Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With...

39
Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion Center ARIES Meeting UCSD January 10, 2002
  • date post

    22-Dec-2015
  • Category

    Documents

  • view

    213
  • download

    0

Transcript of Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With...

Page 1: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

Magnet considerations for ARIES IFE

HTS/LTS algorithms and design options

L. Bromberg

With contributions from J.H. Schultz

MIT Plasma Science and Fusion Center

ARIES Meeting

UCSD

January 10, 2002

Page 2: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

Status of near term magnet development for IFE

High current transport experiment (low energy) Acceleration section (medium energy acceleration)

Superconductor magnet options for IFE High temperature (last meeting) Low temperature

Insulation considerations

ARIES IFE

Page 3: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

HCX (High Current Transport Experiment): 800 m A x 1.6-2 MeV, front-end of:

IRE (Integrated Research Experiment): 200 A x 200 MeV

Possible front-end of HIFSA commercial reactor: 10 kA x 10 GeV)

Present work in magnet community for IFE

Page 4: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.
Page 5: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

HCX cylindrical

Page 6: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

HCXPlate racetrack

Page 7: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

IRE design of array

Page 8: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.
Page 9: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.
Page 10: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

For IFE with HI driver, cost minimization of magnet system is paramount

Cost minimization of final optics by

Epitaxial techniques High Tc to minimize quench requirements and cost of the

cryogenic system (dry system) Optimize structural requirements

LTS magnets require winding, a more costly proposition

Implications of technology program for HU IFE final optics magnets

Page 11: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

IFE magnet construction Epitaxial YBCO films

STRUCTURAL PLATE/ SHELL

HIGH TEMPERATURE SUPERCONDUCTOR

YBCO (123) material

Ni-based structure

CeO2

Silver coating

Page 12: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

LTS magnet designPlate array

Cryostat over magnet assembly; shown over individual plate array for illustration purposes

Page 13: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

BSCCO 2212 layered pancakes on silver(L. Bromberg, MIT, 1997)

Page 14: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

Magnet configuration options

Plate magnets, with superconductor on the surface of structural plates

Two different arrangements are possible (a = 0 or b = 0)

~ r2 (a sin (2) + b cos (2)

sin (2)cos (2)

Page 15: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

Current density sin (2) and cos (2) quad arrays

cos (2) sin (2)

Page 16: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

cos (2 ) plate design optionsHTS

Page 17: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

sin (2) magnet section (not planar due to current return!)

Page 18: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

Current distribution in plate quad arrays

The current density for the two cases is different For the sin (2):

|Kwall| = 4 A x /o (on the x-wall), -a < x < a |Kwall| = 4 A y /o (on the y-wall), -a < y < a

For the cos (2): |Kwall| = 4 A a /o

In both cases, the magnetic field is:| | = 2 A r and the gradient is

| |’ = 2 A

Page 19: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

Current distribution in quad arraysHTS vs LTS supercondutors

HTS superconductor is deposited directly on structural plates

Epitaxial techniques can be used to achieve desired current density for optimal quadrupole field generation Deposition of superconducting material

directly on structural substrate minimized motion of SC and heat generation

Insulation is achieved by depositing thin insulating layer between superconductor and substrate

Conductor surface fraction ~ 0.9

LTS superconductor is wound on structural plates

Epitaxial techniques can not be used due to superconductor stability Stability, quench protection and

manufacturability requires substantial amounts of copper

For NbTi, assume Cu/SC ratio of 1. (Decreases strand current density by a factor of 2).

Stability requires operation at 0.7 of critical

Insulation around each turn; minimal thickness 1 mm. Due to winding difficulties, circular

cross-section conductors preferred. Due to winding, conductor surface filling

fraction ~ 0.5

Page 20: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

Superconductor implicationsHTS and LTS

YBCO HTS material is very sensitive to field orientation (anisotropic): preferred orientation of field is parallel to YBCO tape, (B || ab) Preliminary critical current density: 1010 A/m2

For sin (2) arrays, field is perpendicular to SC For cos (2) arrays, field is parallel to SC

sin (2) requires non-uniform current density, which means that for similar average current density, peak is higher (for comparable thicknesses, current density is higher)

cos (2) requires uniform current density

LTS is isotropic

Page 21: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

The conductor thickness determined by required surface current density K, conductor current Jcond, insulation thickness tins, and number of layers Nlayer

For HTS:

For LTS

Conductor thicknessHTS and LTS

tcond=1

0.9J condK

+2tinsNlayer

tcond=2

0.7J condK

+2tinsNlayer

Page 22: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

Forces in Epitaxial quad arraysHTS and LTS

Because of symmetry, forces are mostly along the substrate (i.e., tension or compression, with almost no shear)

For both sin (2) and cos (2) arrays: F = 4 A2 a3/ o

= F/t (t : thickness of structural plate)

To limit strain in SC to 0.2% (comparable to LTS) Since superconductor material is integrated with structure, equal strains Not additional constraint, since 0.2% corresponds to yield limit in structure. High modulus structural material is desirable

For LTS design with increased conductor thickness due to lower SC current density, need for substantial copper and thicker insulation thickness, the structural performance (percentage of structure in magnet cross-section is decreased compare with HTS

Page 23: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

100

1000

10000

100000

0 5 10 15 20 25

"Un-Critical" Critical Current

Density, A/mm 2

Applied Field, T

Nb3Al

Bi2212Tape

Bi2223Round 4.2K

ITER-Nb 3Sn

Low AC Loss

Bi2223Round 75 K

YBCO Tape 75 K

NbTi

YBCO Tape

At 4.2 K UnlessOtherwise Stated

1.8 KNb-Ti-Ta

Nb3Sn

High AC Loss

ARIES RS

HTS Superconductor options and comparison with LTS(YBCO tape at 4 K and at 75 K)

Courtesy of Lee, UW Madison

Page 24: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

YBCO Current density with field in the “bad” direction (B||c)as a function of temperature

Figure 2 Comparison of Representative Data for YBCO for various fields & temperaturesvs NbTi and Nb3Sn at 8 T and 4.2 K. (M. Suenaga, :The Coated Conductor Issues”, 98HTS/LTS Workshop for High Energy Physics, Napa, CA, Mar, 98)

Page 25: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

Typical Quad magnet design

HTS and LTS

HTS LTS

Structure stress Pa 5.0E+08 5.0E+08Conductor current density A/m^2 9.0E+09 7.0E+08Thickness of beam tube m 0.001 0.001

0.00001 0.00025Thickness of MLI m 0.002 0.002

Bmax T 6 6a m 0.02 0.05Field gradient T/m 300 120A T/m 150 60

Structure thickness m 0.0011 0.0029

SuperconductorCurrent density (K)

sin 2 theta A/m 9.5E+06 9.5E+06cos 2 theta A/m 1.9E+07 1.9E+07

Thickness for 1.e6 A/cm^2sin 2 theta m 0.0011 0.0136cos 2 theta m 0.0021 0.0273

Fraction of cell for magnet, insulation, beam tubesin 2 theta 0.31 0.53cos 2 theta 0.37 0.66

Current, cos 2 thetanumber of turns per layer 10 10number of plates, 10 microns YBCO 212 27Current per turn A 360 7000

Current, sin 2 thetanumber of turns per layer 10 10number of plates, 10 microns YBCO 106 1364Current per turn A 360 7000

Page 26: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

SC for IFE magnetsYBCO for HTSNbTi for LTS

Current density in conductor can be as high as 1010 A/m2 (1 MA/cm2) for fields as high as 9 T, a temperatures between 20 K and 77 K

At 9T (B || c), T < 20K

Temperatures between 4 K and 77 K can be achieved by using dry system (no liquid coolant)

Cooling is provided by conduction cooling to a cryocooler head.

Conductor current density of NbTi is 7 108 A/m2 (4 109*(0.7/2/2) A/m2)

Page 27: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

Fraction of cell not allocated to beam(linear ratio, not area ratio)

HTS

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0.4

0.45

0.5

200 250 300 350 400 450 500 550

Field gradient (T/m)

Magnet, insulation, tube fraction

sin 2 theta

cos 2 theta

Page 28: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

Fraction of cell not allocated to beamField gradient = 300 T/m

0

0.5

1

1.5

2

2.5

0 2000 4000 6000 8000 10000 12000

Current density (MA/m^2)

Magnet, insulation, tube fraction

sin 2 theta

cos 2 theta

Page 29: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

Fraction of cell not allocated to beamField gradient = 300 T/m

0

0.1

0.2

0.3

0.4

0.5

0.6

0 100 200 300 400 500 600

Stress in structure (MPa)

Magnet, insulation, tube fraction

sin 2 theta

cos 2 theta

Page 30: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

Magnet quenchHTS and LTS

Energy in magnet is small (for both sin 2 theta and cos 2 theta):

E = 16 A2 a4 L / 3 o (L is the quad length)

For typical magnets (L ~ 1 m), E ~ 20 kJ/quad

If 100 quads are in parallel, 2 MJ per supply

If single power supply, @400 A and 5 kV discharge, dump ~ 1 s However, at HTS very high current density, magnet will self destroy at this very high current

density

Quench protection for HTS: passive, quench prevention by large thermal mass to avoid flux jumping, large heat release.

Quench protection for LTS: provide enough copper so that the current flowing in the copper during quench does not result in temperature greater than ~150-200 K.

Page 31: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

Edge conditions of the quad arrayCurrent density of outermost quad array boundary modified to control

field profile

Page 32: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

Magnet assembly toleranceHTS and LTS

For HTS with Epitaxial techniques

manufacturing of the YBCO plates with high precision (~ m’ s) Plates can be assembled to tolerances of 5 mils (0.12 mm) Plate supports can be assembled with 10 mil accuracy Magnet racetrack can be placed with 10 mils accuracy (0.25 mm) along

the entire way of the plate (~ 1 m across, 1 m apart).

For LTS with winding techniques: 20 mil gap for insertion of conductor into grove, with random location

within groove. Similar tolerances in assembling plates as HTS

Page 33: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

CoilCooling:

HTS

“Dry” cooling

Assuming that quads are only cooled at each end If only Ni-based material, large temperature raise to midplane of

quad (T ~ 100 K) Cu placed in parallel to structural plates (bonded at the structural

plate edges, to prevent warping) Required cooling cross-section: ~ 5% of plate cross section

0

5

10

15

20

25

30

35

40

0 0.02 0.04 0.06 0.08 0.1

Fraction of Cu

Delta T (K)

1 mW/cm^3

Structure

Copper (for condution cooling)

Superconductor

Page 34: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

Coil Cooling:LTS

NbTi superconductor cooled by pool boiling. Liquid He required for conductor stability (only material with any

thermal heat capacity at 4 K) NbTi conductor will require a cryostat vessel around the magnets

Average additional cross section requirement: 20% of magnet cross section

NbTi magnet require more sophisticated plumbing and manifolding than HTS magnet

Page 35: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

Cryogenic cooling issuesHTS vs LTS

Due to good design (as in the IRE design shown above), operation at 77 K only slightly reduces the required thermal insulation gap Little radiation from 77 K to 4 K Support can be long without implication to radial build Required distance determined by requirement of no contact between 4 K and 77 K

Therefore, the use of HTS is not driven by cryogenic requirements.

Cryogenic requirements substantially reduced by operation at ~ 77 K (by about a factor of 8, scaling from IRE design)

Operation of HTS below 77 K by the use of cryocooler (dry operation)

LTS requires cooling with as low a temperature as possible (4.4-4.8 K), because of large sensitivity of operating current density to temperature

Page 36: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

Advantages of plate quad arrays

The forces are only in tension/compression in the plates There are no loads that are perpendicular to the plates that need to be reacted in shear. Very important consideration for insulator

These forces are easy to react, and do not require the need of large inter-quad structure The only structure is needed to control off-normal loads, gravity and for accurate

positioning of the plates.

The field is a perfect quadrupole field, with total absence (in the ideal conditions) of higher order components Errors due to

misalignment errors in plate manufacture end-fields

Possibility for cheap manufacturing!

Page 37: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

Irradiation considerationsHTS vs LTS

Final focusing magnets will experience high dose:

40 MGray/year from ’s, 8 MGray/year from neutrons (Latkowski)

Very high peaking, ~105 – 106 (Sawan, 1981)

Very high instantaneous dose rate: 105 – 106 Gray/s

LTS materials will require organic insulation (at least hybrid insulation, i.e., ceramic/organic composites), with limited lifetime (107-108 Gray)

HTS design uses mainly inorganic insulation (mainly ceramic), with longer lifetime (maybe 1011-1012 Gray)

Page 38: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

Irradiation testing of insulators at M.I.T.

• Irradiation damage to organics mainly due to radiochemistry, driven mainly by electrons

• Radiation facility consisting of:

• E-beam unit (200 kV, 4 mA)• Capable of large dose rates (up to Grad/s)• Large sample capability (5cm x 5 cm)• Capable of cryogenic testing

Page 39: Magnet considerations for ARIES IFE HTS/LTS algorithms and design options L. Bromberg With contributions from J.H. Schultz MIT Plasma Science and Fusion.

Summary

Design options for plate magnets for IFE quad final optics have been investigated Use High temperature superconductor (YBCO) sin (2) option requires angle elements (non-planar) Epitaxial technique on flat plates, angle elements Large number of elements, easily manufactured.

Design algorithms were developed HTS designs LTS materials limit substantially design space

Irradiation effect need to be determined Lack of large shear between elements allows the possibility of all inorganic

insulation Very large dose rates