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)