in 1982 and are now only about 1.54 times what they were in 1976, but suburban and urban land prices have increased much more dramatically. (The Statistical Abstract does not tabulate these land prices, but it indicates that the cost of shelter increased by about 2.36 times its 1976 cost.) As a compromise, the 1.80 factor will also be used for the right-of-way cost.
only rural
6.5
LOSSES In order to calculate losses, it is necessary to make of the superconductor, the coolant, and the cryogenic assumptions enclosure. about the
properties
6.5.1
Superconductor
Properties
and Current-Dependent
Losses
Conventional Type II superconducting cables are operated at only about oneeighth of their critical current density, for two reasons: (1) it is necessary to stay far below the critical current if the losses are to be kept low and (2) by remaining far below the critical current level, it is possible Por the superconductor to carry the large fault currents that sometimes occur in power transmission networks. If the superconducting cable is to be made using superconducting tapes, it is not necessary to specify a critical current density, because the thickness of the superconducting layers is limited only by mechanical considerations. (It must be possible to bend the tape around the core of the cable without damaging its electrical properties.) The specifications need only give the linear current density (i.e., the critical current per unit of tape width). For Nb3Sn
AC Transmission
65
superconducting tapes, the AC critical current was specified as 2,000 A/cm at 8 K for two cycles at 60 Hz. The actual tapes produced sometimes had values as low as 1,800 A/cm, but they were accepted for test purposes. (For the Nb3Sn tapes, this works out to a critical current density of about 200 x IO4 A/cm2.) BNL chose Nb3Sn over NbTi for testing because of Nb3Sns higher operating temperature, which promised lessexpensive refrigeration. The second critical parameter is the AC loss specified in watts per square centimeter of tape surface area. For the Nb3Sn tape laminates, this value was experimentally determined to be less than 30 VW/cm2 at 500 A/cm and 8 K, even after the laminates were wound into a cable and then removed and individually tested. The losses of the superconducting portion of the tape laminates were only 10 VW/cm2 (also at 500 A/cm and 8 K). It was discovered, however, that the fabricated cable exhibited losses that were proportional to the number of tape edges rather than to the surface area of the tapes, perhaps indicating that the edge losses were the major contributing factor to the current-dependent losses. Therefore, it will be assumed that the total current-dependent losses of the fabricated cable hysteretic, resistive, and eddy current losses) are a equivalent to a loss of 275 uW/cm at 500 A/cm and at operating temperature. Once the mechanism that produces current-dependent losses is better understood, it may be possible to reduce these losses significantly. If the inner conductor diameter is the same as that in the PECO study, 9.56 cm, then the linear current density is 418 A/cm at the maximum steady-state rating of 5,000 MVA. In order to have a cable that is surge impedance loaded, the operating stress would have to be about 10 MV/m. However, if the coolant is subcooled liquid nitrogen, then the cable can be operated at a maximum cable stress as high as 20 MV/m. To take advantage of the higher operating stress and produce smaller cables that would still be matched to their load, it would be necessary to have superconducting tapes that could operate at twice the linear current densities of the conventional Nb3Sn tapes. It is unclear whether cables using the new superconductors can be operated at linear (circumferential) current densities as high as those used in present materials, so it is probably wise not to assume operation at twice those values. If the cables are operated at 20 MV/m and at 418 A/cm on the inner conductor, then the diameter of the outer conductor changes from 12.62 to 10.98 cm (only 13%), and the cable is not as well matched to the load. Moreover, the dielectric losses double. Therefore, a design value of 10 MV/m will be used. The outer conductor diameter will be 12.62 cm, and the 5,000-MVA outer conductor current density will be 317 A/cm. For the PECO 230-kV cable design, the current-dependent loss per phase was 0.451 W/m (1.352 W/m per circuit) at 3,333 MVA. (The loss has been found experimentally to be proportional to the square of the current.)
66
Applied Superconductivity
6.5.2
oEr --
F m
and is found to be 441 pF/m. If the dielectric loss tangent of the insulating polymer and screens is 1.0 x 10d4, the dielectric loss (in W/m) is P = V2wC (tan 6) the loss is 293 mW/m per
For 230-kV, phase-to-phase voltage (132.8-kV, line-to-ground), phase, or 0.879 W/m per circuit.
6.5.3 Cryogenic
Knclosure Losses
The heat inleak depends on the diameter of the enclosure and the amount and type of thermal insulation used. To keep the analysis simple, it will be assumed that the enclosure is the same size as in the PECO study and uses three inches of insulation. The heat inleak consists of three parts: (1) conductive heat leak at the metal vacuum seals at the end of each 62-it section, (2) conductive heat leak at each of the five bicycle-wheeltype supports for the inner pipe, and (3) radiative heat inleak from the outer pipe surface to the inner. In changing the operating temperature from 7 to 77.4 K, the radiative heat flow is reduced by only 0.45%. The conductive heat flow will change by the ratio of X2 (300 K - 77.4 K) to ~1 (300 K - 7 K), where X is the mean thermal conductivity. For alloys, A increases by about 20%, so the conductive heat flow decreases by only 9.7% by It is calculated that the two ends contribute an going to the higher temperature. effective heat leak of 0.245 W/m, and the five bicycle-wheel supports give an additional 0.093 W/m. Calculations of the additional heat leak due to the thermal insulation were done multilayer for three types of evacuated insulation with the following results: superinsulation, 0.267 W/m; perlite, 4.19 W/m; and Cab-0-Sil with 55% copper opacifier, 1.41 W/m. The latter two have a much higher loss but are effective even in a relatively poor vacuum (about 20 urn or less). Since the system is assumed to be operating above the freezing point of nitrogen, cryopumping may be ineffective in producing the vacuum required for superinsulation (less than 1 urn). It appears impossible to produce an economical system at about 77 K that dispenses with the vacuum entirely, because the heat inleak would increase by a factor of perhaps as much as 1,000. On the other hand, if the effective use of superinsulation were to require that the vacuum space be dynamically pumped, the extra cost for energy to run the vacuum pumps and the decreased reliability would tend to favor an opacified powder.
AC Transmission
67
6.5.4
Refrigerator
Efficiency
The refrigerators for the PECO study were assumed to operate at an efficiency of 26% of the Carnot efficiency. The survey by Strobridge indicated that the efficiency of refrigerators (as a percent of Carnot efficiency) depends only on capacity, not on operating temperature. Nevertheless, it will be assumed that the nitrogen refrigerators operate at 30% of Carnot efficiency, or 9.6 W/W. Using the assumptions given above for energy cost and demand cost, and converting energy cost to a capitalized cost of energy for refrigeration, the capitalized cost of removing an additional watt per meter is found to be $30/m ($9.15/ft).
6.5.5 Total Losses If superinsulation is used, the total of the losses at 3,333 MVA is 2.836 W/m per circuit. They are distributed as follows: 21% enclosure losses, 48% current-dependent losses, and 31% voltage-dependent losses.
6.6 CAPITAL COSTS It is necessary to determine which capital cost might change by going from Nb3Sn to the new high-T, materials. Since operation at 77 K will almost certainly require an evacuated enclosure, the cost of the enclosure will not change significantly, although it can be assumed that the enclosure will not have to be dynamically pumped. The cost of the helium is eliminated. Not enough is known about the processing that will be required for the superconducting tapes to determine whether their cost will be greater or less than the cost of Nb3Sn tapes. The present superconducting tapes are made in a relatively simple manner by dipping Nb-l%Zr foil in tin and then reacting it at high temperature. It is then soldered to a copper tape and a stainless steel tape to form a I-mil-thick laminate. The base materials are cheaper for the new superconductors, but the processing may be considerably more difficult and expensive. Therefore, the assumption that the cable cost does not change is purely speculative. The only other major cost that might change would be the cost of the refrigerators. To calculate a refrigerator cost, it would be necessary to know much more about the operating conditions that will be required by the new cables in order to give low losses at useful current levels. In the 1976 PECO study, the refrigerators contributed 12.4% of the total. Therefore, not much additional uncertainty will be introduced by keeping the unscaled cost of refrigerators the same.
68
Applied Superconductivity
6.7
COST OF THE
LOSSES
The calculation of the losses is based on the information in Table 6.2. The total transmission and refrigeration losses, then, are 2.84 kW/km x 106.2 km x 3 circuits x 9.6 W/W = 8.67 MW. There is an additional refrigeration load produced by the six cryogenic terminations on each circuit. This load is expected to be about 2.7 kW, which is equivalent to about one kilometer of cable. So the total loss value is increased by 10% to account for these end effects and other losses, giving 8.67 MW x 1.10 = 9.54 MW.
TABLE 6.2 Losses per Kilometer per Ciiuit at 3,333 WA per Circuit
Source
W/km
The annual loss in MWh is to be based on load values of 9,000 MW and 7,200 MW for six months each. This varying load affects the current-dependent losses only. Instead of 1.352 kW/km, these losses become 1.352 kW/km x [0.9(2) x 0.5 + 0.72(2) x 0.51 = 898 W/km. Then the annual loss in MWh is 2.382 kW/km x 106.2 km x 3 circuits x 9.6 W/W x 1.10 x 8,760 h = 70,190 MWh. The results are shown in Table 6.3.
6.8 CAPITAL COSlS AND TOTAL SYSTEM COST When the assumptions stated above are applied to the PECO study data to get costs in 1987 dollars, the results are as shown in Tables 6.4 and 6.5.
6.8.1
Comparison
Systems
The PECO study also evaluated a high-pressure, oil-pipe-type (HPOPT) system and an aerial/underground system. Using the assumptions that were applied to the AC superconducting system, a similar recalculation of the SOO-kV HPOPT system is shown in Table 6.6. Sixteen three-phase circuits using naturally cooled, cellulose-insulated cables are required to deliver the 10,000 MVA (see Tables 6.7 and 6.8).
System
Losses
The 500-kV aerial/underground system is a five-circuit system. Sixty miles of the system are aerial and use separate towers for each SOO-kV circuit. Each cable of the aerial circuit is a three-conductor bundle. The last six miles of the system are underground in an urban area. The underground portion is also rated at 500-kV and is insulated by gaseous sulfur fluoride (SF6) (see Tables 6.9-6.11).
AC Transmission 69
TABLE
6.3 Capitalized
Transmission and refrigeration Total losses (M!Jja Total annual Losses (MWl~/yr)~ Annual energy cost ($103/yr)c Capitalized ener y cost ($103jd Demand cost (jlOqjave Subtotal ($10 ) Substations Transformer losses (MWja Total annual losses (MWh/yrjb Annual energy cost ($103/yrIc Capitalized ener y cost ($1031d Demand cost (210J IaPe Subtotal ($10 ) Series compensation Total Losses (MWja Total annual Losses (MWh/yrjb Annual energy cost ($103/yrIc Capitalized ener y cost ($103jd Demand cost (j.L05jave Subtotal ($10 ) Total capitalized cost of energy Losses ($103)
9.54
aBased on a LO,OOO-KW Load level, not including start-up (cool-down). bBased on 9,000~MW and 7,200~MW load Levels for six months each. 'Cost of energy = 1.7C/kWh. dCapitalized cost, if based on 40-yr Life and 12.94 present worth factor. eDemand cost is based on $1,20O/kW.
70
Applied Superconductivity
Item
$103
Right of Clearing
EllClOSUte
Manholes
Excavation and backfill Cable Terminations Monitoring systems Cable engineering Capitalized cable maintenance Refrigerators, incl. maint. Substations Series compensation Capitalized refrigeration and cabLe Losses Capitalized substation Losses Capitalized compensation Losses Total
8,910 628 415,048 176 150,661 638,741 17,820 6,312 9,450 23,273 263,772 193,606 30,780 26,889 162,648 407
1,949,181
0.5 0 21.5 0 7.7 32.8 0.9 0.3 0.5 1.2 13.5 9.9 1.6 1.4 8.3 0 100.0
al987
dollars.
6.9 ASSUMPTIONS REGARDING PROPERTIES OF CABLE MATERIALS The new superconducting materials were assumed to be as good at 77 K as the present materials are at 7 K. These assumptions were necessary simply because little is actually known about the AC electrical properties of these materials. Moreover, what is known about the DC electrical properties (i.e., the DC critical current at 77 K) suggests that these materials can be considered for use in power transmission cables only if one is allowed to assume that considerable improvement will occur. The AC loss characteristics of Nb.$n were found not to depend on the bulk current alone, but are better described by a theory that assumes that when the curren;; are below a critical surface current density, the material exhibits a low loss. However, when this critical current density is exceeded, the current is shared between the surface and the bulk of the material, and the losses increase rapidly. Thus, low losses in this Type II material are obtained only at currents much lower than the bulk critical current density, and these losses are sensitive to surface roughness and defects. Therefore, to give the operating (low-loss) current density for Nb3Sn in amps per square
AC Transmission
71
of Costs Cables,
for
Item
Cable Enclosure
Refrigerators Substations Capitalized substation losses Excavation and backfill Series compensation Capitalized refrigeration and cable losses Capitalized cable maintenance Terminations Cable engineering Right of way Monitoring systems
32.8 21.3 13.5 9.9 8.3 7.7 1.6 1.4 1.2 0.9 0.5 0.5 0.3
centimeter is misleading, because if the current were actually flowing in the bulk of the Consequently, the operating current material, it would not have the low loss required. density for Nb$n tapes is usually stated as a surface current density in units of amps per centimeter. The AC losses of the laminated superconducting tapes were found to be much higher in a cable configuration than they were when measured on a single tape, as mentioned above. It has been hypothesized that these higher losses result either from the unusual current flow pattern that occurs in the double-helical layers of superconducting tapes in the cables or from crowding of the current at the tape edges. (Since the tapes were obtained by slitting a sheet, the edges themselves did not provide a conWork on the new superconductors should tinuous path of superconducting Nb$n.) investigate the effects of surface roughness on the AC losses and whether current crowding occurs in tapes in cable configurations. In fact, if it is assumed that new cables will be made using superconducting tapes, various means of reducing currentdependent losses of tapes in cable configurations can be investigated using present superconducting materials. The Nb.$n superconducting tape laminates are made by dipping a Nb-l%Zr foil in Two superconducting layers of Nb$n are tin and then reacting it at 900-1,OOOC. formed. A 2-mil tape of tinned copper and a 1-mil tape of tinned stainless steel are then Two layers of superconducting tapes are soldered to opposite sides of the Nb.$in tape. used to form the double-helix making up each conductor. Therefore, both the inner and outer cable conductors contain four thin layers of Nb$n. The current does not flow
72
Applied Superconductivity
TABLE 6.6 Capitalized Costs of Loasea for 500-kV EPOPT Systems, 16 Circuits
Item
cost
Transmission Total losses (MWja Total annual losses (HWh/yr)b Annual energy cost ($103/yr)' Capitalized ener y cost ($103)d Demand cost ($105jaPe Subtotal Substations Transformer ($10 )
losses
(MWja
Total annual losses (MWh/yrjb Annual energy cost ($103/yr) Capitalized ener y cost ($103jd Demand cost Subtotal ($10 (f105jaPe )
Shunt compensation Total losses (MWja Total annual losses (MWh/ytjb Annual energy cost ($103/yr)' Capitalized ener y cost ($103jd Demand cost Subtotal Total losses ($10 (giOjaPe ) cost of energy
capitalized ($103)
aBased on including
10,000~MW start-up
losses,
not
load
levels
dCapitalized cost, if based 12.94 present worth factor. eDemand cost is based
on $1,2OO/kW.
AC Transmission
73
TABLE
6.7
HPOPT
Cables,
16 Circuits
costa
Item $103 % 1.1 0.1 a.8 1.2 a.8 26.6 1.6
0.8 0.1
Right of way Clearing and toads Pipe Manholes Excavation and backfill Cable Oil Terminations Pressure systems Engineering Capitalized cable maintenance Substations Shunt compensation Capitalized cable Losses Capitalized substation losses Capitalized compensation losses Total al9a7 dollars.
37,989 2,249 310,217 40,819 311,604 940,061 55,049 29,261 4,565 15,246 46,584 256,585 495,497 445,498 216,218 324,270 3,531,712
11
Given ail these caveats, Table 6.12 presents the assumed Some of the values contained in the table are calculated using against above, and thus are not strictly justifiable.
6.10
l,OOO-MVA TRANSMISSION
SYSTEMS
It is, of course, unreasonable to propose that a utility rely upon an entirely new technology to transmit a huge block of power in its own network. To gain acceptance, even after extensive field tests at noneconomical power levels and/or lengths, it will be necessary to introduce the technology at the lowest power level at which it can be technically and economically competitive. Therefore, the capitalized cost of energy for refrigeration was evaluated for some l,OOO-MVA power transmission systems.
74
Applied Superconductivity
TABLE Circuits
Cellulose-Insulated
Item Cable Shunt compensation Capitalized cable losses Capitalized compensation losses Excavation and backfill Pipe Substations Capitalized substation losses Oil Capitalized cable maintenance Manholes Right of way Terminations Cable engineering Pressure systems Clearing and roads
a.8 7.3
The term 1,000-MVA transmission system, applied to a two-circuit system, means that each circuit has a continuous, maximum contingency thermal rating of 1,000 MVA. The design operating rating of each circuit is 500 MVA. Each system consisted of two circuits, either of which would be capable of carrying the full 1,000 MVA should the other circuit be out of service. Unlike the 230-kV system evaluated above, a second contingency must be handled with circuits external to the superconducting transmission system. With the system voltage equal to 138 kV, the effects of varying the linear current density and the maximum cable stress were evaluated. The linear current density determines the diameter of the inner layer of superconducting tapes, while the maximum cable stress determines the insulation thickness. Only the effect on the losses is shown in Table 6.13, but these assumptions also influence the capital cost of the cables and enclosure. Capital costs were not generated. It should be pointed out that it is yet to be determined whether even the lowest linear current density used below can be obtained in a superconducting cable made with the new materials and operating at 77 K. These current levels are just the operating current levels; the quench current levels are 8-10 times higher. The enclosure insulation. diameters are based on a jam ratio of 2.4 and 3 in. of thermal
All of these lOO-km-long systems have a transmission efficiency of better than 99.8% at 100% load factor when only the energy losses associated with refrigeration are
AC Transmission
75
TABLE 6.9 Capitalized Costs of Losses for 500-kV Aerial/Underground Transmission System, Five Circuits
Transmission - aerial Total losses 04W)= Total annual losses (MWh/yr)b Annual energy cost ($103/yr)= Capitalized ener y cost ($103)d Demand cost (310 J )aTe Subtotal ($10 ) Transmission - underground Total losses (MW) Total annual losses (MWh/yrjb Annual energy cost ($103/yr)= Capitalized ener y cost ($103)d Demand cost (210 3 jaPe Subtotal ($10 ) Subscations Transformer losses (MWja Total annual losses (MWh/yrjb Annual energy cost ($103/yrF Capitalized ener y cost ($103jd Demand cost ($10 3 jave Subtotal ($10 ) Series compensation Total losses (MWja Total annual losses (MWh/yrjb Annual energy cost (SlO3lyrF Capitalized ener y cost ($103jd Demand cost ($10 5 jafe Subtotal ($10 ) Total capitalized losses ($103) cost of energy
aBased
on a
including
not
bB,sed on 9,000~MW and 7,200~HW load for six months each. Cost of energy = 1.7CfkWh. if based on 40-yr worth factor.
on
levels
cost, present
life
is based
$l,ZOO/kW.
76
Applied Superconductivity
costa
Item $103 x
Right of way Clearing and access roads Foundations Towers Conductors and devices Engineering Capitalized maintenance Underground section, 6 mi Substations Series compensation Capitalized aerial trans. losses Capitalized undrgnd. trans. losses Capitalized substation losses Capitalized compensation losses Total
139,230 7,923 27,000 74,358 86,400 13,500 58,202 186,480 229,500 29,322 189,425 12,455 216,218 1,339 1,271,252
11.0 0.6 2.1 5.8 6.8 1.1 4.6 14.7 18.1 2.3 14.9 1.0 17.0 0.1 100.0
al987 dollars.
TABLE 6.11 Breakdown of Costs of 500-kV Aerial/Undeground Transmission System, Five Circuits
Item
Substations Capitalized substation losses Capitalized aerial trans. losses Underground section, 6 mi Right of way Conductors and devices Towers Capitalized maintenance Series compensation Foundations Engineering Capitalized undrgnd. trans. losses Clearing and access roads Capitalized compensation losses
18.1 17.0 14.9 14.7 11.0 6.8 5.8 4.6 2.3 2.1 1.1 1.0 0.6 0.1
AC Transmission
77
Value
c2 x 10-5 1.0 x 10-4 2.3 2.3 CL5 <30 275 x LO5 A/cm2 x LO6 A/cm2 ?.Wcm2 pWfcm2 &cm2
Superconducting tapes Max. steady-state current density, inner AC quench current density AC Losses,a single tape Laminate AC Losses,a single tape, after cablin Effective AC Losses,a cable conductor % aAt 500 A/cm.
bActuaL AC losses appear to be a function of the number of edges on the conductor circumference, rather than of the conductor area.
tape
considered. The energy losses due to substations and compensation were not evaluated. For systems of shorter length, the end losses will become more important, and the total system losses will no longer be proportionate to circuit length. There will also be an optimum length before the nitrogen refrigerators reach their most efficient size. From the standpoint of energy use, it appears that the l,OOO-MVA superconducting systems are competitive with conventional power transmission systems. It is necessary, then, to determine how the capital costs scale with power level for the various technologies to see whether there is a net cost savings.
6.11 CONCLUSIONS
6.11.1 Comparison of Electrical Losses and Costs The losses of the different transmission systems are shown in Table 6.14. The superconducting system is unique in that the transmission losses are very much lower than the substation losses. The cost of the 230-kV superconducting system is found to be about 1.53 times the cost of the 500-kV combined aerial/underground system. The naturally cooled 500-kV HPOPT cellulose-insulated system costs about 2.78 times what the 500-kV aerial/underground system costs. It is important to realize that these economic assessments rest upon assumptions about the future cost of energy used to derive a capitalized cost of energy. If the total costs of the systems being compared have large
D
TABLE 6.13 l,OOO-MVA, 138-kV Superconducting Power Transmission Systems, 2 Circuita
Maximum steady-state rms current (A/cm) Linear current, rated power (A/cm) Inner layer diameter (cm) Outer layer diameter (cm) Capacitance (pF/m) Cable jacket outer diameter (cm) Enclosure inner pipe i.d. (cm) Enclosure outer pipe o.d. (cm) Losses per circuit (W/km) Voltage-dependent Current-dependent Thermal heat-leak Total System losses, 100 km MU MWh/yr Capitalized energy and demand costs aThe operating ($103) is half
400 200 3.33 5.37 255.7 7.02 16.85 35.18 la4 224 357 764 1.61 12,750 4,741
400 200 3.33 4.50 383.6 6.23 14.95 33.29 275 238 330 a44 1.78 14,130 5,246
400 200 3.33 4.23 511.4 5.80 14.11 32.45 367 247 318 932 1.97 15,710 5,817
700 350 1.90 4.40 146.1 6.05 14.52 32.85 105 346 324 775 1.64 12,180 4,643
700 350 1.90 3.33 219.2 4.90 11.95 30.29 157 380 287 824 1.74 12,890 4,924
700 350 1.90 2.89 292.2 4.54 10.90 29.23 210 401 272 a82 1.86 13,840 5,281
rating
contingency
thermal
rating.
AC Transmission
79
System
Electrical
S0Ul-X
Loss (%I
230-kV
superconducting
and refrigeration
500-kV HPOPT cellulose Transmission Substations Compensation Total 500-kV aerial/underground Transmission - aerial Transmission - underground Subsrations Compensation Total
differences in the ratios of their respective capital and capitalized energy components, then the relative advantage of one system over another will be sensitive to the energy Also, as mentioned in the introduction, it is possible to reduce a cost assumptions. capital cost by allowing an energy loss to increase, and equally, capital expenditures can The most cost-effective package again depends upon be translated into lower losses. assumptions regarding the future cost of energy (see App. A for one set of assumptions).
6.11.2
Comparison
of High-Tc Superconducting
Cable System
with Nb$n
Cable System
To find out how much is saved by using new superconducting materials operating at 77 K, it is necessary to assume a refrigerator operating efficiency for the helium refrigerator of the 7-K system. The 77-K system efficiency was taken to be 30% of the Carnot efficiency, 9.6 W/W. Although the cables in the PECO study operated at 7 K, temperatures in the refrigerator heat exchangers were as low as 5.65 K, so assuming an efficiency of 26% of the Carnot value (the value used in the PECO study) for a refrigerator operating between 300 and 5.65 K gives 200 W/W. The enclosure losses must also be increased slightly to account for the additional heat leak due to the lower operating temperature. If the energy used for refrigeration is calculated using these assumptions, the sum of the capitalized cost of the energy for refrigeration and losses ($326.7 million) and
80
Applied
Superconductivity
the demand cost ($241.7 million) is $568.3 million, as compared with $26.9 million for the The increase in the total system cost is about 28%. The cost of the 77-K system. helium adds another 3%. Thus, the system cost (including capitalized energy costs) of the 7-K system is about 30% more than the cost of the 77-K system. This savings is about twice that predicted by using the assumptions of the PECO study (loss tangent = 1 x 10e5), but here it is assumed to be ten times larger. (The loss tangent of the polyprop lene laminate insulating tapes used at Brookhaven was measured to be less than 2.0 x 10 -5, but a relatively imprecise thermal measurement of the voltage-dependent electrical losses of the cables indicated an effective loss tangent of about seven times that value.*) Although the current-dependent losses are assumed to be larger by a similar they may in fact be considerably reduced in both high- and low-T, factor, superconductors once the mechanism that gives rise to these losses is sufficiently understood. Because the AC losses of the superconducting tape laminates now in use are only about one-tenth of the cable losses, and as experiments appear to indicate that the cable loss is proportional to the number of tape edges in the superconducting layer of the cable rather than to the bulk properties of the tapes, there is a good chance that these losses can be reduced. On the other hand, if it turns out that the AC losses of cables made with the new high-T, superconductors are higher than for the present materials, the relative advantage of the new cables could become much less significant. The result of these assumptions increases the capitalized cost of the energy for the refrigeration and cable losses to about 22% of the total for the low-temperature transmission system. Consequently, the more efficient 77-K system has a more dramatic impact on reducing the total cost. The large relative advantage of the 77-K system over the 7-K system comes about, then, because of pessimistic assumptions regarding the losses of the 7-K system. In actuality, the next superconducting cables made with Nb3Sn would not have been like the first cables that were produced. The high dielectric loss in the first cables resulted from having to use a plastic laminate for the insulating tapes; polypropylene tapes of the appropriate thickness were not available in time to meet the construction schedule, so thinner tapes had to be glued together with lossy polyurethane. Thus, the dielectric loss would be reduced by a factor of between two and seven in the next cables. Since it is now known that the current-dependent loss does not result from an intrinsically high AC loss in the superconducting tapes, but instead results from some feature of the cable conductor configuration, the current-dependent loss would probably be reduced by a similarly large factor. Finally, the thermal enclosure would be redesigned to lower the thermal losses. All of these changes could be made without significantly changing the overall cost. These changes would also reduce the losses in a 77-K system. However, the relative advantage of the 77-K system over the 7-K system would become much smaller, because the comparison is based on the total system costs of the two systems, where the
*In calorimetric measurements in a large system, very small temperature or pressure errors can produce large uncertainties in the result. Furthermore, it is difficult to separate out heat that is generated in the cryogenic bushings and transported to thermometers monitoring the cable temperature.
AC Transmission
81
total system cost is the sum of capital, capitalized energy, and capitalized maintenance costs. Thus, part of the advantage of the high-l, systems described in this section results from assuming that no further improvements can be made in the next generation of cable designs that use presently available superconductors.
6.11.3
The results obtained for the capitalized refrigeration cost of energy for the l,OOO-MVA transmission systems indicate that evaluating the full cost of systems of that size is jbstifiable. The analysis above determined neither the capital costs nor the capitalized cost of energy for substations and compensation. The most reliable comparative results are obtained when different technologies application. It is are all evaluated with reference to a single, well-characterized recommended that such a study be undertaken for a system application of moderate length (IO-20 km, perhaps). It is desirable to determine the effect of length on the relative advantage of a superconducting system over other underground technologies. At some point, either the end losses or an inefficient refrigerator size will make the superconducting system less economical than alternative methods of moving the power. It would also be useful to determine at what power level a superconducting system becomes a cheaper solution to the power transmission problem.
6.11.4
All cryogenic systems benefit from reducing the cost of the enclosure, and as the enclosure cost is a significant portion of total capitalized costs, improvements in this area can produce substantial overall savings. Recent work12 indicates that enclosures can be produced that are less costly than those used in the cost evaluations above and have comparable thermal losses. A full optimization of a superconducting power transmission system will require careful attention to all components of the system, as well as research and development in areas not directly related to the superconducting material. Much work can progress in parallel with work on the superconductors. Those involved in developing the superconducting materials, on the other hand, need to be conscious of the impact of materials properties on the technical and economic feasibility of a full system. (For example, a rigid superconducting system that required superconducting joints to be made during the installation would probably be very much more expensive than one that used flexible superconducting cables.) Other areas that will require close scrutiny and optimization include the refrigeration cycles and systems and the electrical and mechanical properties of the insulation materials.
82
Applied Superconductivity
6.12 1.
REFERENCES Thomson, W., Lord Kelvin, On the Economy British Assn. Report, p. 526 (1881). of Metal in Conductors of Electricity,
2.
Walldorf, S., and E. Eich, Evaluation of the Cost of Losses for Underground Transmission Cable Systems, IEEE Trans. on Power Apparatus and Systems, PAS102~3355 (1983). Philadelphia Electric Co., Evaluation of the Economical and Technological Viability of Various Underground Transmission Systems for Long Feeds to Urban Load Areas, U.S. Dept. of Energy Report HCP/t-2055/l (Dec. 1977). Forsyth, E.B., and R.A. Thomas, Performance Superconducting Power Transmission System, references contained therein]. Summary Cryogenics, of the Z&599 Brookhaven (1986) [and
3.
4.
5.
Daniels, E.J., R.F. Giese, and A.M. Wolsky (Argonne communication (April 30, 1987). [The text of this App. A. of this report.] Weedy, B-M., Costs, Chap. 8 in Underground Wiley and Sons, New York (1980). Technical Research Assessment Group, Institute Special Report
6.
Transmission
of Electric
Power,
John
7.
Technical Assessment Guide, Electric EPRI PS-866-SR, Palo Alto, Calif. (June Abstract of the United States:
8.
U.S. Bureau of the Census, Statistical Ed., Washington, D.C. (1986). Strobridge, of Standards
1987,
9.
An Updated
Survey,
National
Bureau
10.
Progress through Fiscal Year 1976, Power Transmission Project Report PTP #69, Brookhaven National Laboratory Report BNL-22202, p. 37 (Dec. 27, 1987). Bussiere, J.F., et al., Nb3Sn Conductors Mechanical Characteristics, Advances Publishing Corp., New York (1978). for AC Power Transmission: Electrical and in Cryogenic Engineering, Vol. 24, Plenum
11.
12.
Schauer, F., et al., Prototype of a SemifZexible Multi-Layer Insulated Enclosure for Cryogenic Power Cables and Pipelines, Cryogenic Engineering Conf., St. Charles, 111. (June 1987).
AC Transmission
83
Supplement:
INTRODUCTION The assumptions outlined in App. A are somewhat different than those used in the PECO study. First, it is assumed that both the book life and the operating life equal 30 yr. These periods are not likely to be the same, and in computing the cost of losses the actual operating life should have been used, because the cable will go on producing losses regardless how quickly the accountants depreciate the capital cost of the transmission system. Nevertheless, the 30-yr life will be used below. Second, it is not clear how to evaluate the electricity cost component of the cost of service. All costs are normalized by the amount of energy delivered per year, but the electricity cost appears to be based on the electrical energy going into the system, not on the losses. Only the losses represent an actual cost of service if, as stated, the service is the transmission of electrical energy. The rest of the energy is merely transmitted by the system. What is actually being calculated is the energy price at the delivery end of the transmission system. The cost of transmitting the power is in fact the difference between the price of the energy leaving the transmission system and its The method used in the example will be used price at the point of going into the system. here, but it should be understood that the cost of service is not the cost of transmitting the power, but is instead the cost of the delivered power. Appendix A states that it is possible to judge the likely economic benefit of superconducting technology by dividing the calculated levelized cost of service by the levelization factor and comparing the result to the current price of electricity after transmission. Table A.3 in App. A shows that the price of electricity put into distribution is $O.OSO/kWh, compared with $O.O41/kWh put into transmission. These prices are assumed to be averages for all types of transmission (both overhead and underground), averaged over all the different lengths of transmission corridors. Because 66 mi is an unusual length for an underground transmission circuit, it is probably more valid to compare the values obtained for the three cases with each other than with the national average value. Finally, the example makes the assumption that the electricity should be charged at the price of electricity put into transmission. conventional cables and for the substation and compensation losses. electrical energy for refrigeration, however, should be the distribution change is made in the evaluations below. cost component This is true for The price of the price. This minor
Appendix A recommends that an 80% capacity factor be used. In the calculations of losses used in the main text, loads of 9,000 MW and 7,200 MW for six months each were assumed. Since this is an 81% load factor, the values obtained in Sec. 6.7 will be used here.
84
Applied
Superconductivity
236-kV SUPERCONDUCTING
AC POWER
TRANSMISSION
SYSTEfd
Energy
Delivered (9,000 MW x 0.5 yr + 7,200 MW x 0.5 yr) x 8,760 h&r = 7.0956 Y 10 lo kWh/yr
= 70,956,OOO MWh/yr
Energy
Lost Source Refrigeration and transmission Substations Series compensation Total MWh 70,190 395,000 955 466,145 $103 (1986 dollars) 3,510 16,195 39 19,744
= $2.909
x 109 + $1.974
x 107 = $2.929
Electricity
Capital
Cable
Maintenance ($1,798,530
Cost
x 1.45)/(7.096
*Other
maintenance
costs
are included
in the capital
cost component.
AC Transmission
85
506-kVHPOPTCELLULCSE-INSULATEDNATURALLYCOOLEDSYSTEM
EnergyDelivered h/yr (9,000 MW x 0.5yr+ 7,200 MW x 0.5yr)x 8,760 = 70,956,OOO MWh/yr= 7.0956 10l'kWh/yr x
EnergyLost Source Transmission Substations Shunt compensation Total MWh 1,091,600 521,400 908,400 2,521,400 $103 (1986 dollars) 44,756 21,377 37,244 103,377
ValueofEnergy into System (7.096 lOlokWh/yr x $O.O4l/kWh)+ x $1.034 lo8 x = $2.909 109. $1.034 108 = $3.013 109 x x x
Cable
component.
86
Applied Superconductivity
SOO-kV AERIAL/UNDERGROUND
SYSYEM
Energy
= 70,956,OOO MWh/yr
7.0956
Energy
Lost Source Transmission - aerial Transmission - underground Substations Series compensation Total MWh 440,190 29,618 521,400 3,140 994,348 $103 (1986 dollars) 18,048 1,214 21,377 129 40,768
= $2.909
x 109 + $4.077
x 107 = $2.950
Electricity
Capital
Cable
Maintenance ($4,497,840
*Other
maintenance
costs
are included
in the capital
cost component.
AC Transmission
87
CONCLUSIONS It is clear that this method produces the same ranking as the method of capitalizing the losses. In order to compare the ratio of the costs of the different systems, it is necessary to obtain a cost of service for transmission alone. To do this, the levelized cost of service is divided by the levelization factor, and the cost of the energy put into the system ($O.O41/kWh) is subtracted. This gives the results indicated in Table 6.15. The results are about the same as those obtained by the method used in the main text. Note that after subtracting the cost of the electrical energy, the results have at most three significant figures. The cost of the 230-kV superconducting system is found to be about 1.66 times the cost of the 500-kV combined aerial/underground system. The naturally cooled 500-kV HPOPT cellulose-insulated cable system costs about 2.91 times what the 500-kV aerial/underground system costs.
TABLE 6.15
Cost-of-Service
Systems
($/kWh)
Transmission System
Cost
Levelized of Service
Transmission Only
88
Superconducting
Magnetic
Energy Storage
89
Summary
A low-cost, efficient device for storing electricity could exploit currently underused, base-load generating capacity to meet up to a 15% growth in load, or permit early retirement of inefficient peaking and intermediate generating capacity. Recent economic analyses indicate that large-scale superconducting magnetic energy storage (SMES) could be the lowest-cost alternative among a number of competing utility storage systems and gas-turbine generators for a range of fuel-cost and operating assumptions. The Bechtel Group and G.A. Technologies performed these analyses under contract to Los Alamos National Laboratory and based them on detailed conceptual design studies of SMES devices with energy storage capacities of l,OOO-10,000 MWh. A 0.0083-MWh, lo-MW SMES was designed, constructed, and installed for operation in the Tacoma substation in the Bonneville Power Administration (BPA) utility system. BPA operated the unit on line for about one year, at frequencies from 0.1 to 1 Hz, for more than one million cycles. Although BPA originally intended the unit for dynamic stability control of transmission lines, it used the device principally for dynamic response characterization studies of BPAs northwestern power grid. Of all proposed utility-side storage devices, SMES enjoys the highest round-trip efficiency (90-93%), and its ability to switch from charge to discharge mode in less than a second provides valuable system-stabilization benefits. Section 7 analyzes the potential benefits of high-temperature superconductors (HTSCs), indicating a potential cost savings of about IO%, provided the new HTSC has a critical current density of at least 70 x IO4 A/cm2 and structural properties suitable for coil construction. Operating costs are also reduced greatly. However, operating costs associated with the NbTi system are typically only a few percent of total annual SMES costs. Probably more important than cost reductions, use of the new HTSCs could lead to higher system reliability, which is valued by utilities. In addition, operation of the coil at higher temperatures would reduce somewhat the stresses associated with cooling down from ambient temperature, thereby permitting more flexibility in system design. Two potential problems associated with SMES will not be helped by using HTSCs. First, due to the large economies of scale, only SMES units in excess of 500-1,000 MWh Very few individual utilities could efficiently use appear to be economically feasible. systems of this size. Thus, SMES units will most likely be shared among several utilities or a power pool, resulting in the need to transmit power over extended distances and Second, the forces required to contain the incurring the associated costs and losses. magnetic field are enormous, as is the scale of the machine -- the diameter of the coil in present designs for 5,000 MWh is 1,000 m. In order to contain these forces in a cost-effective way, current designs call for locating the coils in trenches near the Due to these geological constraints, there may be surface, preferably in suitable rock. no suitable sites over large regions of the country, and thus (where feasible) concrete walls might have to be constructed for subsurface support.
90
Applied Superconductivity
7.1
INTRODUCTION
Superconducting magnetic energy storage (SMES) is one of the obvious potential systems that can benefit from the substitution of high-temperature superconductors (HTSCs) for the presently proposed NbTi operating at 1.8 K in superfluid helium. When built on a large scale (5,000 MWh and 1,000 MW), SMES has economic potential in competition with other forms of energy storage for diurnal load leveling in electric utility applications. For diurnal load leveling, SMES has the unusual capability of storing energy directly as electromagnetic energy, without conversion to another form of energy. All other energy storage systems convert electrical energy to mechanical, chemical, or thermal energy and then (when the storage is accessed) back to electrical energy. Thus, SMES enjoys high efficiency by not being penalized for energy conversion: the round-trip, chargeInductive energy storage (energy storage in a discharge cycle efficiency is 90-9396. magnet) was understood long before the advent of practical superconductors. Recent industrial studies2Y3 have made interesting technological modifications to SMES and have proposed using improved NbTi to reduce the capital cost by 27%. The substitution of an HTSC for NbTi requires two main assumptions about the properties of the new materials: (1) the new superconductor has a current density characteristic that approaches the performance of NbTi and (2) the material has good strength and forming properties to withstand the Lorentz force experienced by a conductor in a magnetic field. Unless an HTSC has these important characteristics, the probability of it economically replacing NbTi in SMES is low.
7.2
DISCUSSION
A number of prospective features make the use of HTSCs in SMES attractive. For utility applications, these may be reduced to reliability, which can be much more important than a cost saving for the capital plant. The refrigeration system for operation at liquid nitrogen temperature, for Operating efficiency for SMES could rise to as much as instance, is greatly simplified. 94-95%. Stability of the superconductor would be enhanced by the higher heat capacity of a liquid nitrogen system. The protective energy dump system to exhaust the cryogen Because of the short time could be less elaborate, but it could not be eliminated. required to exhaust the liquid nitrogen, the requirement that the cryogen must be drained from the bottom of the SMES dewar will most likely remain. To exhaust and waste the nitrogen to atmosphere at the top would require a thicker inner dewar vessel to
Superconducting
91
overcome
the
higher
increase
will occur.
density A detailed
Thus,
an offsetting
cost
Stabilizing material associated with the new superconductor might be decreased in quantity because of the matrix materials higher heat capacity at the initial higher temperature, when a transition to the normal conducting state might occur, as compared with the heat capacity in liquid helium. The complexities introduced by an HTSC will require an extensive analysis of SMES to ascertain the engineering design and economic trade-offs. Even some rather profound design modifications may prove to be fundamental to the use of the new superconductors. Perhaps the detailed and intricate superconductor designs used today will become unnecessary because of the increased stability potential. These new ceramic-like materials might be cold-drawn in thin-walled metal tubes to be wire-like, and subsequently bundled together and cooled by conduction or forced convection. The obvious features of SMES that provide cost savings by the use of HTSCs with the foregoing property assumptions are in the areas of refrigeration and cryogenic piping, thermal heat leaks (struts and radiation shields), substation size for auxiliary equipment, coil protection, and helium storage. Reference 3 provides a sound basis from which to examine cost reductions for the items listed in Table 7.1.
7.3 CONCLUSIONS The capital cost from the last Bechtel study for a l,OOO-MW, 5,000-MWh SMES unit was $980 million (1985 dollars). Thus, the HTSC system cost becomes $950 million. Even at best, then, if the savings estimate is off by as much as a factor of two, which is
TABLE 7.1 Potential Savings with 77-K SMES (1,000 MW, 5,000 MWh)
Savings
Item
($106)
Refrigerators struts Thermal shields Coil protection system Cryogenic piping Diesel standby generators Substation Helium storage Instrumentation Total
9.9
92
Applied Superconductivity
is $920 million.
This estimate
into account
replacing
the
Now let us consider the high-tern erature superconductor. density values supplied by Daniels et al. 8 are only a few amps per SMES equivalent fields of 1.6-S T. Because SMES uses current-density of the critical current density, 70 x lo4 A/cm2 at 3 T and 1.8 K for that the HTSC can replace NbTi requires extreme speculation about current density.
At best, the currentsquare centimeter at values of 80-85% NbTi, the assumption its probable ultimate
In the most optimistic view, assuming that IBMs value of 100 x IO4 A/cm2 applies at 77 K and at 1.6-5 T, then for the same SMES plant cost for the superconductor, the HTSC could be allowed to cost close to 22C/g. Appendix A proposes 2.2C/g. If the 100 x lo4 A/cm2 value applies and the cost is 2.2C/g, then the HTSC cost becomes $5.8 million, for an additional saving of $44.5 million. This is not trivial, but it is based on a current density that may far exceed reality. The necessary current density for thz HTSC at a 2.2C/g cost to break compared with NbTi, is 11.5 x 10 A/cm . Even this value far exceeds experimental measurements for bulk material. even, as present
Rather than attempt to undertake an economic analysis as proposed in App. A, an approach has been chosen that gives a comparison on a common basis with other energy storage systems that were studied by Bechtel. This approach gives a relative comparison of units with the same function. What has been used as a tool is the comparative economic analysis by Bechtel National, Inc., reported in Ref. 1. Figures 7.1-7.3 were originally included in Ref. 1. For the 5,000-MWh unit curves, one should draw new curves that will be 3% lower for savings, not including replacement of the superconductor. If a current density of 100 x lo4 A/cm2 is assumed, then the new curves will be 8% lower. Such optimism is based upon current densities not indicated to be attainable in bulk material.
7.4 1.
REFERENCES Luongo, Storage prepared C.A., R.J. Loyd, and S.M. Schoenung, Superconducting Magnetic for Electric Utility Load Leveling: A Study of Cost vs. Stored by Bechtel National, Inc. (March 1987).
Energy
Energy,
2.
Loyd, R.J., T. Nakamura, and J.R. Purcell, Design Improvements Reductions for a 5000 MWh Superconducting Magnetic Energy Storage Alamos National Laboratory Report LA-10320-MS (Feb. 1985). Loyd, R.J., et al., Design Zmprovements and Cost Reductions Superconducting Magnetic Energy Storage Plant, Part 2, Los Laboratory Report LA-10668-MS (April 1986). Daniels, E.J., communication R.F. Giese, and A.M. Wolsky (Argonne (May 26, 1987). National
3.
4.
Laboratory),
personal
Superconducting
Magnetic
Energy Storage
93
Constant dollar fixed charge Fuel cost - gas anrl oil Cost of charging electricity
SMES lines:
- - -
Construction Constrwtion
in rock in soil
----
SMES 5,000
MWh
SYES 10,000 MWh SMES 5,000 MWh SMES 10,000 0.05 1.5 2.0 Capacity 0.1 3.0 Factor I 4.0 0.2 / 5.0 0.3 I 7.0 0.4 I 10.0 MWh
FIGURE 7.1 Revenue Requirement Screening Curves for Various Energy Storage Technologies and Combustion Turbines - Based on Estimated Average Costs for Fuel and Charging Electricity over the Next 20 Years (Source: Ref. 1)
94
Applied Superconductivity
Cormtmt dollar fixed charge Fuel co6t - gar and oil Cat of ch6qing clcctrlcity
rate
SYES lima:
-Construction - - - Construction
in rock in soil
------SSMES
1,000
MWh
SMES
1,000 MWh
Capacity 0.05 L
1.0
Factor
0.2
0.1
I
1.5 2.0
1
3.0
I
4.0
I /
0.3
0.4
,
10.0
5.0
7.0
FIGURE 7.2 Revenue Requirement Screening Curves for Various Enemy Storage Technologies and Combustion Turbines - Elased on Peoje&ed Costs for Fuel and Charging Electricity in the Year 2000 (Source: Ref. 1)
Superconducting
Magnetic
Energy Storage
95
SMES
anes:-
Construction in rock
Construction in soil
SMES
1,000 MWh
Constant dollar flxed charge rate = 10.5% = Sl2lMMKrU Fuel cost - gas and oil = 6ClkWh Coat of charging electricity
FIGURE
Energy Storage Technologies and Combustion Turbines on Increased Costs for Fuel and Charging Electricity (Source: Ref. 1)
8 Motors
Summary
E.J. Daniels Argonne National Laboratory
96
Motors
97
Summary
The attractiveness of superconductors for motors is similar to their attractiveness for generators. Superconductors offer an increase in motor efficiency, which must be balanced against the capital cost premium for the superconducting motor (including its refrigeration system). However, as is pointed out in Sec. 8 and detailed in a supplement to that section, start-up produces large forces in motor windings, and these forces may preclude the use of brittle superconductors. In the review of six different types of motors provided in the supplement, Lipo concludes that a homopolar inductor motor in which the superconducting field winding is stationary (rather than rotating, as in other designs) may be the most suitable design for superconducting materials. Not only would the rotational forces acting on the windings be eliminated, but the practical feasibility of cooling a stationary winding would be greater than that for a rotating winding. As a first approximation, the cost of a liquid-nitrogen-cooled motor was analyzed relative to a conventional 1,500-hp motor. The capital cost of the high-temperature superconductor (HTSC) machine is based on the cost of superconducting generators, relative to conventional generators, reported by Westinghouse Electric Corp. As stated in the main text, the analysis probably overestimates the capital cost of the HTSC motor because it takes no credit for expected cost reductions in the refrigeration system when operating on liquid nitrogen (LN2) instead of liquid helium (LHe). The motor efficiency for the HTSC is based on the expected reduction in losses, as follows:
l
copper
losses, and
copper losses.
losses,
to be negligible,
do not
As the following analysis shows, the HTSC motor would have an expected savings of about 11% compared with the capitalized costs of a conventional motor, because of the increased efficiency of the HTSC motor and the consequent reduction of losses. No credit was taken in capital cost reduction for the HTSC motor due to LN2 refrigerant relative to LHe, although the referenced capital cost multiplier is based on LHe refrigeration. The potential cost savings would exceed 20% if the use of LN2 refrigerant reduces the system capital cost by 20%.
98
Applied Superconductivity
8.1
INTRODUCTION
The recent discoveries of materials that are superconducting at temperatures above the boiling point (77 K) of liquid nitrogen (LN2) may allow the development of power apparatus with significantly higher operating efficiencies and, hence, greatly reduced operating costs. These materials also have the advantages of remaining in the superconducting state at significantly higher magnetic fields than previously seen in Type I and II superconductors. At present, these high-temperature superconductors (HTSCs) appear to be extremely brittle and have a low current density (nominally 100 A/cm ). However, reports of wires and ribbons fabricated from the materials offer hope that potential fabrication problems can be solved. Also encouraging is IBMs announced increase of the current density by a factor of 100. The use of LN2 as a coolant implies immediate economic advantages over the previously required liquid helium (LHe). LN2 is considerably less expensive because the basic raw material is free and the production process is considerably more efficient. In fact, the process is so inexpensive that the operation of HTSC apparatus at LN2 temperatures may well be considered for other technical reasons even if highertemperature superconductors are found. This section presents a first electric motors -- of the new HTSCs. assumptions: 1. Yd:tension of previous HTSC operating region These materials will working configurations Adequate bulk current evaluation of one technological application -This evaluation is based on the following general
LHe superconductors
to the
2.
difficult
to fabricate
into
3. 4.
capability
can be obtained,
and or can be
of the
materials
will be favorable
In addition, the best technological estimates of realistic improvements in operating efficiencies consistent with other engineering constraints are applied where possible. No credit is taken for the higher heat capacities or the greater thermal These latter credits may well offer operating range present at LN2 temperatures. further improvement to the HTSC economic advantage and may provide for technical solutions to some perplexing problems seen in LHe designs. Also, no credit is taken for
Motors
99
the elimination of any iron and the subsequent reduction in losses that may be possible with these materials. Furthermore, the use of magnetically levitated bearings to reduce friction in rotating machines is not considered, since the energy savings are small relative to the uncertainty in the other efficiency estimates. However, the use of such bearings might enhance the operating reliability of large motors and generators -bearing failure is considered a major weakness in current designs. This technology is fit evaluated using a common set of base line economic assumptions (presented in App. A). The total life-cycle costs (TLCCs) are compared for conventional and HTSC applications, and a time to break-even is estimated. Potential problems and research areas for the technology are then summarized.
8.2 APPLICATIONS TO MOTORS During the 1970s and early 19809, a considerable amount of research and development was conducted on superconducting rotating machines, with a strong This research was performed in the emphasis on the large su erconducting generator. United States,- Japan, 85 Europe, and the USSR.7 These projects generally included the design and construction of small prototype units and the design and economic analysis of full-scale units up to 1,200 MVA. In recent years, the reduced demand for increased generating capacity has However, this technology base provides a firm curtailed these research activities. foundation for the development of large generators based on HTSC technology. Indeed the prototypes built during these programs were generally in the range of 5-20 MVA,a-li and this technology should be directly applicable to the construction of AC synchronous motors of 3,000-20,000 hp. Because these prototypes were built to solve technical problems and were not intended to create an optimal economic machine in the above size category, such machines cannot serve as a basis for economic evaluations of either In addition, while motors and generators have a common basis, motors or generators. several significant differences exist. For example, the starting performance of motors is of considerable importance and is complicated in the superconducting case by stability problems in the more common motor types. A more detailed analysis of such problems is given by T.A. Lipo in the supplement to this section. Any superconducting motor or generator would have several important reduction in losses of about SO%, size advantages over a conventional machine: reductions of about 20% in diameter and 60% in length, wei ht reduction of up to 60%, and improved transient response in the generating mode. 12-89 Furthermore, an HTSC machine would potentially benefit from even higher efficiency, reduced size and weight, and reduced capital costs because of the reduced refrigeration load. While generators may benefit from the increased magnetic field strengths that are potentially available in HTSCs, motors may not always benefit from higher fields. In general, preliminary designs indicate that iron or some equivalent magnetic material cannot be as effectively eliminated in motors as in generators; therefore, iron losses are not completely eliminated in this economic evaluation of superconducting motors.
100
Applied Superconductivity
A first-order estimate of the potential cost saving for a superconducting induction motor in the 1,500-hp range, compared with a conventional motor of similar size, can be made by (1) taking the cost of conventional induction motors in this size range ($33.5-49.2/kVA), (2) using the relative cost ratio (see Ref. 2) for the superconductor to conventional generating machines (1.40), (3) estimating the expected efficiencies, and (4) calculating the annual cost of losses of the two motors. The induction motor evaluated is actually an induction-synchronous hybrid; although such machines exhibit certain undesirable control problems, the economic evaluation is easiest for this machine. In addition, the induction motor is the most useful application, and the size chosen is the average for motors above 500 hp. Because capital costs, no reduced thermal induction motors of uncertainties in the ratio of HTSC capital costs to conventional credit is taken for the reduced cost of the LN2 refrigerator and the insulation levels. This should produce a higher estimate for the HTSC capital costs.
Operating costs are estimated from the relative efficiencies of the two motors. The efficiencies for the HTSC case are taken to include the operating costs of the LN2 refrigerator, which are relatively modest compared with the costs of the LHe refrigerator used in the generator design. Hence, the estimated cost of losses is also expected to be conservative. The operating efficiency for a conventional induction motor of about 1,500 hp is taken as 95%. The losses are usually broken into (1) stator copper losses, (2) rotor copper losses, (3) iron or hysteresis losses, (4) stray or unknown losses, and (5) friction and windage losses. The stator copper, rotor copper, and iron plus stray losses are roughly equal and total about 90% of the total motor losses; friction and windage account for the remaining 10%. Estimates indicate that the replacement of a typical 1,500-hp, squirrelcage induction motor by a superconducting machine that retains the same laminated iron stator structure would result in a reduction in losses of 40% (30% from elimination of rotor copper losses; 5% by reduction in stator power requirements, and thus stator copper losses; and 5% by an increase in air gap, and thus a reduction in stray losses). This motor starts as an ordinary induction motor, but it becomes a synchronous motor following transition to the superconducting state. Using this estimate of efficiencies, a TLCC estimate can be made. The results of this evaluation are given in Table 8.1. Figure 8.1 graphs the cumulative relative costs for the conventional and HTSC motor. From this graph, the time to break-even is about five years. In addition, the incremental capital costs are calculated to be $20.98/kVA for the HTSC motor. More importantly, these cost estimates are considered to be conservative, and the savings for the HTSC motor can be expected to be better than indicated. The 1,500-hp motor size represents the average size for motors above 500 hp. Realistically, the most likely candidates for application of HTSC technology are probably in the size range above 3,000 hp. The present economic evaluation should still hold on a dollars-per-kilovolt-amp basis, because the relative loss components are essentially constant and larger motor sizes have higher capacity factors.
Motors
101
Cost Item
Conventional System
cost ($103) Efficiency Annual cost of Lossesa ($103) Present value of total 10sseSb
Capital
($103)
aAnnual cost based on energy at $O.O5/kWh and 50% duty cycle. bPresent value based on 15% discount rate, lo-yr lifetime, and 4% inflation. 'Compared with conventional case.
100
HTSC
: I; 400 y E 0 3020-
::
0
10
Time (yt-j6
FIGURE 8.1 Relative Costs of Conventional and HTSC 1,500~hp Motors (costs are normalized to the cumulative costs of the conventional system in year 10)
102
Applied Superconductivity
As discussed by Lipo in the supplement, the rotating machine that operates as a motor presents several operating and design problems that are not relevant to the application of HTSCs to generator design. These problems are primarily concerned with the starting and control of the motor; however, mechanical stresses also present a significantly greater problem for motors.
8.3
CONCLUSIONS
The potential application of the new HTSCs to electric motors has been evaluated. This evaluation was predominantly an economic scoping study developed from previous work on similar devices using LHe-based technology. This technology shows a strong potential for significant cost reductions using HTSCs when evaluated on a lifecycle basis. Break-even occurs between three and six years, and the analysis is considered to be conservative (i.e., favorable to conventional technologies). These evaluations assume that the new HTSC materials can be made to perform at least as well as LHe superconducting materials in their magnetic, current density, and material properties. Specifically, the AC properties of the HTSC materials have not yet been determined but are expected to be similar to Type 11 superconductors. Even if this is the ease, AC power applications may not be so easily achievable. However, the knowledge gained in applying LHe materials to both AC and DC power devices should reduce the amount of time required to develop useful applications. Several key areas of research appear to have been uncovered by this evaluation. The obvious need of higher current densities and bulk current capability has been previously stated by many researchers. A better understanding of the HTSC physics and material properties is also needed. In particular, experimental and theoretical research on HTSC properties under time-varying magnetic fields must be conducted as soon as possible. If the HTSCs reported to function above 150 K are consistently reproducible, some severe thermal difficulties encountered in earlier designs for superconducting motors may be essentially solved by operating these HTSC materials at LN3 temperatures. Beyond the HTSC properties and fabrication difficulties, further research into superconducting motor designs such as homopolar AC inductor machines, synchronous motors, and induction-synchronous hybrids seems appropriate. Such research will define certain needed properties that may be producible by materials researchers. Furthermore, the control problems encountered in motors using HTSC materials need further research. For example, if methods using power electronics can be developed, the more desirable properties of the AC induction motor may be available for HTSC applications.
Motors
103
8.4
REFERENCES Massachusetts Institute of Technology (J.L. Smith Jr., principal investigator), Superconductors in Large Synchronous Machines, Electric Power Research Institute Report EPRI-TD-255 (Research Project 672-l) (Aug. 1976). Westinghouse Electric Corp. (J.H. Parker Jr. and R.A. Towne, investigators), Superconducting Generator Design, Electric Power Institute Report EPRI-EL-577 (Research Project 429-l) (Nov. 1977). principal Research
1.
2.
3.
General Electric Co. (J.J. Jefferies and P.A. Rios, principal investigators), Electric Power Research Institute Report Superconducting Generator Design, EPRI-EL-663 (Research Project 429-Z) (March 1978). Maki, N., T. Sanematsu, and H. Ogata (Hitachi, Ltd., Japan), Design and Component Development of (I SO-MVA Superconducting Generator, IEEE Trans. on Power Apparatus and Systems, PAS-99(1):185-193 (Jan./Feb. 1980). Kumagai, M., et al. (Toshiba Corp., Japan), Generator, IEEE Trans. on Energy Conversion, Lambrecht, D. Superconducting (Sept. 1981). Development of Superconducting /X-1(4):122-129 (Dec. 1986). of Development MAC-17(5):1551-1559 AC
4.
5.
6.
of
7.
High Efficiency and Low Consumption Material on Magnetics, MAG-19(3):541-544 (May 1983).
8.
Chang, Y.W., et al., Development of a S-MVA Superconducting Generator - Testing and Evaluation, IEEE Trans. on Power Apparatus and Systems, PAS-93(2):496-499 (March/April 1974). Bzura, J.J., F. Abtahi, and L.J. Stratton, Technical Considerations and Ancillary MAG-17(1):880-883 (Jan. 1981). Superconducting Generators: Economics, Technology, IEEE Trans. on Magnetic%
9.
10.
Ashkin, M., et al., Stability Criteria for Superconducting Generators - Electrical System and Cryostability Considerations, IEEE Trans. on Power Apparatus and Systems, P&+101(12):4578-4586 (Dec. 1982). Keim, T.A., et al., Design and Manufacture of a 20-MVA Superconducting Generator, IEEE Trans. on Power Apparatus and Systems, PAS-104(6):1475-1483 (June 1985). Mole, C.J., D.C. Litz, and R.A. Feranchak, Cryogenic Electrical Machines for Ship Propulsion, ASME Paper Winter Annual Meeting, New York (Nov. 1974). Aspects of Superconducting 74-WA/PID-9, presented at
11.
12.
104
Applied Superconductivity
13.
Appleton, A.D., J-8 Superconducting DC Machines - Concerning Mainly Civil Marine Propulsion but with Mention of Industrial Applications, IEEE Trans. on Magnetics, MAC-11(2):633-639 (March 1975). Thullen, P., T.A. Keim, and J.V. Minervini, Multipole Superconducting Motors for Ship Propulsion, IEEE Trans. on Magnet&, MAG-11(2):653-656 1975). Electric (March
14.
15.
Appleton, A.D., et al., Superconducting DC Machines: A I-MW Propulsion System Studies for Commercial Ship Propulsion, IEEE Trans. on Magnet&, MAC-13(1):767769 (Jan. 1977). Brechna, H., and H. Kronig, Three-Phase Induction Motor with a Superconductive Cage Winding, IEEE Trans. on Magnetics, MAC-15(1):715-718 (Jan. 1979). Appleton, A-D., T.C. Bartram, and B.J.C. Grand, Superconducting DC Motors Marine Propulsion, Marine Engineering Review, pp. 8-10 (June 1982). Marshall, R.A., 3000-Horsepower on Magnetics, MAC-19(3):876-879 Superconductive (May 1983). Field Acyclic for
16.
17.
18.
19.
Appleton, A-D., Design and Manufacture of a Large Superconducting Homopolar Motor (and Status of Superconducting AC Generator), IEEE Trans. on Magnetics, MAC-19(3):1047-1050 (May 1983).
Motors
105
INTRODUCTION The prospect of using superconductors (SCs) to power electrical machines has been a tantalizing but elusive goal ever since its discovery by Onnes in 1911. However, recent developments that have raised the critical temperature above that of liquid nitrogen (77 K) have removed the major obstacle to practical application of SCs to rotating machinery, namely the liquid helium refrigerant system. This supplement to Sec. 8 addresses the potential for using high-temperature superconductors (HTSCs) in rotating (as opposed to linear) electrical motors and discusses several approaches to their implementation. Although superconductivity at liquid nitrogen temperature is assumed, the implications of room-temperature SCs are not considered in this supplement.
MOTIVATION
FOR DEVRLCPMENT
OF HTSC ELECTRIC
MCTCRS
It is well known that electrical machinery constitutes the greatest portion of the electrical load in this country. In particular, of the total U.S. electrical consumption of 1,683 x 10 kWh, motors consume 1,081 x 10 kWh, or 64% of the total. In the industrial sector alone, motors account for 76% of consumption.2 While the efficiency of electrical machinery has been rising, the efficiency of squirrel-cage induction motors ranges roughly from 78 to 92% for machines rated between 1 and 100 hp, suggesting that substantial energy savings remain to be achieved. The challenges of using the newly discovered HTSCs in motors rather than generators are, however, complicated by the relatively small size of motors when compared to turbogenerators, which number only in the thousands. There are estimated to be over 50 million motors in use in the industrial and commercial (I&C) sector of the United States, of which 1 million have a size greater than 5 hp.3s4 There are also estimated to be over 7,500 classifications of induction motors in the size range between driver in the end-use market, the first 5 and 500 hp.4 Since cost is such an important motor application of superconductivity will almost certainly come from the family of motors rated above 500 hp, which are normally classified as the form-wound family of machines (as opposed to the random-wound family, which predominates below 500 hp). In support of this observation, the present status of the permanent magnet (PM) motor can be recalled. The PM motor is a close cousin of the superconducting motor since it shares the major attributes of an HTSC machine, namely zero input excitation However, the high cost of permanent magnets have kept PM machines out of the power. low-horsepower motor market, and the same market forces will probably limit HTSC
106
Applied Superconductivity
machines until the machine is large enough to attain an economy of scale for the cooling system. Motors rated above 500 hp are dominated by three types: squirrel-cage induction (27,000 machines installed in the I&C sector), synchronous (roughly 10,000 machines),5 and a relatively small number of DC motors (6,000) used for variable-speed applications.6 The average rating of these machines is 1,500 hp, with an average efficiency of about 95%. While the number of AC machines above 500 hp is small, these machines account for a large portion of the energy consumed by all electrical motors. For example, in 1977 the total energy consumption of AC motors in industry and commerce was estimated to be 540 x 10 kWh. Of this total, 61 x 10 kWh, or 3.6% of all electrical energy produced in the United States, was consumed by machines above 500 hp (7.4% by machines rated above 125 hp).7 As discussed in Sec. 8.2, induction machine losses (stator and rotor copper, iron, stray, and friction and windage) can be reduced by 40% with the application of SC technology. Assuming an energy cost of $O.O5/kWh, the reduced losses would result in savings of about $10,000 for a 1,500-hp motor. If, for simplicity, it is assumed that the same improvement can be obtained in all machines above 500 hp with 95% efficiency, including synchronous and DC, the energy that could be saved by using HTSC machines is 1.22 x 10 kWh, or a dollar savings of It is important to note that these calculations do not include $61,000,000 per year. electrical machines in the electric utilities, government laboratories, or municipal water works, where the use of high-horsepower machines is very substantial. Hence, the savings could probably be safely inflated by at least 50%. Other advantages in addition to energy savings are possible. In particular, the absence of laminated iron in the rotor would make the overall machine lighter in weight. And, since losses are substantially reduced, the power density could be increased accordingly, making for a more compact machine. However, while these considerations are important for special applications, such as land transportation or marine power, they would be of lesser importance to the general industrial market.
APPLICATION CONSIDERATIONS FOR HTSC MACHINES While electric motor operation is not, in principle, substantially different than electric generation, numerous application considerations make motor construction Of particular concern is motor starting performance. In contrast to a difficult. generator, which is brought to synchronous speed by the prime mover, a motor must generally be started with AC power. A reduced-voltage start is usually used to bring the motor up to speed over an interval of several seconds. Often the motor must accelerate a connected load, which greatly increases the acceleration period. Large forces are experienced by the windings in the machine, and heat builds up rapidly due to the high inrush of current. Since the HTSCs have been categorized as brittle, the forces induced by repeated starts could break the coils. In addition, the high starting currents
Motors
107
could
prevent
the
The characteristics of the load also greatly affect the feasibility of using an HTSC machine, again due to the brittle nature of the ceramic material. For example, servo-type applications (steel mill drives, dynamometers, and machine tools) place severe stresses on the windings of machines used for such purposes, and the rotor windings normally must be strongly braced. It is uncertain whether the HTSC rotors required for Also, in these and other applications, the load an AC machine could endure the stresses. torque connected to the motor shaft varies rapidly in almost stepwise fashion. These sudden load changes would cause rapid speed changes, which would, in turn, induce rapid changes of current in the SC field coil. Such sudden changes could bring the coil continuously out of superconductivity, causing extra losses and a severe heat-transfer problem.
HTSC DC MOTORS Probably the most researched SC machine is the DC homopolar machine. The machine employs a rotating disk inside an SC coil (Faraday disk). The disk cuts the flux created by the coils, and the resulting DC voltage induced between the inside and outside Since the machine has, effectively, one turn, edges of the disk is picked up by brushes. The situation can be improved by using a the voltage induced is inherently low. segmented rotor, but the number of brushes increases in proportion to the number of segments.8 While construction of the SC portion of the DC homopolar machine is relatively straightforward, the design of the armature is more difficult (particularly with regard to the brush pickup problem, which requires many brushes supporting a high current density). Sodium-potassium and mercury liquid metals have been used in some prototypes, but without complete success (due to contamination problems). One of the important advantages of an SC homopolar motor is the fact that the machine can be scaled up to very large sizes and high speeds (up to 200 MW at 2,000 rpm) while the conventional DC machine is limited to much smaller values (10 MW at 150 rpm).* Such applications, however, are very specialized (such as ship drives). In general, most conventional DC machines are employed in variable-speed applications, such as in steel mills, where the speed of response is often an important criterion. Since the Faraday disk develops voltage by either rotating the disk at a high speed or, conversely, making the disk have as large a radius as possible, the geometrical shape of this machine is somewhat at odds with the normal requirements (which tend to favor a Nonetheless, the inertia is lower than in a machine with a relatively small diameter). normal commutator-type DC machine (due to the absence of an iron core), and it has low These characteristics are ideal for armature inductance and zero armature reaction. high-performance, mill-type applications. While the opportunities for replacing large, conventional DC machines appear to the future for DC machines in general is not as promising, due to the nagging As a result, variable-speed AC motor drives using of brush maintenance.
be bright, problems
108
Applied Superconductivity
solid-state power converters to change the frequency of the AC supply from a fixed 60-Hz value to a variable frequency have made continuous inroads on the DC motor market for the past 20 yr. This trend is expected to continue until DC motor drives Since the problems of brush constitute a very small, special-purpose market. maintenance are only aggravated with homopolar machines, the long-term opportunities for homopolar machines for anything other than special purposes is not promising.
RTSC SYNCHRONOUS
MOTORS
Because AC machines form the greater part of the market for motors above 500 hp and AC motor drives are largely replacing DC motor drives, the application Probably the most opportunities for HTSC AC machines appear more widespread. apparent application of HTSC is in the synchronous motor, where a rotating SC coil is used to replace the usual excitation from a wound-coil field. As mentioned previously, an important difference between motors and generators is that motors must develop adequate torque to self-start themselves as well as their connected load. Hence, operation directly from an AC supply will require a squirrel cage for starting, since a shorted field winding is incapable of supplying adequate starting The field coil would then be excited upon synchronization in much the same torque. manner as for conventional synchronous machines. Starting could possibly be accomplished by means of a room-temperature copper shell wrapped around tt; superconducting cylinder in much the same manner as in SC synchronous generators. The shell would also serve to shield the magnet from forces due to AC flux variations during starting. Forces would also be exerted on the shell during starting, making the design of this motor structure more demanding. It appears, however, that an iron core rotor would probably still be necessary to create the large forces necessary for adequate starting torque, severely restricting the maximum excitation achievable with the SC field coil and making the overall approach less attractive than for HTSC synchronous generators. Thus, it appears that more conventional stator and rotor structures having conductors embedded in slotted iron cylinders may be the preferred approach for motors, at least when the machine is started off the line. The problems associated with starting the machine could be eliminated if the machine were brought up to speed with a solid-state frequency changer, as shown in Fig. 8.2. In this case, the SC field could be energized with the rotor stationary. With the field then shorted, the field current would remain constant (assuming a constant load), and slip rings or other exciter types could be eliminated completely. Since the rotor always remains in synchronism with the rotating stator field, the starting cage could also be eliminated or substantially reduced, since the shield may still be needed to short circuit the harmonic currents flowing in the stator due to the power converter. Finally, since the starting torque is obtained by synchronous rather than induction motor action, the slotted iron structure could be dispensed with and the excitation of the SC field increased to levels comparable with the SC generator. If a power converter is engaged for the starting process, the same power converter could also be used in a continuously variable speed mode to optimize the
Motors
109
3-Phase AC Supply
Changer
for Accelerating
an HTSC
process (for example, a compressor). However, if continuously variable speed is not required, then upon reaching synchronous speed the machine could be transferred to the line and the converter bypassed. If the rotor cage were not used, system damping could then become a problem. The problem may, however, be eliminated by providing damping from the stator side. One possible method of providing active damping is by inserting back-to-back thyristors in series with one, two, or all three of the stator lines, as shown in Fig. 8.3. The system could be damped by measuring the fluctuations in stator input power or input kVA due to rotor oscillations and controlling the thyristors to reduce these fluctuations. Because power need not be fed to the field coil continuously, elimination of the slip rings and rotating exciters would be an important advantage of an HTSC synchronous machine. However, since the field remains short-circuited, it cannot be adjusted with load, and hence, the power factor cannot be controlled. When the field becomes superconducting and is short-circuited, flux is trapped in the coil and, by the law of constant flux linkages, cannot change thereafter. In effect, when the load changes, field current is induced in the SC coil so as io exactly cancel any change in stator flux linking The machine behaves continuously, even in the the field coil due to the load change. steady state, according to the constant voltage behind transient reactance model.1 The problem of induced field current affecting the power factor is probably not of concern with an HTSC synchronous machine operating from a converter supply. In this case, the machine could be continuously controlled with the frequency converter such that only the component of stator magneto motive force (MMF) orthogonal to the field axis (q-axis) was allowed to change, while the stator MMF in the field axis (d-axis) remained constant. Since control of the power factor is often an important application issue when the machine operates off the grid, it would be useful to develop a means for adjusting the excitation of a machine without recourse to a separate rotating exciter. In the approach shown in Fig. 8.4, phase-back of the same thyristors used for damping in Fig. 8.3 is used to create a small negative-sequence current component, which rotates backward in the air gap of the machine, resulting in a 120-Hz field component as seen from the rotor. Rather than being short-circuited, the SC field coil is connected to a simple diode
110
Applied Superconductivity
AP 3 -Phase AC
SUPPlY
A
Optional
Stator Optional
HTSC Synchronous
Machine
the Speed Osciitions without a Rotor Shell
FIGURE 8.3 Method for Stabilizing of an ETSC Synchronous Generator for Damping
3-Phase AC Supply
Rotor Motor
HTSC Synchronous
FIGURE 8.4 Method for Inducing Excitation Coil without a Rotating Exciter
in an SC Field
The bridge is excited by a smaller coil operating at room temperature. The bridge. rectified induced voltage, in turn, is used to adjust (increase) the field current. Because the presence of the diode and extra coil introduces losses, the field current would decay slowly by free-wheeling through the diodes if the winding is not pulsed repeatedly. However, the losses incurred can be made very small, because the conductor of the additional coil can be made with a few turns of large cross section while the conductor of the SC coil has a much larger number of turns. It is important to note that this method can be used only to increase the field current, since the process of decreasing the current relies solely on the diode conduction drops. Fortunately, in many cases, relatively slow changes in excitation are all that is
Motors
111
While the problem of rapidly reducing the field current could be completely desired. solved with light-triggered devices, this case is, perhaps, awaiting a more practical Also, the process of inducing a current in the field winding implies good solution. coupling between the armature and the field winding, which in turn implies a Hence, the very high field strengths conventional laminated iron stator and rotor. At present, the obtained by the SC synchronous generator cannot be approached. rotating exciter approach (which has its own loss problems) appears to be the most practical method for exciting a machine in which continual adjustment of excitation is required.
RTSC INDUCTION
MOTORS
While superconducting DC and AC synchronous machines have been fairly well researched, other machines also hold promise for the future. Chief among these is the HTSC induction motor.12s13 Because the rotor resistance of an induction motor must be nonxero to develop torque, this concept appears to be a contradiction in terms. The principle of operation, however, is only to start the machine by induction motor torque. When the motor reaches the vicinity of synchronous speed, the current induced in the rotor drops rapidly. As the current drops, the rotor conductors cool rapidly to the point where they become superconducting. The rotor flux at this instant is trapped, and the machine becomes an HTSC synchronous machine. Hence, the starting torque function and the excitation function are combined in the same winding. In contrast to superconducting synchronous machines, which (due to their high field strength) do not require iron in the rotor or even the stator, the iron path of an SC induction motor will still be needed if starting is off the AC mains. Again, starting from a converter supply could alleviate the problem of high inrush currents and perhaps permit an ironless rotor construction. However, the machine then need never operate as an induction motor, and the desirability of this machine structure for such an application is questionable. When the HTSC induction machine reaches the superconducting state and the rotor resistance drops to zero, stability again becomes a problem. As an example of the difficulty expected to be encountered, Figs. 8.5 and 8.6 show the acceleration of a lOOhp induction motor off the AC mains; in Fig. 8.5, the rotor resistance remains constant, while in Fig. 8.6, the rotor resistance drops to zero at 0.98 per unit speed. Continuous strong oscillations in the speed of the superconducting machine can be observed. Again, these oscillations could possibly be damped with thyristors in series with the AC line, as shown in Fig. 8.3. Chief among the problems associated with this machine is the requirement to achieve the superconducting state at precisely the right moment during run-up. If this state is achieved too early, high pulsating torques will appear due to the machine slipping poles; if achieved too late, very little current will be retained in the SC rotor winding, and the machine may still require excitation power from the stator side as it continues to operate under load. Vagaries such as the degree of loading during the start, the linevoltage amplitude during the start, and even ambient temperature will affect the amount of heat generated in the rotor bars during a given start and thereby indirectly affect the instant of superconductivity. The prospect for success with this machine seems very problematical.
112
Applied Superconductivity
1 0.8 Z d s z P 2 .$ z 0" E e % W -0.6 -0.8 -1 0 0.4 0.8 1.2 1.6 2 -0.2 -0.4 0.6 0.4 0.2 0
1.6
Time
(s)
Electromagnetic
0.4
0.8
1.2
1.6
Time
(s)
Motors
113
HTSC INDUCTION/SYNCHRONOUS
HYBRID
Another type of machine that has been proposed is a true induction/ synchronous machine hybrid in which the synchronous rotor is located coaxially within the induction machine, as shown in Fig. 8.7. The induction rotor is connected to the external load, while the synchronous rotor rotates freely. The induction rotor becomes, in effect, a shell within which the synchronous rotor rotates. Each rotor rotates independently, so that the induction machine rotor slips with respect to the synchronous Construction problems associated with machine rotor, which rotates synchronously. supporting the rotating induction motor shell are immediately apparent. Also, it should be noted that since the slip losses are still required if the machine is to drive the load, the rotating field excitation of this type of machine only serves to correct the power factor. While the stator current is reduced somewhat (from about 0.9 to 1.0 pF), the energy saved does not appear to be substantial.
HTSC RELUCTANCE
MOTOR
The HTSC reluctance motor is related to the HTSC synchronous motor in much the same manner as their room-temperature counterparts are related. In this case, the SC coil is not energized from an external source, and torque production is obtained by an The principle can be explained by equivalent saliency effect due to the SC coil. referring to Fig. 8.8. Note that the so-called q-axis is encircled with an SC coil. Assuming that the machine is operating at synchronous speed and no load, the flux produced by the stator is located in the direct axis or maximum permeance axis. When the machine is loaded, the rotor is retarded and moves towards the q-axis. As it does so, current is induced in the SC coil such that the total flux linking the coil remains zero (the value of flux linkage at the instant of achieving superconductivity). Since no stator flux appears to link the rotor q-axis circuit, the machine appears to have a very small permeance or inductance in the q-axis. The current in the SC coil (q-axis) now reacts with the stator flux component remaining in the d-axis to produce torque. Conversely, it
k
h
Stator (Stationary)
Induction Squirrel Cage (Rotates at slip speed with respect to SC field winding)
114
Applied Superconductivity
Qma,
-q-aX,s
t
d-am
of a Superconducting
can be said that the d- and q-axis due to the difference in saliency theory).
to produce reluctance
torque motor
The HTSC reluctance motor is an interesting concept, since external excitation of the rotor windings is unnecessary. Unfortunately, the current induced in the rotor windings is necessarily demagnetizing, thereby ensuring lagging power-factor operation. Good coupling is needed to induce a reasonable amount of SC coil current, so conventional stator and rotor iron structures are mandated. The motor cannot be started without an extra cage or without the help of a variable frequency converter. Finally, the instability problem, already discussed for the HTSC synchronous and induction machines, is also present here. The problem must again be resolved by operating continuously from a frequency converter or inserting inverse/parallel thyristors in series with the AC line (Fig. 8.3). These problems seem to indicate that this machine, while conceptually interesting, is perhaps better suited to lower-power applications (l-25 hp). Further development should probably await the appearance of room-temperature SCs.
HTSC HOMOPOLAR
INDUCTOR
MOTORS
Inductor-type machines are another class of AC machines that have, over the years, been proposed as a possible replacement for conventional synchronous machines, particularly generators.15 The basic principle of an inductor-type machine is to create a pulsating unidirectional field in the rotor. The AC component of this field couples with the AC armature winding field to produce torque. Both radial and axial air gap machines of this construction have been investigated, but apparently not with an SC field coil in mind. Figure 8.9 shows an idealized representation of one possible geometry for an axial air gap machine. The machine is excited by a circular SC field coil inserted between protruding poles staggered on alternate sides of the coil. Pairs of protruding poles face each other across the air gap. The poles are fastened to the rotor and rotate while the
Motors
115
FIGURE
8.9
Axial Al
Motor
SC field coil remains stationary. Also located in the air gap are AC armature windings, which couple with the alternating component of the air gap flux produced by the field. This alternating flux induces currents in the armature windings that react with the field flux to produce torque. The armature windings are also stationary, so that the only member that rotates is the member containing the protruding iron poles. As an alternative, the armature coils could be placed in slots or fastened on the surface of a stationary iron member such that only one set of protruding poles rotate to form the rotor of the machine. In this case, the saliency of the protruding poles is somewhat reduced, but assembly is probably simplified. An obvious important advantage of this scheme over all other HTSC AC machine arrangements is the fact that the field coil remains stationary. Thus, SC coil-cooling problems are simplified enormously. While the machine is not inherently self-starting, starting could be readily accomplished by building a squirrel cage in the protruding poles. Assembly would be simplified and iron losses reduced, since the rotor poles could Since the coils need not be placed in slots, be constructed in tape-wound fashion. assembly of the armature could be simplified as well. It is important to note that the principles of constructing a superconducting reluctance machine could also be used to synthesize the protruding poles by rotating SC coils, which would, in effect, produce saliencies without the presence of any iron For example, the SC coils could be inserted between the protruding poles of member. the rotor to greatly improve the saliency of the structure, and thereby its energyAlso, the protruding iron poles could be dispensed with completely conversion ability. and the SC coil principle used to create an effective rotor saliency similar to that of the Similar principles could also be used to construct a more HTSC reluctance motor. Given the substantial problems facing the design of conventional radial air gap machine. HTSC synchronous, induction, or reluctance machines, the HTSC inductor machine has many useful features that warrant further investigation.
116
Applied Superconductivity
CONCLUSIONS This supplement has summarized the status and future prospects of HTSCs as It has focused on the features of motor operation that applied to motor technology. make the design problems substantially different than those for generators. Problems of developing adequate starting torque, inducing an adjustable field current, and overcoming speed instability have been identified and discussed. It has been suggested that the problems associated with HTSC induction machines will probably preclude their use, while the hybrid, concatenated synchronous/induction machine apparently will not provide sufficient benefits to pursue its development. Although the HTSC reluctance motor may be viable, its inherent low power factor will limit its field of application to smaller machines (where poor power factor is of less concern). The HTSC synchronous motor is more promising. However, the machine may be limited to operating with an auxiliary inverter for starting purposes, unless the problems of designing a machine with adequate starting torque can be worked out. Of all the machines considered, the homopolar DC machine appears to be the most suitable for motor applications and is also Unfortunately, the need for DC motors is at a relatively advanced state of development. small and diminishing. Finally, the potential of the HTSC homopolar inductor AC machine has been presented and discussed. It is suggested that the unique features of this AC machine make it a candidate for a more detailed investigation.
REFERENCES FOR SUPPLEMENT 1. A.D. Little, Inc. (Cambridge, Mass.), Energy Efficiency prepared for U.S. Energy Research and Development Contract CO-04-50217-00, p. 27 (May 1976). A.D. Little, Inc. (Cambridge, Mass.), Energy Efficiency prepared for U.S. Energy Research and Development Contract CO-04-50217-00, p. 26 (May 1976). A.D. Little, Inc. (Cambridge, Mass.), Energy Efficiency prepared for U.S. Energy Research and Development Contract CO-04-50217-00, p. 1 (May 1976). Classification and Evaluation of Electric Motors and Pumps, Report DOE/TIC-11339, pp. 3-15 (Feb. 1980). A.D. Little, Inc. (Cambridge, Mass.), Energy Efficiency prepared for U.S. Energy Research and Development Contract CO-04-50217-00, p. 29 (May 1976). A.D. Little, Inc. (Cambridge, Mass.), Energy Efficiency prepared for U.S. Energy Research and Development Contract CO-04-50217-00, p. 28 (May 1976). and Electric Motors, Administration under
2.
3.
4.
5.
6.
Motors
117
7.
A.D. Little, Inc. (Cambridge, Mass.), Energy Efficiency prepared for U.S. Energy Research and Development Contract CO-04-50217-00, p. 36 (May 1976). A-D., Motors, Generators and Flux Pumps, Bulletin Annex 1969-1, pp. 207-230 (1969).
Motors, under
a. Appleton,
London,
9.
I.I.R.,
Commission
1,
Stevens, H.O., and M.J. Cannell, Acyclic Superconductive Generator 400 Horsepower Generator Design, David W. Taylor Naval Ship Development Center Report PAS-al/14 (Oct. 1961). Mole, C.M., H.E. Hall, and Proc. Applied Superconductivity D.C. Litz, Superconductor Synchronous Conf., pp. 151-157 (1972). Theory of Alternating
10.
Generators,
11.
Adkins, B., and R.G. Harley, The General Chapman and Hall, London (1975).
Current
Machines,
12.
Levi, E., and M. Panzer, Electromechanical Hill Publishing Co., New York, p. 431 (1966). Brechna, H., and H. Kronig, Three-Phase Cage Winding, IEEE Trans. on Magnetics,
Power Conversion,
13.
14.
Bonwick, W.J., and A.L.D. Ah Fock, internally Energized Induction Machines, Proc. International Conf. on Evolution and Modern Aspects of Induction Machines, Torino, Italy, pp. 418-423 (July a-11, 1986). Bateman, Systems, J.T., Trans. A Solid Rotor AC Generator for High Temperature Electrical of the AIEE (Applications and Industry), pp. 400-405 (Jan. 1960).
15.
Supplement: Estimates for High-Gradient Magnetic Separation of Oxygen from Air S.A. Zwick
Argonne National Laboratory
118
Industrial
Separations
and Material
Handling
119
Summary
Sections 9.1 and 9.2 deal primarily with the application of superconductivity to highgradient magnetic separation (HGMS), a technique that has been applied for many years to the separation of solid particles (e.g., dust, pyritic impurities in ground coal, and tiny steel filings in blast-furnace effluent) from gaseous or liquid streams. Ferromagnetic or paramagnetic particles stick to the magnets and can be removed bodily from a stream. A related process, open-gradient magnetic separation (OGMS), may be applicable to paramagnetic gas (e.g., O2 and NO) separation (see Sec. 9.3). Other examples of HGMS applications for separating magnetic impurities from process flows include removal of ferrous contaminants in the food processing, chemicals, and pharmaceuticals industries; desulfurization of coal; and boiler feedwater treatment. Most recently, Eriez Magnetics developed a 4-K HGMS device that has been installed for kaolin processing at the J.M. Huber Corp. To date, this is the only known U.S. industrial application of superconductivity. The efficiency of HGMS systems would be increased further by the introduction of high-temperature superconductors (HTSCs). These systems permit much more intense fields than iron-based magnetic systems but do not entail the Joule heating losses. In Sec. 9.1, a preliminary economic comparison of conventional and 77-K superconducting HGMS systems indicates the following: 1. HGMS with 4-K
The annual operating costs (including capital) for a 4-K superconducting HGMS are about 8% lower than those for a conventional HGMS. Operating costs for power consumption are reduced by 80%. The annual operating costs (including capital) for a 77-K superconducting HGMS are 15% lower than those for a 4-K HGMS and 20% lower than those for a conventional HGMS. The power operating costs for 77-K HGMS are about 7% of the 4-K HGMS costs and about 98% of the conventional HGMS costs.
2.
In Sec. 9.2, a preliminary estimate by B.W. McConnell of the operating cost savings (excluding capital costs) shows a savings of 97% for a 77-K superconducting HGMS system compared with a conventional HGMS system. In addition, McConnell points out that a main advantage of superconducting HGMS is the ability to separate small-diameter and weakly magnetic particles that cannot be separated by conventional magnets. Inasmuch adaptation as HTSCs inherently offer a savings in operating costs, it is clear of existing methods would be economical, as the discussion that a makes
direct
120
Applied Superconductivity
clear. There might also be unexpected advantages in using unconventional designs. Thus, the cost analysis presented assumes that the liquid nitrogen (LN3) coolant in an HTSC HGMS plant would be discarded (as liquid helium currently is, due to handling problems). But LN9 can be piped, stored, and recycled, which should provide additional savings and In addition, it might be possible to apply convenience, as well as economies of scale. unusual magnet designs to the new systems. Section superconducting considered. 9.3 presents a technical discussion OGMS for gas/gas separations, an area of the feasibility of that has not previously using been
121
9.1.1
Introduction
Superconductors, and especially high-temperature superconductors (HTSCs), have a wide range of potential applications in U.S. industry. The applications discussed in this section fall into two broad categories: materials separation and other applications.
9.1.2
Materials
Separation
Over the past 15 yr, U.S. industry has adopted high-gradient magnetic separation (HGMS) for a variety of industrial processing applications. The first industrial application of HGMS is credited to the J.M. Huber Corp. in 1969 for processing of clays.1 Since then, HGMS systems have been commercially offered and installed in industrial applications ranging from minerals processing to removal of paramagnetic contaminants in the food processing and pharmaceuticals industries. Table 9.1 lists some applications for high-gradient magnetic separators. In general, high-gradient magnetic separators offer the potential for higher product purity and reduced operating and maintenance costs relative to alternative chemical, physical, The field strength of most or gravity separation processes. commercially available HGMS systems for industrial applications is about 2 T. Higher field strengths would be expected to lead to higher separation efficiencies and/or higher flow velocities. Typically, however, the incremental increase in separation effectiveness tends to diminish as field intensity increases, whereas the incremental cost of the system increases as field strength increases. The trade-off is compounded by the fact that separation effectiveness decreases at an increasing rate as flow velocity increases. In any given application, there is a trade-off among field strength, capital cost, separation effectiveness, and flow velocity. Commercially available equipment operates ,:iquid streams (both aqueous and nonaqueous), solid dry materials, solid wet materials, and gaseous streams containing entrained paramagnetic particulate matter.6 For example, a 2-T, dry solids HGMS system offered by Carpco (Fig. 9.1), is applicable for a range of applications that includes separation of natural diamonds from garnets, purification of stainless steel powder, removal of paramagnetic impurities from minerals, etc. The field strength of the Carpco machine can be adjusted to a maximum of 2 T by an autotransformer for the magnet-coil input voltage, depending on the magnetic susceptibility of the particles being processed. This allows the user to control flow rates and separation effectiveness as required in the specific application. Most recently, Eriez Magnetics installed separator in an industrial application. the first superconducting The 2-T system is being high-gradient installed at the
magnetic
122
Applied Superconductivity
Magnetic
Separators
Industry
or
Process
Separation Removal of bauxite, limestone, Removal of
Specific
Application
Minerals
processing
of diamonds from garnets weakly magnetic materials calcite, clay, feldspar, manganese, zircon, etc. ferrous contaminants
from glass,
alumina, sands,
and
Removal
of
ferrous
contaminants
from
liquid
flows
Desulfurization of coala Removal of particulate matter Treatment of boiler condensate Recovery Recovery Recovery of of of
from flue gases and process water solid waste chips shredders
Recyclingb
aluminum from municipal titanium and superalloy ferrous metals from car
aAlso
demonstrated
with
bTheoretically
possible
J.M. Huber clay processing plant at Athens, Ga. (the same facility credited with the first commercial application of conventional HGMS). The operating characteristics of this system are presented in Fig. 9.2.
A summary comparison of the principal characteristics of a conventional Z-T the Eriez 2-T superconducting separator, and a separator for clay processing, hypothetical 77-K separator is presented in Table 9.2. According to the Eriez product literature,3 a conventional water-cooled system would have a total power requirement of 300 kW. In comparison, the power requirements necessary to maintain the field strength in the superconducting system are negligible. However, in the 4-K system a helium reliquefier is incorporated to eliminate requirements for liquid helium (LHe) makeup. Thus, the 4-K system is a closed system, for which the parasitic power requirements are 60 kW. The total cost of the 4-K system is about $1.7-1.8 million, of which 12-16% is for the refrigeration system. The cost of the 2-T system for the J.M. Huber clay processing plant is $2 million. However, this design includes reliquefaction capacity for two separators. The Fast of the refrigeration/reliquefaction system is estimated at 25-33% of the total cost. Thus, if the refrigeration/reliquefaction system were sized for a single separator,
123
ture requires P megnebc lace which is muchgreeterthen can be producedby uxw&onal permanent magn&. This magnetic force~stheprcdwtoffiekl imeneityacd therateofchangeofthis field OVB~ distance (magnetic gradient). The saparab& of dry materials on acaminuous basis is acmmpllshed the Carpco by Meter Magnetwdh a combnation of magneticforce and correct feed prBSBmBtt0n. Sulteble magnetic force IS achi by placing a roll made of alternatemagneec end nonmagneticzones (Fig. 1,Section AA) betweenspewally shaped poles (Fig. 1) of a pm.wful electrc. magnet. The efectmmagtwt ix&es magne4tc fields about the magneticlaminations of each mll formmg local regions of hgh magneticintensfty and sharp gradients as sh-wn by hypothetical lines of magneticforce (Fig. 1. Section A-A). Correct feed presentataon achievedby corneying the matenal IS to be separated from the surge hopper into the separetion .?one (Fag. 2) by means of a velocity feed system. This poeWe feed system itxorpaates a &xly feeder unique ,n that it ecceferetes and presentsthe matenel omothemwrq roil without bounceand Into the separation zone at the optimum vefoaty for efficlem sefxreoon. In the eeperat~on zone (Fig. 1). weakly magnebzable matenalIS attractedto the roll by magneticforce. The attrected pan~clesan,carnedbythesurfaceofthemllwtofthestreamof metenaland~ntOareglondlowintensltywhen,theyeltherfallotf or am brushed from the roll. Nonmagnebc particles unaffected by the megneoc field wll follcw a natural profectilepath awayfrom the roll and separabon zone. Middling grens report to an memwdmte location tf a middling pmduct IS dewed.
Figure1
Design Features
High capacity IS made possible by the use of Carpcos patented holkwmll cowtrcbo. The hol!o.v.rolldesign overcomes the engmeenng Iimitatnns [mailmum rc4 length 0.75 m (ZO)] of conwnt~onalsolidcore roll designs. and rolls of onemeterfengmares~.Anaddedbenemist~~~tlon of energy necessary to turn each roll wthln the high magnettc field. Patented velocity feeding(1)makes use of the materlals natural pqectile motion for unhindered passage through the megnetn separation zone. The velocity feeder provides a un~fotmfeedwiUwtbouncingandndscartecingofpaWesover the mll and Yields a s~gmfiiant imprwement onseparatuon efficiency (up to 2U%i compared to conventional feeder designs A weftthn electro-mechanfceI (2) primary feeder
FKX.JRE 9.1
Carpeo
High-Gradient
Magnetic
Separator
(Source:
Ref.
4)
124
Applied
Superconductivity
WIDE
RANGE
OF
APPLICATIONS
Wherever a very high magnetic f,eld and very low power consumption are desired. superconduct,ng magnets will see Increasing use Any liquids containing paramagnet,c elements can be processed. and dry sol,d materials can be exposed to the 50 kilogauss field to investigate changes that may take place in th,s condit,on Current applications Include purification of kaolin clay. separation of finely.ground pyrite (iron sulfide) from coal removal of catalysts from oil. processing of chemical compounds and waste water treatment OPERATION The superconductlng HGMS laboratory model is compact. safe and easy to use Highly trained technical personnel are not required for daily operation The magnet IS basically a circular iron-clad solenoid wh,ch operates at a temperature close to absolute zero (0' Kelvin). cooled first by liquid nitrogen and then by liquid helium A canister packed with a matrix of magnetic stainless steel wool. through wh,Ch the material being processed flows. is placed In thewarm bore in thecenter of the circular coil The warm bore is maintained at room temperature so the material is not affected by the cryogenic cond,tion of the magnet The magnetic field strength can be continuously varied during operat,on by a simple potentiometer The high heat.absorbing capab,lity and low cost of liquid nitrogen make it the choice to reduce coil temperature from ambient to 77' Kelvin Helium gas is then used to blowout the nitrogen before liquid helium is pumped into the chamber surrounding the coil The helium further reduces the temperature to below 10' K. at which polntthemagnet is In a superconducting state A small quantity of liquid nitrogen IS supplied to a chamber at the bottom of the magnet to insulate the liquid helium
Canisters of varying diameter depending upon the capacity desired. are secured in the warm bore in the center of the superconducting laboratorr model HGMS The supports shown for canister insertion and removal are provided as part of the separator In a recent series of tests on kaol,n clay slurry liquid nitrogen usage In the laboratory Unit was 0 18 I,ters per hour and liquid hel,um was consumed at the rate of 10 liter per hour For commerclaloperation a closed loop liquefying system is used to reduce helium consumption to virtually zero
High-Gradient
Magnetic
Separator
125
TABLE 9.2
Summary
Comparison
of High-Gradient
Magnetic
Separation
Systemsa
Item
Conventional system
4-K Superconductor
77-K Superconductor
Power requirements (kW) Field Cooling Total Weight (tons) Footprint (ft2) Capital cost ($106) Annual operating cost ($103jb Annual capital cost ($1031c Total annual cost ($103)
field
strength
is 2 T for capacity
all
three factor
price
= 5c/kWh,
charge
rate
= 26.6%.
the separator unit cost would be about $1.7-1.8 million, with the refrigeration/reliquefier accounting for 12-16% of the separator costs. The costs were adjusted to a single separator and reliquefier to facilitate comparison with a conventional HGMS system and a hypothetical HTSC HGMS system. The advantages 1. of an HTSC HGMS system would be as follows:
Parasitic power requirements for the LHe reliquefier would be reduced, because helium would not be needed or used. Indeed, given the cost differential of helium and nitrogen, the system would likely be redesigned such that any nitrogen boil-off that might occur would not be reliquefied, and parasitic power requirements would virtually be eliminated. Even if it were costeffective to reliquefy nitrogen, the parasitic power requirements would only be about 5% of those for reliquefying helium. The boil-off of liquid nitrogen (LN2) would be expected to be lower than that of LHe, since the system would operate at a higher temperature. Given that nitrogen is relatively inexpensive, an
2.
126
Applied Superconductivity
open-loop system would likely be cost-effective and would eliminate the capital cost of the reliquefier system -- a reduction of about $200,000. Because the primary advantage of a superconducting HGMS is the reduction in power requirements relative to a conventional system, the cost-effectiveness of the system depends on the price of electricity and the operating capacity factor of the system. Given the economic assumptions prescribed for this analysis, the 4-K system is cost-effective relative to the conventional HGMS system at a capacity factor of 50% (Table 9.2) and would be cost-effective at a capacity factor as low as about 20% (Fig. 9.3). The additional reduction in power requirements and the reduction in capital cost for a 77-K system relative to a 4-K system would make the former cost-effective relative to a conventional system regardless of the capacity factor. Superconducting magnetic separation systems have been investigated for a range of applications in industry.2V8-10 In general, the advantages of superconducting magnetic separation systems are as follows:
1.
Reduced power requirements. Reduced weight and volume, due to elimination of the soft iron core and the compactness of windings relative to those of conventional systems. Higher field strengths, which allow for higher processing velocities for a given separation effectiveness or higher separation effectiveness for a given processing velocity.
2.
3.
Offsetting these above advantages are the initial capital cost premium for the superconducting system and the perception of unreliability associated with operating a cryogenic system. However, superconducting systems have begun to be accepted by industry, and if the growth of conventional HGMS systems since 1969 is duplicated by 4-K superconducting HGMS, U.S. industry will represent a significant market for HTSCs as they become commercially available.
9.1.3 Materials Handling and Fabrication Other applications for superconducting magnets in U.S. industry would include materials handling and materials fabrication. For example, circular lifting magnets are used for handling steelworks scrap, loading steel into a melting furnace, etc. Magnetic products are also handled by means of battery-powered magnets mounted on the forks of forklift trucks. It would appear that the primary advantage of superconducting magnets in these applications would be potential reductions in magnet weight, which would allow the overhead crane or fork supporting the magnet and load to carry a greater load, thereby increasing productivity. The reduction in power requirements could be particularly beneficial to forklift-mounted magnets.
127
600
400
5
3 -
I
IVV ,
Note: Levelized Electricity Price = $O.O6/kWh Fixed Charge Rate on Capital = 26.6%
210
6'0
Id0
HGMS
FIGURE 9.3 Annual Operating Costs of Alternative 2-T Systems for Kaolin Processing at 500 gal/min
In materials fabrication, magnets are used to press-fit parts in the automotive industry.* Superconducting magnets could reduce the energy requirements for such operations and could perhaps be applied to a wider range of materials-joining operations that are presently accomplished mechanically (e.g., bolting) and/or thermally (e.g., shrink fitting), if magnets of higher field strength were available.
9.1.4 1.
References
Oberteuffer, J.A., and B.R. Arvidson, General Design Features of Industrial High Gradient Magnetic Filters and Separators, in Industrial Applications of Magnetic Separation, Y.A. Liu, ed., IEEE Catalog #78 CH1447-2 (1978). Cryomagnetics, Tenn. (undated). Inc., A Guide to Superconducting Magnets and Systems, Oak Ridge,
2.
128
Applied Superconductivity
Inc.,
Magnetic
Sepamtor,
Brochure
OTB-726,
High-Intensity Inc., Meter Magnet 6550, Jacksonville, Fla. (undated). and C.H. Chilton, Chemical Book Co., New York (1973).
Induced-Roll
Magnetic
Separator,
Engineers
Handbook,
21,
Roy, N.K., M.J. Murtha, and G. Burnet, Recovery of Iron Oxide from Power Plant Fly Ash by Magnetic Separation, in Industrial Applications of Magnetic Separation, Y.A. Liu, ed., IEEE Catalog #78CH1447-2 (1976).
7.
Merwin, 1987).
R. (Eriez
Magnetics,
Inc.,
Erie,
Penn.),
personal
communication
(July
a. Doctor,
R-D., C.B. Panchal, and C.E. Swietlik, A Model of Open-Gradient Magnetic Separation for Coal Cleaning Using a Superconducting Quadrupole Field, Recent Advances in Separation Techniques - III, N.N. Li, et al., eds., American Institute of Chemical Engineers Symp. Series, 82(250):154-166 (1966). Watson, J.H.P., and D. Hocking, Magnetic Sepamtor, IEEE Trans. The Beneficiation of Clay Using a Superconducting Magnet&, MAG-II:1956 (1975).
9.
10.
Price, C.R., and W.F. Abercrombie Jr., Practical Aspects of High Gradient Magnetic Separation, in Industrial Applications of Magnetic Separation, Y.A. Liu, ed., IEEE Catalog #76 CH1447-2 (1976).
129
9.2.1 Introduction The recent discoveries of materials that are superconducting at temperatures above the boiling point (77 K) of LN2 may allow the development of apparatus with significantly These higher operating efficiencies and, hence, greatly reduced operating costs. materials also have the advantage of remaining in the superconducting state at significantly higher magnetic fields than previously seen in Type I and II superconductors. At present, these high-temperature superconductors (HTSCs) appear to be extremely brittle and have a low current density (nominally 100 A/cm). However, reports of wires and ribbons being fabricated from the materials offer hope that Also encouraging is IBMs announced potential fabrication problems can be solved. increase of the current density in thin films by a factor of 100. The use of LN2 as a coolant implies immediate economic advantages over the previously required LHe. LN2 is considerably less expensive, because the basic raw material is free and the production process is considerably more efficient. In fact, the process is so inexpensive that the operation of HTSC apparatus at LN2 temperatures may well be considered for other technical reasons even if higher-temperature superconductors are found. This section presented an evaluation of one technological application of the new Magnetic separation, while presently applied to very HTSCs -- magnetic separation. specialized situations using both conventional conductors and LHe superconductors, has The potential ability of HTSCs to remain in the not seen wide application. superconducting state at high magnetic fields makes this application particularly attractive. This evaluation is based on the following general assumptions: 1. Extension of previous designs using LHe superconductors HTSC operating region is possible, These materials will prove no more difficult working configurations than existing applications Adequate bulk current carrying capability to the
2.
3.
can be obtained.
In addition, the best technological estimates of realistic improvements in operating efficiencies consistent with other engineering constraints are applied where
130
Applied Superconductivity
possible. No credit is taken for the higher heat capacities or the greater thermai operating range present at LN2 temperatures. These latter credits may well offer further improvement to the economic advantage of HTSC systems and may provide for technical solutions to some perplexing problems seen in LHe designs. Also, no credit is taken for the elimination of any iron and the subsequent reduction in losses that may be possible with these materials. This technology presented in App. A. summarized. was evaluated Potential problems using the baseline and research areas economic assumptions for the technology are
9.2.2
Discussion
Magnetic separation in which magnetic dipole moments are induced in paramagnetic particles by a highgradient magnetic field has proven useful in many industrial and utility applications. Both field intensity and high gradient are required for Various methods of applying successful particle separation by this method. superconducting coils to this process have been reported in the literature. A superconducting by the saturation
l
restricted
magnet generates a higher magnetic of iron. This can be used to advantage is to separate that cannot be small-diameter separated by
The main advantage magnetic particles magnets. The flow configuration. The matrix rates
can
be
increased
through
fixed
matrix
volume
can be reduced
for a given
flow rate.
Reference 1 presents a preliminary economic analysis for pilot plant or smallscale production units based on open-cycle cooling (i.e., the cryogenic fluid is added to the system as required and is not recaptured). This is obviously much more economical with nitrogen than with helium. For a 2.4-L conventional separator operating at 2 T, the continuous power requirement is about 160 kW. At $O.OG/kWh, the operating cost of this system is $4/h at a 50% duty cycle. A similar superconducting system cooled by liquid nitrogen would require 0.2-2.0 L/h depending upon the complexity (and cost) of the cryogenic container. If the least-expensive container (and thus, the highest rate of nitrogen usage) is assumed, the cost is calculated to be $0.12/h, a 97% savings in operating cost. Table 9.3 summarizes these calculations. A comparison of the initial capital costs of a conventional and an LHe superconducting system for matrix volumes up to about 5 L indicates that the superconducting system is slightly less expensive. This conclusion is highly dependent on the cost of the superconducting material, for which no data are presented in Ref. 1. Note that operation with LHe, at a cost 50 times that of LN2, would result in higher operating costs than those for the conventional unit, on the basis of the assumptions made in Ref. 1. On the
Industrial
Separations
131
Costs of Open-Cycle
Magnetic
Separatorsa
Cost
Item
Conventional System
Electricity (kW) Cost of electricity ($/kWh) Liquid nitrogen (L/h) Cosr of liquid nitrogen (S/L) Total cost of operating ar 50% capacity (S/h) Savings (%I
of separator: = 2 T.
160 0.05 0
4.00
aRating
density
matrix
volume
= 2.4
L and magnetic
flux
other hand, an LN2 system (as indicated above) would have a considerable operating cost advantage. A total life-cycle cost analysis is not given, since the operating costs so strongly dominate the results in this ease.
9.2.3
Summery
end Conclusions
Magnetic separations technology represents a potential application for the new HTSCs. This evaluation was predominantly an economic scoping study developed from previous work on a conventional separator. This technology shows a strong potential for significant cost reductions using HTSCs, when evaluated on the basis of operating cost. Several key areas of research The obvious need for higher current previously stated by many researchers. material properties is also needed. appear to have been uncovered by this evaluation. densities and bulk current capability has been A better understanding of the HTSC physics and
Magnetic separations technology (HGMS or OGMS) is perhaps the easiest technology in which to apply HTSC in place of LHe superconductors. Indeed, it appears to be quite cost-effective under less than optimal design conditions and requires a relatively modest superconducting magnet operating under steady DC conditions. This technology could see applications in mining separations, waste treatment, coal beneficiation (by removing sulfur before combustion to decrease SO2 emissions), and perhaps removal of NO, from boiler flue gases using OGMS technology.
132
Applied Superconductivity
Z.J.J., A Superconducting High Intensity Magnetic Separator, IEEE Trans. on Magnetics, MAC-11(5):1594-1596 (Sept. 1975).
2.
Marston, P.G., The Application of Superconductivity to Magnetic Separation, IEEE Trans. on Magnetics, MAC-11(2):602-603 (March 1975).
Gerber,
3.
R., Some Aspects of the Present Status of HGMS, IEEE Trans. on Magnetics, MAC-18(3):812-816 (May 1982). Gradient Magnetic Separation, Mining
4.
5.
Jungst, K.P., et al., Magnetic System for a Superconducting Magnetic Separator, Cryogenics, 24(11):648-652 (Nov. 1984).
Industrial
Separations
133
9.3.1
Introduction
High-gradient magnetic separation has been commercially applied to separation of magnetic particles from solid and liquid streams. Because of the higher field gradients that could be achieved with superconducting magnets, the technical feasibility of applying open-gradient magnetic separation (OGMS) technology to separation of gaseous species from gaseous streams (in particular, 02 from air and NO from flue gas) is considered in this section. High-gradient magnetic separation is a term that has been applied for many years to the separation of solid particles (dust, pyritic impurities in ground coal, tiny steel filings in the blast furnace effluent,ltnyd magnetic impurities in kaolin) from other materials in a gaseous or liquid stream. Open-gradient magnetic separation is a continuous process that achieves a spatial separation in the open, unobstructed magnet bore. In OGMS systems, paramagnetic species are drawn toward the bore wall while diamagnetic species are repulsed from the field toward the center of the bore. The desired separation is achieved by physically splitting the process stream at the exit of the magnet bore. Strong reasons exist for recovering 02 and NO from gas mixtures. Concentrated 02 has a multitude of industrial and research uses, ranging from forced drafts in steel production to medical applications, while NO is an unwanted constituent of exhaust gases that can interact with organic compounds in the air to create smog. (However, NO is a desired constituent in nitric acid production.)
9.3.2
OGMS Systems
for Separation
of Gases
Magnetic susceptibilities for several common gases are presented in Table 9.4.* The data indicate that the gas species 02 and NO are strongly paramagnetic; this is in contrast to most other ases, which are weakly diamagnetic, and to a few that are 9-11 Hence, 02 and NO might be separated economically from gas weakly paramagnetic. mixtures, such as air or flue gases, by passing the mixture through a magnetic field having a strong gradient. The paramagnetic component is drawn in the direction of the field gradient (or toward the magnet) and diffuses through the remainder of the gas, which is weakly repelled from the magnet. In typical HGMS (radius of about 25 urnI superconducting windings applications, a matrix or mesh of fine stainless steel wire is placed in an intense, uniform magnetic field enerated by that may carry a current density of 10 A/cm f . The field
134
Applied Superconductivity
of (293 IC)
Gas
Gas
Gas
At-
Hz H20b He Ne N2 NH3
-4 -13 -2 -7 -12
-18
3,449
-18
A unit
of magneti susceptibility, x, is Positive equivalent to 10-8 erg/mol*G2. entries denote paramagnetism and negative entries diamagnetism.
a.
magnetizes the wires, which in turn severely Finite-sized magnetic particles of comparable The mesh is then demagnetized, by removal separation cycle.
the field locally (about lo3 T/cm). stick to the wires in passing.7913-22 the field, and washed for the next
The techniques applicable to small-particle separation are not easily adapted gas separation, so paramagnetic gases must be split off in other ways. Commercial separation of gases from a flow is ordinarily accomplished by chemical, centrifugal, cryogenic processes or by the use of selectively permeable membranes. Separation ions from other components of a flow by ma netic drag (JxB forces) in 83 magnetohydrodynamic channel has also been proposed.
to or of a
All of these methods involve cumbersome machinery and extensive operating If OGMS methods could be applied, a much simpler and more economical systems. approach would be possible, because the components used to generate a magnetic field are always physically, chemically, and electrically distinct from system elements involved in the gas stream. Thus, it is not necessary to break down a plant periodically to clean or regenerate magnetic components, as in an HGMS system. OGMS technology can be used to continuously concentrate paramagnetic gases. The gases flow into one end of the magnetic separator and out the other end through a nested set of concentric pipes to split the paramagnetically enhanced gas mixture in the
135
outer pipe from the depleted mixture in the center pipe. The fractional separation achieved in a single stage could be enhanced by repeating the OGMS process downstream of the splitter, until a desired degree of purity is reached. Experiments would be needed to determine whether various combinations or recombinations of the split streams might be worthwhile.
9.3.3
of OGMS gas systems could be increased further by the introIn HTSC OGMS systems, intense fields could be generated without the complicated equipment and piping used in LHe superconducting plants. They would allow much higher field gradients than iron-based magnetic components, and they would not entail Joule heating losses. duction TH$$6q&-26 The magnets used in the HTSC OGMS gas system are apt to be some modification of magnets currently available for open-flow separation, from which we may infer An open-gradient coal-beneficiation system recently studied at possible benefits. Argonnea used a Nb3Sn, 4-K superconducting quadrupole magnet that carried a current of 900 A and provided a field gradient of 61 T/m (3.6 T maximum B) along the sides of a central bore of 12.5 cm. The estimated force on pyritic coal particles was 2.7 times conventional values, and the system was operated at a cost (due principally to the cooling system) estimated at 75% lower than the operating cost of iron-based magnets. A principal drawback to HTSC OGMS plants is the current lack of information about HTSCs, which inhibits practical planning for OGMS systems. However, if reasonable assumptions are made, the characteristics of some generic plants can be estimated. By proceeding on this basis, it should be possible to assess the practicality and promise of HTSC OGMS methods and provide the necessary groundwork for actual plants when adequate data become available.
9.3.4 1.
References
for Section
Gerber, R., Some Aspects of the Magnetics, MAC-18:812 (May 1982). Goodling,
Separation
2.
Drehmel,
Control,
1979). 3. Kolm, H.H., MAG-12~450 Oberteuffer, IEEE Trans. Oder, Trans. Research Needs (Sept. 1976). in Magnetic Separation, IEEE Trans. on Magnetics,
4.
J.A., Engineering Development of High Gradient on Magnetics, MAC-12~444 (Sept. 1976). Theory and
Magnetic
Separators,
5.
Applications,
IEEE
136
Applied Superconductivity
6.
Magnetic
Separator,
IEEE Trans.
7.
Stekly, Z.J.J., and J.V. Minervini, Shape Effect of the Matrix on the Capture Cross Section of Particles in High Gradient Magnetic Sepamtion, IEEE Trans. on Magnetics, MAC-1 2474 (Sept. 1976). Handbook of Chemistry and Physics, 51st Ed., Chemical Rubber Co. (1970-1971).
a. 9.
SkZectionBes Diamagnbtisme et ParamagnOtisme; Garter, Foi+x, G., Constantes C.-J., and L.-J. Smits, Relaxation Pammagnktique, in Tables de Constantes et Donnies NumCriques, Vol. 7, Union Int. Chimie Pub., Paris (1957). Van VIeck, J.H., The Theory Press, London (1932). Vonsovskii, S.V., Magnetism, of Electric and Magnetic Susceptibilities, Oxford Univ.
10.
11. 12.
trans.,
Wiley,
of Magnetism,
F. Cap,
13.
Boucher, R-F., and A.C. Lua, Loadability of High Gradient Matrix, IEEE Trans. on Magnetics, MAC-l&1662 (Nov. 1982).
Magnetic
Gas Filter
14.
Cowen, C., F.J. Friedlander, and R. Jaluria, High Gradient Magnetic Field Particle Capture on a Single Wire, IEEE Trans. on Magnetics, MAC-II:1600 (Sept. 1975). Cummings, D.L., The Motion of Small Paramagnetic Particles in a High Gradient Magnetic Separator, IEEE Trans. on Magnetics, MAC-12471 (Sept. 1976). Gerber, R., and M.H. Watmough, Magnetic Magnetics, MAC-I&1671 (Nov. 1982). Hollingworth, M., and J.A. Finch, Trans. on Magnetics, MAC-l&1674 Iannicelli, Magnetics, Separator Equations, IEEE Trans. on
15.
16.
17.
in Longitudinal
HGMS, IEEE
18.
Separation,
IEEE
Trans.
on
19.
Kolm, H.H., et al., High Intensity Magnetic Filtration, in American Institute of Physics Proc. on Magnetism and Magnetic Materials, No. 5, H.C. Wolfe, ed., New York (1971). Lawson, W.F., and R.P. Treat, Particle Trajectory Trans. on Magnetics, MAC-18 (Nov. 1982). Lua, AX., and R.F. Boucher, An Investigation Magnetic Gas Filtration, IEEE Trans. on Magnetics, Observations in Dry HGMS, IEEE
20.
21.
of Efficiency MAC-I&1659
137
22.
Watson, J.H.P., Theory of Capture of Particles in Magnetic High-Intensity IEEE Trans. on Magnetics, MAC-ll:1597 (Sept. 1975).
Filters,
23.
Del Casal, E.P, and J.J. McAvoy Jr., Magnetofluidynamic Gas, J. AIChE, 21:615 (May 1975). Jungst, K.P., et al., Magnet System Cryogenics, 24:648 (Nov. 1984). for a Superconducting
Separation
of a Binary
24.
Magnetic
Separator,
25.
Marston, P.G., The Application of Superconductivity Trans. on Magnetics, MAC-11:602 (March 1975). Watson, J.H.P., Superconducting (Aug. 1983).
to Magnetic Separation,
IEEE
26.
Mining, 2~121
27.
Doctor, R.D., et al., Investigation of Open-Gradient Magnetic Separation for Rlinois Coal, presented at 2nd International Conf. on Processing and Utilization of High-Sulfur Coals, Carbondale, Ill. (Sept. 27-Oct. 1, 1987).
138
Applied Superconductivity
Simplified estimates are presented here for the effect of high-gradient magnetic separation (HGMS) of oxygen from air. The same approach can be used for separation of NO from flue gases if the other gas species are assumed to have negligible magnetic susceptibilities, which is reasonable if no oxygen is present in the flue gas. (Oxygen and NO are the only gases having any appreciable magnetic activity.) Therefore, notes on the application of the oxygen system to NO are included in the discussion.
FLOW EQUATIONS The equations of continuity and motion for oxygen (subscript o) diffusing through nitrogen (subscript n) in air, under the influence of an applied magnetic may be taken as Po p, = = n,kT, n,kT,
(= 0)
steadily field B,
v.n,vo v*nnvn
=
= =
0, 0,
VP, +
m,n,v,*Vv,
ncV(~c*B)
~,,~ncnOQn,,~nO(~n - v,),
m,n,vn*Vvn (=
0)
In the equations of motion, the last term re resents the Boltzmann collision integral (giving drag forces acting on the two gases 14 ), in which m,, = mnmo/(mn + m,) is the reduced mass for N3-09 interactions, Q,, is the cross section for molecular momentum transfer in these interactions, en0 = (tkT/m,,) 12 is the mean speed of impact at temperature T, and k is the Boltzmann constant (1.381 x lo-l6 erg/K). Ordinarily, O2 is paramagnetic, which means that oxygen molecules on average develop a magnetic moment p. = x,B along B, where x0 is their magnetic susceptibility. As indicated in Eq. 9.1, this paramagnetism tends to draw oxygen into regions of strong field. For diffusion processes at moderate velocities, the left side of the force equations (which gives the rate of change of momentum per unit volume) may be neglected in comparison with other terms. In general, the drag forces must be retained in the equations. However, we show below that these terms are proportional to the
139
an equilibrium diffusion currents, and so they will also drop out at equilibrium: configuration can be obtained by channeling the flow properly. Separation estimates are made on this basis.
MAGNETIC PROPERTIES At moderate magnetic field intensities, oxygen is paramagnetic from below 50 K to well over room temperature (taken as 293 K in Table 9.4), its magnetic moment per molecule u. increasing with B. At very low temperatures or strong fields, u. saturates and approaches a constant (essentially ferromagnetic) value. The mechanism operating here is that oxygen molecules ordinarily At temperature equilibrium, in a weak field, permanent magnetic moment. collisions randomize the directions of the magnetic moments but leave a predominantly along E3,in the lowest energy state. The molecules precess about direction. Strong fields, however, will tend to line up nearly all the molecules The net moment u. then no longer increases with B. have a spin-flip fraction the field along B.
Quantum mechanical calculations by Van Vleck and othersj- indicate that the mean magnetic energy per molecule u-B of paramagnetic gases in a magnetic field B may be expressed as a sum over effective molecular states J of the form <y*B>
where:
kT 1 oJ(T) J
fJ(gJuB9/kT),
oJ(T) = occupancy
gJ = splitting factor for state J, and MB = eW2mec = 9.26 x 10b21 erg/G. Here, uB is the Bohr magneton. The index J is an angular-momentum quantum number appropriate to the effective state, which depends on the way the actual levels couple when a magnetic field is present. For reference, we note the limiting behavior coth(z) - l/z + z/3 for small z,
as z -c -,
(9.3)
as as z + 0, z + m.
!40
Applied Superconductivity
These limits cross at about 22 = 3/(J + l), which is in the asymptotic Hence the value of fJ(z)/fJ(-) at the crossover point in 1 _ (I/J)~-~/(J + 1) = 80% for J = 1 to 3.
VALUES FOR O2 AND NO Molecular oxygen is a triplet 32 configuration in the ground state, with energy levels very narrowly separated in comparison with usual values of kT and occupancy factors oJ = 1. The three energy levels combine to give two effective states: One corresponds to the net orbital quantum number A, here 0 (named a 2 state) with g, = 1; the other corresponds to the net spin quantum number S, which is 1 for oxygen (giving multiplicity 25 + 1 = 3, making the state a Yriplet) with gS = 2. The magnetic moments for these two angular momentum states in effect precess independently about the field direction and add to provide the total moment. For oxygen, therefore, Eqs. 9.2 and 9.3 give c,0**13 = kT [f*(lrBB/ZkT) - 8uBB2/3kT - 2UBB + f9(2ygB/2kT)j fat for B + 0 6 + = kT fl(uBB/kT) case) (9.4)
(patamagnetic (saturation)
From the estimate following Eq. 9.3, about developed at the value of B where the estimates would be 3kT/4uB, or about 340 T at 300 K.
80% of the saturation moment 2ug is in Eq. 9.4 cross. The field at this point
Nitric oxide is a doublet 2~ configuration with orbital quantum number A = 1 (a x state) and spin S = l/2, in which the energy levels are widely separated. In the presence of a magnetic field, the doublet levels react independently. The lower level, with J = A - S = l/2 and gJ = (J - S)/[J(J + 1)112 = 0, is nonmagnetic, i.e., fJ(0) = 0. The upper level, with J = A + S = 3/2, is strongly pa amagnetic. In weak fields, S links to and precesses about A, giving gJ = (J + S)/[J(J + l)] 172 . Hence the effective moment becomes
<uNOB
- kT [o,12(T) - o~,~(T)
f1/2(0)
+ o~/~(T)
(A + 2S)2pB2B2/3kT
In strong fields, the link between A and S is broken, and both formula for the moment in Eq. 9.4 applies in this case, yielding QNO*B - kT o~,~(T) - o~/~(T) [fn(lrBB/2kT) + f8(2uBB/2kT)I
141
The limits
for NO from - 4
(uElO!B)
for 6 * 0
fat B * -
where 0312(T)
= (1 - em* + xe-x)Ixx
+ xe-%),
as calculated by Van Vleck for the doublet energy separation AE of nitric oxide.3 The occupancy factor 0~12 for the upper state is zero at T = 0 and 1 at T = -, reaching 0.84 at T = 293 K.
DIFFUSION
RELATIONS
To make estimates of physical quantities, it is convenient to convert from the in n (total particles per variables no, nn, vo, and vn of Eq. 9.1 to mean flow relations cubic centimeter) and v (average particle or molar velocity) and relative diffusion equations in J (the diffusion current) and x (oxygen mole fraction). The latter variables are defined by *=n and J=J,=n,(v According to Eq. 9.7, Jo + J = (nova and nJ = nn,(v
0 v) = ,( 0 + *)vo ,(,v, + V) = nonn(vo V). 0
0 + 9
nv = 0 0 + nnv
p = p,
+ p = nkT, (9.7)
-VI,
J = (v
- v),
x = no/n.
+ V)
- (no
+ ,)V
= 0,
or
J = -Jo, (9.8)
The last term here may be recognized as the velocity difference of motion for O2 and N2. When the equations for the two gases Using Eqs. 9.1, 9.7, and 9.8, one finds v*nv = v*,v, J =-D Vn, J-J = V*novo + v*n,v, where = 0, Vn = x V(uo*B)/kT, (Ficks dx/dt law)
D = kT/nmocnoQo = -
and equation).
(9.9)
- V*xnv = -nv*Vx
(diffusion
142
Applied Superconductivity
LIMITING MAGNETIC
EFFECTS
By using Eq. 9.9, together with suitable boundary conditions and a prescription for B, it is possible to give details of oxygen diffusion with various channel arrangements. An example of this type of analysis is discussed in the next section. For order-of-magnitude separation estimates, however, a much simpler approximation is adequate. From Eq. 9.8, one can identify J as the relative diffusion current of 02 through Assume that flow channels for the air near the magnet are such that diffusion N2. currents are precluded by the geometry. (This requires basically that the flow remain oriented normal to the field direction near the magnet.) Then the drag terms in Eq. 9.1 will be zero, and the equation for oxygen will reduce to = vno n,V(u,*B)/kT, giving no = ;;o,u,*B/kT of weak (9.10) and
by integration, where ii0 is the value of no at zero field. For the cases strong fields (paramagnetic and saturated moments), Eq. 9.4 gives Ln no/ii, = (1.3 = (4.6 x LO+)(B/l x lO-3)(B/1 T)2/(T/300 T)/(T/300 K)2 K)
(panmagnetic)
current
Now, the magnetic field at radius r about of density J (in amps per square centimeter) B = 2na2J/re = 63 T (J/LO6 A.cm2)(a/l
(9.12)
(The conversion factors here are 10 A/cm = 1 G+cm, and 1 T = lo4 G). If HTSC wire current densities greater than lo6 A/cm2 were possible, Eq. 9.12 suggests that it should be possible to reach fields of 20-30 T at one side of a channel having a cross section of a few millimeters near a wire with a radius of 1 mm. (Actual values would depend on the ultimate structural and electrical characteristics of the superconducting material.) According to Eq. 9.11, this would yield on the order of a few percent enrichment of the oxygen. The nitrogen density would be relatively unaffected. By splitting off this oxygen-enhanced flow and repeating the process, a reasonably enriched stream of O2 should be achievable. Thus, about 50 fractionations at 3% increase of O2 per step would yield s mixture of half oxygen and half nitrogen. An important advantage of using superconducting wires is that the geometry of For example, a possible configuration for the the HGMS system becomes more flexible. separator would be to run superconducting wires along the inside of the splitter channel tube. Slots in the tube would provide regions of high field and field gradient, through which the oxygen would diffuse into the splitter.
143
ESIIMATION OF DIFFUSION RATES Whether an equilibrium oxygen distribution is reached, as assumed above, To estimate the diffusion rate, depends on the flow velocity and channel arrangements. we consider a sample problem in which the air flow is along a channel of width a, and the field B is normal to the direction of flow but varies across the channel. Air enters the channel at one end. The nonuniform field acts on the oxygen, causing it to diffuse toward the wall, where B is greater. The point of the calculation is to estimate the mean distance along the channel needed for this to occur. Diffusion effects are considered to be weak, so that the mean flow velocity v = ku, taken along the z-direction, is constant and the variation of air density, which may be in the channel is defined as s = In n/n, is small compared to 1. The field strength approximated by B2 = 82b(y)
so that
B = j B bl*. of z, might represent to, the channel. the field in the channel
(9.13) due to a
This relation, which is independent wire outside of, but running parallel The flow relations Vn = n;E VxB2/kT, and .J = D Vn(1 - x) = V@, giving from or
11,= nD [(l
- 2,s
- xl,
(9.14)
v2* = v*J
-ii~
axlaz.
At the channel walls, the normal component of J must vanish, and hence the boundary conditions are that the potential function II, must have a maximum or minimum in y at y = o and y = a. The simplest way to satisfy the boundary conditions is to expand b, s, and C in functions, such as cos mny/a (m = O,l,Z,...), that become constant as y + 0 or a. Thus, we may take b = 1 b, and x = l ~~(7,) The equation cos mnyla. that - z)Ebm. (9.16) co9 mnyla, s = : + Bj; 1 b, cos mny/a, (9.15)
% - (iiu/nD)xA
144
Applied Superconductivity
Solutions
%I = j;(l where
- :)6b m + Cme-rmz f
=rn = [(iiu/211D)~
+ (rni~/a)~]~~
iiui2nD.
In Eq. 9.17, the root r, was chosen so that x, would remain finite for large z. The constants Cm are arbitrary, but if we consider x to be a constant at the channel inlet (z = 0), then x, must vanish there for m not equal to 0. Therefore, C, = -j; (1 - &3bm,
and x has the solution x = j; + j;(l where - :)6 I[1 - e-m] b, cos mny/a, (9.18)
According to Eq. 9.17, the roots rm increase with m, eventually reaching the limit rm - mn/a. Hence the exponentials in a in Eq. 9.18 die away most rapidly for m > 1, and the relaxation of x to the diffusion solution (proportional to b plus a constant) depends essentially on the first term, m = 1. But the nature of the solution depends on the value of the constant o = iiuaj2nnD - ua/BxD. For G <( 1, rl - n/a and the rate of relaxation is independent of u. This corresponds to a very thin channel or slow air flow. If G >> 1, then r1 = (n/a)/[a so that the solution 3: = l/rI + (l&l) would relax - ua2/n2D 1121 _ l?D/ua2 (9.19)
corresponding to a thick channel and/or rapid flow. As the analysis indicates, this conclusion is more or less independent of the form of B. However, the downstream concentration of oxygen (or air density) approaches a profile that mirrors that of B2.
REFERENCES
1.
FOR SUPPLEMENT S., and T.G. Cowling, Mathematical Univ. Press, London (1951). and E.N. Lightfoot,
Chapman, Cambridge
2.
Transport
Phenomena,
Wiley,
New
145
3.
and Magnetic
Press, London (1932). 4. 5. Vonsovskii, S.V., Magnetism, Vol. 1, R. Hardin, trans., Wiley, New York (1971). F. Cap, trans., Pergamon
Wagner, D., Introduction to the Theory of Magnetism, Press, New York (1972).
Summary
Larry R. Johnson Argonne National Laboratory
146
147
Summary
Section 10 makes the following observations about the demand for high-speed train service in the United States and the potential market for magnetically levitated trains: 1. Speed has always commanded a premium in transportation service, resulting in intercity travel that is dominated by automobiles for short distances and air travel for long distances. However, both highway and air traffic congestion are factors that have caused a reexamination of the proposed solutions (more cars, highways, air flights, and new airport construction) to satisfy the increasing demand for intercity business and personal travel. Transportation petroleum consumption now exceeds domestic oil production, and the situation will continue to worsen as U.S. petroleum reserves continue to decline while transportation demand grows. Electrically powered, high-speed trains address both the petroleum supply and congestion issues. The two principal technology options for high-speed ground the very-high-speed, steel-wheel-on-rail transportation are systems (such as the Japanese bullet trains and the French TGV) and the magnetically levitated (maglev) systems for which advanced prototypes have been developed in Japan and West Germany. Steel-wheeled systems are limited to a maximum speed of about 170 mph because of the loss of traction beyond those Maglev systems have demonstrated speeds of 300 mph. speeds. While the investment costs of steel-wheel systems would be lower than those of a maglev system, their maintenance is expected to be higher due to the close track tolerances that must be maintained. The Japanese maglev system uses low-T, superconductors in an electrodynamic (repulsive) system, while the Germans have an electromagnetic (attractive) system that does not require the use The perceived advantage of the attractive of superconductors. system (using conventional technology rather than superconductors that have to be cooled to 4 K) may have been erased with the potential for superconductors that can be maintained at 77 K using liquid nitrogen. The repulsive system has the further advantage (over an attractive system) of a large air gap (6 in. versus 0.5 in.), which relaxes the design requirements and is dynamically stable, eliminating the need for gap sensors and accelerometers to provide vehicle stability.
2.
3.
4.
148
Applied Superconductivity
5.
High-T, superconductors should reduce the cost of the primary suspension system considerably; more important, they will provide improved system reliability and reduced complexity. Numerous market studies have recently found a market niche (generally at intermediate distances between cities, e.g., 100-600 mi) for very-high-speed trains in the United States. A dozen states are examining intercity corridors in which either maglev systems or TGV-type systems appear to be the best technology solution for increasing travel demand, when economic, energy, and environmental factors are all evaluated from a totalsystem perspective. The only high-speed ground transportation systems available for use in the United States are foreign technologies. Both the Germans and the Japanese are sufficiently serious about exporting their systems (including maglev) that they are conducting feasibility studies at their own expense, including a $l.Z-million German-funded study in Texas for the Dallas-to-Boston corridor. High-speed ground transportation projects appear likely to begin in Pennsylvania, Florida, Texas, Nevada, Michigan, and Illinois; as other intercity corridors become congested, more states are likely to follow. Applying high-l, superconductivity to maglev trains offers the opportunity to resolve a critical national transportation need while contributing to the economic competitiveness of the nation.
6.
7.
149
10.1
BACKGROUND
Speed, throughout history, has commanded a premium in the provision of transportation and consequently has greatly influenced the choice of mode for both passenger and freight movement. In the United States, intercity passenger travel is dominated by automobile and air travel, which require relatively high energy intensity to combine personal convenience with speed. Similarly, air and truck travel dominate the transport of valuable freight needing timely delivery. The air and automobile/truck modes have required massive public and private expenditures. Demand for high-speed passenger and freight service will continue to grow As a result, further increases in as both population and economic activity increase. investment for intercity transportation will be required. The question now is whether a form of high-speed rail transportation is a legitimate option among the alternatives advanced to satisfy future intercity travel demands in the United States, and, in particular, if magnetically levitated trains can be enhanced by using high-criticaltemperature (high-T,) ceramic materials such that a magnetic-levitation (maglev) system would prove to be the preferred choice among high-speed rail alternatives. Current interest in high-speed rail alternatives is stimulated principally by two factors. First, increasing air-traffic congestion is a problem that will not be easily resolved because of the frequent controversies surrounding continued airport expansion and new construction. Second, in spite of the current availability of oil in the world market, energy problems are likely to be significant during the next decade. Transportation petroleum consumption in the United States now exceeds domestic oil production, and this situation will continue to worsen as oil reserves are further depleted Electrically powered, high-speed ground and transportation demand continues to grow. transportation addresses both of these issues.
10.2
ADVANCED
GROUND
TRANSPORTATION
OPTIONS
10.2.1
Conventional
Trains
The most advanced revenue-service trains in the world today use conventional steel wheels on steel rails. The French Tres Grande Vitesse (TGV) achieves speeds of up to 170 mph, while averaging 130 mph between Paris and Lyon. The electrically powered Japanese bullet trains (Shinkansen) have a design speed of 160 mph, although
150
Applied Superconductivity
operationally they have had a maximum speed of 130 mph. These bullet trains average better than 100 mph over the 600-mi route between Tokyo, Osaka, and Hakata. British Railways operates diesel-powered high-speed trains (HSTs) on several long-haul routes with top speeds of 125 mph and average speeds between cities that frequently exceed 90 mph. Because HSTs are defined as those that can operate at speeds greater than 125 mph, the United States is conspicuously absent from any list of HST operations. However, several examples of current fast train service in the United States can be Amtraks Northeast Corridor between Boston and Washington, D.C., operates cited. passenger trains with maximum speeds of 120 mph between New York City and speeds of about 80 mph. Amtrak also operates a Washington, D.C., and average successful gas-turbine-powered Turbotrain service in New York State, with maximum speeds of 110 mph and average speeds of about 70 mph between Albany and New York City. In spite of the relative success of the HSTs in France and Japan, serious problems limit the use of conventional technology (steel wheels and rails) in achieving faster First, at speeds of about 120 mph and speeds and, in turn, attracting more passengers. Federal Railroad greater, there is significant wear on both wheels and rails. Administration (FRA) standards for conventional freight trains and commuter trains allow relatively large discrepancies in level between one rail and another -- 1.25 in. for 80-mph operations. However, the FRA standard drops to 0.5 in. for 120-mph operations. In comparison, the French TGV standard is 0.16 in. for the 170-mph portions of the system. Satisfying the standards for high-speed service is not impossible, but it is very expensive. Second, the speed limit of conventional trains driven by wheel traction is limited by the frictional forces that develop between the wheels and rails. The use of linear propulsion motors, in which the wheels would be used only for suspension, would make However, the high maintenance costs maximum speeds of 180 mph technically feasible. associated with the required track alignment, smoothness, and curvature specifications might render this approach economically unattractive. Third, the serious problems associated with noise and vibration of steel-wheeled trains have been major factors influencing the Japanese to examine other alternatives for intercity transportation.
10.2.2
Levitated-Vehicle
Technology
Several research groups throughout the world are investigating the use of levitated vehicles in conjunction with linear motors. During the 196Os, France, Britain, and the United States conducted large-scale tests of air cushion vehicles (ACVs) designed for intercity speeds of 160 mph. However, development of tracked ACVs is virtually at a standstill because of the growth of interest in magnetic levitation. Two principal methods of achieving sufficient magnetic vehicle are actively being developed. Electromagnetic conventional iron-core electromagnets to provide an attractive gap (about 0.5 in.) between the vehicle and the track. The force to levitate a large suspension (EMS) uses force, with a small air EMS technology is being
151
developed by the Federal Republic of Germany. The Transrapid 06 is a maglev prototype vehicle designed for a maximum speed of 250 mph, with test speeds of 220 mph having been recorded. A 13-mi test facility built at Emsland (a planned second loop will extend the system to 20 mi) is designed to become part of an eventual revenue-service corridor. The Japanese are pursuing an electrodynamic-suspension (EDS) system that employs superconducting magnets to achieve a large (4- to 6-in.) gap, using a repulsive force between the track and the train. Test facilities, including a 4-mi test track, have been built at Miyazaki, Kyushu, in southern Japan. Long-term plans are to build a 25-mi test track. A small prototype (ML500) vehicle has set a world speed record of 320 mph (517 km/h), while larger versions (MLUOOI and MLU002) are being designed to run at 260-300 mph. The NbTi superconductors are cooled to 5.2 K at 2 atm by liquid helium in cryostats. During the many tests that have been conducted, no quenching of the magnets has occurred during the operation of the vehicles.
Advantages
of Magnetically
Levitated
Vehicles have caused both Japan and West ground transportation. Several of
Discernable advantages of magnetic levitation Germany to invest in this new technology for advanced the prominent benefits are as follows:
l
Magnetic levitation overcomes the principal limitation of Speed. wheeled systems, the loss of traction at high speeds. This is a major problem for HSTs between the speeds of 150 and 170 mph. With maglev systems, speeds of 300 mph become feasible. Operations. Noncontacting operation means that inclement weather (rain, snow, or ice) will pose significantly fewer problems to safe and timely operation of the maglev system than such weather poses to wheeled systems. Because of the noncontacting suspension, the maintenance costs associated with maglev transportation will be considerably less than those for conventional rail systems. There is virtually no mechanical wear on either the track or the suspension Although operational data are needed to system of the train. quantitatively verify this, a number of independent studies have come to this same conclusion. No contradictory evidence has been found on this point. Dependability. greatly increase The lack of moving parts for a maglev train should the dependability and reliability of the system.
Maintenance.
Maglev systems, being electrical, do not depend on Energy. dwindling U.S. petroleum supplies; the electrical energy can be provided by hydroelectric generation, coal, or nuclear power. On a passenger-mile basis, the energy intensity of a maglev train would
152
Applied Superconductivity
of
one-fourth
that
of
intercity
aircraft
or
Because of the combination of higher achievable Economics. speeds, which increases the ridership potential, and the greatly reduced maintenance costs of a noncontacting system, a maglev train may have the greatest potential among the intercity train options to operate on a revenue-sustained basis.
Environmental Concerns. Noise and vibration, which reportedly a major concern with the Japanese bullet trains, should considerably less from a noncontacting, maglev system.
are be
Advantages
of High-T,
Superconductors
for Magnetic-Levitation
Technology
Because an attractive (EMS) system uses the developed technology of conventional electromagnets, this system may appear to be closer to commercialization than the repulsive (EDS) system, which uses superconducting electromagnets. However, a definitive conclusion is not easily determined, except that the initial capital costs of the primary suspension system would probably be lower with an EMS system when compared to an EDS design. However, the EDS levitation system is dynamically stable, requiring The EMS system is inherently unstable, no feedback controls to maintain clearances. necessitating gap sensors and accelerometers to regulate the power to the electromagnets to maintain stability. High-T, superconducting the EDS design. 1. superconducting magnets would magnets. The new superconductors offer several advantages may provide additional over low-Tc advantage to
The current achievements in high-T, materials may significantly reduce the costs, not only compared to present-day superconductors, but even compared to attractive force systems. First, the high-T, feature would permit the use of low-cost liquid nitrogen refrigerant, rather than liquid helium, resulting in a savings on the order of 20 to 1. Second, the total refrigeration system could be simplified, with an attendant reduction in the amount of insulation required around the magnets, thus increasing the flexibility (and further reducing the costs) of the design. High-T, superconducting magnets make higher magnetic fields possible, resulting in increased track clearances for the EDS design, which in turn permit the relaxing of design requirements for the guideway. Because of the larger clearances, the system is more tolerant of discrepancies in rail level and problems associated with inclement weather. Only superconductors provide the levitation forces that create large track clearances at reasonable costs.
2.
Magnetic Levitation
for Transportation
153
3.
Lighter-weight superconductors will reduce the weight of the primary suspension system (and, in turn, vehicle weight) of an EDS system compared to an EMS design using iron-core electromagnets. The weight reduction should further reduce the size and weight of the propulsion system. For any given superconductor with a lower T,, the applications that could initially use these materials would be the ones with the least-restrictive current density requirements. In the early 197Os, an evaluation by Philco-Ford Corp., SRI, and MIT established an operating current density of 0.4 x lo4 A/cm2, which is less than that for most other superconductor applications (see Table 2.1). However, current Japanese des$n goals call for an operating current density of 20 x lo4 A/cm . It appears that the threshold current density design criteria for a commercial vehicle would be about IO4 A/cm2. As the current density of low-T, superconductors increases, the size and weight of the magnets can be reduced.
4.
Applicability
of Magnetic-Levitation
Technology
to U.S. Travel
Needs
Recent studies of U.S. intercity travel demands have indicated that magnetically levitated trains are being seriously considered in a number of corridors. A market niche appears entirely feasible for intercity trips with distances of about 100-600 mi. Trips of less than 100 mi will still be dominated by automobile travel, although there may be a maglev market for some business travel. Trips much longer than 600 mi will still tend to be served principally by airplane. Active (funded) interest in high-speed intercity train service, including the potential for maglev systems, is described below. This overview of projects is current as of 1986. 1. Pennsylvania. The Pennsylvania High Speed Rail Commission voted unanimously to narrow the technology options for a Philadelphia-Harrisburg-Pittsburgh passenger rail corridor to two: a 160-mph HST and a 250-mph maglev system. The new service, it is estimated, could have 4-12 million riders annually and provide up to 292,000 person-years of jobs during its construction.
Nevada.
2.
The city of Las Vegas is conducting studies of the of HST service (specifically including maglev for the corridor connecting Las Vegas and Los
Michigan. Hearings have been held on high-speed rail passenger service in the Detroit-Chicago corridor. Two technical reports addressing technological options, including maglev systems, have been prepared -- one by the Michigan Department of Transportation and the other by the Advanced Rail Consortium.
154
Applied Superconductivity
4.
Ohio. A liO-mph HST system linking Cleveland, Columbus, and Cincinnati is being pursued by the state. The new system is planned for a dedicated right of way, rather than using existing trackage. The eight-year construction effort would add 60,00075,000 jobs. Both the Chicago-Milwaukee and Chicago-St. Louis Illinois. corridors have been examined as part of a study of high-speed rail These two corridors continue to be options in the Midwest. viewed as extensions of the Detroit-Chicago HST line in the ongoing Michigan studies. Florida. A recent study concluded that an HST in the MiamiOrlando-Tampa corridor could be operated without state or federal subsidies if creative financing was used to take advantage Both TGV-type and of development rights around stations. maglev systems are being examined. Texas. The Houston-Dallas corridor is being examined in a $1.2-million study funded by the Federal Republic of Germany and a consortium of ten German companies. The 250-mi corridor is thought by the German representatives to be the best in the United States. This corridor would be the first leg of a Texas triangle that would eventually link San Antonio to the other two cities. Missouri. The state is studying the feasibility of upgrading passenger rail service in the Kansas City-St. Louis corridor to 8%%~55n&&\* So-mph trains. New York. HST service between New York and Montreal (via Vermont) using TGV technology has been studied. Operation and maintenance costs are expected to be covered by passenger revenue and ridership is projected to increase six-fold. Three thousand construction jobs would be created, and an additional 1,500 continuing new jobs would be added for the HST operation. A feasibility study of a loo-mph electrified rail New Mexico. service between Santa Fe and Albuquerque has been conducted. A ticket price of $5-6 appears to be sufficient to pay operating costs for the 66-mi trip. True high-speed operation on a longer (300-mi) corridor between Los Alamos, Santa Fe, Albuquerque, and Las Cruces is also under examination. The state has begun a preliminary study of HST Washington. service between Portland (Oregon), Seattle, and Vancouver, B.C. If the conceptual study is favorable, a larger study of demand, markets, and needs will be conducted.
5.
6.
7.
a.
9.
10.
11.
155
12.
Georgia. The state is planning a feasibility study for rail passenger service between Atlanta and the seacoast city of Savannah, a 270-mi corridor that includes Macon. French representatives have visited Georgia transportation officials to promote TGV technology.
13.
Northeast Corridor. Japanese National Railways has presented a proposal to the Coalition of Northeast Governors to study the feasibility of maglev technology for the corridor encompassing Boston, New York, and Washington, D.C. The study would highlight the similarities and differences between the northeast corridor and the Shinkansen Tokaido corridor.
Opportunity
Development
More than a dozen states are funding feasibility or engineering studies of highThe choice of technology options is usually speed rail transportation between cities. between advanced rail systems -- such as the French TGV, the Japanese bullet trains, and indeed, the United States is viewed as an even the German ICES -- and maglev systems. export market by some of these countries, with each of them studying the potential for their technology, including maglev systems, in U.S. markets. Research in this country was essentially halted in 1975, but it has been aggressively pursued in Japan and Germany, where full-size maglev vehicles have operated at speeds in excess of 300 mph. England, Canada, and even Rumania now have maglev research programs. High-T, superconductivity is not an enabling technology, but it is an enhancing While Japan has the lead, for now, in EDS technology one, especially for maglev trains. for maglev applications, the United States can use its lead in basic research on high-T, superconductors to leapfrog the current EDS technology with a more-reliable, lessAlthough the United States is behind in terms of maglev test expensive system. facilities, a conventional maglev prototype system could be constructed in this country by drawing on the large body of work conducted in the United States during the 1970s and updating it with the more recent advances in other countries. The development of a Mod 1 maglev system using conventional superconducting magnets could take place the development of commercial simultaneously with quantities of high-T, superconductors that would meet the specifications of an advanced maglev system. The Mod 1 design would allow replacement of the liquid-helium-cooled EDS system with a liquid-nitrogen-cooled advanced EDS technology. Most of the states that are examining HST technology are proposing systems that would be implemented in the mid-1990s and provide service well into the next century. Thus, the timing appears right for focusing the advances in high-T, superconductors on one of the nations critical transportation needs.
156
Applied Superconductivity
10.3
BIBLIOGRAPHY Ground Transport (Queens Univ., Kingston, Ontario), SuperSystem, Las Vegas/Southern California Corridor, Phase 2: - Development Status of Major Maglev Subsystems and for Las Vegas Dept. of Super-Speed Train Development, No. PRO-421) (March 1986). Rail Association Yearbook, High Speed Rail Assn.,
Canadian Institute of Guided Speed Ground Transportation Maglev Technology, Task 5 Critical Components, prepared CIGGT Report 86-10 (Project
Philco-Ford Corp., Aeronutronic Div., Conceptual Design and Analysis of the Tracked Magnetically Levitated Vehicle Technology Program (TMLV) - Repulsion Scheme, Executive Summary, U.S. Dept. of Transportation Report DOT-FR-40024 (Task 4) (Feb. 1975). Proc. International 1986). Propelling Passengers Rhodes, R.G., Press (1981). Thompson, Upsurge Conf. on Maglev and Linear Drives, Vancouver, B.C., Canada (May
(Aug.
L.S., High-Speed
Rail, Technology
Review
(April
1986).
Railroading
Office
of Technology
Appendices
Appendix
A: Economic Assumptions
Appendix
B: Superconductor
Performance
157
158
Applied Superconductivity
SUMMARY To facilitate the first-cut economic analysis of high-temperature ANL prepared a base set of economic/financial assumptions to serve appendix presents those assumptions. superconductors, as guidelines. This
Our purpose was to provide a set of economic assumptions that would be plausible, consistent, and simple to use in assessing the effect on project economics of recent technical advances. The assumptions are also compatible with the assumptions made by the electric utilities, and they are reasonable from industrys perspective. The economic analysis is conducted in current dollars, assuming no escalation of real cost, and is guided by the Tax Reform Act of 1986.
ECONOMIC
EVALUATION
Introduction Determination of the profitability of superconducting technologies would require a side-by-side comparison of those technologies relative to competing technologies Such an analysis would provide an explicit within representative utility systems. treatment of the effect of superconducting technologies on system load profiles and the resultant capacity/expansion benefits of a technology that promises higher efficiencies. Although a more detailed analysis may ultimately be required, the following assumptions are intended to provide a plausible, consistent, and simple first-cut estimation of the expected profitability of a typical project. Example evaluations are provided following the tables of assumptions (Tables A.l-A.31 described below. Tables A.1 and A.2 present the levelized fixed charge rate on utility capital and industrial capital (i.e., capital on the customers side of the meter), respectively. The bases for these fixed charge rates are also shown. The book life of utility capital is assumed to be 30 yr, while the book life of industrial capital is 10 yr. If one believes, for example, that a utility-owned device must be replaced before 30 yr have passed (i.e., the book life is less than 30 yr), then one should determine the equivalent capital cost (in 1986 dollars), including replacement at the end of the equipment book life, to provide a project book life of 30 yr. Thus, if a storage facility had a book life of 15 yr, it would be
Appendices
159
TABLE
A.1
Ownershipa
Item
Value
Levelized fixed charge rate on capital (X) Inflation (X1 Weighted average cost of capital, discount rate (X) Effective tax rate, federal and state (Z) Investment tax credit (X) Book life (yr) Tax life (yr) Depreciation (% of declining balance) Capacity factorsb Generation (I) Transmission (X) Storage Levelization factor for fuel, electricity, and O&M
65
aoc
As designed 1.45
aAssumptions are consistent with the Tax Reform Act of 1986. bEquivalent operating hours at 100% capacity. 'Based on dedicated lines.
replaced at the end of year 15 to provide a project life of 30 yr. cost of the 30-yr project would be calculated as follows:
ECC
Icc
ICC(1 +
(1 + rP
e)
where: ECC = equivalent capital cost for 30-yr project life, ICC = initial capital cost of equipment for 15-yr life,
160
Applied Superconductivity
TABLE
A.2
Industrial
Ownership
Item
Value
Levelized
fixed
charge
rate
on capital
(X)
26.6 4 15
Inflation (X) Weighted average cost of capital, discount rate (%) Effective tax rate, federal and state Investment tax credit t%) Book life (yr) Tax life (yr) Depreciation Capacity factors
(I)
i8
0 10
As designed 50 1.20
e = inflation rate (4%), r = discount rate, and n = year of replacement. The second term is the present value of replacing the equipment in year 15. On the other hand, if one believes that the device has a longer actual life than the book life, one would not make this adjustment. These lives are typical of business planning. Table A.3 presents the price of electricity in several circumstances (e.g., onpeak and off-peak). These prices were derived from the available data on U.S. average prices in 1986, and they are given in 1986 dollars. The cost-of-service calculations illustrated in examples the levelized cost of service in 1986 dollars. 2 and 3 show how to
calculate
Appendices
161
TABLE
A.3
Electricity
Item
Value
value
0.041
into of
storage storage
price
al986
dollars.
bTo simplify the analysis, the value of losses at any point in the utility system is taken as $O.O4l/kWh, which is the long-run average value of losses. In reality, the short-run marginal value of losses would approach the cost of fuel saved by changes in losses. The cost-of-fuel component of electricity price is taken as $O.O17/kWh. The long-run marginal cost of losses would approach the incremental cost of new base-load capacity (about $0.060O.O70/kWh).
162
Applied Superconductivity
Example Calculations
Example 1: Calculation
of Incremental
The incremental capital cost (in dollars per kilowatt) is the equivalent capitalized cost of the operating cost savings of one technology relative to another. For example, if a nominal 300-MW superconducting generator had an efficiency of 99.596, it would incur losses of about 1.5 MW. If a standard generator had an efficiency of 98.5%, it would incur losses of 4.5 MW. The incremental capital cost (maximum capital cost premium) for the superconductor would be evaluated ss follows for utility applications: Loss of standard - loss of superconductor
3 MW x 8,760 h/yr x 0.65 (capacity = 1.7082 x lo7 kWh/yr
factor
1.7082 x 10 kWh/yr x $O.Oll/kWh (from Table A.3) x 1.45 (levelization factor from Table A.l) = $1.016 x 106/yr $1.016 x lo6 + 0.187 (equivalent capital) (levelized fixed charge rate) = $5.43 x lo6
$5.43 x lo6 + 300 MW = $lS.lO/kW Thus, if the standard generator had a capital cost of $SO/kW, the superconducting generator cost could be no more than $SE.lO/kW to be competitive.
Example 2: Calculation
of Transmission
Cost of Service
Consider a transmission line. One wishes to estimate the cost of service (dollars per kilowatt-hour) that service being the transmission of electrical energy. Suppose the design capability of the transmission line is 100 MWe, the total installed cost (i.e., capital cost) of the transmission line is $30 million, and the electrical energy loss (including the electrical energy consumed for refrigeration) is 4% of the annual energy input to the line. The cost of service would be calculated as follows:
l
Electricity into system = 100 MW x 8,760 h/yr x 0.80 (capacity = 7.3 x lo8 kWh/yr
Electricity out of system = 100 MW x 8,760 h/yr x 0.80 = 7.008 x lo8 kWh/yr Capital cost component = $30 x lo6 x 0.187 (fixed charge rate from Table A.l) + 7.008 x lo8 kWh/yr = $O.OOS/kWh
Appendices
163
Electricity cost component = 7.3 x lo8 kWh/yr (into system) x $O.O41/kWh (from Table A.3) x 1.45 (levelization factor from Table A.l) i 7.008 x 10 kWh/yr (out of system) = $O.O82/kWh Cost of service = capital cost component + electricity cost component SO.O08/kWh + SO.O62/kWh = $O.OlO/kWh
If the system had operating and maintenance costs, such costs would also be included in the cost of service. For example, if the above transmission system had an annual O&M cost of Sl million, the O&M cost-of-service component would be: $l,OOO,OOO/yr x 1.45 (levelization = $O.O02/kWh Thus, the total cost of service factor from Table A.l) I 7.008 x lo8 kWh/yr
would be SO.O72/kWh.
In Sec. 6, R.A. Thomas and E.B. Forsyth emphasize that the above method refers to the delivered cost of electricity from a particular generating station (one charging $O.O41/kWh at the bus bar) over a line of fixed but unstated distance. One could also subtract the bus bar price of electricity from the delivered price (SO.O72/kWh in the example) and divide the result by the length of the transmission line to get a transmission cost-of-service per mile, as was done in Sec. 6.
Example 3: Calculation
Consider a device for storing electrical energy. One wishes to estimate the cost of service, the service being the discharge of electrical energy when it is needed. The following must be estimated: the total installed cost (i.e., capital cost), denoted by K; the annual consumption of electricity by auxiliaries (including refrigeration), denoted by C; and the schedule for charging (off-peak) and discharging (on-peak) the device. The cost of service (CoS) would be calculated as follows: COS = I(K x FCR) + [(B x Pi) + (C x Pa)]LF) where CoS = cost of service (S/kWh), K = capital FCR cost of facility rate, electricity into storage (kWh), (J), +A
= fixed charge
B = annual Pi = off-peak
off-peak
price of electricity
164
Applied Superconductivity
consumption
for
auxiliaries,
such
as
losses and
($/kWh),
LF = levelization A = annual
quantity
of electricity
delivered
out of storage
(kWh).
The terms A, B, and C depend on the design and are to be determined by the analyst. Consult the tables appropriate to the side of the meter on which the storage is placed for values for FCR, Pi, Pa, and LF.
Appendices
165
II
Appendix B: Superconductor
Performance
SUMMARY To provide a benchmark for the evaluation of high-temperature superconductors, ANL staff completed a series of measurements; this appendix presents the results. Because these results were obtained from one sample, they are physically consistent. The cost of was estimated at 2.2e/g ($lO/lb), and its density was taken to be ;;g;92cu307-x These series of measurements provided a starting point for the analyses subsequently conducted, which evaluated the expected performance of the new materials in specific applications. This appendix also presents the charge given to authors by this reports ANL organizers.
PARAMETERS
FOR
Experimental Results To provide a benchmark for evaluating the technical and economic performance of higher-temperature superconductors, the following properties have been experimentally determined by ANL from a single sample of YBA2Cu307_x. Although the attached data represent what we believe to be currently achievable in terms of performance for higher-temperature superconductors, these performance parameters may not be adequate for economic feasibility in specific applications. Therefore, your analysis should identify and characterize remaining impediments to the adoption of the new technology, given the experimental data provided. The experimental
l
Fig. B.1. This figure plots current density (in amps per square centimeter) versus magnetic field (in teslas) for seven temperatures (in degrees kelvin). Fig. 8.2. The data are the same as in Fig. B.l, except that current density is plotted against temperature for five magnetic fields.
166
Applied Superconductivity
C-J,
(77 K, H = 0) = 233
A/cm2
.3
0
0
Key
q
J, (73.0
K) K) K) K) K) K) K)
0 .
l 0
0
. l
0
n
0 .
000 .
* J, (68.7
J, (64.7
l 0 0 l 0 l
J, (77.0
M&ne+ic
FIGURE B.l Current Density
Field, H F)
of Temperature (!Source: Ref. 1)
Analytical
Considerations In your analysis, 1. we ask that you consider the following general questions:
What are the advantages and limitations of operating at higher temperatures with a refrigerant such as subcooled liquid nitrogen, assuming technical performance of the superconductor equivalent to that of low-temperature (4-K) superconductors? What are the limitations imposed higher-temperature superconductors? by the experimental data for
2.
3.
What are the performance requirements density) such that the higher-temperature economically competitive? assume a cost
density
material
of Z.%C/g and a
Appendices
167
4 A
Kev
0
J, (0.0
T)
R 4
4
0
J, (1.0 T) T) T)
44 ;
0 40
.
o J, (0.5
0
0 4_
A J, (5.0
0
.O
I
Oa 44
cl o
J, (10.0 T)
0 0
10
20
30
40
50
a::4
60
70
80
0 J, (12.0 T)
90
Temperature,
T (K)
FIGURE B.2 Current Density vs. Temperature as a Function of Magnetic Field (Source: Ref. 1)
Reference 1. Capone, D.W., et al. (Argonne Div.), personal communication National Laboratory, (June 1987). Material Science and Technology
Addendum
The information in Addendum I is from Department of Defense Superconductivity Research and Development (DSRD) Options: A Srudy of Possible Directions for Exploitation of Superconductivity in Military Applications, issued by the Department of Defense, July 1987. This plan was prepared by the ad hoc DSRD Working Group consisting of the following: Fernand Bedard, NSA; Tad Berlincourt, OSD; Gerry Borsuk, NRL; Charles Craig, OSD; Kenneth Davis, ON!?; John Dimmock, AFOSR; Arthur Diness. ONR; Edgar Edelsack, ONR; Donald Gubser, NRL; Dallas Hayes, RADC; John Hove, IDA; Bobby Junker, ONR; Michael Littlejohn, ARO; John MacCallum, OSD; Martin Nisenoff, NRL; Clyde Northrup, SDIO; Robert Pohanka, NRL; Richard Reynolds, DARPA; Kay Rhyne, DARPA; Michael Stroscio, ARO; Harold Weinstock, AFOSR; Nancy Welker, NSA; Ben Wilcox, DARPA; Stuart Wolf, NRL; Nicholas Yannon, RADC; and Richard Yesensky, SDIO.
169
170
Applied Superconductivity
I.
INTRODUCTION
associated (LTS)
with
traditional is relatively to
1000 LTS supermagnets, patient, resonance magnets, capability particle machines advanced. record levitate are are
enough
a human magnetic
lifesaving
1000
21 feet
particle the
the worlds
machine, of
accelerator. of unprecedented
An experimental
which
holds
(more
allow
above
LTS sensors
sensitive and
known)
in widespread applications. is
defense
the
technology
worlds areas
engineering known.
parameters
the
well
Addendum
I:
171
for
the newly
discovered the
high
unknown. risky
Accordingly,
prognostications plan,
at best,
such
subject includes
to great ambitious
Although goals,
it
is
could
activities at that
adherence outlined
Indeed, better
the be
might
as a map of
territory
worth
itinerary.
As greater engineering
and as sound
applications
and demonstrations
confidence
and according
to more rigid
to
the
above
limitations, of
this Defense
document
sets
forth
a five-year
Department
and Development
realized
for
military (sensors,
applications
including
Josephson-junction hybrid
and superconductor-semiconductor
electronics)
172
Applied Superconductivity
scale
(magnets,
rotating
energy
energy
program
management of
experience
science
applications of the
ceramics,
rationale
DSRD, and
describing which
development,
projects the
be included level of
proposed relevant
additional program
activity
and funding),
and (d)
summary of this
report)
activities
and (b)
some possible
and funding
DSRD activity.
It to
is
important the
to emphasize
at
the
outset
that
DSRD will of
seek
integrate
scientific
and technical
capabilities
academic, and is
industrial, to close
laboratories coordination
endeavor
pledged
both
federal
In
so that agencies
other
into
an integrated
federal
program.
encompass
a spectrum
of
activity
extending
from
through
exploratory
development
to demonstration.
Addendum
173
Contracts
will
be awarded activities,
for
efforts
ranging
from
single
through
multi-investigator
processing,
and of
altogether small
awards effort.
a relatively
fraction
It reservoir
is
intended of
that
DSRD provide
and expertise of
by DOD program projects. literally, embryonic worldwide unrealistic be dictated program areas plan
in support proposed
program
HTS is
diversity be
of
would
that
could
with
certainty
should
be viewed
only if
problem HTS is to
which
ultimately
addressed
be determined understanding
emerges
in areas of at
which
incomplete spectrum
and are of
to be representative
the wide
174 AppliedSuperconductivity
applications
areas. of
It
is
immediately
apparent all of
that the
mutual profit
interest from
will evident
that
many of
successfully
developed, sectors
will
in the
civil will
programs
by DOE, NSF, DoC, and is already facilitated Far East. in effect, by DOD
with
efforts is being
developments in Europe
and in the
Addendum
175
SMALL-SCALE APPLICATIONS 1. Optics and Infrared for focal plane arrays-detection/signal detectors/AD 6
Sensors
conversion/memory
optical
mirrors
temporal)
Specialized
superconductor/semiconductor Displays-semiconductor/superconductor command and control Homodyne/heterodyne Optical beam steering for lasers/wavelength Wave applications detection
systems
Q-switching 2. Microwave
shifters
and Millimeter
waveguides
Q-switched
millimeter
and structures
176
Applied Superconductivity
3.
Novel
Superconductor/Semiconductor Devices
Superlattice
and
Quantum-Coupled 4. Magnetic
Components/Detectors
confinement
SQUIDs Josephson
(Magnetic junction
field (JJ)
Inertial/Geomagnetic
Systems
Gyrotrons Field Aircraft Switches Power Supplies, (Helicopter) and energy Motors, Electrical Storage Generators, Batteries
Devices
Lasers/Nuclear 6. Electromagnetic
Simulations Guns/Launchers
Addendum
177
POTENTIAL
NAVY
APPLICATIONS
Multiplexers A/D converters Signal 2. Microwave Mixers Amplifiers Phase shifters lines processing and Millimeter Wave
Transmission 3.
Submarine Surveillance
High-performance, LARGE-SCALE APPLICATIONS 1. 2. 3. 4. Ship Propulsion for and Microwave Lasers and
Power and
Electron Storage
Pulsed
Power
178
Applied Superconductivity
POTENTIAL AIR FORCE APPLICATIONS SUPERCONDUCTIVITY SMALL-SCALE APPLICATIONS 1. IR - Sensors Multiplexers Digitizers Signal 2. mm Waves processors
OF HIGH-TEMPERATURE
Addendum
I:
179
POTENTIAL NATIONAL SECURITY AGENCYAPPLICATIONS OF HIGHTEMPERATURE SUPERCONDUCTIVITY ANALOG 1. Microwave/Millimeter Wideband Low-noise Multi-GHz Wave lines mixers processors arrays and amplifiers
antenna
A/D conversion
signal range
multiplexing mixers
dynamic
A/D conversion
A/D converter 4. Sub H-F Wideband, All DIGITAL 1. 2. 3. 4. 5. Back-plane Back-plane zero digital high
dynamic
range
ELF receivers
magnetic
sensors
resistance
dispersionless
180
Applied Superconductivity
II.
forty
years
ago,
first
low temperature
research
on a broad contract
laboratories helium
research of
a multitude funding,
in superconductivity.
superconductivity
generation
professors
Early bolometers
addressed radiation,
superconducting primitive
digital
computer
gyroscopes.
In the
superconductivity
thereafter
DOD-funded isotopic
a firm theory
microscopic 1950s
In the of
late
the
in physics.
Also,
researchers
contributions
Addendum
I: Military
181
steady
advance
of
temperatures
at which
superconductivity
be observed.
In the contributions
early to
1960s the of
DOD-funded of
researchers Josephson
science
the
and other
magnetic
sensors
submarines
and for
the detection
frequencies related
to the microwave
been
significant
technol.ogy-transfers
sector
in magnetic astronomy.
of
electronics digital
DOD programs
have
achieving
electronic very
In a joint
significant but
development
compact
Subsequent
DOD contracts
with
a small
demonstrated
on high-performance activity is
the worlds
oscilloscope,
182
AppliedSuperconductivity
as a diagnostic circuits,
tool most of
for
development will
of
which
ultimately
systems.
been
at
the
in a number of
aspects
of
test-bed
as replacements
transferring the
a shipboard ships
propellor.
In an Air
Force
lightweight,
superconducting to provision of
directed the
by superconducting microwave
magnets.
systems. efforts
magnetic
resonance
electric All of
insightful
investments
in superconductivity
Addendum
183
Cited efforts
above
are
just
a few examples
of
the
very
vigorous
DOD
to provide, which
by means of cannot
superconductors,
superconductivity military
in operational of
attributable required
has been
DOD has
actively
In one very
the heat
exchanger,
orifice,
and
contained refrigerator
chip. achieved of
micro-miniature
for is
finding
in missile
guidance chips.
systems
the highest
Now that
refrigerator
will
sensors
successful out.
HTS materials
scale,
the
era
of
the
that
military past
forty-odd of
years
finally
come in
knowledge
to exploit
HTS materials.
184
Applied Superconductivity
III.
temperature the
superconductors
are
ceramics, and
and
many of ceramics
as structural processing
bulk
and films,
crystal, the
and micro
structures
DOD is
in an excellent expertise
capitalize science
past
and present
in ceramics
engineering.
In the Program,
late
the major
Ceramic stimulus
which of
development ceramics of
high
domestic its
industry.
involvement
research
and fabrication of
at a variety strong
services chemistry,
has
efforts
in the
processing
as exhibited in the
by comprehensive
and in-house
Force
extensive
materials
science
expertise a very
coupled sound
with
DODS
provides
technical temperature
in a major
R&D effort
in high
superconductivity.
Addendum
185
IV.
emphasis
of
DSRD is
on exploitation will
of
effort
may be capabilities.
processing there is
capability, of executing
the
shortest the
development
initially It is
discovery
essential
as the
be executed
Of
and lowest-noise at
have
to be operated efforts
on the
continue
to be appropriate
for
such applications.
Included of
within
the
scope
of
searches
superconductors,
approaches
to processing
for
186
Applied Superconductivity
structures
(films),
bulk
materials
(wires,
cables, of of
rods,
bars,
monolithic
physical small-scale
behavior, (sensors,
shielding, In each
electromagnetic there is
instance
DOD presence.
applications for
both
small-scale that
it be
is
to composition,
crystallographic
Also
and microstructure.
required
is
and superconducting-state and mechanical for for higher properties. engineering building in the Such
essential capable
theory
materials
performance.
Such a search component unlikely applications instructive of that DSRD. the have insight,
for
is
also
moving
progrossion
in earlier-generation applications.
Addendum
I: Military
187
the
earliest
material
of
choice silicon
which in
supplanted of
by silicon. there
While are
supreme
applications
many special
military special
and quantum well In small-scale was from nitride. soft In was from to
required the
the
progression stannide
niobium-zirconium material,
niobium-titanium
Add to
evident space
compositional to be assured
identified
will
doubtless
While other
DOD will
surely
benefit
significantly
from of
efforts materials
of
organizations
(DOE, NSF, DoC, NASA) in areas theory, it and search is essential arenas. for
materials, activity is
in these
so highly theory,
as a
188
Applied Superconductivity
means
to
provide
focus
in directions Weight
of
greatest are
perceived paramount
impact in many
considerations of NASA),
and DOD has other and thermal dictate that shock, DoDas
requirements
as to radiation characterization
specific
investigations
be pursued.
manner,
in the
area
of
materials foci.
its
discipline.
to progress processing
have been
pursued
over
such armor,
as radomes,
IRdomes, Thus,
DSRD processing
encompass
crystal will
growth, address
importantly, desired
composition, in the
defect
surface
and properties
Addendum
I:
Military
189
geometty
required of the
and in conjunction required the full will characteristics benefits be closely but, of
with
materials
activities and of
with
agencies
industry, determined
above,
will
focussed DOD.
in directions
by the
requirements
experience
of
DOD in the
a firm
further to further
electronics-technology-based IBM, AT&T Bell in HTS, it simply then Laboratories) might, for at first
consideration, to
wait step
those
organizations the
commercial
This are
approach
appropriate, of overlapping
although
there
significant sensor
areas
one need to
compare
to military
alternatively, processors
consider
acoustic
anti-submarine counterparts) It
signal
(which
have no civil
the unique
requirements.
190
Applied Superconductivity
that of
DOD has
for
funded
effort
at
the
National
Bureau
Standards
(DoC)
laboratories electronics.
to address
specialized
DOD needs
in superconducting
includes which
activity
to
large
scale highenergy
arise
electric for
systems;
systems;
magnets; beam
compact for
gyrotron
has
requires
designed
winding
and winding
DSRD for of
significant
involvement
superconducting recognized,
wires,
however, margin,
an area greatest
in which experience,
investment,
greatest
different between
systems
applications.
and every
be made to profit
Addendum
191
V.
A.
Characterization
of
and Search
for
High Temnerature
Superconducting
Materials
there
is
sound of this
justification HTS it is
excitement early
regarding the
applications impact of
to perceive
remarkable will
to chart
a course
which
allow
generation
generation
higher
aggressive of
efforts.
needed. are
which
devices
and systems.
At the
data
comparisons will
existing of
models, theories,
development chart
help
the path
still-high-performance
later-generation
materials.
we catalog
the or,
essential
HTS materials
and to
192
Applied Superconductivity
performance large
materials. first
of
this
and the
materials
which of
together, prodigious
represent
a matrix
is
in magnitude
its the
characterization class of
for
superconductors of a
a period
and the
corresponding a road
base
has provided
charting
the
investigations
of
Clearly, will of
because
of
the
intense
activity
be difficult exploration
indeed no matter
the most effective highly spectrum on selected applications follow directly variety mentioned such flexible, of
doubtless
level
in depth to
as most
pertinent
concern
the
a course,
directing
focus sensors,
impact of
DOD requirements
and the
high-current-density proposal.
applications
Listed
below
are
the
types
of
measurements will
and theoretical
investigations
which
be undertaken
Addendum
193
progress
optimization,
in engineering
applications,
and in the
still-higher-temperature
materials:
1.
Transition of
T,,
of
chemical
isotopic operating
and of ranges
and provide
for
search
for
higher
T,
2.
Enernv nan.
of
temperature,
magnetic of of (or
field,
design
parameters based
Giaever) highly
as expected, will
be introduced
might possible
provide
opportunities
concepts
energy
electronics characteristic
levels
temperature
superconductors.
3.
Magnetic
field
nenetration field,
denth.h,
as a function These
of data begin
temperature, determine
magnetic
at what specimen
phenomena
194
Applied Superconductivity
to appear.
This
successful circuits,
design
of
superconducting magnets.
and film-based
4.
Josenhson of field,
junction junction
(JJ)
tunneling
nhenomena voltage,
as
functions magnetic
structure,
temperature, orientation,
temperature.
electronics sensors,
processors), sensors,
radiation oscillators.
and voltage-tuned
5.
of
with
electromagnetic state
The
zero
property
superconducting
correct small,
At finite losses
at large of from 60 Hz
oscillating
influence
the
operating applications)
greater properties.)
These configuration
material
processing
parameters,
microstructure,
195
6.
of
HTS materials
and will
devices explore
radiation. photonic
investigations analagous
to those structures.
already This
will
be carried
concepts
for
electronic-photonic
make use
semiconductor-dielectric-superconductor
structures.
7.
Thermodynamic
DroDerties, as functions
viz., of
specific
heat
and
both
temperature measures of is
provide which,
in turn, might
a measure to
be introduced densities
at high index
Condensation relative
thus
an important for
a material
suitability
supermagnet
applications.
a.
Critical
magnetic
fields
(viz.: field;
lower
or
magnetic-flux-penetration-onset or upper-field sheath critical limit field, all These operating applications. of
upper vortex
or upper-field of the
temperature available
fields
and in a
regimes
superconducting supermagnets
For example,
196 AppliedSuperconductivity
operate while
at fields
of
the
order
of
one-half
to
two-thirds
of
Hc2,
cavity that
resonators
in the absent.
Meissner
supercurrent
9.
Approaches
introduction pinning
of
material vortices as a
supercurrent
supermagnet defects,
performance. will
such
and voids
be considered.
10.
Determination
of
the
point
a material
surface.
11.
of occur
flow.
creep,
and
iUlllDS.
These
transient
are
requires
jumping nitrogen
twenty liquid
greater
at liquid
temperature if
Accordingly, densities
support nitrogen
at high
197
their
energies potential
must be wells to
to allow
as vortex
trapping
and thermomechanical of to winding large purposes, (by currents mechanical data are
aspects. with
generated
stresses. required
Hence,
engineering
design
on tensile,
moduli
and by acoustic Because to periodic supermagnets applications for high well and are of
are
importance,
effects
in magnet
electronic either
applications, with or
superconducting in composite
in contact
materials. in must be
adequate
thermomechanical thermal
designs, all
appropriate
components, because
normal, expansion
and
However,
considerations small-scale
apply (sensors
to both
large-scale applications.
and electronics)
198 AppliedSuperconductivity
effects.
All
sources
of
thermal systems
superconducting above)
data
engineering aspect
provides activity,
for
adequate
an ancillary
this deserve
magnetocaloric on the
consideration for
a basis serves
a magnetization
in which Prepared
as the working
substance.
superconductors
14.
effects. of high
circuits
migration effects
onset
conditions
both will
superconducting
applications
be investigated.
15.
Atomic
level at the
Full is
of
HTS of superior
materials existing
optimization of
materials
progress of
development very
powers are
following x-ray
effective
approaches crystal
required:
neutron electron
structure electron
determinations, microscopy,
microscopy,
scanning
Addendum
199
magnetic
resonance,
electron diffraction,
spin
Auger
spectroscopy,
electron
tunneling
16.
Chemistry.
In-depth
crystal
chemistry
plays of
the the
role
in materials
synthesis
understanding useful of
to
structure with
with chemical
properties interactions, of
concerned
vapor,
ambients.
Oxygen and to
atomic the
diffusion
information
critically for
needed fixing
determine stoichiometry
requirements as well
the
long
compositional
stability. polycrystals
equilibria reaction
establish
chemical
and materials
be in contact
and subsequent
use.
17.
Effects
of
ionizing have
radiation. to operate
In a number of in ionizing
applications
HTS materials
will
radiation
200 AppliedSuperconductivity
It
is
thus
essential account
designs range of
to
transient
upsets
radiation
massive employing
damage.
There
possibility desired
processing
as a means for
enhancing
properties
in some instances.
18.
Experimental
with
(GLAG) macroscoDic HTS materials and development undertaken with with are
Evidence II
superconductors.
However,
and systems of
cannot
experiment
GLAG theory of
Landau kappa value parameters resistivity,&); properties. anticipated, dependent& coherence critical for (electronic
transition specific
and normal
heat
electrical
a strongly
markedly
length, fields of
and hence
properties), anisotropy
probably of the
inclusion
layer-like
Addendum
201
or
linear have
character profound
of
This
latter utility
could
impact
ultimate
materials.
19.
Experimental of
comoarison
with of the
microscopic
theories.
Tests
of (BCS)
compatability with
Bardeen-Cooper-Schrieffer are of critical in the suggests may not from may help of
theory
shift
temperature pairing
interaction spectra,
be
quasi-particle particularly
to the for
hallmarks its
invite
are
and for
resonance nature of of
to extensions concepts,
to creation
and to
still-higher-temperature
materials.
20.
Electronic-energy-band
structure
Electronic-energy-band experimental
determinations of the
are
understanding rise
remarkable
electronic
to HTS and to
suggest
further
HTS search.
Of particular the
side,
as means to determine
electronic
202 AppliedSuperconductivity
are of
measurements cyclotron
of
photoelectron and of
and,
if
possible, such
resonance
Alphen
effect.
be larger
appears of
magnetic
order
de Haas-van for
in fact
characterizing
and the
Not to be overlooked which of may relate HTS materials, because of of the of are
special properties,
behavior which,
magnetic earth If
ions, not
a variety
types
ordering. between
and not
such
could fully
lead
Conversely, useful
understood
new design
alternatives.
Addendum
203
Provosed
Budnet
CHARACTERIZATION
&!!
Budget 6.1 6.2 6.3A Total 0 10 Category FY88 10 0 0 16 m 13 3 0 15 m 9 6 0 16 FY91 8 8 0 14 FY92 6 Total 46 25 0 71
204
Applied Superconductivity
B.
Processing
1.
Introduction. from
stage
of
derivative
generally based
copper
oxide-barium of this
oxide-yttrium program it
system. these
that
often
ceramics,
be the
ones
focussed
upon for it is
applications.
that
development. involve
future
materials
as nonoxidic
polymers.
We may have
learn
how to
and this
places available
processing
intended
paradigm is
addressed of
in the
materials
science
and
definition
structure-property
Addendum
205
for
each
particular says
system structure
under of a or
This
paradigm size
scale, to
whether its
surface, defines
bulk the
end-use
paradigm of
is
processing
understanding
to develop
prepare
(manufacture) (structure,
product
displaying state).
character processing
Ceramic
sensitivity scale,
character the
to processing effects of
control
operations bodies
meant to of
surfaces,
or produce
or specimens
processing of ceramic
operations materials,
profoundly with
influence cost is
derivative technology
concerns. on techniques,
Manufacturing processes, of
very
interest on the of
Depending with
more monolithic
concern in bulk
are
both
and in film
forms.
The processing
206
Applied Superconductivity
methods separately
are
somewhat below.
different
for
these
cases
and are
considered
2.
Thin
Film Materials.
Devices
and Circuits
a.
Introduction.
for are
thin those
film used in to
industry. thick)
priority
the
ability
(and
metallic
defined
character multilayered
to process
to be used the
to detect
signals format.
obtained
A
circuit chips
fabrication whose
circuits
controllable
and reproducible
and which
are
corrosion
resistant,
extended on a lesser
at or near in the
semiconductor
from earlier,
a systematic
high-temperature technology
principle. long-term
successful
development magnitude.
demand a very
considerable
Addendum
207
circuit consisting of
technology of
Often
requires
as many as a dozen and insulating layer is layers deposited, and it with HTS the of of is
conducting
dozen
steps.
As each geometry
then processed the next electronic deposition compatible these films, layer.
a desired to
and then
Thus,
establish
a technology
integrated of the
circuits,
superconducting layers
insulating both
and techniques
superconducting challenge,
into the
layers.
superconducting necessarily
films require
will the
have
applications of to
in
junctions. coat
For example,
a high-frequency losses
resonant
films
can
as shields
ambient
electronics originating
surrounding Films
machinery,
there that
special
inherent
character
208
Applied Superconductivity
these
materials.
for the
these
devices
are for
in some
on
vacuum in the
deposition
techniques, structures
however, using
been
used
to produce
similar
more conventional
lower-transition-temperature
materials.
b.
Thin-Film
the
recently are
discovered
highsystems, a the
four-component Accordingly,
deposition of
a challenge.
should
be explored the
and evaluated
for
including
following:
Sputtering Thermal
etc.)
Electron-beam Laser
epitaxy epitaxy and other growth (RTA) techniques to optimize techniques for assisting/
film
Organometallic
Addendum
209
techniques
should
be tried film of
using
single
desired
arising the
between post
annealing substrate
diffusion
addition,
techniques
associated techniques
with
should
this
program. become
At the
or two, depositing
should these
apparent
which
high-temperature
materials
the greatest
potential.
C.
Materials are
of
Films.
As thin
films
of
these
must be extensively
characterized,
electrically
following which
characteristics should
be determined
stability character
210
Applied Superconductivity
Mechanical Adhesion to
properties substrates
including
thermal
shock
response
Electrical include
and magnetic
properties
which
should
be measured
density fields
Temperature Microwave
and critical
spectrum
These
chemical,
physical,
properties
and correlated of
films.
optimize techniques.
properties
as functions
d.
Device to
focus
has been
given
should loops,
The techniques
structures
Addendum
I:
211
(modified)
lithography
These
techniques
with
respect
to
their
and the
layers
required
fabrication.
it
will
to
various isolating
insulating
oxides, from
metallization insulating
The properties
of
these
should
Pinhole Prepared
free
to
insure
positive that is
by technique
preparing Thermal
superconducting comparable
expansion
212 AppliedSuperconductivity
different selective
etching
rates
than of
that
for
HTS films
processing
multi-layered
of
insulating chemical
layer
materials
will
the high of
be determined
portions
program.
of
these
there contacts
dependable Small
small
resistance milling
such
to depositing or
contact
implanting
to provide be selected
films,
which
must be determined
during
early
program.
of
Josephson
structures
should in a thin
The The it
a narrow of
advantage only
Josephson deposition.
a single of of
thin
are
that
it
Josephson placed
junction
and that of
requirements
on the dimensions
Addendum
213
constriction.
Typically,
and length
of
the It
the easy
order to
devices
the
for
current, device
requirements iunction,
phenomena,
structure. regions
consists thin
by a very
barrier,
a semiconductor, 1 to 10 nanometers
or a normal thick.
The tunneling is
a current
across
interface the
crucially A very
coherence high
length
quality of
to a distance In the
surface.
case
coherence
has been is it
to be of lattice there
challenge
quality
properties
surface.
quality for
T, electrodes
requirement
temperatures
or high
temperature
post-deposition
214
Applied Superconductivity
to
obtain
the high
transition
It the the of is
is
straightforward electrode
to use while
elevated elevated
processing processing
deposited.
fabrication
all-high-Tc the
how serious
an obstacle might
coherence the
temperatures tunnel
be for both
junctions
with
electrodes
T, materials.
tunnel the
junctions
devices
have
following
characteristics
should
be studied:
of
the
electrodes
and of
the
The superconducting temperature The sub-gap procedures The critical dependence Specific capacitance field current
gap of
the
device
and its
of
the
device
and its
temperature
of
the
dependence
physical, device
and thermal
Addendum
I:
215
Document
the of the
variation
of
critical procedure
current
density
as a
currently devices
available with
device
processing
minimum crossvalues of
areas current
exhibiting
minimal
and capacitance
to Josephson
devices,
electronic inductors,
circuits capacitors,
require filters,
signal etc.,
couplers,
fabricated
ground a thick
electrical the
components
Develop
superconductor-thick
insulator-superconductor
structures Explore various loss the insulator associated materials with to minimize the insulator stability of the
chemical,
physical
and thermal
structures that the lowest step electrical edge loss insulator can
continuous
coverage
216
Applied Superconductivity
selection
of
the
most
promising
thin
film
a multi-layered
will be developed.
superconducting Specifically,
integrated the
following
Select
deposition films
technique
for
the
the
reproducibility
fabricating to of
densities
approaching
structures
containing circuit
Josephson components;
electrical,
and phyical
characteristics If necessary, the electrical, film modify the processing technology to optimize of
chemical
characteristics technology
integrated
with trade-off
the
of
the
above
be conducted electronic
superconducting conventional
sensors, semiconductor
technologies.
Addendum
I: Military
217
activity
will
guide
projects payoffs
in directions in military
determined capability.
to offer
3.
material
electric fields
critical
current
densities operating
presence
magnetic the
at reasonable
This
requires
introduction, lower
of
Lorentz of
flow
electric
dissipation.
be placed in for
on development of the
processing
appropraite desired
operating
Moreover robust
the brittle
to withstand processing
ceramics of
concepts
bulk (or
of strong, provide
the
other
or metallic
superconducting
windings of
motor, ceramic
Toughening
superconducting of
the
fiber/matrix failure. A
bridging
catastrophic
218
Applied Superconductivity
of
other
processing be employed
techniques
(e.g.,
dynamic depending
component
microstructure
and ultimate
application.
processing supercurrent
which
have
(corrosion
characteristics Examples of
required such
for
devices given
requirements
Electromagnetic
Motors, generators, energy storage, mm-wave tube magnets Magnetic resonance imaging
to optimize diffusivity of
in the
composition
ultimate
Characterization
similar
Addendum
I: Military
219
In some ways, greater continuing techniques. such effort proposed challenge strong While would in the
of
bulk
is
films.
and 6.2
programs effort is
to
explore
explicitly programs
applications
as those
following
sections.
of
HTS materials
every
ceramic affects
processing
evolution have
character to uniformity,
a material.
Special
attention
to be paid agglomerate
preparation, density
gradients, phase
final and
impurities,
equilibria,
For the
solids
route,
the
relationship
on starting of bulk
cables,
through with
utilization metallic
recently of
developed appropriate
approaches
shaped
alloys
220
Applied Superconductivity
composition, desired
which
are
subsequently This
reacted work is
to
yield
ceramics stages
of of
composition
and shape.
in early
development.
fluids
route,
glass-forming
and glass particle importance direct While are suitable of processes for these to
vapor interest is of
processing
from
feed
in film masses
family
might
for
coatings,
finishing, methods
Investigations of very
importance
fine-grained
materials, (but
prepared vapor
by spraying for
in bulk
deposition (growing
and advancement
techniques
them directly).
4.
Single
Crystals. methods,
crystal methods,
growth flame
techniques, process,
melting,
hydrothermal etc.,
be explored Similarly,
potential battery of
bulk
single
described for
in 2, single
above crystal
be sol-gel
precursors
for
films.
Addendum
221
ProDosed
Budnet
PROCESSING
RSD
&!!I
Budget 6.1 6.2 6.3A Total 0 13 Category fYJ$ 6 7 m 9 15 1 26 FY90 6 17 A 27 FY91 5 15 12 32 FY92 4 13 14 31 Total 30 67 32 129
222
Applied Superconductivity
C.
Small
Scale
Annlications
and Demonstrations
exhibit
unique
properties
which
allow systems.
sensors following:
and electronics
properties
Zero
DC electrical (into
resistance the
resistance materials),
1OOGHz range
Exclusion
of
magnetic
flux
from limit).
the bulk
of
the
material
(up
to a materials
dependent
Quantum effects very zero flux non-linear resistance quantiration effects. tunneling tunneling
AC Josephson
list list
of of
phenomena, generic
an
electronics
of
at temperatures electronics
above
Refrigerators weight
are
in comparison there
In addition,
Addendum
223
fluids that of
whose
latent
heat
of
vaporization either
is a small
than same
helium. power
This
permits
dissipation
a system to these
compared cooling
helium
cooled there
advantages
to operate
semiconductors
and superconductors
same cryogenic
environment.
Future severely
semiconductor
integrated delays
circuits
will
by propagation
in interconnects.
integrated
ultrasmall
picosecond-scale
superconductors increases
upon magnetic
exclusion,
systems
digital
interconnects
Passive
wave components
High performance
analog
filters
224
Applied Superconductivity
many potential
of
HTS which
depend
upon
quantum effects
following:
Magnetic
switches
and circuits
contributions
operational
areas:
Devices
Circuits
Onerational
Applications
Magnetic
sensors
mm wave components
space space
communications radar
Addendum
225
IR sensors
A/D converters
Logic/memory
antenna
There
are
a number of
generic
materials-based to applying
questions
which
that
nanometers
to tens
microns
can be full
sheets
lines;
deposition, technology
advances
research
in several of
including:
characterization interfaces;
high-quality of
the measurement
fabrication
interconnects.
226
Applied Superconductivity
of
the
usable the of
temperature critical
domain
for
the must
considered,
density films of
as a function
temperature behavior of
the
critical, speed. of
because It is
often films is
to mean high
imperative connection:
form
electrical
interface
metallurgy
which
is
compatible element.
both there
materials of
context
to be used.
to devices, junction
in speed their
devices
higher
in higher
devices
while
low power
Nb and NbN superconductive in this connection insulators such junctions for devices. may use at is the
importance
in very
applications
temperatures
even higher
than present
semiconductor
Addendum
227
materials
for
junctions quantum
sensing
(superconducting number of
systems
opens
up a large
effective
or by all
intercommunicating processing.
so as to realize
to
a number of Although
specific
smallpreclude has
uncertainties
these
applications, that
prospects
earliest
realization.
1. for
Magnetometers measuring
useful Gauss
magnetic SQUIDs.
superconducting lo-11
Systems
have
fielded
10-D to
Gauss and have beer. used The unique response today feature
in both is the
situations. with
a frequency
tens
principal this
drawback
limited
wider of the
has been
cooling helium
Essentially If
present
systems
liquid
can produce
SQUID sensors
228 AppliedSuperconductivity
of
substantially a number of
the very
that
present
ones
offer, to be
then
applications
may prove
practical.
Gradiometers be fielded electrical reliability detection systems. used for with power
for
submarine dewars
could
hold-time thus
or very
refrigerators logistics.
greatly
enhancing could
reduce detection
the
sensors,
based,
could (e.g.,
be
vehicles
trucks, Arrays
as well devices
be deployed,
so as to realize
known advantages
advantages unmatched of
of
sensitivity
strength
reasonable (if
exceptionally
low power,
necessary)
mission
It will
be the use
goal
of
this
activity field
to develop
in magnetic lo-lOG/E.
sensing
instruments
sensitivity
Sensors junctions
of
weak sensing
links
versus is
field
elements
Addendum
229
e.g.,
are
tunnel
junctions Experimental is
new materials? if
one type
superior will
application
be characterized.
been
known,
will
be experimentally
verified.
Noise
Measurements
- After
the they
feasibility will
of
DC and/or
RF
established, at all
be thoroughly of interest
to noise
frequencies 27K.
operating
at temperatures
above
SQUID sensing
element
will
be
or semiconducting to
operating
Such a system
simulated,
fabricated,
and tested.
230
Applied Superconductivity
ProDosed
Budnet
a!!
Budget 6.1 6.2 6.3A Total 0 4 Category FY88 2 2 -z 6 FY89 1 2 r 3 FY90 1 1 0 0 FY91 0 0 0 0 FY92 0 0 Total 4 5 _tl 13
Addendum
231
2.
Hybrid
Svstems. circuit
do not,
designer
accustomed.
gates
semiconductors
but wherein of
in passive system
devices. which,
can even
perform alone.
technology
can accomplish
and relatively is
easy
use
of
systems.
achieve are
performance, cooled
designers As the
semiconductors.
demands on the
become
transmission of wiring
levels
approach long
present
low temperature high narrow speed, lines, current (c<lns possibly densities (at on
pulses
two planes. of
In addition,
critical than
appear
present) resistance
consider
distribution permit
zero reduced
1Opm width.
could
232
Applied Superconductivity
switching large
noise
power
distribution multichip
of system.
very
(-1OOOW) supply
to a large
plane improved
array if
systems the
using
can be
preliminary
accomplished
with
speed the
and
interface
to both
electronics compared
wave receivers,
wherein
cooled
produce the
interconnects bandwidth
a superconductive
an added
improvement.
may be required;
with
HEMTdevice.
exploit
being
by supercomputing of tens to
architects of
for
envision
separate
to commonly systems
accessed fast
demand very
with
resolution.
The merits
Addendum
233
less
expensive, of
slower the
processors performance
to
achieve
hundreds
times rate
processors.
uniaue
devices, network
to build
switch
between of
these such
room-temperature parallel
large
systems
and hydrodynamic
to name a few.
of
a switching such is
network, arrays
there for
are which
as antenna for
required
systems
enhanced interface.
HTS switches
of
to
semiconductorand unique
higher
performance
It will improvements
of
this
effort
to assess
the
performance systems.
at effect
the of
chip
on system e.g.,
parameters and
have
impedance,
structures
be fabricated
234
Applied Superconductivity
Analysis
will
the
systems
where e.g.,
can
be obtained with
technology,
superconducting where
be modeled
to determine
If materials
the
results
of
this of
study the
are
positive,
a detailed
investigation will
demonstrate
hybrid
performance,
designed,
simulated,
fabricated,
and tested.
Addendum
I:
235
Prooosed
Budnet
a!
Budget 6.1 6.2 6.3A Total 0 4 Category FY88 1 3 FY89 0 3 -5 1 m 0 2 FY91 0 0 Li 0 0 FY92 0 0 u 23 Total 1
236
Applied Superconductivity
3.
progression
from
the
centimeter of the
wave region electromagnetic to decrease. performance a serious performance adequately. sensing aperture not
tends
electronics systems
thus
problem systems
designer.
presently such
demand cooling
Systems
as communications radars,
passive
arrays, radars
array
antennas,
technology
realistic
improvement.
If present
the
electrical
performance
which
metal
superconductors at higher
demonstrate temperatures,
improvements
a phased
amplifiers, for
can these
be cooled
Since, i.e.,
A/D a
data
oscillator
power would
least
of
Addendum
237
This
would
greatly
reduce number of
the
spurious
emission for
and the
allow
a larger
receiver
channels
same size/power.
where system
power could
is
at a premium,
the small to
a cryogen the
and weight.
Of course, refrigeration
makes very-low-power
an attractive
alternative.
for
are
very-low and
bandwidth, capability.
power,
very-high-performance
The goal enhancement evaluation made ; this transmission performance semiconducting will use data
will of of
be to demonstrate
a high items
performance
will
include
such
as antenna
lines,
mixers, using
digitizers, superconducting
improvements components
be investigated. analysis
This
by theoretical presently
available
materials
to guide
application
materials
important
developmental
238
Applied Superconductivity
Proposed
Budnet
mm WAVE RECEIVERS
ad
Budget 6.1 6.2 6.3A Total 0 6 Category FY88 2 4 -z 6 FY89 1 2 J 5 m 1 1 1 2 m 0 0 0 0 FY92 0 0 is 19 Total 4 7
Addendum
I:
239
4.
Infrared
Sensors.
present
infrared
sensors at
already long
require
cooling
in order
to achieve tunnel
good
Superconductive bandwidth
arrays of
the
transforming sampling It
currents
digital
format
adequate
very
demanding.
to use
superconductive a number of
to multiplex analog
convert
signals.
superconductive power,
a be
heat little
consequence. by a liquid
For modest or a
no need
for
sensor
data
are
in digital
format,
it
is
feasible the
to data
information
serially
thereby output
transferring high
standard
room temperature
with
efficiency.
advantages very
of
sensors
and
low power,
sensitivity,
and small
240
Applied Superconductivity
in this high
development performance
is
to determine sensing
the
of
infrared
systems. with
as sensors Moreover,
in comparison
be investigated
in conjunction These
semiconducting include:
and superconducting
sensors.
registers. cryogenic
In addition, system
between
be investigated.
the
critical with
components lower
available
temperature
Addendum
I: Military
241
Provosed
Budnet
INFRARED SENSORS
a!!
Budget 6.1 6.2 6.3A Total 0 5 Category FY88 2 3 r 4 FY89 1 2 1 6
m
1 2
m
0 0 1 2
FY92 0 0 0 0
Total 4 7 Ji 17
242
Applied Superconductivity
5.
Digital
Svstems films
(Lonicl.
A number of
properties
of for digital
and devices
junction using
switch
using
speed
very
compact,
faster
may be increased
It
is
well
known that
using
lines performance.
throughout Present
a computing
system with
superconductors, lines If of
between the
devices
and chips,
no resistive
distribution
gap,
material
property,
can be used
power
locallv, envisioned
on chip. to
Finally,
in an era and
configure
a system,
currents circuits
to destructive looks
superconducting
approach
especially
Addendum
243
resistance eliminate
combined greatly
should enhance
Given electronics of
the
above
characteristics
the
implications the
for computing
digital power
are
remarkable.
a Cray I contained
neon days
a system of
cryogen,
an efficient of unregulated
refrigerator
consuming is the
10 watts of
6.5cm
5 watts. power
a large
memory.
be dominated of unregulated
by the power.
refrigerator, (The
consuming
several
CRAY II,
uses
about
65KW of
refrigeration
150KW of
power.)
volume, that
and power
are
at a premium,
such
performance. radar,
such
aperture
processing,
superconductive
technology
offers
244
Applied Superconductivity
of
this
project
is
of following
functional
superconducting
parameters:
Device Fully
time
delay
superconducting to support
logic
families
will
a general logic
purpose will
computing
be investigated. sunerconductinn
use present
temperature
initially, will
Logic
be simulated of
selection will
be fabricated e.g.,
Selected will be
functions, designed,
multipliers,
registers,
simulated,
fabricated,
and tested.
Addendum
245
ProDosed
Budnet
DIGITAL
SYSTEMS
(LOGIC)
&!I!
Budget 6.1 6.2 6.3A Total 0 2 Category FY88 1 1 1 4 FY89 1 2 1 5 m 1 2 m 0 0 _1. 1 0 0 FY92 0 0 4 12 Total 3 5
246
Applied Superconductivity
6.
Digital
Systems
(Memoriesl. which
of be
superconductive applied
devices
units
low power
and high
in the loop,
in a superconducting non-volatile
is
no power
as long
as the of nearly the CPU system rate can be units quite the and
allows
a high of the
memory,
with
superconducting bulk
semiconductor also
memories
the next
Cooling their
semiconductor
reliability.
It
is
the
goal
of
this
project the
to demonstrate parameters:
feasibility
of
superconducting
RAM with
following
Access Sizez4K
time bits
-z 1nS
w/cell 227K
Addendum
I:
247
for the
fast
will
be designed of available
to
superconducting
to
will
be simulated,
Small
be fabricated
and tested
acceptable
a test
simulated, contain
fabricated 1K bits.
and tested.
will
248
Applied Superconductivity
ProDosed
Budnet
DIGITAL
SYSTEMS (MEMORY]
a!!!
Budget 6.1 6.2 6.3A Total 0 3 Category FY88 1 2 FY89 1 3 -L 5 r 4 FY90 1 2 1 1 FY91 0 0 0 0 FY92 0 0 1 13 Total 3 7
Addendum
I:
249
7.
Devices.
device,
Josephson to
the voltage
by directly to
injecting
which
ingeniously have
relationship
i.e.
junction
analog
digital,
3hree-terminalf enhance, it
reasons, at switching
important
speeds power
of 2 lOpsec, gain.
dissipates
a power
of
this
effort
is
to develop
a practical device
logic with
on a true
three
approaches of higher
investigated
the
operation.
structures if
any deliver
250
Applied Superconductivity
be directed realization
toward of
conception
of
completely
to
the
useful
three-terminal
superconducting
Addendum
I:
251
ProDosed
Budnet
SE!
Budget 6.1 6.2 6.3A Total 0 2 Category FY88 1 1 0 2 m 1 1 0 3 m 1 2 0 3 FY91 1 2 1 2 FY92 0 1 Total 4 7 _1. 12
252
Applied
SuperconductivitY
a.
like
Systems all
Demonstration
Vehicle. into It is
others, a useful
possible for
feasibility stage.
use without
or component inputs
problems,
with
be addressed
by designing it
adequately
it
is
the
goal
of
this
using
structures of
described
above,
and/or
above
will
be
evaluation
be made of
that system
non-
Addendum
I:
Military
253
Provosed
Budnet
a!!
Budget 6.1 6.2 6.3A Total 0 0 Category FY88 0 0 0 0 FY89 0 0 1 2 FY90 0 0 Ir 12 m 0 0 ~ 0 0 zs 25 Total 0 0 39 39
254
Applied Superconductivity
9.
Refrigeration.
Presently
most
of
has centered
a transition like
cryogenic point of
would
liquid
neon with
nitrogen of
at 77.3K
vaporization to 50 times
50 times hold
which load
time
the
in open closed
systems
efficient
weight, on the
requirement
In space
80K.
As
higher
temperature coolers
developed
thermoelectric
may be adequate
some applications.
It
is
the
goal
of
this
effort
to develop of the
so that
advantage
previously of
helium
A careful will
evaluation
77K range.
refrigerators to maximize
be designed,
built
efficiency
and reliability
and weight.
Addendum
255
Cooling active
capacities
will
range
with
a few all-
devices
such
5 watts
A separate cooling
capacity
be the
for
use with
hybrid
systems heat
contribute cryostats
a significant will
load.
helium
what
to optimize
them for
radiative from
be
handle
10 mwatts will be
watts.
performance
temperature applicatiom of
superconductors thermoelectric
be evaluated.
256
Applied Superconductivity
Proposed
Budnet
REFRIGERATION
a!
Budget 6.1 6.2 6.U Total 0 3 Category FY88 1 2 A 5 m 0 2 r 2 m 0 1 0 0 m 0 0 0 0 FY92 0 0 A 10 Total 1 5
Addendum
I:
257
proDosed
Budnet
SMALL-SCALE
APPLICATIONS
SUMMARY
a!!
Budget 6.1 6.2 6.3A Total Category FY88 11 18 A 29 FY89 6 17 11 40 FY90 6 13 18 37 FY91 1 2 22 2.5 FY92 0 1 26 27 Total 24 51 83 158
258
Applied Superconductivity
D.
Large-Scale
ADDlications
and Demonstrations
applications primary
of
superconductivity i) of
of
concern of ii)
to all
technologies:
by themselves
larger
RF systems, isolation.
shields
these
rather materials
the
lower for
Nb-Ti
wires not
these
technologies
have
not
all)
will
be required to
pellets
Similar of fabrication
problems
procedures
superconductors, with
expected
the new HTS materials. made significant programs, power progress e.g.,
development pulsed
electric
airborne
generators,
resonator of
particle
accelerators presents
on this
DOD experience
reduced
refrigeration
requirements
implicit
in their
Addendum
I:
259
superconducting
materials
of
greatest
are
type
II
an intrinsically which
critical
field surface
superconductivity
layer),
and a large
condensation critical
(corresponding
Hc)
to a large related to
thermodynamic
magnetic current
field,
which
is
in the
is field
the
H,l, magnetic
which flux
equilibrium of a type
conditions, II
first
superconductor. type II
unsuitable is
resonator
there
some evidence
that
in cavity of this
resonator scientific
rests
heavily
resolution
The new HTS materials values relatively These facts (a value high of over
extremely reported)
high and
Hc2
H, values that
2 Tesla limits
reported).
suggest
intrinsic
amount of
260
Applied Superconductivity
that
by the new superconducting the new HTS materials high-power and Hc are they applications also have of
large. (sO.lT);
values
RF cavities parameters
important limit
applications,
but
do not
the maximum
As already extrinsic
emphasized of
the
critical
current
density, It is
J,,
is
an by
property
a bulk
(voids,
which
processing.
These
magnetic
a large
anisotrophy J, it
large-scale H. a wide
fields, over
be evaluated fall of
fields. of
reports research
Jc with is needed.
H; hence,
property
be important which
to ascertain J,(H)
of
enhance that
and which
defects plays
the role
crystalline
anisotropoy
in limiting to
at higher will
temperatures result
J,(H),
activation increased
pinned creep
temperatures
jumping.
Addendum
261
will
of
course
limit
the
maximum attainable
fields
of
supermagnets
operated
at higher
temperatures.
development depends
of
large
scale
applications careful
critically.upon
in concert of the
revolutionary
particular
demonstration in early of
on progress
achieved
development
components will
be strongly that an
industry is
so as to
base
available
production
development
completed.
of
supermagnet
structures solenoid
will
not
be most
standard
Consideration structure
be given
at the
and at the
cavities instances
have
a common to
be necessary
262
Applied Superconductivity
produce numbers
dense of
materials
with
having
minimal
surface will
defects.
be explored sprayed
ceramic
plasma
coating
techniques,
and thin
techniques.
From among the many important superconductivity, indictions of several where benefits are
DOD large
scale below
of
summarized
by incorporation
years;
not
a rating not
in terms
priority
and should
be so construed.
1.
(near specific
term).
Shields
to confine are
fields large
regions of
in space
needed
applications electric
are
to minimize
environment.
circuitry
are
used
Shields
are in
importance energy
confinement systems.
and isolation
kinetic
and directed
energy
weapon
with etc.,
Ceramic
structures.
Testing
Addendum
263
characteristics
and
design
concepts
disturbances the
H, T) require
testing. shielding
applications,
Hsh(max), smaller
be signficantly
performance
parameters,
the required
geometry.
for
H,h(
are
determined, areas
system are in
Obvious
systems,
shielding. and
development
motors
under
development
development could
projects
prototype
shields gains
be replaced
in performance
and/or
economy.
Cost
of
this
would
$4M for
materials
includes of
the near
term character
activity.
264
Applied Superconductivity
Proposed
Budnet
SHIELDS
a!!
Budget 6.1 6.2 6.3A Total 0 3 Category FY88 2 1 FY89 1 2 _1. 4 1 3 FY90 0 1 1 2 m 0 0 0 0 m 0 0 r 12 Total 3 4
Addendum
265
2.
SuDermaenets
for
Microwave of
(near and
Development
high-power, is of major
beam interacts
magnetic
field. of
electromagnetic field
radiation
tunable
to coincide
at 35 (100) Fields
Tesla. typically
supplied are
magnets.
used
supermagnets at 4.2K.
at significantly consequent
temperatures
as high
as 77K) with
convenience
and economy
in refrigeration.
development
efforts
of
this
on conductors
matched
wave tubes
already
modest
modest
considerations anticipated.
mature,
and so early
can be
Cost
of
the
effort
would
be $94 for
materials
magnet development,
distributed
as shown in the
266
Applied Superconductivity
Provosed
Budnet
WAVE SOURCES
a!.?
Budget 6.1 6.2 6.3A Total 0 1 Category FY88 1 0 m 1 1 -Q 2 r 3 m 1 1 1 3 m 0 1 1 1 FY92 0 0 3 10 Total 3 3
Addendum
267
3.
Supermagnets
for the
Shin
Pronulsion
Systems engaged
(mid in
E far
term)_.
Since of
early
development large
as alternatives power
to
the a
reduction to
from
required
Under this
superconducting installed
motors
Cheasapeake include
flexibility A 40,000
design,
HP superconducting of 4 in weight
of
a factor
40,000 with
refrigeration
more efficient,
more convenient.
of for
this
program
will of
HTS
replacement
in existing As greater
developmental expertise is
electric gained
designs active
and/or as
shielding on shields
the above.
to be developed
268
Applied Superconductivity
As
very-large-scale to the
become
feasible
will
wherein
force,
presence
a magnetic
the
without
a propeller.
of
this
program
would for
development
as shown in the
following
Addendum
269
Proposed
Budnet
SHIP
PROPULSION
SYSTEMS
S.&f
Budget 6.1 6.2 6.3A Total 0 4 Category FY88 2 2 r 4 FY89 1 2 1 5 m 1 2 J 5 m 0 2 1 3 FY92 0 1 s 21 Total 4 9
270
Applied Superconductivity
4.
Sunerconductinn of
Mannetic
Enernv
Storane
terml. in the
large-scale
support
from DOE and have simply significant that levels and This use for
primarily
as well.
inductor
an extended
as needed,
approach
systems,
vehicle on factors
significant
has already
made in this
DOD applications those could large toroidal appear envisioned utilize domestic for quite
of
include utility
smaller
systems
public novel
designs include
would
concentric magnetic
structures externally
internally
high
of
military
will
doubtless
specifically heavy
tailored emphasis
application, of interest
places
applications pulsed
at an early
stage. for
power applications
energy
weapons.
Addendum
271
for
would
be $6M for
materials as
development,
distributed
table.
272
Applied Superconductivity
Proposed
Budget
&d
Budget 6.1 6.2 6.3A Total 0 1 Category FY88 1 0 0 1 m 1 0 0 2 FY90 1 1 1 4 FY91 1 1 -z 5 m 1 2 Total 5 4 Li 13
Addendum
I:
Military
273
5.
Electromagnetic
Launchers
(mid
tern). the is
successful. is that
velocity pulse
a traveling
a prime
a storage
several
switches pulsed
(possible
switches),
supermagnets), of
applications
materials
can be used
in weapon
fusion. further
advantages this
may significantly
development.
to
a materials
program effort
for include a
launch
would
superconducting
potential in the
acceptable
superconducting
274
Applied Superconductivity
for
this
program for
would
be $gM for
materials as shown in
design
assessment
distributed
Addendum
275
ProDosed
Budnet
ELECTROMAGNETIC LAUNCHERS
a?!
Budget 6.1 6.2 6.3A Total 0 1 Category FY88 1 0 FY89 1 1 JJ 2 1 5 m 2 2 -z 6 FY91 2 2 J 5 FY92 1 1 A 19 Total 7 6
276
Applied Superconductivity
6.
Directed
Energy
Weapons for
term).
of
energy
weapons
deployment
lighter-weight, magnetic
resonant
as suitable
losses they
zero of
at
that
cavities.
Development successful
has proven
for
and design
problems
reliability,
and reduced of
complexity the
temperatures
make development
HTS materials
desirable.
problems
to be faced
here
are
different
and field it is
structure that
problems.
theoretical Processing to
material
defect
surfaces. will
techniques used to
structures
be similar
those
electromagnetic
shields,
by ceramic expected.
processing
This
development design
development
development,
distributed
as shown in the
following
Addendum
277
Proposed
Budnet
DIRECTED
ENERGY WEAPONS
&!I
Budget 6.1 6.2 6.3A Total 0 1 Category FY88 1 0 FY89 1 1 -!J 2 r 4 FY90 1 2 m 0 2 -2. 4 1 3 FY92 0 1 1 14 Total 3 6
278
Applied Superconductivity
7.
Magnetic
Bearings for
(mid
terml. or is over
Bearings
are
components worn-out
any rotating
translating a major
system. military
When
damage is noise
inevitable is a
lubrication.
bearing
in the development
submarines.
rotating shaft
will
confined
without will
high
shaft
or vibration
as in cryocoolers.
bearings spaces.
require Refrigeration
fields
of
modest
magnitude
in
must be developed
restricted achieved,
has already
term payoff
of
the
program
would
be $5M for
materials
design
and tests,
distributed
as shown in the
Addendum
I:
279
Provosed
Budnet
MAGNETIC BEARINGS
&I
Budget
Category
FY88 1 0 0 1
FY89 1 1 0 2
FY90 0 1 r 2
m
0 1 1 3
FY92 0 1 1 3
Total 2 4 2 11
280
Applied Superconductivity
a.
Mine Sweeping is
(mid
term)_.
Exploding
frequently field
intense from
a moving
application
to medium term.
of
the
would
be $SM for
materials
studies
distributed
as shown in
Addendum
I:
281
Provosed
Budget
&!I
Budget 6.1 6.2 6.3A Total 0 1 Category FY88 1 0 1 4 FY89 1 2 1 5 FY90 1 2 FY92 0 0 -L 2 FY92
0 0
Total 3 4 r
0 0
12
282
Applied Superconductivity
9.
Pulsed effort
Power Systems
(far of
term).
The Air
Force
major
to development power
magnets
pulsed
generators. in developing
has achieved
as the
which for
applications. from
pulsed
magnet
systems
of is
In a pulsed
these This
means fully
diameter
superconducting
filaments. of
capacity of
temperatures.
filaments
in a stabilized specific
military
program of
will
particular include
concern
program
an assessment
cryogenic
requirements considerations
refrigeration
approach.
Addendum
283
This
rlequire (design
about
$lOM for
develolpment in the
development,
following
284
Applied Superconductivity
Proposed
Budnet
&
Budget 6.1 6.2 6.3A Total 0 2 Category FY88 2 0 0 2 FY89 2 0 0 4 m 1 3 t 5 FY91 1 2 2 4 FY92 0 1 J 17 Total 6 6
285
10.
(far
Extremely been
low
magnetic submerged
modulation to the go
provide
data-bit be
submarine. scenario.
communication
would
a go-no
Because preferably refrigeration change rotating cryogenic project $lZM for table. is in the and
magnet -the
is
needed
--
operation
economics. modulating
be carefully would
Funding tests,
materials in the
design
distributed
286
Applied Superconductivity
Proposed
Budnet
ELF COMMUNICATION
Sk!
Budget Category FY88 FY89 m FY91 FY92
Total 0 1 6 7 A 4 19
1 0 0 1
2 1 0 3
2 3 -z 5
1 2 J 6
Addendum
287
ProDosed
Budeet
LARGE-SCALE
APPLICATIONS
SUMMARY
&
Budget 6.1 6.2 6.3A Total 0 16 Category FY88 13 3 FY89 12 11 3 26 FY90 10 18 lo 38 FY91 5 13 22 40 FY92 2 Total 42 53 53 148
a
18 28
288 AppliedSuperconductivity
11.
Other
Annlications. of
There
are
several of
other
applications should here. heavily discussed assessments continued greatest possible wigglers), fields), also Most
superconductivity,
discussed
fields, aspects of
on materials above.
already
of large
prospects interest
meaningful free
include:
synchrotron
radiation
the bending
magnetohydrodynamic testing,
energy
sources
and ore
or materials
VI.
DSRD BUDGETRECOMMENDATIONS
yctlined
in the This is
DSRD
approach. for
in the
requirements been
have
already budget
presented. are
individual
estimates forth
combined last
budget report.
estimates Although
as set the
on the
individual uncertainty
elements it the is
are
subject represents
believed
total
estimate to bring
of
amount of of maturity
funding
which
will
HTS to a state
Addendum
289
suitable
for
incorporation
in a variety projects.
of
military-systems-
specific
advanced
development
reasons
this
budget
figure
will
sought
in the of
generally
as a result
by other some of
by industry, projects
the proposed
to be carried of some of
ultimate
payoff other
pursued.
dictate
factors should
considered, be regarded is
estimate subject to
here
iteration,
as HTS technology
290
Applied Superconductivity
AGGRESSIVE
TECHNOLOGY-LIMITED
$4
Budget 6.1 6.2 6.3A Total Category FY88 40 28 _I! 68 FY89 40 46 22 108 FY90 31 54 32 117 FY91 19 38 56 113 FY92 12 30 58 100 Total 142 196 168 506
Addendum
291
FY89 16 26 40 26 108
4.B m
15 27 37 38 117
FY91 16 32 25 40 113
m
14 31 27 t8 100
292
Applied Superconductivity
FY88 1.9
5.1 3.7
5.9 10.1
1.5
5.0
5.0
10.3
22.9
25.3
1.6
16.5
16.5
Addendum
293
ARMY NAVY AIR FORCE NSA DARPA SD10 BTI _ 5.2 u_7.4 10.5 5.0 15.7 18.5 10.0 28.6
_
0.2 0.2 0.1
TOTAL 6.1+6.2+6.3A 19 A 3 5s a 1 70 A 4
Addendum
II
The information in Addendum II is from Military System Applications of Superconductors, issued by the U.S. Department of Defenses Defense Science Board, October 1988.
295
296
Applied Superconductivity
EXECUTIVE SUMMARY
In 1911 a Dutch scientist discovered a class of materials which, at temperatures near absolute zero, could conduct electricity with no resistance and therefore zero loss of power. In spite of the revolutionary potential of this superconducting material, the difficulty in producing engineered materials and in maintaining low operating temperatures precluded practical applications for many decades. The recent dramatic discoveries of high temperature superconducting materials (up to 125 degrees Kelvin) have prompted an intense international surge in superconductivity research and development. This surge of research and development activity, particularly that of the Japanese, combined with the promise of revolutionary performance improvements in many applications prompted President Reagan to establish a national program in high temperature superconductors. The Defense Science Board was tasked to study the military system applications of superconductors. The attached report presents the findings of this study. The Task Force found a number of superconductivity applications that could result in significant new military capabilities, including electronics and high power applications. In particular, superconducting materials could enable significant military improvements in:
.
.
greatly increased sensitivity for improved detection and enabling increased detection in clutter
$tarinP Infrared Focal Plane Array sensors incorporating superconducting electronics permitting significant range and sensitivity increases over current scanning IR sensors . . . and Ultra-Fast Dw for radar and optical sensors wer M~tprs an Genera for ship and aircraft propulsion leading to: decreased displacement; drive system flexibility; increased range; or longer endurance on station
. . .
m for high power microwave, millimeter-wave or optical generators (e.g., free electron laser); capability for powering quiet propulsion systems capable of launching hypervelocity projectiles for antiarmor weapons and close-in ship defense weapons Mamtetohvdrodvnamic (MHD) Prooulsion enabling ultra quiet drives for submarines, torpedoes, and surface ships
Addendum
II :
297
As these examples illustrate, superconducting materials have potential for significant military applications. It is important to note that many of the applications have high value for commercial and scientific applications as well. However, an extensive program of basic and applied research and materials development will be necessary to make these applications possible. The present R&D level in the U.S. is below critical mass to achieve the desired applications in a timely way. By comparison, the Japanese effort in superconductors is substantially greater than that of the aggregate U.S. commercial and government effort. If these trends continue, the U.S. may fall so far behind in this field that defense and important commercial applications will be achieved only by using foreign source materials and designs as they become available to the U.S. It is the judgment of the DSB that such dependency on foreign sources is an unacceptable position for the U.S.
We have recommended a significantly expanded superconductor R&D program for the Department of Defense which increases the 1989 effort by 50 percent and triples the current effort by 1992. The Task Force members believe such an aggressive program is required to assure U.S. leadership in the many high leverage superconductivity applications. This recommended R&D effort is balanced between exploitation of old (LTS) materials and development of new (I-ITS) materials. It includes avigorous program of building engineering models that will demonstrate the substantial performance advantages achievable with superconducting materials. The demonstration programs recommended include engineering models of a space surveillance system, mine detector, hypersonic tank gun, undersea MHD propulsion system, and a millimeter-wave radar. Most of these efforts involve old (LTS) materials. To achieve the very real cost, weight, and logistic benefits of the new (I-ITS) materials in these applications, substantially more progress must be made in the U.S. R&D program, particularly in the development of new material processing techniques. We have also recommended the development of improved militarized cryogenic devices, because even the new HTS materials will require cooling. In the near future we do not anticipate room temperature operation of superconducting materials. In summary, superconductor materials represent a major opportunity to significantly improve performance in important defense missions as well as in commercial applications. To achieve these benefits, we will need to make substantial, focused increases in R&D over a sustained period. While U.S. superconductivity research is competitive with that of other countries, we cannot count on our commercial developments providing this capability for defense. In fact, U.S. industry is already well behind Japanese industry in the development of superconductivity applications.
298
Applied Superconductivity
SECTION1
INTRODUCTION
The recent dramatic discoveries of high temperature superconducting materials (up to 125 degrees Kelvin as shown in Figure l-l) have prompted an intensive international surge in superconductivityresearch and development. As a result, the Defense Science Board was tasked (terms of reference, Appendix A) to study the military system applications of superconductors. The tasking specifically requested: . . . . A review of the current understandings of superconductor physics, as well as the status of materials properties and their processing. Anevaluationof possible military systemapplicationswith tial for significant new capabilities and cost savings. emphasis on their poten-
Identification of commercial or scientific applications of interest to DOD. Identification of supporting technology necessary for realization of military applicatiOIlS.
Task Force membership, Appendix B, heard a variety of presentations, which are listed in Appendix C. The Task Force concluded that there are some very significant military superconductivity applications which could result in enhanced military capabilities. Some of these are ready for engineering models using low temperature superconductors. However, an extensive program of research and materials development will be necessary to make these applications possible in I-ITS. Given the current level of foreign investment in this area, there is a substantial possibility that the United States will fall behind in superconducting materials processing capability.
Addendum
299
1301
L
I TlBaCaCuO BiCaSrCuO
4?
II A
x 1
120 -
YBaCuO
_-
LN 2
LaBaCu04
Nb3Sn \ \ Nb3Ge 7
~_zJ.s$;qq-;;;
1930 1950 1970 1990
1910
YEAR
DISCOVERED
300
Applied Superconductivity
SECTION 2
FINDINGS
The Task Force focused on the following aspects of superconductor technology: . . . . Superconductivity theory, technology, and materials (both low temperature and high temperature technologies) Selected supporting technologies (cryogenics and high strength materials) Military applications The level of U.S. and foreign research expenditures in high temperature superconductivity
STATUS OF SUPERCONDUCTING THEORY, TECHNOLOGY, AND MATERIALS A superconducting state is characterized by zero dc electrical resistance and zero internal magnetic field. Materials technology and theory are discussed for two classes of superconductor applications: electronics and high power systems. Electronic applications typically incorporate thin superconducting films and integrated circuit structures. High power applications use a composite, multi-filamentary wire. The technical usefulness of superconductors is limited to temperatures not exceeding 0.5 to 0.7 of the superconducting transition critical temperature (Tc), as shown in Figure 2-l. The figure depicts the critical surfaces which bound the current densities, magnetic fields, and temperatures which can be achieved by superconducting materials. The niobium-containing materials were identified and studied by 1950. The current high temperature materials are exemplified by the surface indicated in dashed lines (YBaCuO). Projection of a point in the temperature-magnetic field plane on such a surface defines the highest possible (critical) current density, Jc. It should be noted that, due to their early state of development, there is great uncertainty in levels of current density which can be achieved for the new high temperature materials. LOW TEMPERATURE SUPERCONDUCTORS (LTS) LTS materials of interest have Tcs grouped either around 10 degrees K (niobium metal, and niobium alloys) or between 15 and 23 degrees K (mostly niobium compounds). These metallic materials conduct electricity in all directions, thus simplifying the fabrication of conductors. They are used predominantly in polycrystalline form and are capable of sustaining very high current densities.J The phenomenon of low temperature superconductivity is well understood using the theory developed by Bardeen, Cooper, and Schrieffer (BCS). The publication of the BCS full
1 J = lo6 to 10 A/cm* low magnetic fields and 104 to 16 in high Gelds at liquid helium (LHe) temperature, in T = 4.2 degrees K).
Addendum
301
(Amps/cm)
302
Applied Superconductivity
microscopic theory of superconductivity in 1957 was a significant contribution to modern physics. This theory led to the prediction of the Josephson effect, a non linear tunneling process exploited in many electronic devices. This theory also helped guide the developmentof materials suitable for useful magnets. For electronic applications at liquid helium temperatures, all-refractory niobium devices and circuits are in a mature state. In contrast to Japan, the U.S. has a limited industrial base for fabricating superconducting LSI circuits for digital application and currently has no design capability for digital circuitry. For electronic applications above 4.2 degrees K and within the lowest threshold of portable closed-cycle refrigeration (T = 8 to 10 degrees K), the material of choice is niobium nitride. The technology to produce NbN devices and LSI circuits is not mature but demonstrated to be feastble. Further investment in the industrial fabrication base is required to attain required tolerances and yields. Ductile niobium titanlun~ alloys allow composite conductors to be fabricated for magnet and machinery applications involving dc magnetic fields up to 6-8 tesla. This technology is mature. The ac applications are limited by unavoidable conductor losses. Niobium tin wires can sustain dc magnetic fields up to 12-16 tesla but are brittle. Although a U.S. industrial base for this technology exists, applications are limited by poor mechanical properties and the cost of supporting structures. Non-ductile superconductors which sustain fields up to 20-30 tesla are known but the feasibility of manufacturing practical wires has not been fully demonstrated. All materials manufacturing capabilities in the U.S. reside in small companies. Few applications are being pursued at the very high field levels due to the excessive mechanical support required to withstand the enormous forces associated with such magnetic fields. HIGH TEMPERATURE SUPERCONDUCTORS (HTS) The I-ITS materials currently under development are cuprates (oxide compounds of copper, an alkaline earth metal and other elements) discovered in 1986 with critical temperatures in excess of 30 degrees K Due to the very high level of worldwide research effort, the number of these materials and their confirmed Tcs are growing. The highest critical temperature recorded to date is 125 degrees K, in a thallium-based cuprate. The most researched material is YlBaz0307 (YBCO) and its derivatives where Y is replaced by another rare-earth element. Their common Tc is 90-95 degrees K These I-ITS are capable of conducting electricity, however, they exhibit strong directional electrical conduction properties. These materials have the ability to sustain extremely high magnetic fields without loss of superconduction.z Although single crystals of YBCO are capable of sustaining very high current densities, presently available wires and Shns of the new I-ITS materials sustain only low current densities in zero applied magnetic field at 77 degrees K.2 These current densities fall off rapidly in increasing magnetic fieldsP
(Estimated at 50 to 300 tesla near T = 0 degrees K and, io YBCO, up to 20 tesla at the temperature of liquid nitrogen (LNz), T = 77 degrees K) 3 (Je = 3 x Id A/cm2 at fl( in zero field)
4 (Jc = 10 to 10 A/cm2 at 77K in zero fdds and 0.5 x lo6 A/cm2 in tields up to several tcsla.)
Addendum
II:
303
In contrast to the LTS case, theoretical understanding of I3TS is still poor. The phenomenon was not predicted, and may represent a new physical effect. Greater understanding would further enable the development of useful materials and devices. Further basic discoveries, both experimental and theoretic& can be expected in the next few years. These should have a beneficial (if unforeseeable) impact on development projects. I-ITS materials may prove useful for electronics in the form of both single crystal and polycrystalline films. Several thin film fabrication techniques have been demonstrated; others are being researched. For electronic applications, HTS films currently exhibit properties which require further study, such as high mkrowave losses. The loss mechanism is partially understood, so suitable loss reduction appears feasible. Current films are also characterized by high electronic noise. Noise mechanisms are being investigated and eventual reduction is expected. Significant effort is required to pursue the capability for design, fabrication, and optimization of todays HTS materials. The HTS materials manufacturing base must be aggressively developed in order to provide the basis for the wide range of potential military and commercial applications. Appendix D outlines the range of materials and mamrfachtriug issues which must be addressed for HTS materials. The U.S. is at a significant disadvantage, particularly with respect to Japan, in that our manufactuting capabilities in advanced ceramics is limited. Most essential is research to determine the nature of current-limiting weak links and to attain high critical currents in polycrystalline materials. Further development of thin film processing capabilities for hybrid (semiconductor plus superconductor) structures is needed. While the materials technology for brittle conductors poses very challenging technicaI and economic problems, solutions to these problems appear feasible. The initial high temperature superconductivity developments in basic research materials, and manufacturing sciences needed for military applications are also relevant for commercial applications. In both cases, a strong industrial base for mamrfacturing materials and components is needed for near-term, cost effective deployment of both commercial and military devices.
304
Applied Superconductivity
STATUS OF SUPPORTING TECHNOLOGIES CRYOGENIC COOLING One technology that has limited the previous use of superconductors has been the availability of reliable, long-life, cryogenic coolers. An extensive review of current cryogenic technology is contained in Appendix E. The older materials required cooling to 10 degrees Kelvin or lower. The new materials with critical temperatures between 90 and 125 degrees Kelvin are likely to require cooling to 40 to 80 degrees Kelvin. Since cryogenic cooler efficiency and complexity are strongly dependent on how closely absolute zero is approached, the new high-temperature materials should greatly ease the cooling problem. At the present time, coolers (to 4 degrees K) for commercial ground-based applications are available. The availability of low temperature, ruggedized, long-life coolers for military and space environments is much more limited. At higher temperatures (40 to 80 degrees Kelvin), a number of low-power (up to 1 watt cooling capacity) military designs exist which are suitable for electronics applications. The principal deficiency is the lack of proven long life, high-capacity, militarized coolers for high-power applications. HIGH STRENGTH MATERIALS Achieving the high current densities and magnetic field levels of many high power superconductor applications will require significant advances in high strength materials. As shown in Figure 2-2, the stress exhibited by magnetic fields above 15 to 20 teslas exceeds even the capabilities of graphite fibers. In fact, mechanical stress levels, not current density, are the principal limiters to high power HTS applications. Any movement to higher field levels will exceed available structural materials capabilities. To make possible such applications, further development of composite superconductinfligh strength materials is needed. It appears, however, that certain mechanical constraining structures may also be used. (see Appendix F).
Addendum
305
NiCd BATTERY
0.0001
I
1
1 1
I
1
I I
I
I I
10
30
100
300
MAGNETtC
FIELD -
TESLAS
Id4
STRESS -psi
lb
FOR
106
lb7
lb6
= o.,
SOLENO;;D;;2KNESS
306
Applied Superconductivity
MILITARY APPLICATIONS OF SUPERCONDUCTOR!3 INTRODUCTION Superconductor technology, both for the older low temperature materials and the new high temperature materials, has very wide applicability in both military and civilian products. As shown below, superconductor technology can be applied to a number of important electronic and high power applications.
Magnetic Field Sensors IR Sensors MicrowavehmWave Sensors DCto UHFSeason Analog to DigitalConvertors Analog SignalProcessors DigitalData Proeessws
Superconductor technology will support a number of important military systems including ballistic missile submarines, ballistic missile defense, anti-armor warfare, advanced air-to-surface missile, and anti-submarine warfare. ELECTRONICS Overview Military and space systems place the greatest demands on the performance of electronic devices, components, and systems. In this performance-driven field, superconductive electronics can have a major impact on sensor, signal processing, and data processing systems. The impact of superconductors in electronics is based on several unique attributes which make possible: . . . . Ultra-low loss/dispersion transmission lines and filters; High speed, low noise, and low power Josephson junction active devices; Superconducting quantum interference devices (SQUIDS) used for magnetic and electromagnetic sensing; Monolithic integrated circuits for both analog (microwave and millimeterwave) and digital components.
Uniauelv. ultra hieh sueed. low noise and low Dower can be realized simultaneously in
suaerconductors,
IR Sensors Superconductivitys major impact on IR sensors is the reduced power required to cool the signal processing and data extraction components supporting large focal plane arrays. This allows realization of large staring arrays with greater sensitivity and range than the lower performance scanning sensors which they would replace. Superconductors also have the potential
Addendum
II:
307
to
at longer wavelengths,
spatial resolution,
Future space based IR focal plane array (IRFPA) sensors will incorporate a large detector array, electronic multiplexing circuits, and a data harness from the cryostat to ambient temperature electronics. Because of the large number of detectors required, signal processing for these sensors is a major technological bottleneck CMOS A/D converters could consume several kilowatts of power. Equivalent LTS A/D converters, cooled to the 10 degrees K temperature of the IR detectors, could reduce this power requirement by 90 percent. LTS A/D convertors dramatically reduce cooling power requirements and significantly reduce system weight and size (Figure 2-3). The availability of I-ITS A/D converter technology will require development of active devices in the new materials systems. Such devices will be important for IR detectors operating at higher temperatures whether semiconducting or superconducting (e.g., HgCdTe at 77 degrees K). The key is to deploy low power A/D converters which can operate at the detector temperature. Projected on-chip power dissipation for these superconductive A/D converters is linear with temperature, but reduction in cooling power more than compensates for this due to signal processing power requirements. Some very large infrared imaging arrays may only be feasible with superconductive A/D converters. Microwave and MMW Sensors Low noise, low power monolithic receivers incorporating superconductors will have improved range and resolution. These improvements are especially important for space surveillance and communications. While the improved noise figure alone does not justify superconductors for earth or ocean imaging, it is impractical to deploy focal plane arrays of conventional detectors at MMW. With superconductor technologies, multi-band MMW imaging arrays may be feasible. These arrays could provide an all-weather capability, as well as cloud and smoke penetration not available in visible and IR systems; improved spatial and Doppler resolution, and the capability to detect low signature targets. Multi-element MMW focal plane arrays improve signal collection efficiency proportional to the number of array elements. In a low background space communication link, the improved receiver noise (Figure 2-4) can increase range, reduce power requirements to a level where solid state transmitters become attractive, and/or reduce antenna size and weight. These attributes should enhance system lifetime, autonomy, and security. High temperature superconductor technology could improve the viability of very wideband communications systems. A major impact of I-ITS will be lightweight passive microwave/MMW components for phased array spacecraft antennas and improved MMIC components such as oscillators.
308
Applied Superconductivity
Figure 2-3 SUPERCONDUCTING FOCAL PLANE PROCESSING REDUCES POWER, WEIGHT, AND COMPLEXITY
Higher speed and lower power dissipation allows faster update times and larger number of array elements using superconducting processing Operation of superconducting processor is consistent with cooled detectors
Use of fiber harness reduces heat load. weight, and size Applications for wide area IR or visible surveillance
Addendum
I I:
309
Figure 24 NOISE FIGURE (F) AND NOISE TEMPERATURE FOR VARIOUS DEVICES AND NATURAL LIMITS - 1964
2.30 -
2000
1.30 - l00o 0.70 0.30 0.15 0.075o" 2o" 100 !i" 2O lo 100 200 MHz 500 1 GHz 2 5 10
GH2
20
310
Applied Superconductivity
DC to UHF Sensors
The extremely low noise inherent in a superconducting SQUID amplifier permits the use of a very small antenna at UHF frequencies and below while retaining both high sensitivity and wide bandwidth. Applications would include ELF/VII communication systems and advanced HF receivers as well as compact high gain UHF antennas.
Magnetic Sensors
Low temperature superconducting SQUID magnetometers and gradiometers are already highly developed and commercially available. These low temperature SQUID devices are under evaluation by the Navy for ASW applications and mine detection. Figure 2-5 compares the performance of superconducting SQUIDS with conventional magnetic sensors. At higher temperatures, thermally-generated noise power will be inherently larger than at lower temperatures, eliminating some detector noise-limited applications. The complexity of the logistical support would be significantly reduced which would make currently unattractive remote sensing applications much more practical.
Signal Processing A/D Converters
Superconducting A/D converters offer significant advantages in speed and power efficiencies over semiconductor devices. Figure 2-6 shows that the predicted performance of superconductive A/D converters exceeds projections of conventional converters in both speed and, linear resolution. The development of HTS A/D converters will substantially increase their utility due to reduced cooling, lower overall power dissipation, and improved performance. The significantly lower power will make it possible to deploy multi-channel A/D converters for analysis of multi-GHz spectral systems.
Delay Line Signal Processor
Analog signal processors based on tapped delay lines provide wideband signal processing for w-ideband radar and communications systems. Superconductor devices have the capability to perform waveform chirp, convolution, correlation, spectral analysis, and matched filtering with bandwidths as high as 20 GHz Use of new HTS materials will allow integration with semiconductor devices to expand functional performance. The reduced cooling requirement will open deployment opportunities in many wideband radar and intercept systems.
Addendum
I I:
311
Figure 2-5 COMPARISON OF SUPERCONDUCTIVE MAGNETIC SENSORS WITH CONVENTIONAL MAGNETIC SENSORS
CONVENTIONAL
MAGNETIC SENSOR
SENSITMTY
IO TO 1O-s Gauss
-u--l-=
MEASUREMENT CAPARILITY Field Strength Only at 10 Gauss; Not Capable of Measuring Source Strength and Full Field Gradient; Capable of Measuring Source Strength and Location
312
Applied
Superconductivity
TECHNOLOGY
BITS (N-)
JJ (Counting Type)
l2*
6
12
10
SILICON BIPOLAR
6-8
so0
8000
TRENDS:
FURTHER
IMPROVEMENT
Addendum
II:
313
Digital Signal And Data Processing LTS digital signal processors provide more than a factor of ten increase in processing speed over conventional GaAs logic for the same complexity and at lower power dissipation (Figure 2-7). Computers can also realize significant performance benefits through application of LTS The shortest response time measured to date is approximately two picoseconds technolo g or 2x10- seconds. This, combined with signal swings of a few millivolts, produces high performance logic circuits that consume very little power. Fujitsu has reported a four bit microprocessor based on Josephson technology that is ten times faster and consumes 0.002 times the power of a gallium arsenide version of the same microprocessor. (See Appendix G). A critical problem in high performance computing is heat removal. For a given heat removal technology the space required per circuit at the systems level increases as the power per circuit increases. Therefore, the transit time or the time to transmit logic signals within the system becomes increasingly important in determining overall system performance. The very low power requirement of a low temperature Josephson technology system is partially offset by the power required to maintain the low temperature. However, the very low power required by even a large system permits the sytem to be built in a very small volume. The greater device speed combined with the shorter transit time will provide greater computational capability in LTS for equivalent total power dissipation (including cooling). At the present time, approximately ten circuit families based on the Josephson technology, have been reported. All of these circuit families are of the Kirchhoff or threshold logic type. This type of circuit performs logic by summing currents and, as a result, is sensitive to parameter variations. Josephson technology has evolved in this direction because the Josephson device is a two terminal device without power gain. There is no fundamental reason to believe that a superconduc!ing equivalent of the transistor is not feasible. If such a device can be developed that has power gain with very high performance, it would revolutionize the use of superconducting technology in computing systems. HTS materials show great potential because they may be much more compatible with semiconductor materials than the older superconductors, thus enabling hybrid approaches (superconductors with semiconductors). It should be emphasized, however, that achieving the potential of three terminal devices will require the invention of a fundamentally new device.
314
Applied
Superconductivitv
1000 -
z
; Y 5
Fi F
100 -
2 i n JJ SUPERCONDUCTOR
POWER
&W/Gate)
Addendum
315
HIGH POWER APPLICATIONS High power applications will primarily exploit the intense critical magnetic fields that can be generated by high current density superconductor materials. There are many promising application areas for such materials. Superconductor magnetic fields can store large amounts of energy for extended periods of time, They can provide compact, high magnetic field sources for rotating electrical machinery and offer the promise of unconventional electric drive systems for military platforms. Employed in weapons, they can accelerate projectiles to exceptionally high velocities and, as control magnets for electron-beam tubes they can provide high-power sources of millimeter and visible wavelength energy. All of these possibilities include important military applications. Magnets -- Applications The earliest superconductor technology applications were high-field magnets used in particle accelerators and energy storage inductors. In operation, energy is fed into the inductor slowly, stored for an arbitrarily long period, and then released on demand. A modest-sized superconducting storage inductor (3x10 joules) has already been used by the Bonneville Power Administration. A larger scale storage system ( 1012joules), could meet the high-peak power demands of a ground-based free electron laser or a space-based directed energy weapon. Magnetic energy storage devices with high critical fields could reduce the size and weight of these storage devices as shown in Figure 2-8. Near term applications of high-field magnets could include superconducting beam control magnets employed in gyrotrons and free-electron lasers for the generationofhigh-power levels at microwave, millimeter and optical wavelengths. Electrical Machinery The quickest payoff in high-power applications will come from the exploitation of superconductor materials in rotating electrical machinery. Substantial weight savings can be realized by eliminating magnetic circuit materials and customary field windings. Already, an experimental 3-megawatt superconducting D.C. motor has been built for ship propulsion and tested at sea. This motorwas 33 percent smaller than the equivalent conventionally air-cooled A.C. motor. Substantially greater motor size reductions are possible with conventional LTS materials. A superconducting homopolar DC. motor of 40,000 h.p., employing superconducting shielding, could be built at about one fourth the size and weight of a contemporary A.C. motor. The decreased size and weight and increased electrical efficiency reduce fuel requirements and lead to an overall reduction in propulsion system demand on the ships resources. A superconducting generator, which may be located remotely from the ship drive motor, will provide an efficient, flexible ship propulsion system. The effect on a destroyer-class ships performance would be to reduce ship displacement by 14 percent and increase its range by 30 per Critical magnetic fields have been measured up to 20 tesla (T) at 4 degrees K for conventional LTS and 30 T at
77 degrees K for HTS.
316
Applied Superconductivity
ENERGY/VOLUME
B2 z-
B O-=L.W
1 5 20 30
W&Wh/m3
0.15 3.8 60 135
NiCd BATTERIES
Addendum
317
cent.
If the propulsion system were mounted in an external pod, the ships displacement be decreased by 25 percent and its cruising range increased by 40 percent.
could
As illustrated in Figure 2-9, high temperature, high field materials would allow further decreases in weight and size. At this point, the propulsion system would be a negligible fraction of overall ship displacement, and multiple redundant drive systems could be installed. While the first high-power propulsion applications are likely to be in ships, Figure 2-9 suggests that high-field superconductors could also provide light-weight generators and motors for armored vehicles and, more speculatively, for aircraft propulsion. It must be emphasized that if these systems are to come about, the necessary cryogenic support systems must be developed to withstand the rigors of an operational environment. Other superconductor propulsion sysems are clearly foreseeable. In Japan Magneto Hydrodynamic (MHD) drives have been built and tested at scale-model level by Kawasaki Heavy Industries. By 1990, Mitsubishi Heavy Industries, in partnershipwith Toshiba and Kobe Steel, plans to have a 120-ton displacement ship with MHD drive in operational test. In addition to surface ships, MHD drives can also find use as quiet propulsion systems for submarines and torpedoes. Speculating about further term applications, an MHD collector-diffuser and an MI-ID magnetic nozzle may make feasible a scramjet propulsion system for space bodies traveling in an ionized medium. In this concept (Figure 2-lo), a combustor operates between two MHD sections, one of which adjusts the flow velocity and temperature to be suitable for combustion, and the other of which provides thrust augmentation. Excess inlet gas energy is removed by the forward section and re-injected as electrical energy into the stream by the aft section. This will require particularly compact and lightweight magnetic field sources. Launchers The electromagnetic (EM) mass accelerator concept is some 25 years old and has been explored intermittently. Recently, SD10 has supported EM rail-gun technology for use as a projectile accelerator. The EM accelerator is of interest because it is capable of propelling a large mass to a very high velocity. Unlike chemical propulsion systems, the achievable terminal velocity is not limited by the speed of exploding gas, but rather by the speed of a traveling electromagnetic pulse. Hence, a projectile could be accelerated to act as an effective kinetic energy weapon. EM accelerator applications include: launching close-in ship-defense projectiles against cruise missiles, launching torpedoes from submarines, or as a hypersonic antiarmor weapon which, because its velocity could exceed the sound velocity in protective armor, would be an assured penetrator. Superconductor materials, whether LTS or I-ITS, will increase the feasibility of EM launchers as military weapon systems for many of the reasons previously stated in other applications -- lower weight, smaller volume, and higher efficiency. Superconductor materials would be used in the prime power generator, in the energy storage system and in the high speed switch which could employ superconducting thin films.
318
Applied Superconductivity
SHIP PROPULSION
CONVENTIONAL AIR-COOLED
SUPERCONDUCI-OR DC-5TESL.A
18,000 LB 45 FC DIAMETER
REALIZABLE
WITH CURRENT
TECHNOLOGY
PROJECTED
40LB 1J FC DIAMETER
SPECULATIVE
Addendum
II:
Military
System Applications
319
MAGNETIC
DIFFUSER
COMBUSTOR
MAGNETIC NOZZLE
FORWARD
MHD
Ah
MHD
MAGNETIC NOZZLE
THRUSTER EXHAUST
320
Applied Superconductivity
High-current density superconductor materials could make feasible a new concept, the superconducting augmented rail gun (an inverse rail gun), in which most of the launch energy is stored in a superconducting field magnet. This eliminates the requirement for a pulsed, highenergy system to achieve the desired acceleration levels. The problem of high current, high speed sliding contacts, which is common to all rail guns, would still remain. An EM coaxial launcher, which requires no physical contact between the projectile and launcher, has been proposed to get around the sliding contact problem. In a version known as the superconducting quench gun, it could operate very effectively with compact, high-field intensity superconducting elements. In this concept, all the launch energy is stored in a multisection solenoid barrel (Figure 2-l 1). The magnetic field of each section is very tightly coupled to its immediate neighbor and to the superconducting projectile coil. The projectile coil is accelerated by the solenoid magnetic field and, as it passes the mid-point of each section of the solenoid barrel, the superconducting current flow in that section is quenched. The quench gun directly converts magnetic energy into kinetic energy at high efficiency. Figure 2-11 MAGNETIC LAUNCHERS
PROJECTILE MASS = M
MAGNETIC
ENERGY
IN
PROJECTOR
2 PO
7rr2
PROJECTILE
VELOCITY
V =
SOLENOID
STRESS
S =
-= 4
PO/R - r)
Addendum
II :
321
U.S. AND FOREIGN RESEARCH EXPENDITURES IN HIGH TEMPERATURE SUPERCONDUCTMTY With the discovery of high temperature superconductivity, substantial R&D efforts have been undertaken in the U.S., Europe, Japan, and very likely in the USSR. It is very difficult to make estimates of national R&D efforts. 1988 estimates of U.S. and foreign high temperature superconductivity research, as drawn from CIA and NSF inputs to the Task Force, are as follows:
#OF U.S.
Government Industry Japan UK FEUICC West Germany
l
95 M US 25 20 15
See Appendix H for more detailed information. The above estimates for Japan do not include sala& of the researchers. Au other funding numbers do include such costs.
U.S. Government funding details are contained in Appendix H. It is estimated that in 1988 approximately 500 professionals are supported by U.S. Government funding. Most of the U.S. industrially-funded research is concentrated in a few large research laboratories (e.g., IBM, AT&T, etc.). In addition, several start-up companies have been formed. The rest of U.S. industry is investing relatively little and maintaining a wait-and-see attitude. The intensity and emphasis of the Japanese effort is notable. Both basic research and rapid industrialization are emphasized. Single crystal materials with significant current carrying capacity at 2 tesla fields have already been achieved. In contrast to the U.S., Japan is already applying significant effort toward the industrialization of both LTS and I-ITS. According to a recent OTA report,
Japanese companies have been more a&e in pursuing the commercial potential of HTS. They have more people at work many of them applications-oriented engineers and business planoers charged with &inking about ways to get HTS into the marketpIace...As the Scientitic race becomes the commercial race, Japanese fm could quickly take the lead Indeed, they may already be doing so!
The European efforts are mainly concentrated in universities and emphasize basic research. At the present time, it seems clear that high-temperature superconductivity research is geographically widespread and that the U.S. is not the principal focus of research.
6c .
322
Applied Superconductivity
SECTION 3
CONCLUSIONS
Based on these Emdings, the Task Force came to the following conclusions: 1. The new high-temperature superconductors are of great significance because of their high operating temperatures and magnetic fields. 2. The discovery of high temperature superconductors has rekindled interest in low temperature applications which have not been exploited. 3. There are Superconductor applications of potentially significant military impact, as shown in Figure 3-l. 4. To make these military applications possible, intensive research and development in the following areas will be required: . Expanded efforts in superconductor theory and basic research should provide the fundamental understanding of the new materials to guide applied research. Such basic research (theory and experiments) could also lead to the scientific breakthroughs which will make the speculative applications feasible. Thin HTS film fabrication, with emphasis on lower processing temperatures, perfecting surfaces/interfaces, reducing RF surface losses, minimizing electronic noise, and increasing environmental stability, including radiation hardness. HTS composite films/conductors/wires with emphasis on increasing current densities in high magnetic fields to useful levels, minimizing persistent current creep and AC losses, and attaining requisite mechanical strengths and flexibility. Militarized cryogenic coolers with long lifetimes and increased reliability, especially portable, miniaturized coolers. High strength structural materials for magnet support systems.
. .
5. DOD sponsored developments in basic research, materials, and manufacturing processing will provide direct benefit to commercial manufacturing organizations. 6. Some applications of great military significance could be embodied in engineering models in the near future. The following programs, which combine a high degree of significance with a reasonable expectation of technical success, could be started in parallel with the efforts to develop improved high temperature superconducting materials:
Addendum
323
Figure 3-1
LTS HTSLTSHTS -
Low Risk Medium Risk LowRisk Medium Risk (Phased-Array Antenna) Low Risk (Analog) Medium Risk (Digital) Medium Risk (Analog) Speculative (Digital) Low Risk (Ship) High Risk (Armored Vehicle) Speculative (Aircraft)
LTS HTS -
MOTORS AND GENERATORS Ship, Aircraft, and Advanced Vehicle Propulsion with: Decreased Displacement; Increased Range/Longer Endurance; and Drive System Flexibility; lMAGNETS/ENERGY STORAGE High-Power Generation for Microwave, Millimeter-wave or Optical Generator (e.g., FEL) ELECTROMAGNETIC LAUNCHERS Hypervelocity Projectiles for Anti-Armor Weapons and Close-in Ship Defense Weapons MHD PROPULSION Ultra Quiet Drives for Submarines, Torpedoes, and Surface Ships .MHD DIFFUSER/MAGNETIC NOZZLE High Altitude Hypesonic Propulsion; Orbital Power Generation
LTS -
LIS HTS -
LTS HTS -
LTS HTS -
FEASIBILITY
KEY:
324
Applied Superconductivity
Space Surveillance System. Build an IR focal plane array demonstrating high resolution and low power consumption by combining detectors using existing extrinsic silicon materials with signal processors employing LTS materials. In parallel, a 6.2 program could develop sensor elements with HTS materials. Mine Detector. Build and demonstrate a magnetic field sensor with LTS materials suitable for use as a mine detector. In parallel, a 6.2 program could develop sensor elements with FITS materials. Hypersonic Tank Gun Build and demonstrate an electromagnetic projectile launcher using LTS materials. This launcher should achieve hypersonic velocities capable of penetrating reactive armor and modem composite armor. Undersea MI-ID Propulsion. Build and demonstrate a small scale MHD propulsion systems with LTS materials. This engineering model would be designed to power a torpedo. Later models would be scaled up for submarine applications. Millimeter-wave Radar. Build and demonstrate a millimeter-wave radar. This radar would embody HTS materials in its filters, transmission lines, phase shifters and possibly the reflector.
7. Foreign investment in superconductivity research and development is increasing rapidly and significantly exceeds that of the U.S. Japan is currently spending considerably more than the total U.S. effort in superconductivity research and has targeted superconductivity as an important commercial area.
Addendum
325
SECTION 4
RECOMMENDATIONS
Based on this evaluation, the following recommendations are made: . DDR&E should implement a focused plan for superconductivity basic research, (theory and experiments) materials development, and application demonstrations. This plan should include cooperation with industrial organizations in order to build a strong industrial base in the area of superconductivity. This plan should also incorporate substantial funding which increases over the next several years. A model funding profile is shown in Figure 4-l. The Services, SD10 and DARPAshould implement an aggressive plan for early exploitation of high-temperature superconductivity in electronic applications, including sensors and data processing, as well as weapon and propulsion systems. Initial emphasis should be placed on electronic applications. A suggested funding profile is included under the high;temperature 6.3 lines of Figure 4-1. To facilitate the earliest military applications of superconductivity, the Services, SD10 and DARPA should build a number of engineering test models exploiting existing low temperature materials. Estimates for funding of these efforts are shown in Figure 4-l under the last two 6.3 lines.
326
Applied Superconductivity
Figure 4-1 SUGGESTED DOD SUPERCONDUCTMTY (Dollars IIIMillIons) gg 6.1 6.2 Basic Research including Theory Applied Research on Processing of New Materials, Manufacturing Sciences, Cryogenics, and High Strength Composites Engineering Demonstrations of Electronics Applications of New Materials (e.g., Magnetic Sensor, IR Sensor, and Microwave Antenna) Engineering Demonstrations of High Power Applications of New Materials Early Exploitation of High Power Engineering Test Models Using LTS (e.g., Quench Gun, MHD Torpedo for Quiet Propulsion) Early Exploitation of Electronics Engineering Test Models Using LTS (e.g., digital signal processing, squids, millimeter-wave sensors) TOTAL 17 22 FUNDING*
89 20 50
90 20 60
91 25 70
92 25 70
93 30 75
63
13
10
20
30
40
50
63
10
20
63
22
30
50
70
80
70
63
10
10
20
20
15
79
120
160
215
245
260
Addendum
327
OF DEFENSE
2OSOl
WASHINGTON,
4 IJEC 1907
MEMORABDDM FOR CBAIBMAB, DEFENSE SCIENCE BOARD
- Assess the current understanding OS the physics involved in high To superconductivity and the status of materials processing technology neoessary to produce devices. Where possible, project likely advances in critical temperature, critical fields, stability o? materials , and ease of manufacture in various configurations (thin films, uires, etc.). - Enumerate possible system applications and their military impact. Attempts should be made to quantiiy system periormance improvements, and where possible, cost savings. - Order the potential system applications in terms OS necessary superconduator oharacteristics and in terms of military capability. - Identiiy those system applications that will be unique to the military and those which , as they are developed for commercial interests, will assist the military. Suggest ways in which ue might develop the uniquely military components and systems, as well as ways in which ue might cooperate with industrial development.
328
Applied Superconductivity
Identify what supporting technologies (those not directly related to superconducting materials) will need development in order to realize each system application. should Make reooomendations on how DOD, and in particular pursue development in these areas. DARPA,
Secretariat representative. It fs not anticipated that your inquiry nfll need to o into any particular matters within meaning of Section 20 % of Title 18, D. S. Code.fi
The Dfreotor of DARPA and &he Deputy Under Secretary of Defense for Research and Advanced Technologywill co-sponsor this Task Force. Mr. Walt Morrow, Jr., has agreed to serve as Chairman. Dr. Kay Rhyne of DARPA will be the Executive Secretary. LCDR George A. Mikolai, DSN, will be the DSB
the
Addendum
329
MEMBERSHIP
CHAIRMAN
Mr. Walter E. Morrow, Jr. MIT J..incoln Laboratory
VICE VAN Dr. Wti J. Perry H&Q Technology Partners MEMBERS Dr. Philip B Allen State University of New York at Stony Brook Dr. Richard N. Herring Ball Aerospace Dr. Donald C. MacLellan MlT Lincoln Laboratory Dr. Aleksander 1. Braginaki Westinghouse R&D Center Mr. Joseph Logue Consldtant Dr. W&am H. Press Harvard University Dr. Arnold H. Silver TRW Space and Technology Group
CUTNE
SECRETARY
TARY ASSISTANT LCDR George A. Miiolai, USN DSB/OUSD(A) GOVERNMENT ADVISORS Dr. Donald Gubser Naval Research Laboratory
Office of Deputy Under Secretary of Defense Research and Advanced Technology Dr. Bernard Paiewomsky Office of Secretary of Air Force
Dr. Clarence Thornton Director, U.S. Army Electronic Technology Laboratory Dr. Harold Weinstock Air Force Oftice of Scientific Research
330
Applied Superconductivity
TODIC
Dawn of the Third Electronics Revolution Based on Superconductivily Proassiq Fabrication and Properties of Helium Temperature Superconductors
Developments of Theories and their Role in Exploiting Superconductivity Status of Superconducting Gradiometw
Opportunities in Electromagnetic Launcher and pulse Power Systems Using High Temperature Superconductors Superconductive Sensor aadd Signal Processing Technology
Addendum
331
Name/Organization
Dr. Femaud Bedard National Security Agency March 17-19.1999 Dr. Michael SupercynsLi David Taylor Research Center Dr. Harold Weinstock AFOSWNE Dr. Charles Hogge Air Force Weapons Laboratory Dr. Gerald Iafrate AmY Dr. Gerald P. Dbmeen HoncyweU, Inc. Dr. Richard Withers MIT Liicoln Laboratory William Duggleby, Donald Lundy CIA
April 1915, 1999
Topic
Electronic Applications of Superconductitit)
Superconductivity
Machinery
Air Force Research and Development Plan for Superconductivity Air Force Superconductor Applications
High
Temperature Superconductivity... Issues, Novel Concepts, and Army Applications Corporatq Overview on Superconductivity
Supercon-
Dr. Glenn Penaisten Alpha Partners Inc. Dr. Alex Malozemoff IBM
Issues
332
Applied Superconductivity
INTO
Section 2 of the DSB Report summarizes the status of superconducting materials. The purpose of this Appendix is to define in a concise manner the more important I-ITS materials and manufacturing issues in the two classes of applications: military electronics and high power systems. The R&D directions with the greatest need of DOD support are also identified. A few general issues and ensuing R&D directions are common to both classes of applications and these are discussed first. Exploration of these general problems will be best addressed by the search for new materials. This activity is most appropriate for academic, long range, lowlevel research programs, both theoretical and experimental. The National Science Foundation and the DOD Offices of Research are suitable conduits of support for a majority of such programs. 2. General Issues 2.1 Operating temperature: in almost all applications the operating temperature is 0.5 to 0.7 To The desired operation at liquid nitrogen (77K), and eventually room temperature, mandates a search for higher Tc materials with a goal of attaining about 15OK. The successive I-ITS discoveries which raised Tc from about 80K to 12SK in the past two years suggest that further Tc increases might be expected. Both thallium and bismuth compounds that are more stable in air than YBCO have been found with critical temperatures greater than 1OOK. Development of an applicable theory of FITS superconductivity mechanism could provide clues for novel, higher Tc materials. 2.2 Range of superconducting interactions: in all oxide HTS materials this range is extremely short and often comparable to the distance between adjacent atoms. Consequently, small defects and imperfections disrupt superconductivity. For example, the transfer of high electric current density between grains (crystallites) of such ceramic materials is most severely limited by imperfect grain boundaries. Defective film surfaces and interfaces make prospects for the operation of I-lTS Josephson tunnel junctions uncertain. 2.3 Directionalproperties: all I-ITS cuprate crystals have highly directional (anisotropic) magnetic field and electric current carrying capability. There is a potential for developing devices which exploit the anisotropic nature of the materials. The search for HTS materials having field and current capabilities independent of direction (isotropic) is still desirable. The most recent discovery of such a material, a Ba-K-Bi-oxide, with a Tc approaching 30K indicates that work along these lines might be a fruitful part of such an R&D program.
Addendum
II :
333
2.4
HTS theory akve@nent. As noted in section 2 of this report, theoretical understanding of I-ITS is poor. Although R&D can proceed without a full theoretical basis, this lack of understanding inhibits the development of new materials and processes. Theoretical research efforts should be supported as a part of DODSHTS program. of ProawingMethodr. Since the new superconducting materials with critical temperatures greater than 90K have four or more components, their deposition as thin films has been challenging. A large variety of techniques must be evaluated for appropriate properties.
2.5 R+tement
All cuprate I-ITS materials discovered and investigated to date exhibit qualitatively a similar behavior, so that the priority objectives/directions for applied R&D are relatively common to all such materials. Present issues and deficiencies define the practical goals for the investigation and optimization of properties and processes which are listed below for the two classes of applications. The foremost requirement in these investigations is to acquire basic understanding of the material deficiencies. 3. HTS Materials for Electronics Superconductors, both LTS and FITS, are used in electronics mostly in the form of thin tilmand layered film structureswhich also incorporate nonsuperconducting films of insulators, semiconductors and normal conductors. Material requirements are to a large degree common to low and high power electronics. The identified obstacles to I-ITS utilization are the high radio frequency (RF) losses, high electronic noise and the film surface/interface superconductivity degradation. The weak links between grains can be either utilized (in granular films) or eliminated by epitaxial, single crystal film deposition. 3.1 Objectives for Electronic Materials Investigation: 3.1.1 Maximize critical current density and flux pinning in epitaxial HTS superconductor 6lms for active devices to attain and exceed 10 A/cm2 at 77K. Maximize flux pinning in granular HTS films in order to obtain lo6 A/cm2 at 77K. 3.1.2 Reduce low-field RF loses in HTS epitaxial films and film substrates to at least two orders of magnitude lower level than in copper at the same frequency and temperature. Increase the RF surface magnetic penetration field to approach the superheating critical field. 3.1.3 Reduce the low frequency (l/f) noise energy in HIS epitaxial and granular films to a target level of less than 10m30 Joule/Hz at 77K and 1 Hz. 3.1.4 Attain superconductivity at HTS film surfaces and interfaces comparable to that inside the film and demonstrate tunneling into these surfaces.
334
Applied Superconductivity
3.15 Determine mechanism of inter-grain coupling and opto-electronic librium effects in granular HTS films.
nonequi-
3.1.6 Explore effects in HTS films which might lead to new functional devices, especially three-terminal transistor-like devices. 3.1.7 Search for optimized substrate materials and buffer layers for HTS films (epitaxiaL!nonepitaxial, having low t-f loss, chemically and thermally compatible to minimixe interdiffusion and strains). 3.2 Objectives for Electronic Materials Fabrication Process Development 3.2.1 Develop HTS film deposition methods insuring: (a) an ultra-precise composition control, (e.g., 02 content) (b) stabilization of highest-Tc structural phase, (c) epitaxial, single crystal film growth with specified crystal orientations, (d) granular film growth with controlled inter-grain coupling, (e) reproducibility, versatility and high throughput at a lowest capital investment level. 3.22 Attain lowest possible deposition and processing temperatures for HTS films and multilayers with a target temperature not to exceed 500 C. 3.23 Attain integration of HTS superconducting films with semiconductors, sulators and metals to fabricate hybrid electronic circuitry. in-
3.2.4 Develop micron and sub-micron scale patterning of HTS films and multilayers. 3.25 Develop techniques for fabrication of electrical contacts in integrated cuits.
4.
cir-
In high power, low frequency and direct current (DC) applications, the most important embodiment for the LTS and HTS superconductor is a composite conductor in the form of a multifilamentary wire, tape and cable. Composites must include normal metal matrix or cladding for cryostabilixation and mechanical support. The biggest obstacles to fabricating such HTS conductors are the presence of weak links between grains of bulk, polycrystalline materials which limit the current density to very low values, the anisotropy of critical field and current density, the brittleness of the HTS ceramics and the incompatibility (reactivity) with economically viable stabilizers: copper and aluminum.
Addendum
II:
335
Material
Properties Investigation
4.1.1 Increase critical current density in bulk polycrystalline aggregates (to a target level of at least 16 A/cm* at 77K in magnetic fields of 10 to 12 tesla and at least lo6 A/cm* below 1 tesla) and minim& glassy behavior by atmining clean, defect-free grain boundaries. 4.12 Determine detailed phase diagrams, oxygen diffusivity data and other physico-chemical properties of Pffg compounds. 4.1.3 Determine mechanical properties (elastic constants, Youngs modulus. hardness etc.) and thermal properties of HTT3 single crystals and polycrystalline aggregates over a wide temperature range. 4.1.4 Determine and optimize electromagnetic and thermal properties, especially alternating current (AC) losses and stability, of HTS composite conductors having the minimum required current density at specified field intensity. 4.2 Objectives for Conductor Fabrication Process Development. 4.2.1 Develop methods for fabricating crystallographically oriented (textured) HTS wires and tapes with clean, defect-free grain boundaries. 4.22 Develop methods for fabricating/pulling single crystal HTS fibers, fiber or tape coatings and fiber or tape substrates in suitable diameter or thickness ranges. 4.2.3 Develop methods for fabricating HTS composite wires and tapes with economically viable normal metal stabiliaation. 4.2.4 Develop methods for fabricating HTS composite wires and tapes with mechanical composite reinforcement sufficient for handling and high-field operation of brittle ceramic fibers and coatings. 42.5 Develop HTS magnet conductors in geometries other than wire or tape (e.g. a Bitter-type configuration). 4.2.6 Develop methods for joining HTS composite wires.
4.2.7 Explore alternative methods of I-ITS conductor fabrications to optimize reproducibility, quality control and ease of manufacturing at lowest unit cost and capital investment. Efforts 4.2.3 to 4.2.7 will deserve high priority after a sufficiently high current density in wires, long fibers or tapes is demonstrated.
336
Applied Superconductivity
CRYOGENIC TECHNOLOGY
The three important characteristics of the superconducting state are the critical ternperature (Tc) (the temperature below which the superconducting state occurs), the critical magnetic field (Hc), and the critical current density (Je). It is well known that temperatures above Tc quench superconductivity. Perhaps it is less well known that the superconducting state is also quenched by an external magnetic field, H > Hc, or an electrical current, J > Jc or a combination of the three parameters. The highest temperature possible for the transition to the superconducting state, Tc, occurs when both H and J are zero. For useful values of H and J, the operating temperature of the superconductor must be less than T e. A useful guideline is that the absolute operating temperature, T, should be about one-half the transition temperature. This temperature level provides useful values of H and J, as well as a margin below Tc to accommmodate any localized transient heating of the superconductor that might occur. (The actual relationships among T, H, and J are complex, but the above rule is nonetheless a handy guide.) Accordingly, the ordinary (Type II) superconductors with Tc in the 20K range could typically be usefully operated at 1OK. Since no stable, liquid refrigerant exists at lOK, the operating temperature is usually lowered to 4K, the temperature of liquid helium. By the same rule, the new class of superconducting compounds, with Te at about 9OK, would be useful superconductors if cooled to 4%. Cooling to the temperature of liquid hydrogen, 2OK, may not be unreasonable because hydrogen is much less expensive than the 40K liquid refrigerant, Neon. Similarly, the yet undiscovered room temperature (300K) superconductors, in order to accommodate a useful magnetic field and electrical current, would be cooled to 15OK.For practical reasons, the temperature of liquid nitrogen, 77K, would probably be chosen. Therefore, superconductors, old, new, and yet undiscovered, tures (4K, 2OK, and 77K, respectively) to have useful properties. require cryogenic tempera-
The balance of this appendix discusses our present ability to create such a cryogenic environment. The discussion is divided into ground-based systems and space-based systems. These two categories are then divided into large-scale and small-scale systems. It will be seen that cryogenics is a mature technology and poses no particular obstacle for operating any superconductor, old or new, at any desired temperature. Table 1 provides a list of some current, extensive reviews of cryogenic cooling technology.
Addendum
337
Table 1 REVIEWS OF CRYOGENIC STATUS 1. Smith, Joseph L., Robinson, George Y.. and Iwasa Yukikazu, Survey of the State of the Art of Miniature Cryocoolers for Superconducting Devices, prepared under the offtce of Naval Research, Contract NOOO1483K0327. (Not published in the open literature) 2. Daunt, J.G. and Goree, W.S., Miniature Cryogenic Regrigerators, under contract NONR-263(70), July 1969. 3. Crawford, AH., Specifications 1970.
4.
of Cryogenic Refrigerators,
Cryogenics,
Applications of Closed-Cycle Cryocoolers to Small Superconducting Devices, NBS Special Publication 508, Eds. Zimmerman, J.E. and Flynn T.M., April 1978.
5. Refrigeration for Cryogenic Sensors and Electronic Systems, NBS Special Publication 607, Eds, Zimmerman J.E., Sullivan, D.B,, and McCarthy, S.E., May 1981. 6. Walker G., Crvocoolers, Vols. I and II, Plenum Press, 1983.
338
Applied Superconductivity
SECTION 2 CRYOCOOLERS A cryocooler is a refrigerating system capable of achieving temperatures in the cryogenic range, generally considered to be less than 12OK. Cryocoolers are often rated by the available refrigeration capacity measured in watts. To be meaningful, however, it is necessary to specify not only the refrigeration capacity but also the temperature at which the refrigeration is available. A cryocooler having a capacity of 1W at 4K (Liquid helium temperature) is very different than a cryocooler having a capacity of 1W at 77K (Liquid nitrogen temperature). Thus, the level of refrigeration (4K - helium, 20K - hydrogen. 77K - nitrogen) is specified, as well as the capacity at that level (usually in watts). Another important parameter is the power input or work required to achieve refrigeration, usually measured in watts of input power per watt of useful refrigeration (Winput/wcoolbg). This figure is closely related to the efficiency of the refrigerator. The efficiency of presently available cryocoolers range from a minimum of less than one percent to a maximum of near 50 percent. Efficiency depends more on the scale of the machine than on the temperature level or thermodynamic cycle employed and, therefore, the size of the ctyocooler is important to this discussion. The small machines used for electronic applications have the lowest efficiencies. This is because nearly all the refrigeration generated is consumed in cooling the low-temperature parts of the machine itself. The surplus or useful refrigeration available from these units is very small--only fractions of a watt. Applications requiring a larger useful refrigeration load use larger, more efficient systems. The highest efficiencies are found in large machines used for liquefiers and range from 20 to 50 percent of the Carnot value. (The Camot value is the thermodynamic limit of the best that can be done.) Space-based cryocoolers differ from ground-based systems in several important respects. Aerospace cryocoolers must be able to withstand the high acceleration and vibration spectrum of a rocket launch, and have the ability to operate in any orientation and in a zero or low gravity. The reliability and long-life of aerospace cryocoolers assume an importance not found in ground applications. Thus, a useful distinction is to classify cryocoolers according to ground-based or space use. Ground-based systems are divided into large and small applications. Space-based systems are usually small--lOW or less.
Addendum
II :
339
LARGE SYSTEMS Industrial uses of cryogenics are spectacular and commonplace, and most frequently at the 77K (nitrogen) or 90K (oxygen) levels. Uses include food freezing (McDonalds uses %SOM/yr of liquid nitrogen), sewage treatment (many cities use oxygen produced on site from liquid air to speed up treatment), breathing oxygen for hospitals obtained from liquid oxygen storage, and the production of chemicals (anti-freeze) and steel from liquid oxygen. Cryogenics is a routine industrial tool currently found in the Yellow Pages of the telephone directory. Such industrial cryogenic systems as these are routine product lines of Air Products and Chemicals, Inc., Linde Division of Union Carbide, the AiResearch Co., and others. Hydrogen for industrial and aerospace use is routinely stored as a liquid at 2OK, strictly as a convenience. Texas Instruments Stafford, Texas plant uses the countrys largest commercial liquid hydrogen storage tank to supply hydrogen gas for semiconductor processing. NASA and the USAF each have l,OOO,OOO gallon liquid hydrogen storage tanks, supplied by a network of hydrogen liquefiers. Liquid hydrogen (20K) technology is off-the-shelf from such firms as Air Products and Linde. The status of large-scale liquid helium facilities is perhaps even more surprising. Most large high-energy accelerators use helium-cooled superconducting magnets simply because: (1) it is cheaper to do so, compared to the electric power required for normal conductor magnet; and (2) to achieve higher levels of magnet performance. Table 1 shows a partial list of these facilities. At present, these facilities have a combined capacity of 82.5 kW at 4K. When complete, the Superconducting Super Collider (SSC) will bring this total to 114 kW of refrigeration capacity at 4K. Large-scale helium refrigerators are produced by Koch Process Systems, CTI Cryogenics, CVI, Air Products, Linde. and others. It is a fact that the large-scale production of helium temperatures is a routine, commercially available technology. SMALL SYSTEMS The first major requirement for small ground-based cryocoolers was brought about by the need to refrigerate ground-based parametric amplifiers to 4K for use in the satellite communicationnetwork. Several units meeting this requirement were developed and built by A.D. Little, based on the Gifford-McMahon (G-M) cycle,with a Joule-Thomson (J-T) circuit.These units produced a few watts at 3.8 - 4K. Further development in amplifier performance led to an amplifier which would perform satisfactorily at 20K. As a result, the major market for the Gifford-McMahon, closed-cycle, 20K cooler evolved. Cryogenic Technology, Inc. has produced close to 1,000 of these units, which are in continuous operation in the satellite communications network. This basic Gifford-McMahon cooler is also produced by Cryomech, Inc.
340
Applied Superconductivity
NAME
CRYOGENIC CAPACITY 24 kW at 4 Kelvin 5000 Uhr Liquid He 83000 L Liquid He Storage 254,000 L Liquid N2 Storage 10.5 kW at 4K 500 kW at 77K
NAME
ermilab 1 Tevatron
e
(JET)
Livermore, CA
MFTF-B: 1.05 x lo6 kg magnets at 4.358 900 m2 of cryopanels at 1.35K 10 days to :ool system to 48
Oxon, UK
YES
500 W at 3.8K 20 kW at 77K 300 W at 1.75K 700 W at 4.OK 10.7 kW at 80K 1.4 kW at 4K 50000 kg at 1.6K 120000 kg at 4.5 k 20000 kg at 80K 380000 kg at 3.8H 20 days to cool system to UK
Experimen1 or&upra I European Fusion okomak ir ntemational Fusion Su il erconducting Magnet est Facility (IFSMTF) rookhaven National b (BNL) lectron-Proton Cotcider HERA Deutscher
I
France
1988
Oak Ridge, TN
1983
Upton, NY
1985
24.8 kW at 3.8K
Hamburg, FRG
w391)
NO
Addendum
II:
341
and by Air Products, Inc. These two companies, as well as Cryogenic Technology, Inc. (CII Cryogenics), have built a number of G-M units for specific applications with the addition of a J-T loop to provide a final stage of refrigeration at 4K. Units of this type have also been furnished by Cryosystems, Inc. and by Cryogenic Consultants, Ltd. in England. Installations include cooling of computer systems and cooling of superconducting magnets for magnetic separation processing and NMR experiments. The next major use to evolve was that of cooling infrared (IR) detectors. The initial requirements for IR detectors were 0.25 - 2W at 80 K. A number of manufacturers become involved in producing refrigerators for this level of refrigeration. Thousands of open cycle J-T units have been produced as well as several thousand integral Stirling and split Stirling refrigerators. These units are used for cooling military IR detectors and are produced both in the United States and abroad. For instance, each Bradley Fighting Vehicle uses eight separate cryocoolers for IR systems. This requirement for large numbers of coolers for infrared detectors in the military system led to development of a common module cryocooler meeting specific size, weight, and performance requirements. These units are manufactured by a number of companies abroad in order to serve their own government defense systems. These companies, in addition to CII Cryogenics and Air Products, Inc., include Hughes Aircraft, Texas Instruments, H.R. Textron, and Magnavox in the U.S.; Telefunken Co. in Germany; LAir Liquide and AB.G. Semca in France; Hymatic in England; Philips in Holland; Ricer Ltd. in Israel; and Galileo Corporation in Italy. The third major commercial use is that of Cryopumping. Cryopumping produces a high vacuum by condensing residual gases on cryogenically cooled panels. A number of G-M and Stirling cycle refrigerators were installed on cryopumping systems in the early 1970s. However, the market did not fully develop until coolers were required for semiconductor production. The general range of cooling required for cryopumping systems is 50 - 65W at 80K and 5W at 12 - 1X. Closed cycle refrigerators for cryopumps in this range are produced by CH Cryogenics, Air Products, and CVI, Inc. In addition, the major vacuum equipment companies produce their own refrigerator systems. These include Balzers High Vacuum Varian, Inc., and Sargetn Welch, Inc., in the U.S.; LAir Liquide in France; Leybold-Heraeus in Germany; and Osaka Oxygen Industries, Suzuki Shokan Ltd., Ulvac Cryogenics, Inc., and Toshiba Corp. in Japan. Although there are no companies producing many refrigerators meeting the requirements of IW at 4K with a reasonable efficiency and size suitable for cooling small superconductive devices, the major manufacturers listed above have the capability to develop such systems. Ground-based cryocoolers, large scale or small scale, for use at tually off-the-shelf. (See Table 2.)
4K,
342
Applied Superconductivity
42% 1-4w
lO-2OK l-5W
2X&&lK
4wBr6ow Cryopumpiq
X X X X
X X X
X X
CVI, Incorporated Hughes Aimaftt Co. MMRTu%ologics MagnaMx Sargent Welch Texas Instruments H.R. Textroa
X X
X
X
X
X
X
Addendum
343
Cryogenic cooling and storage have been used in space instruments for over twenty years. The cooling needed is typically for temperatures below the boiling point of liquid nitrogen (77K), and for heat loads of ten watts or less. The primary need for cooling is for IR detectors for astronomy, and for surveillance; X-ray and gamma-ray detectors have also been cooled. The storage of cryogenic fluids for the atmosphere ofmanned spacecraft and for power production in fuel cells is a major use dating back to the late 1950s. Many additional cryogenic applications continue to appear. Three basic refrigeration methods are used to meet these cooling requirements: (1) passive thermal radiation to space, (2) storage of cryogenic fluids or solids, and (3) active refrigerators. Passive radiation to space is a simple and reliable method of producing small amounts of cooling. This method has limited applications because the amount of cooling obtainable is very small at cryogenic temperatures--typically fractions of a watt. The practical limit is that the radiator becomes very large and heavy for larger loads. Storage of fluids or solids has been the mainstay for cryogenic cooling in space. This method employs highly insulated storage tanks. The cooling temperatures obtained range all the way down to less than 2K in the case of superfluid helium storage. Many cryogenic materials have been used to produce cooling by using the heat of vaporization to absorb heat loads. The practical limitation of cryogenic storage is that the size and weight become large for long-duration missions and for high heat loads. Active, closed-cycle refrigeration systems (cryocoolers) do not suffer from the severe size and weight limitations of the other methods. Electrical power is used to produce continuous cooling. The limitation of space cryocoolers is availability. Such coolers have been under development for about 30 years. Although some types have been developed and successfully flown, other types still require significant development. The space cryocoolers under development usually fall into three categories: regenerative or Stirling coolers, reverse Brayton coolers, and J-T coolers. Table 3 gives a summary status. Development of a long-life space cryocooler has been an elusive goal because of fundamental problems relating to contamination and wear. Stirling coolers were used in space to cool gamma ray detectors to about 80K in the P78 satellite. These coolers were built by the Philips Corporation and employed a linear drive mechanism to achieve several years of operation. Degradation in performance occurred during the mission in the form of steadily rising cooler temperature.
344
Applied Superconductivity
A few
watts
at 75K is feasible
65W2W
Concept Test
(JTJ)
Linear Stirling Tactical Stirling Small scale (l-2W), 10K . . 10K is about the lower temperature limit achievable by regenerative cycles. Some systems in test: In Test In Test Qualified cycles 65Kl2W 65WlW In Test Off-the-Shelf
10K and lower feasible for recuperative Brayton 1OWlW IOWlW 15WlW
Addendum
II :
345
Avery promising development of a Stirling type refrigerator is taking place in Great Britain at Oxford University and the Rutherford Appleton Laboratories in conjunction with British Aerospace Corporation. Operational times of about 20,000 hours have been achieved, and this machine has been slated for use in several space systems. The limitations of Stirling coolers are that temperatures below about 20K are difftcult to achieve with acceptable power input requirements. Brayton coolers are being developed for temperatures below 1OKand higher cooling loads. The AiResearch Corporation is developing such a cooler using multistage turbomachinery. The Arthur D. Little Company is developing a Brayton cooler that uses positive displacement (piston and cylinder) principles. Both the Stirling and Brayton coolers have moving parts in the cold regions of the machines that generate vibrations. These vibrations are often unacceptable in sensitive space instruments. Additional benetitsaccrue to the J-T approach in comparison to the Stirling and Brayton machines as a result of the fact that J-T coolers produce liquid cryogens. High, short-term peak heating loads, and variable heating loads, can be absorbed at the constant temperature of the boiling liquid refrigerant. This is not possible with other systems that produce only a cold gas, and is sometimes very important to space instrument cooling.
346
Applied Superconductivity
High temperature superconductors are brittle materials with low tensile strength but good compressive strength. Superconducting magnets utilizing such materials thus require designs that minimize the stresses that develop as a result of the j x B body forces. It may be possible to form a superconducting composite using a high strength material to withstand the stresses associated with large magnetic fields. However, the highest strength material known graphite fiber, lacks the strength required for the high fields anticipated. A stress-free condition can be achieved if the forces are counteracted by a stiff constraining medium. An example is presented for a cylindrical solenoid geometry consisting of laminated superconductor, constrained by a stiff composite shell (Figure F-l). The materials selected for the laminate should have a high elastic modulus and high yield strength so that the stress generated in the superconductor is small. The laminate should also include a protective layer of metal adjacent to the superconductor to minimize the effects of quenching. By applying a shrink-fitting high strength alloy jacket to the laminate, the high compressive strength of the superconductor can be exploited. Figure F-l STRUCTURAL SUPPORT FOR HIGH FIELD APPLICATIONS
SUPERCONDUCTOR/CERAMIC/METAL
Addendum
347
A JOSEPHSON
4 BIT\ MICROPROCESSOR*
SEIGO KOTANI, NORIO FUJIMAKI, TAKESHI IMAMURA, AND SHINY H&XJO FUJITSU LIMITED 10-l. MORINOSATO-WAKAMIYA, ATSUGI, 243-01, JAPAN
A 4-bit microprocessor - the first microprocessor to use Josephson devices -- will be described. The circuit was fabricated using 2.5 -pm all-niobium Josephson technology, utilizing Nb/AlOJNb Josephson junctions. 5011 Josephson junctions are contained on a 5.0 x 5.0 mm die. The microprocessor was operated with clocks of up to 770 MHz under worst-case conditions, dissipating 5 mW. Josephson gates offer superb high speed operation and low power dissipation. More than ten types of logic gates have been reported. We proposed a gate, which we called the modified variable threshold logic ( MVTL) gate, for Josephson LSI circuits. The MVTL OR gate was 46 x 31pm2 and the unit cell was 112 x 79pm2; each unit cell was composed of three gates, two MVTL QR gates and one single-junction m gate, so this unit cell performs an (A + B) (C + D) logical operation. 25pm diameter Josephson junctions were used in these gates. We demonstrated its high-speed,) and applied it for a Hi-bit arithmetic logic unit (ALU)? We then designed the microprocessor. This was the first instance of applying Josephson devices to a microprocessor, sowe wanted to verily the feasibility of the chip in comparison with a typical microprocessor constructed with semiconductor devices. We selected chip functions that were similar to those of the Am 2901 microprocessor made by Advanced Micro Devices Inc.) This microprocessor has come to be regarded as the standard four-bit microprocessor slice. The fastest operation of this microprocessor has been achieved using GaAs devices; a 72 MHz clock with a 2.2 W power dissipation.) Figure 1 is a block diagram of our microprocessor. It has a dual memory set which is used as a 16-word by 4-bit two-port RAM with a RAM shifter, an eight-function ALU, a Q register with a Q shifter, and several controller. This circuit is driven by three-phase power; 01. &. and 83.6) Their waveforms are sinusoidal with dc offsets and their phases fare separated by 120 degrees. The three-phase power has the advantage of preventing the racing phenomenon in the logic circuit. and a Josephson timed inverter (II) can be fabricated easily if the power is used as the timing signal. Dual-rail logic was adopted in the ALU and controllers of the
348
Applied Superconductivity
microprocessor, and complement signals are made from the input signals by TIs powered by 01. Decoding operations are run in gates powered by 61. reading memory data by $z, and modifying- and writing data by b. _ Figure 1 BLOCK DIAGRAM OF MICROPROCESSOR
A Address
4
A Decode o--L Function 4
6 A Memory Cell
Aegislcr I/O -0 -0
In this microprocessor, the critical path is the route of the carry signal transmitted from LSB to MSB in the ALU and then the sum signal transferred from the ALU to the RAM. After detail design, the number of gates which have to switch sequentially along the path proved to be 41, with an interconnectin line length of 15 mm. MVTL gates operate with a sub-ten ps gate delay in actual circuits, 5, and the propagation delay in interconnecting lines is about 8 ps/mm. By rough estimation, the critical delay time seems to be 0.5 ns and the duty ratio of the sinusoidal power is l/2, so the maximum clock frequency was estimated to be 1 GHZ.
Addendum
349
The process is summarized inTable 1. This fabrication process is almost the same as that reported previously. *) The circuit consists of Nb/AloJNb Josephson junctions, Nb wiring, MO resistors, and SioZ insulators. Both the minimum junction diameter and line width are 2.5 urn. The interconnecting lines ars 4 urn wide. The critical current density of the fabricated Josephson junction was 2300 A/cm , this being slightly higher than the optimum design value (2100 A/cm*). The measured operating margin was L34% for the h4VTL OR gate and A32% for the unit cell.
Table 1 PROCESS SUMMARY
JOSEPHSON JUNCMON JUNCTION MINIMUM DL4hWIER MINIMUM WIDTH INSUIATOR RESISTOR WIRING
Figure 2 is a photograph of the fabricated chip with unit locations. The specifications of the chip are given in Table 2. The basic gate is MVTL, as mentioned above, and the total number of gates is l&01. Six power pads are provided, three for sinusoidal power and the others for dc offsets. The number of signal pads is 32; 8 for address, 14 for data, and 10 for control. 52 ground pads are used to suppress the deviation of the chip-ground level from the packageground level.
Table 2 CHIP SUMMARY
DIE SUE MEMORY # AL.UFUNCIIONS RASIC GATJ5 # GATES # JUNCITONS POWER &PHASE SINUSOIDAL) # SIGNAL. PADS # PWRIGND PAiS CLOCK FREQUENCY
5011 hW 32 s8 77OMIiz
350
Applied Superconductivity
PHOTOGRAPH OF MICROPROCRSSOR
All experiments were performed with the chip immersed in liq+uidHe. All functions and source combinations were confirmed with an operating margin of -16% at a clock frequency up to 100 MHz which was limited by the maximum clock of the word pattern enerator. The operation along the critical path of the chip was tested using the high-speed p2s e generator, and Figure 3 shows the results obtained at the maximum clock frequency (770 MHz). The same waveform was obtained for the MSB signals of output and memory I/O, thus confirming correct operation. The reason for the amplitude difference in these waveforms is that the timing of the sum signal arriving at the memory is later than that of the signal at the output controller. The power has a sinusoidal waveform, so the bias power of the memory I/O gate at its switching timing The power was consumed at the OR gate supply resistors, and the value was obtained from the bias level. The gate power dissipation was 3.6 pW/gate. and the total power of the chip was 5 mW.
Addendum
I I:
351
AT 770 MHz
$1
$2
@3
Power
>
We verified that the Josephson microprocessor operated with a one-order three-orders less power than a semiconductor microprocessor.
The
present r-ch
puting System, conducted under a program set by the Agency of Industrial Science and Technolcgy, Ministry of IntcmationaI Trade and Industry.
352
Applied Superconductivity
References: 1) Fujimaki, N., Kotani, S., Hasuo, S. and Yamaoka, T., 9 ps Gate Delay Josephson OR Gate with Modified Variable Threshold Logic, Japan J. Appl. Phys.. ~01.24, p.Ll; Jan, 1985. 2) Kotani, S., Imamura, T. and Hasuo. S.. A 2.5~~ Josephson Technical Digest; Dec., 1987. OR Gate, IEEE IEDM
3) Kotani,S., Fujimaki. N., Imamura,T.. and Hasuo, S., A 1 ns Josephson ISSCC DIGEST OF TECHNICAL PAPERS, p.60-61; Feb., 1987.
4) Mick, J., Am 2900 Bipolar Microprocessor Family, IEEE Proceedings Annual Workshop on Microprogramming, ~56-63; 1975.
of the Eighth
5) Hendrickson,N.. Larkins. B.. Bartolotti, R., Deming, R., and Deyhimy, I., A GaAs BitSlice Microprocessor Chip Set, IEEE Proceedings of the GaAs IC Symposium, p.197200; Oct. 1987. 6) Fujimaki, N., Kotani, S.. Imamura, T., and Hasuo, S, Josephson &bit Shift Register, IEEE J. of Solid-State Circuits, vol.SC-22, p-886-891; Oct. 1987.
FOR
noTokyomo ofthe
Report Memorandum M
1152
JAPANESE
In JFY 1988 the Japanese Government will spend over Summary: 9,049 million yen ($72.4 million) and Ja anese corporations will spend at least 11,511 million yen (% 92 million) on superconductivity RLD. Much of the government budget and seventy percent of the corporate funds will be for high temperature superconductivity. Largely separate from these funds, Japanese corporations have contributed considerable amounts of money to join the new International Superconductivity Technclogy Center (ISTEC), a private foundation engaged in RLD on high temperature superconductivity, initiated with guidance from MITI. The total of such funds so far committed, though not necessarily completely handed over, is 5,420 million yen ($43.4 million). (Caution should be excercised in interpreting ISTEC funds as 1) in contrast to the government and corporate funds reported herein, the ISTEC funds largely consist of one-time initial donation, much of which will probably be used for capital expenses, and 2) in that ISTEC is in principle open to international participation. See section on ISTEC below for further details.)
354
Applied Superconductivity
GOVERNlENT Hr. Makio Hattori, Director for Ratecial Research and Agency (STA) of Development, RLD Bureau, Science and Technology Japan has provided the following information on Japanese Governments FYI88 budget for superconductivity RLD as approved by the National Diet on April 7, 1988, with corresponding figures for FY87 given for comparison. Corporate RLD figures are from a recently concluded survey by the NSY Tokyo Office. (Figures for both government and private sector are in million yen and include both low temperature and high temperature superconductivity. Current exchange rate is 125 yen to the dollar.): (In Million Yen) FY1987 FY1988 Science and Technology Agency (STA) 0 (11 Multicore Project: 2.044 (2) RLD on Superconducting coils (Japan Atomic Energy Research Inst.): 678 1,109 (3) ERATO Project on Uaqnetoflux Logic Research (Research 377 Development Corp.: 395 97 75 (41 Others --_______________--__-__-__--_--__-~_-__-___________ Subtotal of STA: 1,583 3,192 ($25.5 mil.1 Ministry (11 of Education, Science and Culture
(MONBUSHO)
Grants-in-Aid for Scientific Research: (*Undecided) (563) [*Note: Proposals for grants are still under review, and to be decided in May ]
RLD Equipment Procured by Funds under Special Accounts for 527 National Schools: 1,775 ---_---__-_____-_-__-~~~~~~-~~-~------~-~~~-~-~----527 Subtotal of Monbusho: 1,775 ($14.2 mil.) Plus "Grants": (563) (?I
(2)
Trade
and Industry
(HITI) 1,123
RID Projects on Basic Technologies for Future Industries: 71 National RLD Program (the Large Scale Project): 350 RLD on Energy Conservation Technologies (the Moonlight Project: Primarily for Development of Lou Temperature or Traditional Superconducting Power Generator Systems):
100 Surveys/Studies on Matters RE Superconductivity: Specific International Joint Research Project: Others: Subtotal of HITI: (HOT) 295 (MPT) 0 0 35
440
1,652 182 14 17
_---________________-____~~~~~~~~~_-~___-_--________
556
Hinistry of Transport
597
R&D on Superconducting Telecommunications Systems: 358 57 ---_-_------------__--~~~~~----~~~-~~~~~~~~~~~~-~~~----~ GRAND TOTAL: 2,960 9,049 ($72.4 mil.) (Plus Honbusho Grants: (?I (563) (Honbusho's scientific research grants will be decided in nay.)
With respect to flITI's budget, HIT1 officials contacted by the NSF Tokyo Office expect that most of the funds allocated for the'" Basic Technologies for Future Industries" program will be allocated through the New Energy and Industrial Technology RLD Organization ("NEW NEDO") for contract research by industries, including about 400 - 500 million yen expected to be provided to the International Superconducting Technology Center (ISTEC) for contract research although exact yen amounts are yet to be decided.
356
Applied Superconductivity
CORPORATEEXPENDITURES A recent survey by the NSF Tokyo Office of Japanese corporate superconductivity R&D indicates that 41 leading corporations spent 8,671 million yen ($59.96 million @ 144.61 yen per dollar) in JFY 1987 and that they expect to expend 11,511 million yen ($92 million @125 yen to the dollar) in JFY 1988. Seventy percent of the corporate funds are for high temperature superconductivity. ISTEC The International Superconductivity Technology Center, a newly established private foundation initiated under the guidance of MITI, to conduct R&D in high temperature superconductivity, so far has attracted 45 Japanese companies as full members, meaning that they may partake in the activities of both the etc.) and the Labroatory. The Center (symposia, workshops, initial one time donation for joining both is.2 million yen for the Center and 100 million yen for the Laboratory. The annual fees for the Center and Laboratory are 2 million yen and 12 In addition, fifty Japanese million yen respectively. corporations have joined just the Center, for which the initial donation and annual fees are 2 million yen and 2 million yen respectively . Thus, in total Japanese corporations have committed 5,420 million yen (943 million @125 yen/dollar) including 4,690 million yen in donations and 730 million yen in first year dues. the initial donation may be .spread However, over two years at 60% and 40% respectively. A portion of the initial donation will be used for capital expenses such as buildings and equipment. Other portions will be used for the research, though just how much in each category is uncertain. Corporate funds for ISTEC are largely separate from corporate In addition, ISTEC is in expenditures as reported above. principle open to membership by foreign companies. So far one foreign affiliated company, IBM Japan, has joined the Center (as opposed to the Laboratory). No foreign company has as yet joined as a full member, though some are reportedly considering doing so. Comment: The figures for the government and private sectors represent substantial increases over JFY 1987 levels. For the government sector they are even higher than forecast in September last year by the main Japanese agencies. Host of the government increases can be assumed to be for high temperature superconductivity, with the exceptions of the Moonlight Project and the Magnetic Levitation Railway Systems. 1An undefined
Addendum
357
amount of new money for the Uoonliqht Project will be used high TC superconductivity, although the b;lk will be for conventional superconductivity.) For the corporate sector, of the expenditures are for high temperature superconductivity. (The results of the NSF corporate superconductivity survey will be made available in nay.1
for 709
NOTE:
This report was compiled by nasanobu Miyahara, Affairs Advisor, NSF Tokyo Office, through
with several Japanese government
and academic
358
Applied Superconductivity
FOCUS ON
Oceangoing Ship Planned
REPORTS
ON JAPANESE
DEVELOPMENTS
IN SUPERC0NDlJCtlVll-V
JAPAN _anumt
Volume 1, Number 2
May.
1998
bad will be capable of gcnuating g.ow llewcml of poww @apatuJ s@ is g knots). Mitsubishi Havy Musuies will be rcrponsiblc Ior one cl rhc supuconducting magnus. wilh Tcshiba supplying the aher. Kobe Steel will pmrtde helium rclrigemlion equipmcnt.
Mwcb
II,
pn&ccd 6;mclting and dten rapidly coaling BSCCO (Bismuth. Suomium. Cakium. Copper. Oxygen) matils.
Addendum
II :
359
Remarks
360
Applied Superconductivity
fiiiiii%i~%~i%%ii~6%j~it;
r,----------,--,---J
Final report on Industrial Superconducting Technology Development --December 1987
I
JL
M
I--@ T
I
INTL CENTER FOR INDUSTRIAL SUPERCONDUCTING TECHNOLOGY SUPERCONDUCTING ENGR. LAB (ISTEC) Ninety Corporate Members
,
R A D CONSORTfUM FOR SUPERCONDUCTING GENERATOR AND EQUIPMENT AND MATERIALS (SUPER-GM) Eight year project; 13 Electnc Utilities and Electric Eauioment Manufacturino Comoanies are Consortium Members
0
E
--GB-*I:T
SPECIAL HTS RESEARCH PROJECTS -Group Research centered at Tokyo University Research centered at Tohoku University
M 0 T
Addendum
361
HIGH TEMPERATURE
SUPERCONDUCTIVITY
FUNDING
($M)
Department of Defense ArmY Air Force Navy SD10 DARPA NSA BTI Department of Energy Basic Energy Fusion Energy High Energy and NuclearPhysics Defense Depts Office of Conservation & Renewable Energy Office of Fossil Energy National Science Foundation Department of Commerce NASA Department Department
lb.5
0.3 3.7 6.8 2.0 1.5 0.2 2.0
XL9
1.9 4.6 10.3 12.9 20.7 0.5
__
4.85
0.2
0.2
0.28
0.3
29 (NBS)
lu
39
u.5 2.8
!u of Transporation of Interior
Q.5 __
0.l
!u !ufi
362
Applied Superconductivity
GLOSSARY OF TERMS
AC - Alternating current AJD Converter - Analog-twligital convexter. This device is used to transform analog signals into digital signals. such as electric current, or spatial derivatives thereof are dependent
Coherence L..eegth - The characteristic size of a Cooper pair. CMOS - Complementary metal oxide semiconductor Cooper pair -The paired electrons that are believed respoesible for the phenomenon of low temperature superconductivity and may play a role in high tempcraturc superconductivity. DARPA - Defense Advanced Research Projects Agency dB - Decibels DC - Dired current DDR & E - Director, Defense Research and Engineering DRAM - Dynamic Random Access Memory ECCM - Electronic counter-countermeasures ELF - Extra Low Frequency. A slightly lower frequency than VLF; less than 10 KHz EM - Electromagnetic EM Launcher - An electromagnetic device used to accelerate projectiles to high speeds. A railgun is one example. EO - Electra-optic FEL - Free Electron Laser FET - Reid Effect Transistor GaAs - Gallium arsenide. The fastest conventional electronic devices are GaAs devices.
Addendum
363
Gauss - A measure of magnetic tield strength. The Earths geomagnetic field averages about one-third of a Gauss. GHz - Gigahertz or billions of cycles per second. I& - Magnetic critical field, or the magnetic field above which superwnduditity is quenched.
HEMT - High electron mobility transistor. HF - High Frequency hp - Horsepower HTS - Hi Temperature Superwndudor
IC - Integrated Circuit IPPA/IRPPA - Infrared Focal Plane Array. 3-3c0THz IR - Infrared radiation
JC- Critical current density, or the amount of current a superconductor is able to carry. JJ - Josephson Junction (see below) Joscphsoa Junction - A superconducting electronic device. K - Degrec~ KelvtL Zero degrees Kelvin is equivalent to -273 degrees Celsius or 459 degrees Fahrenheit. LSI - Large Scale Integration LTS - Low Temperatie Superconductor
EPAs are used for IR sensors. The frequencies for infrared are
MHD - Magnetohydrodynamic. MHD Drive has been proposed as an advanced superconductive propulsion technique.
MHP - Magnetohydrodynamic
power
MHz - Megahertz or millions of cycles per second. Micron - One micrometer or one one-millionth of one meter. MMIC - Miieter-wave integrated circuit
MMW - Millimeter-wave. This corresponds to frequencies of 40 - 100 GHz. MOPS - Millions of operations per second.
364
Applied Superconductivity
NPB- Neutral Particle Beam. POM - Program Objectives Memorandum. R&AT - Research and AdvaDccd Technology RF - Radio Frequency. This is a broad frequency band, from 100 MHz to 100 G&
of electromagnetic wave, includiq microwave and milEmeter-wave. encompassingseveral types
RPM - Revolutions per minute SAW - Surface Awustic Wave SD10 - Strategic Defense Initiative Offk SECDEF - Secretary of Defeose SMES - Superconducting Magnetic Energy Storage SQUID - Superconducting Quantum Interference Device. This device is used extensively in magnetic sensors, such as mine detection devices. SSTS - Space SmveiIIaoce and Tracking System Tc - The transition temperature or temperature at which a given material becomes superconducting. TN - Noise temperature TesIa - Ten thousand Gauss are equivalent to one Teska. Large permanent magnets typically have field strengths of 1 - 5 TesIa. THz - Terahertz or trillions of cycles per second.
UHF - Ultra Hi
Frequency, 03 - 1 GHz
USD(A) - Under Secretary of Defense for Acquisition VLSI - Very Large scale Integration VHSIC - Very Hi Speed Integrated Circuit Corresponds to frequencies of 10-u) KHz.
Addendum
II :
365
BiCaSrCuO - Biiutb
Diftkult to produce
f_aBaCuOd - Lanthanum Barium Cuprate, one of the fust high temperature Nb - Niobium, a low temperature supexwnductor. NbC - Niobium coppex, a low temperature compound. NbN - Niobium nitride., a low temperature compound. mGe - Niobium germanium, a low temperature compound.
superconductors.
Nbgi - Niobium silicoa, a low temperature compound. MD - Niobium tin, a low temperature compound.
TlCaBaCuO - Thallium Calcium Barium Cuprate, one of the most recently discovered bigb temperature compounds. It has a Tc ofl25 K. YBaCuO - Thesuperconducting material currently receiving a bigb degree of research attention. compound is YBa2Cu307-x, also known as YBCO. The full
l-2-3 - This sequence of numbers refers to the chemkaJ composition of tk above mentioned cuprate superconductors, YBaCuO.