CO2 Cooling for an ATLAS upgrade

1
CO2 Cooling for an ATLAS upgrade
(or the CO2 conspiracy)

• Plant requirements
• CO2 Properties
• CO2 Consequences of the properties
• CO2 From LHCb to ATLAS

2
Requirements for ATLAS

1. Cool many distributed heat sources spread over
   large volumes
2. Low material in the detector
3. Small temperature gradients over long distances
4. Radiation hard
5. Reliability

Will try to convince you that CO2 has it all!
3
CO2 properties p-H diagram
4
C3F8 properties p-H diagram
5
Run conditions _at_ -25C
C3F8 CO2
Pevaporation 1.7 bar 17 bar
?T for ?P-0.1bar 1.4 C / -1.5C 0.2 C / -0.2 C
?T for ?P-1.0bar 12 C / -20 C 1.8 C / -1.9 C
?H for evaporation 100 J/g 280 J/g
Flow for 100 W 1.0 g/sec 0.4 g/sec
Volume flow 0.6 cm3/sec 0.4 cm3/sec
Major difference for CO2 with respect to C3F8
cooling is the increase by a factor of 10 of the
evaporation pressure for T-25C.
6
High Pressure

• CO2 plant must be able to withstand about
  100bar (C3F8 in ATLAS 20bar)

low temperature gradients (dP/P small because P
is high) low tube diameter (because we can
allow large dP) low tube thickness (because of
low tube diameter) high tube flexibility
(reduce mechanical stress) - need more pipe at
your heat sinks - higher pressure higher tube
thickness
Some b.o.e. calculations to follow to get some
feeling for the numbers.
7
Back-of-the-envelope-calculation

• What design consequences would it have to replace
  C3F8 with CO2?
• Calculations assume CuNi pipes, just as for SCT
• Cool approximately 100W / cooling loop
• Disclaimer
• calculations need more refinement
• calculations must be supported by measurements

8
Pipe diameter pressure drop
?P pressure drop
Fm mass flow
? viscosity
R pipe radius
L pipe lenght
Pressure drop for a 1g/sec CO2 flow is of the
order of 0.08 bar. Even with a factor 10 more
flow temperature differences of a couple of
degrees can be expected!
9
Pipe diameter pipe thickness
d pipe diameter
T pipe wall thickness
S tensile strength
Pmax max pressure
With the same pressure as in the C3F8 system we
could reduce the pipe thickness by a factor four
A CO2 system has to withstand much higher
pressure, so we have to multiply the pipe
thickness again by a factor of six
T 0.12 mm
10
Material CuNi pipes for CO2 and C3F8
II CuNi for C3F8 at 20 bar
I CuNi for CO2 at 100 bar
11
Pipe flexibility stress
Low diameter piping is much more flexible goes
like R5. So no funny bends needed in your
structure to absorb mechanical stress.
12
Heat absorption
P required cooling power H heat transfer
coefficient in Wm-2K-1 (calculate/measure it
O(5000Wm-2K-1) d pipe diameter ?Tchange in
temperature
To absorb 10W you would need 4 cm of large
diameter CuNi pipe, versus 14 cm small diameter
pipe if CO2 is used.
Solutions RD - snaky cooling pipes at
cooling contact - large contact area with
module - or.
13
CO2 From LHCb to ATLAS

• Carbon copy the LHCb plant
• (by then) tested technology
• Cold input lines
• Relatively small return lines
• Design single-stage freezer
• Warm input lines
• Need oil-less CO2 compressor (!)
• Need relatively big return lines

14
CO2 plant compressor vs pump based
liquid
Pressure bar
2-phase
gas
Enthalpy kJ/kg
15
NIKHEF CO2 effort

• Interest from NIKHEF ATLAS group
• physicist (Fred Hartjes, Auke-Pieter Colijn, Els
  Koffeman)
• physicist/engineer (Bart Verlaat expertise)
• LHCb cooling work finishes in near future
• We plan to setup CO2 test-bed at NIKHEF
• for NIKHEF detector RD (Gossip)
• for ATLAS upgrade development support
  back-of-envelope calculations of this talk with
  measurements

16
(No Transcript)