November 29 th 2007 Thermomag-07 A CARE-HHH workshop on Heat Generation and Transfer in Superconducting Magnet 19-21 November 2007 Paris WS-Summary based

November 29th 2007

Thermomag-07A CARE-HHH workshop on

Heat Generation and Transfer in Superconducting Magnet

19-21 November 2007Paris

WS-Summarybased on common “Close-out” by

session chairs

B. Baudouy, A. Siemko, D. Tommasini, R. van Weelderen

THERMOMAG 07 WS-Summary Rob van WeelderenNovember 29th, 2007

Goals of Thermomag-07Minimizing and evacuating heat is one of the main challenges for the next generation of superconducting magnets for high intensity particle accelerators such as the IR magnets for the LHC luminosity upgrade and the fast cycled magnets for FAIR, PS2, SPS+

The WS aims at reviewing the present knowledge on heat transfer in superconducting magnets and identifying a common thermal design basis

Identify the state of the art on

o Cooling techniques (fluids and regimes)

o Heat transfer mechanisms

o Modeling of heat transfer from coils to cooling system

o Heat transfer experiments

Identify a common set of thermal design criteria


Participation TOTAL 33

CERN 13 INFN 4 GSI 4 CEA 3 EPFL 3 Wroclaw Univ. 1 ENEA 1 KEK 1 Twente Univ. 1 JINR 1 EFDA 1

All researchers directly or indirectly working on the subject were present


Monday, 19 November 2007

09:30->14:00 Morning: Introduction and heat generation (B. Baudouy)

o 09:30 Welcome by Antoine DAËL (CEA)

o 09:40 Introduction to the workshop by Bertrand BAUDOUY (CEA)

o 09:50 Cryogenics for superconducting magnets by Luigi SERIO (CERN)

o 10:30 Thermal design criteria for various cooling schemes by Rob VAN WEELDEREN (CERN)

11:10 Break

o 11:30 Beam induced losses by Elena WILDNER (CERN)

o 12:00 Cable and magnet losses by Luca BOTTURA (CERN)

12:30 Lunch

14:00->18:00 Afternoon : Heat transfer (D. Tommasini)

o 14:00 Mechanisms of heat extraction through cable insulation by Bertrand BAUDOUY (CEA)

o 14:30 Cable in conduit and thermal budget at Nuclotron by Alexandre KOVALENKO (JINR, Dubna)

o 15:00 Nb3Sn versus NbTi in He II by Davide TOMMASINI (CERN)

o 15:30 Heat and mass transfer in superfluid helium through porous media by Hervé ALLAIN (CEA)

16:00 Break

o 16:30 Modeling of quench levels induced by steady state heat disposition by Dariusz BOCIAN (CERN)

o 17:00 Modeling of cable stability margin for transient perturbations by Pier Paolo GRANIERI

17:30 Discussion


Tuesday, 20 November 2007

09:00->14:00 Morning : experimental results (R. van Weelderen)

o 09:00 Transient Thermohydraulics measurement in cooling channels for the Iseult magnet by Philippe BREDY (CEA Saclay)

o 09:30 Design criteria for cable in conduit conductions in relation with expected disturbances by Jean-Luc DUCHATEAU (CEA)

o 10:00 First results of experiments at WUT by Maciej CHOROWSKI (WUT)

o 10:30 Experience at CEA (30') Jaroslaw POLINSKI (CEA Saclay)

11:00 Break

o 11:15 Experience at CERN by David RICHTER (CERN)

o 11:45 Experience at KEK (30') Nobuhiro KIMURA (KEK)

o 12:15 Diversity of heat transfer requirements for FAIR magnet applications by Marion KAUSCHE (GSI)

12:45 Lunch

14:00->18:30 Afternoon : round table & closeout 14:00 Round table (1h30') Andrzej SIEMKO (CERN)

15:30 Break 16:00 Close out (30') Davide TOMMASINI (CERN)


General considerations (1)The subject is not new, but is taking a NEW importance for projects like FAIR, and the LHC upgrades (injectors & IR)

Heat deposition modeling (E. Wildner/CERN, INFN, LARP)

“need improved feedback from magnet designers, and thermal simulation criteria”)

(IR-)Magnet structure cooling (R. van Weelderen/CERN, LARP)

“values for 50 W/m up to 100 W/m are still in the constructible range, major limit at coil to bath thermal pathway”

Fast ramp magnets for PS2, SPS+, FAIR (L. Bottura/CERN, M. Kausche/GSI, A. Kovalenko/JINR)

“A reasonable internal heat load target per unit length for future superconducting ring accelerator magnets is in the range of 5 W/m to 10 W/m”

“with hollow conductors 100 W/m is achievable”

CERN

7

The “debris-cone”

Triplet

23 m

Absorber

1.7 m19 m

IP

CERN

8

Simulation codes

Fluka "FLUKA: a multi-particle transport code",

A. Fasso`, A. Ferrari, J. Ranft, and P.R. Sala,CERN-2005-10 (2005), INFN/TC_05/11, SLAC-R-773

Geant Nuclear Instruments and Methods in Physics Research A 506

(2003) 250-303, and IEEE Transactions on Nuclear Science 53 No. 1 (2006) 270-278.

Mars Mokhov, N. V. The MARS code system user's guide.

Fermilab-FN-628, Fermi National Accelerator Laboratory (1995).

CERN

9

Results, what we consider Cable

We make the binning for the scoring so that it corresponds to a maximum volume of equilibrium for the heat transport (cable transverse size, with a length of around 10 cm, value to be confirmed)

Total power deposited in the magnets Important to know the volume of the magnet (the

model has to be realistic)

The power deposited per meter of magnet

N.B. For the total energy deposited we need a ”realistic” design of the magnet

R. van Weelderen, ThermoMag, Paris 19-20.11.2007 10

CLASSIFICATION OF HEAT EXTRACTION PATHS

Heat transfer

∆Tcoil: typically 80-90 mK available down from 2.17 K max∆Tcoil-freeA (radial): typically 60-70 mK available around 2.050 K∆TfreeA-bHX (longitudinal): typically 80-90 mK available around 1.98 Kabout 160 mK remains for heat transfer to cold source and up to cold compressors


Specific Conductive Cross section (cm2/[W/m m3/4]])

Specific Conductive Crossection cm2Wmm43; perWm, perconductionpath lenght43

40

60

80

DT for heat extraction mK1.85

1.9

1.95

2

TbathK0.6

0.8

1cm2Wmm4340

60

80

DT for heat extraction mK

Aspec is in the range of 0.3 to 1.1 cm2/[W/m m3/4]


CLASSIFICATION OF HEAT EXTRACTION PATHS

Example:

NED dipole/Q1 LHC inner triplet upgrade, 100 W/m, up to 5 m longitudinal heat extraction length, Tbath~1.935 K, ∆T~85 mK, Aspec ~0.55:

--> A~470 cm2 to be made in the yoke Assuming 15% of the cold mass volume taken up by

the coil, which is what needs to be condcuted our radially over ~ 0.05 m at Tbath~2.020 K, ∆T~65 mK, Aspec ~0.75:

--> A~0.21 cm2 to be provided in the collar & yoke laminations every 10 cm

Conclusion: values are still in the “constructable” range

The tri-lemma of the optimum pulsed superconducting cable design (PERITUS DELINEANDI OPTIMORUM DUCTORUM)

(courtesy of P. Bruzzone, ECOMAG-05)

AC Loss

Current Distribution

Heat Balance Protection and Stability

Pulsed Field

Conductor

… and cost !

Summary - 1/2

It is always beneficial to minimize AC loss, compatibly with protection, stability (transient heat balance) and current distribution

Summary - 2/2 The best compromise of AC loss, current distribution, heat

transfer, and cost can only be found in conjunction with the specific needs of the accelerator system and magnet design

A reasonable internal heat load target per unit length for future superconducting ring accelerator magnets is in the range of 5 W/m to 10 W/m Higher values are not economically interesting Lower values may bear too much complication in the cable

design

The above target may be largely exceeded, for specific applications and locations, and over short lengths


General considerations (2)

For Rutherford cables for a long time the only systematic experience on thermal transfer from cable to helium bath was in CEA, limited to 1.9 K, and some activity in KEK

Now there is very recent activity at KEK, CERN and Wroclaw University of Technology

Work on:

Ceramic insulating schemes (CEA)

Classical high-conductive schemes (RAL),

Highly porous kapton wrapping schemes (CERN, WUT)

B. Baudouy Mechanisms of heat extraction through cable insulation CARE-HHH Thermomag-07

Electrical insulation

▫Historical insulation : 2 wrappings ◦First wrapping in polyimide with 50% overlap

◦Second wrapping in epoxy resin-impregnated fiberglass with gap

▫The LHC insulation work : 2 wrappings◦First wrapping in polyimide with 50% overlap

◦Second wrapping in polyimide with polyimide glue with gap

▫Current LHC Insulation : 3 wrappings◦First 2 wrappings with no overlap

◦Last wrapping with a gap

▫Innovative insulation for Nb3Sn magnet

◦First wrapping 50% Courtesy of F. Rondeaux (CEA)Baudouy [1], Meuris [2] and Puigsegur [3]


NED : Innovative insulation

Courtesy of F. Rondeaux (CEA)

▫One wrapping with 50% overlap

▫Heat treatment of 100 h at 660 °C

▫10 MPa compression only !

▫5 conductors heated

LHCSSC

Increasing permeability

ΔT=5 mK @ 150 mW through 5 conductors

Baudouy [8]


NED : Conventional insulation

LHCSSC

Increasing permeability

ΔT=5 mK @ 150 mW through 5 conductors

▫Glass-fibre epoxy insulation developed by RAL

▫Determination of λ and Kapitza resistance◦Λ 4 times lower than kapton

◦Rkapitza identical

NED Ceramic insulation

NED conventional in

sulation

Canfer [7]

Baudouy [8]

Davide Tommasini Nb3Sn versus NbTi in HeII THERMOMAG-07, Paris 19 November 2007

Enhanced porosityEnhanced porosity


Tests Results/1Tests Results/1

Vertical compression 10 MPa


Heat transfer testsHeat transfer testsPower per cable edge (normalized to LHC inner layer ~ 2 mm) per meter of length

1.8

1.85

1.9

1.95

2

2.05

2.1

2.15

2.2

0 200 400 600 800 1000 1200 1400

Power [mW]

Tem

per

ature

[K

]

Enhanced (measurement D.Richter, to be published)LHC (measurement B.Baudouy Cryogenics 39, 1999)

SSC (measurement B.Baudouy Cryogenics 39, 1999)


General considerations (3)Activities in thermal modeling, needing experimental support and validation

•Heat and mass transfer through superfluid porous media (H. Allain/CEA, B. Baudouy/CEA)

“working experiment, modeling reasonable but in every regime still issues to be resolved”

•General magnet cooling, steady state and transient (M. Chorowski/WUT, R. van Weelderen/CERN)

“Based on ANSYS ICEM, CFX and dedicated Helium modules: Superfluid helium conduction module under development, code comparison with analytical & literature data has started”

•Supercritical Helium cooling modeling (FAIR)

•Modeling of quench levels by steady state beam loss heat load (D. Bocian/CERN, A. Siemko/CERN)

“Equivalent resistance network model: Validation of model with magnets at 4.5 K within 20%, validation of model at 1.9 K not completed”

•CICC stability (J. L. Duchateau/CEA)

“Stekly criterion not adequate for conductor design for fusion applications, the less copper, the higher the stability limit”




J.L Duchateau Thermomag November 2007 26

AssociationEuratom-CEA

The Stekly criterion in question

TF ITER conductor prototypemanufactured by Nexans

a non copper section Anoncu, a copper section Acu and an helium section AHe. In a project like

ITER the optimum composition of the conductor components is calculated through the so-called design criteria. The recent review of the ITER project has led to some interrogation about the systematic use of the Stekly criterion to calculate the copper section of the ITER PF NbTi coils.



The Stekly criterion in question

The Stekly criterion imposes that the copper section of the cable has to be adjusted such as

<1 to be in the so-called well-cooled region with being the Stekly parameter:

The Stekly criterion expresses that when the strands are taken at a temperature above T c by a

disturbance, the CICC can be stable and can recover by evacuating the power generated in copper as it is in communication through heat transfer with an infinite bath whose temperature is at T0 if the criterion is respected.

In practice in case of NbTi, the application of the Stekly criterion can lead to very high copper to non copper ratio increasing the price of conductor

)TP(ThAJρ

)TP(ThAJρα

0cc

2

0cc

cu2cucu



Conclusion

No, Stekly criterion is not adequate to design conductors for fusion application.

Contrary to Stekly criterion, it has been demonstrated that, for a given composite allocation, the stability limit in energy for a disturbance (100 ms, long length of CICC) is a decreasing function of the copper content :

The less copper, the highest the stability limit !

The particular role of He and comp is highlighted thanks to a simplified approach which demonstrates that the critical energy is essentially linked to the current sharing temperature in poor cooled regime. The crucial role of h is linked to these two parameters.

Copper is necessary for intrinsic dynamic stability but also for short (1 ms) mechanical disturbances applied to small length of CICC (1 cm).


General considerations

• THERMOMAG is the first workshop dedicated to heat transfer in superconducting accelerator magnets

• not yet a fully synergic activity between teams

• the subject is not new, but is taking a NEW importance for projects like FAIR, and the LHC upgrades (injectors & IR)

• for Rutherford cables for a long time the only systematic experience on thermal transfer from cable to helium bath was in CEA, limited to 1.9 K, and some activity in KEK

• very recent activity at KEK, CERN and Wroclaw University of Tec

• activities in modeling, needing experimental support and validation

• activities in development of new insulation schemes/materials


Thermal design criteriafor accelerator magnets

we believe the following principles can be used as guidelines• introduce thermal design in early stage • lattice magnets designed for < 10 W/m (includes ~5 for beam losses) to be economical• in HeII (50-100 mm aperture order)

• up to 50 -100W/m, for short (up to ~ 40m) magnet strings; hard limit basically cable insulation• insulation porosity dominates heat extraction from cable through insulation • below 9T temperature margin may not be an argument for Nb3Sn

• in supercritical helium or two phase (50-100 mm aperture order)• with Rutherford cables 30 W/m, however further limited by cable insulation• with hollow conductors 100 W/m are achievable • for heat loads > 2 W/m look for alternatives in Nb3Sn, or NbTi CICC, or need of specific studies


Desirable work• critically review and organize heat transfer experience & data

• characterize heat exchange in supercritical He

• fundamental experiments in narrow channels/voids in all regimes

• investigate role of porosity in supercritical helium

• investigate applicability and benefits of high thermal conduction insulation especially in supercritical He

• continue the effort in development of insulation schemes/materials

• continue and consolidate the effort in modeling

• Strengthen communication beam loss calculation teams with magnet and thermal designers

• design experiments to validate models in the different regimes

all this can be done ONLY by an efficient network of collaborations


Proposed initiatives• we define a “community list”

• web site on heat transfer with integrated database

• each laboratory writes a short report of on-going activities by Jan 31st

• this is circulated within the “community list” with feedback for an organized work with tentative list of “deliverables”

• by Feb 28th we agree on a proposal for “deliverables”

• we will then try, where applicable, to get formal agreement

• we issue a “status report” by June 30th 2008

• we meet again in autumn 2008

A dedicated coordinator could improve the effectiveness of the community

Documents

November 29 th 2007 Thermomag-07 A CARE-HHH workshop on Heat Generation and Transfer in Superconducting Magnet 19-21 November 2007 Paris WS-Summary based