concentric vacuum insulated transfer lines: …
TRANSCRIPT
CONCENTRIC VACUUM INSULATED TRANSFER
LINES: FEEDBACK FROM ATLAS AND CMS
NEWLY INSTALLED SYSTEMS
P. Tropea, N. Bacchetta, T. Blaszczyk, C. Bortolin, J. Daguin, T. Eisel, J. Godlewski,
M. Gutt-Mostowy, P. Petagna, H. Postema, B. Verlaat, L. Zwalinski
30 June 2014- Forum on Tracking Detector Mechanics - DESY
OU
TLIN
E
Outline
30 June 2014
2 P. Tropea
Concentric transfer lines for CO2 cooling
• Why?
• How?
Insulation
Common approach: ATLAS & CMS choices
Overview of ATLAS Transfer lines
• Rigid transfer lines
• Flexible transfer lines
CMS Transfer lines
• On detector
• Off detector
Common issues:
• Design
• Installation
• Testing
Conclusions
2P
AC
L B
AS
ICS
2PACL: 2-Phase Accumulator Controlled Loop
3
Pump
1 4
5
3 2
6
7
Gas
Pre
ssu
re
Liquid
Isothermal line
Enthalpy
2-phase (Evaporation)
Evaporation in
detector
3
4 5 6
2
1 7
Heat
Plant
Detector
Heat exchanging
transfer line
CO2 evaporative cooling:
• Atlas IBL
• CMS Pixel Phase I
30 June 2014
P. Tropea
TR
AN
SF
ER
LIN
ES F
OR
CO
2 2
PA
CL
: INT
RO
Why do we need saturated liquid?
30 June 2014
4 P. Tropea
Gas
Pre
ssure Liquid
Isothermal line
Enthalpy
2-phase(Evaporation)
Tliquid
T2phase Vapor return
P control
Heat exchange at the plant
Pipe length
Gas
Pre
ssure Liquid
Isothermal line
Enthalpy
2-phase(Evaporation)
Tliquid
T2phase
Vapor return
P control
Heat exchanger near detector
Detector inlet is saturated liquid
Pipe length
Plant
Detector
Transfer line Heat
exchanger
Heat
exchanger
Detector inlet can be 2-Phase
(bad flow distribution)
Plant
Detector Transfer line
OP
TIM
IZIN
G T
HE
RM
AL E
XC
HA
NG
E
The possible solutions
30 June 2014
5 P. Tropea
Technology Pro Cons
Local heat
exchanger at the
detector inlet
Local control Bulky, heavy, not accessible
Common insulation
for inlet and return
Can be applied on
existing pipes
Distribution of pipes in the
bundle strongly influences the
performances, heat from
environment can still reach
inlet, bulky
Concentric: inlet into
return
Compact, full control of
liquid no matter heat
sources around
New installation (for CMS)
Concentric
insulated tube Bundled
insulated
tubes
TH
ER
MA
L SH
IELD
ING
I
Foam insulation?
30 June 2014
6 P. Tropea
LHCb-Velo system
Armaflex: k=0.04 W/m*K Dries out and looses flexibility, shall be carefully glued
Aerogel Spaceloft 9251: k=0.013 W/m*K Dusty and difficult material to handle, but good performance if super well protected from humidity. All parts enclosed in polyethylene bags, sealed, then adhesive vapor barrier around: QA during installation shall be super strict…
CMS Tracker system
TH
ER
MA
L S
HIE
LD
ING
II
What about vacuum insulation?
30 June 2014
7 P. Tropea
CMS • 0.4 watt/m for -196’C for a vacuum jacket of 63mm (25.5mm
insulation thickness)
• 6 watt with vacuum compared to 87 Watt with foam (but vacuum
reference at -196’C)
• YB0 lines will have a reduced diameter (28mm)
Atlas IBL • 0.55 Watt/m for -196’C for a vacuum jacket of 63mm
(32mm insulation thickness)
• 55 watt with vacuum compared to 800 Watt with foam (but vacuum
reference at -196’C)
Vacuum insulation performs a magnitude better than
foam, is more reliable and is smaller in diameter
0 10 20 30 40 50 60 706
8
10
12
14
16
18
20
22
59
73
87
101
115
129
143
157
171
Insulation thickness (mm)
Surf
ace t
em
pera
ture
(ºC
)
T
f luid=-20 ºC, T
amb=22 ºC, Ki=0.04 W/mK, Length=15 m, Angle=0º
Heat
pic
k-u
p (
W)
Surface temperature
Heat pick up
12mmOD, t=1mm
0 10 20 30 40 50 60 70-5
0
5
10
15
20
25
-2
332
666
1000
1334
1668
2002
Insulation thickness (mm)
Surf
ace t
em
pera
ture
(ºC
)
T
f luid=-40 ºC, T
amb=22 ºC, Ki=0.04 W/mK, Length=85 m, Angle=0º
Heat
pic
k-u
p (
W)
Surface temperature
Heat pick up
0.625" OD, t=0.049"
0.75" OD, t=0.049"
800W at 32mm
CMS
Atlas IBL
TH
ER
MA
L SH
IELD
ING
III 30 June 2014
8 P. Tropea
Summary of pro & cons
Foam insulation: Vacuum insulation:
Advantages
Disadvantages
In principle cheaper... Expensive
Large heat leak: allowing high T set points with no
additional load
Small heat leak: ask for additional load in order to
operate at high T
Disadvantages
Advantages
Risk of leaking seals (Water condensation) No seals
Large diameter (ex: -30 C /12mm pipe +2x32 mm
insulation =76mm external envelope)
Small diameter (ex: -40 C/12 mm pipe - 28mm external
envelope for vacuum jacket)
Hard to apply in crowded areas, access needed
everywhere
Only access needed at welds
More heat leak (larger primary chiller)
Small heat leak (smaller primary chiller)
TH
E D
IFF
ER
EN
T C
HO
ICE
S
A common approach with different boundaries
• CO2 @ -40 C
• Main transfer line an industrial
cryogenic solution
• Distribution to IBL an in-house
development as only flexible lines
were possible.
• Flex lines routing is as cabling
• Flex lines are actively pumped
ATLAS
• All paths concentric
• All passive vacuum insulated
along main routing
• An industrial approach by
cryogenic transfer line technology
• Customized design inside the
detector for reduced diameter
30 June 2014
9 P. Tropea
CMS
• CO2 @ -20 (but design of
insulation for -40)
• 3 separate paths, different sizes
and granularity
• all vacuum insulated, passive
CO
NC
EP
T
ATLAS IBL Transfer lines
30 June 2014
10 P. Tropea
Vacuum insulated
concentric transfer tube
Concentric split & end
of vacuum shield
Vacuum insulated
Concentric flexible
detector loops
LAR station
To plants in
USA-15
cavern
Junction box containing valves,
sensors and dummy load
Manifold box
Vacuum system for
flex line insulation
Vacuum line
Muon
In Atlas 2 technologies:
1 Concentric main transfer line (21.3x2.11mm) in a 63.5mm stationary vacuum tube (100m)
14x Flexible concentric detector loops (1.6x0.3mm) in a 17mm flexible bellow tube (11m)
TR
AN
SF
ER
LIN
E S
OLU
TIO
NS I
ATLAS IBL Main Transfer line
30 June 2014
11 P. Tropea
~80m transfer line
(Vacuum, 63.5 mm)
Detector
Cooling plant in USA15 MLI
10mmOD / 8mmID
concentric CO2
liquid tube.
21.3mmOD /
17.08mmID CO2
vapor return tube
50mmOD / 47mmID Vacuum
tube for 12m section
63.5mmOD / 60.5mmID Vacuum
tube for 80m section
Industrial solution (Demaco)
TR
AN
SF
ER
LIN
E S
OLU
TIO
NS II
ATLAS IBL flexible detector loops
30 June 2014
12 P. Tropea
• The routing towards the IBL has limited space in the Inner detector end plate.
• Need to be like cabling in order to fit.
• Routing through inaccessible areas
• Only 1 solution: Flexible and Diameter < 18mm = Vacuum
Flexible vacuum insulated
detector loops
Purpose of the concentric flexible detector loop.
1. Prevent unwanted boiling of the flow distribution capillary
2. Make it flexible as a cable for integration purposes
3. Avoid condensation on outer surface
SIZ
ING
OF F
LE
X L
INE
S
Concentric flex line considerations
30 June 2014
13 P. Tropea
0 5 10 15 20 25-2
0
2
4
6
8
10
12
14
16
IBL temperature and pressure profile. MF=0.8g/s, Tsp=-40ºC, Q=101.8, xend
=0.41
Branch length (m)
Delta T
(`C
) &
Delta P
(Bar)
1 2 3 4 56 78 9 10 11 12
Stave TFoM: 13ºC*cm2/W
Pixel maximum temperature:
-24.4ºC
dT Tube wall (ºC)
dT Fluid (ºC)
dP Fluid (Bar)
dT Pixel Chip (ºC)
0 5 10 15 20 25-2
0
2
4
6
8
10
12
14
16
IBL temperature and pressure profile. MF=0.8g/s, Tsp=-40ºC, Q=101.8, xend
=0.5
Branch length (m)
Delta T
(`C
) &
Delta P
(Bar)
1 2 3 4 56 78 9 10 11 12
Stave TFoM: 13ºC*cm2/W
Pixel maximum temperature:
-24.2ºC
dT Tube wall (ºC)
dT Fluid (ºC)
dP Fluid (Bar)
dT Pixel Chip (ºC)
Capillary
No Boiling
Boiling starts
Boiling starts
Simulation without concentric heat exchange Simulation with concentric heat exchange
Outer sleeve -
spacer
10 layers MLI
Ø4x0.5mm
Stainless steel
tube
Ø16mm ID Outer
flexible corrugated
hose Vacuum
interspace
Ø1.6x0.3mm Stainless
steel capillary
Flex line dimensions: • A Ø1.6x0.3mm flow distribution
capillary inside a Ø4x0.5mm tube.
• Length 11m long
• Vacuum hose Ø16mm bellow hose
• MLI inside glass fibre sleeve
• A 4mm stainless steel tube is considered as flexible
Concentric
capillary
AT
LA
S F
LE
X L
INE
S I
Flex line design
30 June 2014
14 P. Tropea
A-A ( 1 : 1 )
B-B ( 10 : 1 )
A
A
1.4301 X5CrNi18-10112-040101.iptEnd of the Flex Line Part 111
Fa.Witzenmann RS341S00-UA22S DA12 RS341S00-UA22S-DN12.iamFextube12
novotek 1522-3SO1.4301 X5CrNi18-101522-3SO.iptKF M. ROHRANS.304 KZ,
NW16 L=30 D=17,2 / D =14
13
1.4301 X5CrNi18-10112-040102.iptEnd of the Flex Line Part 214
1.4301 X5CrNi18-10112-040103.iptpipe elbow 315
Inner tubes.iamInnertubes17
1.4301 X5CrNi18-10VA-Rohr DA 1.6-DI 1.iptpipe 1.617.1
1.4541 X6CrNiTi18-10VA-Rohr DA 4-DI 3.iptpipe 317.2
super insulst ionsuper insulat ion.iptsuper insulst ion17.3
St cklisteBemerkungNormAbmessungenWerkstoffZeichn. Nr.BenennungSt .Teil
B
B
DIN EN 20 216 - A2 (420 x 594)
Hauptprojekt ion Gewicht .........: -
mmDimensionen :
Ma e ohne Toleranzangabe nach
DIN ISO 2768 m K
Werkstoff
Zeichnungsnummer / EDV Nr.:
Software......: Inventor 2011
Blat t : Gesamtzahl: 1 V12.b1
Teil:
Flex Line
Flex Line
Flex Line
CO2-ATLAS-BELLEProjekt
Max-Planck-Institut f r Physik(Werner-Heisenberg-Institut)
gezeichnet
geprueft
geplot tet
Tag Name
31.01.2014 S.Vogt
M nchen
+0,1 -0,1
Ma stab
112-040100.idw
1:1
11000mm
51 42
7.1
3
st itch up the 10 layers Super- insulat ion
144,8 mm + 105mm = ca.250
144,9
orbital welding 3
orbital welding 3orbital welding 3
orbital welding 17,2 orbital welding 17,2 orbital welding 6
35
28
6
R10
19
44
59
KF16
17,2
10
55
12,5
16,5 14 17,2
30
96
10
orbital welding 4
1.6mm braze joint
Presure seal welded for:
test pressure: 160bart ightness with 100 bar over 48h: nomodificat ions to the pressure gauge Welde cleanliness: inside with
protect ive gas / outside brushed orbital welding TIG where possible
7.2
7.3
Nut Gasket
Sealing
glands
Rotatable joint to align connection to manifold box
Flex line dimension
Liquid CO2 in 2-phase CO2 out
Vacuum connection
Special reversed VCR to
disassemble concentric joints
(outside of VCR pressurized)
AT
LA
S F
LE
X L
INE
S II
Laboratory testing
30 June 2014
15 P. Tropea
Goal of tests: see if surface temperature is above dew-point
Test set-up with coiled hose (Maximum wall contact) 4mm tube with MLI and glass fibre as
continuous thermal spacer
Surface temperature
measurements to spot cold
contact bridges
AT
LA
S F
LE
X L
INE
S III
Flex line production
30 June 2014
16 P. Tropea
Detector side splitting CO2 tubes with MLI
Prepared flexlines Ready for installation
Concentric weld
IAT
LA
S F
LE
X L
INE
S IV
Flex lines installed in Atlas (June 2014)
30 June 2014
17 P. Tropea
Flex line manifold in PP2 sector 5 Flex line installation Flex line terminal box near IBL
Almost identical: cooling lines and cables
Extra length storage
INT
EG
RA
TIO
N
CMS Pix Phase I upgrade CO2 cooling
30 June 2014
18 P. Tropea
UXC 55 USC 55
Manifolds
Cooling plant cores and accumulators
YB0
R507 Chiller
1
2
3
4
2 plants, 15 kW each, -20 C nominal
1) Preliminary layout done, to be installed with detector in YETS 2016
2) Installed in Jan 2014
3) In construction, installation August 2014
4) Design on-going, installation September 2014
TE
CH
NIC
AL S
OLU
TIO
N
CMS vacuum insulated transfer lines
30 June 2014
19 P. Tropea
Plant
BPix
Plant
Fpix
Manifold
Bpix
Manifold
FPix
USC55 UXC55
2x
concentric
transfer
lines
4x4 concentric
cooling loops,
common vacuum
jacket
4x4 concentric
cooling loops Vacuum jacket
max 80 mm OD
Liquid line
(12 mm x1)
2-phase
return
(33 mm x3)
Vacuum jacket
(max 80 mm OD)
Liquid line
(6 mm x1)
2-phase return
(12 mm x1)
Vacuum jacket
28 mm OD
Liquid line
(6 mm x1)
2-phase return
(12 mm x1)
FROM
CALCULATIONS
& TESTS
SIZ
ING
EX
AM
PLE: M
INIM
IZIN
G R
ET
UR
N Δ
P
CMS on detector transfer line sizing
20 P. Tropea
0 5 10 15 20 25 3019.7
19.75
19.8
19.85
19.9
19.95
20
20.05
20.1
20.15
Length [m]
Pre
ssure
[bar]
L = 30.0m | d = 8.00mm | m = 10.0g/s | 50% Vapor | Vmax
= 2.02m/s | DPmax
= 0.45bar
Friedel
Gronnerud
Chisholm
Lockhart & Martinelli
Bankoff
Muller-Steinhagen & Heck
30m line results into
ΔPmax = 0.45bar ≈ 1°C
• The CO2 transfer line regulates the detector
inlet liquid temperature via internal heat
exchange with the detector’s returning flow.
MINIMIZING RETURN Δp IS THE GOAL
• Max required mass flow up to 10g/s at main
transfer line.
• 2-Phase pressure drop models indicate loss
of 1°C for max flow rate
FROM
INTEGRATION
ISSUES
& CALCULATIONS
of SUPPLIER
Vacuum jacket
28 mm OD
Liquid line
(6 mm x1)
2 phase return
(12 mm x1)
J. Noite
CM
S Y
B0 T
RA
NS
FE
R L
INE
S
30 June 2014
PP1 +14 Connection box
Vacuum ports
Sector -6
P. Tropea 21
This was great! Rather do differently…
3D model almost perfect on YB0 Verification with mock-up for zones
outside detector
Swagelok VCRs: not a single leak Small size vacuum ports in difficult
locations
2 routings not needing on-site welds 2 routings needing on-site welds, included
vacuum jacket…
Welding QA Preparation of supports in advance
PP1 by-passes
Installed in January
2014, pressure and
leak tested
YB
0 T
O M
AN
IFO
LD
TR
AN
SF
ER
LIN
ES
22
30 June 2014
P. Tropea
Construction on-
going
Installation mid-
August
…so we changed
Full scale mock-up with PVC pipes to
update 3D model
Sections coming with pre-vacuumed
jackets, only small volumes to be pumped
on-site
High pressure leak testing test bench for
pressure decay method being built
T. Pakulski, N. Smiljkovic
UX
C55
/US
C55
– T
R. L
INE
S
FR
OM
PLA
NT T
O M
AN
IFO
LD
30 June 2014
23 Nicolas Szilasi (Louvain), CMS E&I office
Order placed last week
Installation mid-
September
1
3
2
LE
SS
ON
S L
EA
RN
T
Integration
30 June 2014
24 P. Tropea
Mock-ups or 3D model or both?
• 3D model very precise in the proximity of the detector, looser
precision at the interface cavern/detector in some cases
• Mock up with PVC piping super useful also for verification of
handling procedures
Passive or active vacuum?
All parts installed so far in “big sizes”, coming from industrial
standards, passive vacuum
• 10 years theoretical guarantee, periodic check recommended
• Better not to rely on vacuuming on site, use bake-out and pre-
vacuuming whenever possible
• No active components, no interlock
Atlas flex lines with active vacuum
• Constant performance monitoring
• Smaller envelope
LE
SS
ON
S L
EA
RN
T
Pressure and leak testing on site
25
CO2 transfer lines are designed for high service pressure. CMS PS =
110 bar, ATLAS PS = 110 bar.
- PED: 1.43*PS = testing pressure
- CERN safety: pressure tests to be executed when pipework in their
final destination
- Gas test: cavern evacuation (noise issues)
- Specific equipment needed (pressure reducer and gauges)
30 June 2014
P. Tropea
Leak detection
- He in sniffer mode (but test in
vacuum rather than in pressure:
VCR connectors not loaded
properly!) – not always accepted
- Pressure decay: time consuming,
need proper test bench
J. Daguin, N. Frank
Conclusions
30 June 2014
26 P. Tropea
Both Atlas and CMS have invested in vacuum insulated
transfer lines for their CO2 cooling systems
- Increased insulation capacity wrt volume
- Significant increase of reliability
“Industrial” solutions applied where possible, same
cryogenic standards, no big surprises and good results!
Some attention & innovative design (see flex lines) when
routing is on detector (integration, accessibility of
connections/vacuum ports).
So far only positive feedback!
Thanks for your attention!
SPARE
30 June 2014 P. Tropea 27
CM
S: O
VE
RV
IEW
OF T
HE P
RO
JE
CT
CMS CO2 cooling
28
30 June 2014
P. Tropea
FPix Plant core
(USC55)
Bpix Plant core
(USC55)
CO2
Storage
FPix Accumulator
(USC55) BPix Accumulator
(USC55)
FP
ix
BP
ix
Manifold
Boxes
(UXC55)
Filt
er
Filt
er
Concentric
SS
pip
es –
insula
ted (v
acuum
) Concentr
ic S
S p
ipes –
in
sula
ted (
vacuum
)
PT PT
Filt
er
Filt
er
CM
S –
YB
0 C
ON
CE
NT
RIC
Details of pressure drops for YB0 concentric
30 June 2014
29 P. Tropea
0 5 10 150
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Lam TurTur
Tur
Tur
Tur
Tur
Tur
Tur
Tur
0 0.12 0.24 0.36 0.48 0.6 0.72 0.84 0.96 V(m/s)
Pre
ssure
dro
p (
bar)
Mass flow (g/s)
Liquid pressure drop of the CMS YB0 liquid feed tube
Fluid=CO2, T=-20 ºC, Length=15 m, Angle=0º, Roughness=0 mu, Po=10 bar
6mmOD, t=1mm
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.05
0.1
0.15
0.2
0.25
0.3
0.35
0.4
0.45
Tur SlugSW
SWAnnu
Annu
Annu
Annu
Annu
Annu
SW
0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6QE(kW)
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
Pre
ssure
dro
p (
bar)
Vapor Quality (-)
2-phase pressure drop of the CMS YB0 vapor return tube
Fluid=CO2, MF=10 g/s, T=-20 ºC, Length=15 m, Angle=0º
Tem
pera
ture
dro
p (
ºC)
Concentric 12mmOD, t=1mm with 6mm center tube
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 10
0.01
0.02
0.03
0.04
0.05
0.06
0.07
LamStrat
Strat
SW
SW
SW
SW
SW
SW
SW
000000000QE(kW)
0
0.02
0.04
0.06
0.08
0.1
0.12
Pre
ssure
dro
p (
bar)
Vapor Quality (-)
2-phase pressure drop of the CMS YB0 vapor return tube
Fluid=CO2, MF=3 g/s, T=-20 ºC, Length=15 m, Angle=0º
Tem
pera
ture
dro
p (
ºC)
Concentric 12mmOD, t=1mm with 6mm center tube
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 50
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
Lam Lam Lam
Tur
Tur
Tur
Tur
Tur
Tur
Tur
0 0.04 0.08 0.12 0.16 0.2 0.24 0.28 0.32 V(m/s)
Pre
ssure
dro
p (
bar)
Mass flow (g/s)
Liquid pressure drop of the CMS YB0 liquid feed tube
Fluid=CO2, T=-20 ºC, Length=15 m, Angle=0º, Roughness=0 mu, Po=10 bar
6mmOD, t=1mm
Low flow (3 g/s)
Liq
uid
feed
Va
por
retu
rn
High flow (10 g/s)