Ultralight carbon fiber structures evaporative tests Claudio BORTOLIN

Ultra-light carbon fiber structures: evaporative tests Claudio BORTOLIN (CERN) Martin DOUBEK (CTU, Czech Technical University, Prague) Andrea FRANCESCON (CERN) Manuel GOMEZ MARZOA (CERN) Romualdo SANTORO (CERN) 4 th September 2012 M. Gomez Marzoa ALICE Cooling Meeting - 4 th September 2012 1

Contents 1. Heater analysis § NTCs vs. thermographic picture analysis 2. Single-phase water tests: D 08 prototype 3. Evaporative tests: D 08 prototype § Comparison with water single-phase tests § Temperature distribution 4. Conclusion 5. Prototype and test facility optimization M. Gomez Marzoa ALICE Cooling Meeting - 4 th September 2012 2

Heater power distribution analysis § The D 06 prototype with single-phase water tests: presented at WG 4 Meeting the 27 th July 2012 Ø Two warmer regions were seen towards the centre of the stave at the sides. Ø Possible causes: q Lack of thermal contact plate-heater v Manufacturing difficulties 8 L min-1, 0. 5 W cm-2 v Gluing defects q Heater power dissipation maldistribution? § A single heater and the D 04 prototype heater will be powered up: Ø Check temperature distribution Ø Deviation of measurements thermocamera/NTCs M. Gomez Marzoa ALICE Cooling Meeting - 4 th September 2012 3

Heater power distribution analysis Case: Single heater I [A] V [V] P [W] 0. 15 4. 2 0. 63 3 0. 25 7. 4 11. 2 2 2 -0. 7 -0. 5 -1. 6 -2. 7 -0. 5 -7. 0 -1. 6 1. 3 1 3. 92 3 -0. 1 1 1. 85 3 0. 35 2 ΔT-3* ΔT-2* ΔT-1* [°C] 1 *ΔT-n = Average_T_NTC – Average_T_Thermo. Pic M. Gomez Marzoa ALICE Cooling Meeting - 4 th September 2012 4

Heater power distribution analysis Case: D 04 heater I [A] V [V] P [W] 0. 15 4. 4 0. 66 3 0. 25 7. 4 10. 6 2 2 -0. 4 -0. 5 -1. 8 -0. 4 -0. 7 -1. 6 -1. 0 -0. 8 1 3. 71 3 -2. 0 1 1. 85 3 0. 35 2 ΔT-3* ΔT-2* ΔT-1* [°C] 1 *ΔT-n = Average_T_NTC – Average_T_Thermo. Pic M. Gomez Marzoa ALICE Cooling Meeting - 4 th September 2012 5

D 08 prototype: description Pipe OD [mm] 1. 5 Pipe thickness [mm] 0. 035 Pipe ID [mm] 1. 43 Carbon paper sleeve thickness tcs [mm] 0. 03 CF tangential coverage β [deg] ~ 360 Pitch p+w [mm] 7. 5 Fiber width w [mm] 1. 5 p [mm] 6 Angle fibers with pipe axis α [deg] 23 IN OUT M. Gomez Marzoa ALICE Cooling Meeting - 4 th September 2012 6

D 08 prototype: water tests Case: D 08, 0. 3 W cm-2 Case: D 04, 0. 31 W cm-2 M. Gomez Marzoa Q Δp [L h-1] [bar] v [m s-1] TH 20 [°C] ΔTH 20 ΔTHeater [°C] 3. 0 0. 19 0. 52 15. 1 2. 4 9. 8 5. 0 0. 25 0. 86 14. 8 1. 5 9. 0 8. 0 0. 46 1. 38 14. 7 0. 7 8. 0 12. 0 0. 74 2. 08 14. 7 0. 6 6. 8 v [m s-1] TH 20 [°C] Q Δp [L h-1] [bar] ΔTH 20 ΔTHeater [°C] 3. 0 0. 13 0. 52 14. 7 2. 9 12. 2 4. 9 0. 22 0. 85 14. 7 1. 9 12. 0 8. 1 0. 38 1. 40 14. 6 1. 2 10. 8 12. 3 0. 76 2. 13 14. 5 0. 8 8. 0 ALICE Cooling Meeting - 4 th September 2012 7

D 08 prototype: water tests Case: D 08, 0. 5 W cm-2 Q Δp [L h-1] [bar] v [m s-1] TH 20 [°C] ΔTH 20 ΔTHeater [°C] 8. 0 0. 43 1. 38 14. 7 1. 5 16. 0 12. 0 0. 76 2. 08 14. 8 0. 6 13. 5 Temperature along stave: D 08, 8 L min-1, 0. 3 W cm-2 § Assuming same power density across stave, D 08 performs better than D 04 § Cannot cool at 0. 5 W cm-2 and needs optimization M. Gomez Marzoa ALICE Cooling Meeting - 4 th September 2012 8

Water tests: conclusion M. Gomez Marzoa ALICE Cooling Meeting - 4 th September 2012 9

D 08 2 -phase C 4 F 10 tests @DSF § Inlet vapor quality: § Superheating at stave outlet: T = const x = const p [bar] § Mass flow rate calculation: 1 where L is latent heat [k. J kg-1]: 3 2 § Usually: 3’ 4 Qstave [W] M. Gomez Marzoa h [k. J kg-1] ALICE Cooling Meeting - 4 th September 2012 10

D 08: water vs. C 4 F 10 @0. 3 W cm-2 Water C 4 F 10 Q Δp. St v -1 [L h ] [bar] [m s-1] TH 20 [°C] ΔTHeater [°C] 3. 0 0. 19 0. 52 15. 1 2. 4 9. 8 5. 0 0. 25 0. 86 14. 8 1. 5 9. 0 8. 0 0. 46 1. 38 14. 7 0. 7 8. 0 12. 0 0. 74 2. 08 14. 7 0. 6 6. 8 x. In [m s-1] x. Out [m s-1] TC 4 F 10 m Δp. St [g s-1] [bar] -Out [°C] ΔTHeater [°C] 0. 16 0. 08 0. 92 16. 8 8. 0 0. 20 0. 07 0. 08 0. 75 14. 0 5. 5 0. 40 0. 06 0. 08 0. 42 13. 4 5. 6 0. 60 0. 2 0. 06 0. 31 13. 4 6. 0 Evaporative cooling system performs as good as single-phase water M. Gomez Marzoa ALICE Cooling Meeting - 4 th September 2012 11

D 08: water vs. C 4 F 10 @0. 5 W cm-2 Water C 4 F 10 Q [L h-1] Δp. St [bar] v [m s-1] TH 20 [°C] ΔTHeater [°C] 8. 0 0. 43 1. 38 14. 7 1. 5 16. 0 12. 0 0. 76 2. 08 14. 8 0. 6 13. 5 m [g s-1] Δp. St [bar] x. In [m s-1] x. Out [m s-1] ΔTSH [°C] 0. 4 0. 17 0. 06 0. 65 13. 4 13. 0 0. 6 0. 26 0. 05 0. 46 13. 4 14. 0 0. 8 0. 33 0. 03 0. 36 13. 4 14. 5 ΔTHeater [°C] § Good selection of mass flow rates and agreement between thermographic pictures and NTCs over the heater. § Knowing the vapor quality at the outlet is very important. M. Gomez Marzoa ALICE Cooling Meeting - 4 th September 2012 12

D 08: C 4 F 10 tests discussion Ø Two cases did not perform as expected: Case: 0. 3 W cm-2 m [g s-1] Δp. St [bar] x. In [m s-1] x. Out [m s-1] TC 4 F 10 - 0. 8 0. 28 0. 04 0. 26 13. 3 Out [°C] ΔTHeater [°C] 14. 0 § Low vapor quality at the stave entrance: saturated liquid entering stave? § Low vapor quality at stave outlet: single phase flow? Case: 0. 5 W cm-2 m [g s-1] Δp. St [bar] x. In [m s-1] x. Out [m s-1] TC 4 F 10 - 0. 2 0. 09 0. 08 1. 20 21 Out [°C] ΔTHeater [°C] 28. 0 § Low vapor quality at the stave entrance: saturated liquid entering stave? § Mass flow rate too low: superheated vapor at stave outlet M. Gomez Marzoa ALICE Cooling Meeting - 4 th September 2012 13

Conclusion § Almost the same cooling performance is achieved with single-phase water cooling circuit as when using evaporative C 4 F 10 for the same prototype. 1. There is not a big increase of the HTC wall-fluid using evaporative C 4 F 10 Heat Transfer Coefficient [W/m^2 K] Ø C 4 F 10, two-phase: 10000 9000 8000 7000 6000 5000 4000 3000 2000 1000 0 Ø Water, single phase: Q [l/h] V [m s-1 ] Shah (1982) Gungor-Winterton (1986) Re [-] HTC [W m-1 K-1] 3. 00 0. 52 653 1650 12. 00 2. 08 2612 8076 Evaporative C 4 F 10 means 0 0. 1 0. 2 0. 3 0. 4 0. 5 0. 6 0. 7 0. 8 0. 9 1 Vapor Quality [-] 2. ΔT wall-water: through the HTC, establishes the margin of improvement by using a better cooling system for this setup: CFD Simulations M. Gomez Marzoa HTC wall-fluid [W m-2 K-1] Tmax Silicon [o. C] 1646 43. 02 5000 39. 25 10000 38. 22 ALICE Cooling Meeting - 4 th September 2012 14

Optimization lines 1. Stave optimization: § Pipe inner diameter: can be smaller than 1. 5 mm (but less contact area!) § More rigid piping: PEEK (avoid deformations, pinching, ensure contact) § D 08 prototype shows no better thermal performance with evaporative flow Ø Improve weak parts of model (thermal contact, gluing…) Ø Structure thermal analysis/simulation helpful § Avoid connectors: leaks, extra pressure drop. Ø Proposal: single pipe w/ 180 deg elbow. Ø In/Out connector: select useful pipe diameter. 2. Setup optimization: § A by-pass will be added to the circuit in DSF in order to be able to work with smaller mass flow rates (especially microchannel) Ø For this reason, a coriolis flow meter will be moved in DSF Ø Need for subcooled liquid before the flow meter! § Sensors calibration (see backup slide). M. Gomez Marzoa ALICE Cooling Meeting - 4 th September 2012 15

Ultra-light carbon fiber structures: evaporative tests Claudio BORTOLIN (CERN) Martin DOUBEK (CTU, Czech Technical University, Prague) Andrea FRANCESCON (CERN) Manuel GOMEZ MARZOA (CERN) Romualdo SANTORO (CERN) 4 th September 2012 M. Gomez Marzoa ALICE Cooling Meeting - 4 th September 2012 16

Backup D 08: C 4 F 10 tests discussion Q Pd Heater T 1 P 1 Subcool_1 p 2 ∆p. Lam Tsat-p 2 xin-stave T 3 [g s-1] [W cm-2] [o. C] [bar] 0. 16 0. 2 0. 4 0. 6 0. 8 [o. C] [bar] [°C] [-] p 3 ∆p. St T 3' Error@3 Superheating_3 [o. C] [bar] [°C] h. Out xout-stave [ [°C] [k. J kg-1] -] 0. 30 20. 2 3. 07 9. 1 1. 80 1. 27 13. 2 0. 08 16. 8 1. 74 0. 06 12. 2 - - 99. 37 0. 92 0. 30 20. 4 3. 10 9. 2 1. 82 1. 28 13. 5 0. 08 14. 0 1. 75 0. 07 12. 4 1. 6 - 84. 40 0. 75 0. 50 20. 4 3. 10 9. 2 1. 84 1. 26 13. 8 0. 08 21. 0 1. 75 0. 09 12. 4 - 8. 6 125. 74 1. 20 0. 30 20. 3 3. 08 9. 1 1. 81 1. 27 13. 3 0. 08 13. 4 1. 75 0. 06 12. 4 1. 0 - 53. 94 0. 42 0. 50 20. 4 3. 06 8. 8 1. 93 1. 13 15. 1 0. 06 13. 4 1. 76 0. 17 12. 5 0. 9 - 74. 71 0. 65 0. 30 20. 3 2. 87 6. 8 1. 96 0. 91 15. 6 0. 06 13. 4 1. 76 0. 2 12. 5 0. 9 - 43. 81 0. 31 0. 50 20. 3 2. 85 6. 6 2. 02 0. 83 16. 5 0. 05 13. 3 1. 76 0. 26 12. 5 0. 8 - 57. 59 0. 46 0. 30 20. 3 2. 79 5. 9 2. 03 0. 76 16. 6 0. 04 13. 3 1. 75 0. 28 12. 4 0. 9 - 38. 75 0. 26 0. 50 20. 3 2. 78 5. 8 2. 11 0. 67 17. 7 0. 03 13. 3 1. 78 0. 33 12. 8 0. 5 - 49. 08 0. 36 Where; § Subcooling = TSAT@p 1 – T 1 (entrance of stave). § T 3’: saturation temperature at p=p 3. Used to calculate superheating at the stave outlet (if superheated vapor present). § Error in temperature measurement at point 3: calculated as ε = T 3 -T 3’ M. Gomez Marzoa ALICE Cooling Meeting - 4 th September 2012 17

Backup An estimation of the uncertainty of measurements: Q Pd Heater T 1 P 1 Subcool_1 p 2 Tsat-p 2 T 3 [g s-1] [W cm-2] [o. C] [bar] 0. 16 0. 2 0. 4 0. 6 0. 8 0. 30 20. 2 3. 07 [o. C] [bar] [°C] h. Out xout-stave h 3 Lsat x. Sat-p 3 T 3' Error@3 -1] [k. J kg-1] [k. J kg [-] [o. C] [bar] [°C] [k. J kg-1] [-] 9. 1 1. 80 16. 8 13. 2 p 3 ∆p. St 1. 74 0. 06 111 14. 92 107 1. 04 12. 2 - 99. 37 0. 92 15. 09 107 1. 01 12. 4 1. 6 84. 40 0. 75 0. 30 20. 4 3. 10 9. 2 1. 82 13. 5 14. 0 1. 75 0. 07 108 0. 50 20. 4 3. 10 9. 2 1. 84 13. 8 21. 0 1. 75 0. 09 114 15. 09 107 1. 08 12. 4 - 125. 74 1. 20 0. 30 20. 3 3. 08 9. 1 1. 81 13. 3 13. 4 1. 75 0. 06 108 15. 09 107 1. 01 12. 4 1. 0 53. 94 0. 42 0. 50 20. 4 3. 06 8. 8 1. 93 15. 1 13. 4 1. 76 0. 17 108 15. 26 107 1. 01 12. 5 0. 9 74. 71 0. 65 0. 30 20. 3 2. 87 6. 8 1. 96 15. 6 13. 4 1. 76 108 15. 26 107 1. 01 12. 5 0. 9 43. 81 0. 31 0. 50 20. 3 2. 85 6. 6 2. 02 16. 5 13. 3 1. 76 0. 26 108 15. 26 107 1. 01 12. 5 0. 8 57. 59 0. 46 0. 30 20. 3 2. 79 5. 9 2. 03 16. 6 13. 3 1. 75 0. 28 108 15. 09 107 1. 01 12. 4 0. 9 38. 75 0. 26 0. 50 20. 3 2. 78 5. 8 2. 11 17. 7 13. 3 1. 78 0. 33 108 15. 60 107 1. 00 12. 8 0. 5 49. 08 0. 36 0. 2 1. At point 3, calculate h for saturated liquid and vapor using p 3. With p 3 and T 3, the point is superheated vapor and h 3 can be calculated. If temperature measurement was fine, , and: In the real case, x 3 > 1. The deviation is the % of total error resulting frpm measuring p, T and calculating the enthalpies (Ref. Prop). M. Gomez Marzoa ALICE Cooling Meeting - 4 th September 2012 18

Backup An estimation of the uncertainty of measurements: Q Pd Heater T 1 P 1 Subcool_1 p 2 Tsat-p 2 T 3 [g s-1] [W cm-2] [o. C] [bar] 0. 16 0. 2 0. 4 0. 6 0. 8 0. 30 20. 2 3. 07 [o. C] [bar] [°C] h. Out xout-stave h 3 Lsat x. Sat-p 3 T 3' Error@3 -1] [k. J kg-1] [k. J kg [-] [o. C] [bar] [°C] [k. J kg-1] [-] 9. 1 1. 80 16. 8 13. 2 p 3 ∆p. St 1. 74 0. 06 111 14. 92 107 1. 04 12. 2 - 99. 37 0. 92 15. 09 107 1. 01 12. 4 1. 6 84. 40 0. 75 0. 30 20. 4 3. 10 9. 2 1. 82 13. 5 14. 0 1. 75 0. 07 108 0. 50 20. 4 3. 10 9. 2 1. 84 13. 8 21. 0 1. 75 0. 09 114 15. 09 107 1. 08 12. 4 - 125. 74 1. 20 0. 30 20. 3 3. 08 9. 1 1. 81 13. 3 13. 4 1. 75 0. 06 108 15. 09 107 1. 01 12. 4 1. 0 53. 94 0. 42 0. 50 20. 4 3. 06 8. 8 1. 93 15. 1 13. 4 1. 76 0. 17 108 15. 26 107 1. 01 12. 5 0. 9 74. 71 0. 65 0. 30 20. 3 2. 87 6. 8 1. 96 15. 6 13. 4 1. 76 108 15. 26 107 1. 01 12. 5 0. 9 43. 81 0. 31 0. 50 20. 3 2. 85 6. 6 2. 02 16. 5 13. 3 1. 76 0. 26 108 15. 26 107 1. 01 12. 5 0. 8 57. 59 0. 46 0. 30 20. 3 2. 79 5. 9 2. 03 16. 6 13. 3 1. 75 0. 28 108 15. 09 107 1. 01 12. 4 0. 9 38. 75 0. 26 0. 50 20. 3 2. 78 5. 8 2. 11 17. 7 13. 3 1. 78 0. 33 108 15. 60 107 1. 00 12. 8 0. 5 49. 08 0. 36 0. 2 2. At point 3, vapor quality at the stave outlet (calculated using an energy balance) indicates that the fluid is in the two-phase region. If that is the case, then: However, a is read instead. The difference remains stable for most of the cases: § Calibration systematic error? § Incorrect setting of temperature sensors? M. Gomez Marzoa ALICE Cooling Meeting - 4 th September 2012 19