Challenges and Limitations of Thermosiphon Cooling What We Learned from CMS…

Challenges and limitations of thermosiphon cooling What we learned from CMS…

Philippe Brédy CEA Saclay DSM/Dapnia/SACM/LCSE

Symposium for the Inauguration of the LHC Cryogenics, CERN, 31/05 & 01/06 2007 • The thermosiphon principle (example of two phase )

– Self sustained natural boiling convection • Heating applied (Heat to be removed) –Boiling – Weight unbalance between legs of a loop – Induced flow limited by friction

Single channel Loop

PhB, Symposium for the inauguration of the LHC Cryogenics, CERN, 31/05 & 01/06/2007 2 • Rising heat exchangers with various geometry • A mass flow and a flow quality deduced from geometry and heat load • An upper tank –thllhe well-namedhd phase separator- is needdded – to recover gas and to supply supply or condensation of gas (cryocooler) – for separation of the two phases

ΔP flow (m,x,geometry) Ù ΔP (x, hi)

Hyp: homogenous model (x << 10%)

PhB, Symposium for the inauguration of the LHC Cryogenics, CERN, 31/05 & 01/06/2007 3 Advantages vs Inconveniences

•A passive creation offlf flow • NdNeed a minimum hhheight – no pumps or regulation valves – Pressure head to create flow

• Autonomy in case of external • Circuit geometry must avoid cryogenic failure (volume of liquid any high point or strong in the phase separator) silitiingularities – separation of the phases • Minimization of the liquid –risk of vapor lock volume – use of the phase separator • No possible external action as a back -up volume for liquid –and a po ssible fru stra tion for the operator ! •A quasi-isothermal loop - mainly function of height • Pre-cooling before starting the ThS effect

PhB, Symposium for the inauguration of the LHC Cryogenics, CERN, 31/05 & 01/06/2007 4 Examples of applications on large superconducting magne ts w ith LH e/GH e loop at Tsat ATLASSCS( CS (LHC) … CMS (LHC)

ALEPH (LEP)

SMS G0 (JLAB)

CLOE (CORNELL) PhB, Symposium for the inauguration of the LHC Cryogenics, CERN, 31/05 & 01/06/2007 5 PANDA, R3B (GSI)… CMS

Phase separator cryostat

• 220 tons at 4.5 K • 174 to 500 W at 4, 5 K • 5 modules Coil cryostat • 86 parallel exchangers

X 2 X 5

PhB, Symposium for the inauguration of the LHC Cryogenics, CERN, 31/05 & 01/06/2007 6 CMS thermosiphon

Phase separator cryostat (892 l GHe+LHe)

First global calculation : x = 3,5 % ( 6,5% in SD) m = 200 g/s ( 400 g/s in SD) CiCryogenic chimney Thermosiphon loop (350 l)

5 LHe/GHe return pipes ∅ LHe supply pipe ∅ 60.3 x 1.6 48.3 x 1.6 Top Outlet manifolds (∅ 45 x 2. 5)

Heat

CB-2 CB-1 CB0 CB+1 CB+2 exchangers (∅i 14)

Inlet manifolds Bottom (∅ 45 x 2. 5)

PhB, Symposium for the inauguration of the LHC Cryogenics, CERN, 31/05 & 01/06/2007 7 CMS R&D

Height between LHe tank outflow and liquid level

Altitude 0 e

(3.04)

Upstream line (4.56) Venturi flowmeter

Collecting manifold (∅ 0.04)

Experimental test loop 7 heat exchanger scaling CMS tubes (5 .00 ) design

Hydrostatic head Distributing 9.22 manifold ( ∅ 0.04) (level at 50 %)

PhB, Symposium for the inauguration of the LHC Cryogenics, CERN, 31/05 & 01/06/2007 8 CMS R&D (scaling values) (on experimental test 200 180 loop ≈ 1/10 CMS) 160

140

/s) 120 100 • A strong mass 80

transfer in comparison Mass flow (g 60 Experimental and with heat deposit 40 20 homogenous model

• A large margin 0 • Homogenous model 0 50 100 150 200 250 300 350 400 450 0,14 Total deposited power Q (Watt) correct still above 14% 0,12 • h > 1600 W.m2.K 0,10 ity • ΔPr/Δps ∼ 1 ll 0,08 • ΔT height ~ 60 mK 0,06 • x < 3% (CMS nominal) Qua Vapor 0040,04

0,02

0,00 0 50 100 150 200 250 300 350 400 450 Total deposited power Q (Watt)

PhB, Symposium for the inauguration of the LHC Cryogenics, CERN, 31/05 & 01/06/2007 9 CMS R&D • Self-sustained circulation with the heat deposited

(on experimental test loop)

4.80 K 150 W total mass flow vs different heat loads 125 W 4.72 K

100 W

4.64 K 75 W

4.56 K

50 W 4.48 K temperatures 25 W 0 W 4.40 K

time

PhB, Symposium for the inauguration of the LHC Cryogenics, CERN, 31/05 & 01/06/2007 10 CMS on site • Self-sustained circulation with the heat deposited

(on site, at the beginning of a slow dump with dynamic heat load)

current

mass flow

PhB, Symposium for the inauguration of the LHC Cryogenics, CERN, 31/05 & 01/06/2007 11 CMS on site • Stabilisation and increase of the mass flow by heating the return lines

Effec t on the co il temperatures t - 0.01 K temperature decrease on coil

Mass flow Effect on the mass flow - Increasing and stabilization of mass flow rate

PhB, Symposium for the inauguration of the LHC Cryogenics, CERN, 31/05 & 01/06/2007 12 CMS R&D

• Smoothdbhlh sensitivity near and above the critical point

PhB, Symposium for the inauguration of the LHC Cryogenics, CERN, 31/05 & 01/06/2007 13 Limitations and Extensions - Circuit ggyeometry : - Feedinggp of several parallel - A minimum height circuits with different heat deposits is possible (in CMS - Risk of high point or design, in ratio of 5) singularities where gas could (feeding and collecting pipes must be be separated from the liquid designed to be as isobaric) - Minimization of pressure drop (singularities) - Star ting the na tura l circulation is achievable before the liquid presence (between 15 -A quasi-adiabatic supply line and 20 K for CMS design)

- Good separation phase in the - Heaters on the return pipes upper tank could be use d to lim it (what could be its minimum size?) instabilities due to gas/liquid separation at low velocity (low heat l oad or over-siiizing )

- Flow quality could be chosen well higher than the traditional limit of 5 % PhB, Symposium for the inauguration of the LHC Cryogenics, CERN, 31/05 & 01/06/2007 14 Conclusion

Confirmed by these tests and operation measurements, thermosiphon loop stays a convenitient way tito insure thiditthe indirect coo lifling of large equipment and must be taken into account during a design study without preconceived fears.

PhB, Symposium for the inauguration of the LHC Cryogenics, CERN, 31/05 & 01/06/2007 15 References

• 1. L. Benkheira, B. Baudouy, and M. Souhar, Heat transfer characteristics of two-phase He I (4.2 K) thermosiphon flow, International Journal of Heat and Mass Transfer, (2007) 50 3534-3544 • 2LBkhi2. L. Benkheira, BBdB. Baudouy, an dMShd M. Souhar, FlbiliFlow boiling reg imes an dCHFd CHF pre ditifdiction for HIHe I thermosiphon loop, Proceedings of the 21th International Cryogenic Engineering Conference, to be published, Ed. G. Gistau, (2006) • 3. P. Brédy, F.-P. Juster, B. Baudouy, L. Benkheira, and M. Cazanou, Experimental and Theoretical study of a two phase helioum high circulation loop, Advances in Cryogenics Engineering 51, AIP, Ed. J. G. W ei send , (2005) 496-503 • 4. L. Benkheira, M. Souhar, and B. Baudouy, Heat and mass transfer in nucleate boiling regime of He I in a natural circulation loop, Advances in Cryogenics Engineering 51, AIP, Ed. J. G. Weisend, (2005) 871-878 • 5. B. Baudouy, Heat transfer near critical condition in two-ppphase He I thermosiphon flow at low vapor quality, Advances in Cryogenic Engineering 49, AIP, Ed. S. Breon, (2003) 1107-1114 • 6. B. Baudouy, Pressure drop in two-phase He I natural circulation loop at low vapor quality, International Cryogenic Engineering Conference proceedings 19, Ed. P. S. G. Gistau Baguer, (2002) 817-820 • 7. B. Baudouy, Heat and mass transfer in two-phase He I thermosiphon flow, Advances in Cryogenic Engineering 47 B, AIP, Ed. S. Breon, (2001) 1514-1521

PhB, Symposium for the inauguration of the LHC Cryogenics, CERN, 31/05 & 01/06/2007 16