35
First look at the cooling requirements for the UT detector LHCb CO2 cooling kick-off meeting 28 May 2014 CERN Simone Coelli I.N.F.N. MILANO 1

First look at the cooling requirements for the UT detector

  • Upload
    isanne

  • View
    41

  • Download
    0

Embed Size (px)

DESCRIPTION

LHCb CO2 cooling kick-off meeting 28 May 2014 CERN. First look at the cooling requirements for the UT detector. Simone Coelli I.N.F.N . MILANO. Summary. Overview Cooling power requirements for the upgrade of the UT detector UT Detector cooling open issues - PowerPoint PPT Presentation

Citation preview

Page 1: First look at the  cooling  requirements for the UT  detector

1

First look at the cooling requirements for the UT

detector

LHCb CO2 cooling kick-off meeting28 May 2014

CERN

Simone CoelliI.N.F.N. MILANO

Page 2: First look at the  cooling  requirements for the UT  detector

2

• Overview

• Cooling power requirements for the upgrade of the UT detector

• UT Detector cooling open issues

• Open points and questions

Summary

Page 3: First look at the  cooling  requirements for the UT  detector

3

• Four Planes of Silicon Strip Sensors• with ASICs readout• Support with integrated cooling

STAVE ELEMENTS

16 +16 +18 +18 = 68TOTAL NUMBER OF STAVES

3 STAVE FLAVOURS:8 + 8 + 52 = 68

LHCb UT detector upgradeOVERVIEW

Page 4: First look at the  cooling  requirements for the UT  detector

4

THERMAL REQUIREMENTS

• POWER

ASICs READ-OUT Power electrical dissipation expected = 3200 W+ SELF-HEATING + POWER FLEX DISSIPATION =>4000 W OF DISSIPATED THERMAL LOAD TO BE REMOVED

Considering a margin to take into account the box and the transfer lines insulation..=> 4 to 5 kW cooling system power

• CO2 INLET TEMPERATURE

SENSOR OPERATING TEMPERATURE ALWAYS UNDER - 5 °CASICs OPERATING TEMPERATURE NOT EXCEEDING 40 °CStudies have been done on the thermal transfer, work is ongoing to limit the thermo-mechanical stress over the silicon sensor (100mm*100mm)

Þ Inlet CO2 liquid supply T from - 10 °C to - 35 °CWith the possibility to regulate the inlet fluid temperature

Page 5: First look at the  cooling  requirements for the UT  detector

5

OTHER REQUIREMENTS

The detector is openablemade in two halves

=> the cooling connections should be attached on the two sides (symmetric flow division)

=> having the possibility to open the semi-detector assembly (about 800 mm) possibly relying on the transfer line flexibility

Actual UT

CLOSE POSITION(DATA-TAKING)

OPEN POSITION(MAINTENANCE)

Page 6: First look at the  cooling  requirements for the UT  detector

6

UT Detector cooling open issues

Page 7: First look at the  cooling  requirements for the UT  detector

7

• SNAKE PIPE DESIGN• BASELINE (BEST THERMAL PERFORMANCE)

Stave with a bended pipe embedded in the stave support passing underneath the ASICs

• 2 STRAIGHT PIPES DESIGN• OPTIONAL CONFIGURATION

Stave with two straight pipes running vertically parallel to the longitudinal stave axis

INTERNAL STAVE GEOMETRYHalf lenght stave represented here

Page 8: First look at the  cooling  requirements for the UT  detector

8

Sensor attachment

Has an effect• on the stave thermal profile• on the thermo-mechanical behaviour (to guarantee allowed

deformation and stress)

Several sensor mechanics options un study

example

Page 9: First look at the  cooling  requirements for the UT  detector

9

CO2 Boiling flow direction (upward/downward)

• To be decided what is the best option• Can be simulated using the CO2 code CoBRA?• Should be tested using a full scale prototype,

UTbX half-plane representation, for the snake pipe option

Page 10: First look at the  cooling  requirements for the UT  detector

10

1 2 3 4 5 6 7 8 90

102030405060708090

HALF PLANEFLOW DISTRIBUTION

APPROXIMATE STAVE POWERDISTRIBUTIONTHE STAVES HAVE DIFFERENT THERMAL LOADS

GEOMETRY OF THE SERPENTINE FOR THE CENTRAL STAVEIS DIFFERENT:• 4 MORE BENDS• > TOTAL LENGHT

COOLING DISTRIBUTION SYSTEM DESIGN:

• DEDICATED COOLANT PRESSURE DROP AT EACH STAVE INLET

• INLET FLOW TROUGH CAPILLARIES WITH A DEDICATED CALIBRATION

• SYSTEM TEST FOR THIS SUBASSEMBLY TO VERIFY THE DESIGN

IN THIS SITUATION USING EVAPORATIVECOOLING:THERMO-HYDRAULIC INSTABILITIES CAN ARISE

Total cooling power ~ 500 W

Page 11: First look at the  cooling  requirements for the UT  detector

11

Exhaust manifold :• Stainless steel• 8/9 welded pipe

fitting connections

Supposing snake pipes and an upward flow:

L3

L1

L2

Coolant inlet distribution:• Capillaries for

liquid flow distribution

• L1• L2• L3• Optimized

lengths

manifold conceptual sketch

Page 12: First look at the  cooling  requirements for the UT  detector

12

flow re-balancing system using capillariesconnection to a common manifold?=> half detector: 34 CO2 supply lines3 inlet DP flavours: 4 * L1 + 4 * L2+ 26 * L3

manifold conceptual sketch

Exhaust outlet line

Supposing snake pipes and an upward flow:

CO2 supply line

Page 13: First look at the  cooling  requirements for the UT  detector

24 FEB. 2014 13

A POSSIBLE HALF PLANE FLOW DISTRIBUTION TEST SET-UP

THE MOST UMBALANCED SITUATIONTO BE TESTED TO DEMONSTRATESTABILITY OF THE SYSTEM

Þ ONLY CENTRAL STAVE POWER ON

Þ TRACI COOLING SYSTEM COULD BE USED (POWER 100 W)

THERMO-HYDRAULIC CHARACTERIZATION OF THE SNAKE PIPE

Page 14: First look at the  cooling  requirements for the UT  detector

14

Detector geometryone of the two half-detector unit

~ 1500 mm 45 mm

225 mm

45 mm

Page 15: First look at the  cooling  requirements for the UT  detector

15

OPEN POINTS & QUESTIONS

• UT Detector responsibility (definition of points)• Transfer line is part of the cooling plant?

• Failure analysis: proposals• VELO/UT cooling plants: inter-connection or separated systems?

• How to proceed for the design• Cooling system simulation for design and optimization, CoBRA• Max pressure design for the piping system (MDP=10 Mpa starting

assumption)

Page 16: First look at the  cooling  requirements for the UT  detector

16

Back-up slides

Page 17: First look at the  cooling  requirements for the UT  detector

17

• Prototype testing• Single stave => blown system• Half-plane =>TRACI system• Full scale ..?

Page 18: First look at the  cooling  requirements for the UT  detector

Electronics power dissipation=> Thermal loads

A1.T8 13 mW

A1.T7 17 mW

A1.T6 23 mW

A1.T5 34 mW

A1.T3 135 mW

A1.T4 58 mW

A1.T1No Grid distribution*= total 261 mW

A1.T2No Grid distribution*= total 171 mW

ASICs dissipated thermal power:0,768 mW / ASICThis half stave has on it a total 44 ASICs including both upstream & downstream faces=> 44 * 0,768 = 33,8 W total

Data-Power Flexbusdissipated thermal power:+ 2 W upstream face+ 2 W downstream face

SENSORS self-heatingthermal power:

upstream face downstream face

«central staves»• Number = 8• Cut-out for the the

beampipe• with 8 ASICs read-out• Higher thermal load

Page 19: First look at the  cooling  requirements for the UT  detector

19

Flow regulation valve

Page 20: First look at the  cooling  requirements for the UT  detector

20

UT DETECTOR THERMAL MANAGEMENT

SOME CONSIDERATIONS ON THE EXTIMATED TERMAL GRADIENTS COMING FROM THERMAL FINITE ELEMENT ANALYSIS

• ANALYSIS : STEADY-STATE FOR NOMINAL OPERATING CONDITION• BASED ON THE CENTRAL STAVE GEOMETRY MODEL – IT IS THE MAX POWERED STAVE• EXPLOITING AN IMPROVED DESIGN: USING ALL CARBON FOAM CORE• CALCULATION CLEARLY IS IN IDEAL CODITIONS: PERFECT ADHESION OF GLUE AND THERMAL

CONTACTS• CONDUCTIVE MATERIAL PROPERTIES USED ARE A GUESS• SENSITIVITY ANALYSIS HELP TO SEE THE EFFECTS OF INPUT CHANGE ON THE TEMP. SOLUTION• => SAFETY MARGIN SHOULD BE USED IN COOLING DESIGN TO ACCOUNT FOR SIMULATION

APPROXIMATION

TEMPERATURE REQUIREMENTS REMINDER

• SENSOR OPERATING TEMPERATURE ALWAYS UNDER -5 °C• ASICs OPERATING TEMPERATURE NOT EXCEEDING 40 °C

Page 21: First look at the  cooling  requirements for the UT  detector

21

• THE THERMAL F.E. MODEL TAKES IN ACCOUNT ONLY THE THERMAL CONDUCTION• ACROSS THE STAVE MATERIALS• FROM THE THERMAL POWER SOURCES

• ASICs DISSIPATION• SENSOR SELF-HEATING • FLEX-BUS DISSIPATION

• TO THE COOLING PIPE EMBEDDED INTO THE STAVE AND SERVICING THE TWO STAVE FACES

• THE STAVE DESIGN GOAL IS TO MANAGE ALL THE THERMAL LOADS• EXPLOITING THE BEST THERMAL CONDUCTION IN THE STAVE UNTIL THE COOLING PIPE• THERMAL CONVECTION TOWARDS THE FLUID IN THE PIPE • PHASE CHANGE USING THE EVAPORATING COOLANT CO2

• THERE IS AN ADDITIONAL THERMAL CONVECTION WITH THE GAS FLOWING IN THE DETECTOR BOXTO CONTROL THE BOX ENVIRONMENT

• GAS INLET TEMPERATURE SHOULD BE LOWER THAN THE MEAN DETECTOR TEMPERATUREOTHERWISE THE DETECTOR COOLING SYSTEM WILL HAVE EXTRA WORK => T INLET GUESS: AROUND 0 °C=> MEAN SURFACE TEMPERATURE OF THE STAVES WILL BE CALCULATED FOR AN EVALUATION OF THIS EFFECT• GAS FLOWRATE MUST BE SUITABLY CHOOSEN - DEPENDING ON THE BOX AIR-TIGHTNESS

• REQUIREMENT: THE UT BOX DEW-POINT SHOULD BE ADEQUATELY LOW (LESS THAN - 40 °C)• THIS SETS THE MAX HUMIDITY ADMISSIBLE INSIDE THE BOX TO AVOID CONDENSATION ON THE COLD

PARTS• THIS DICTATES REQUIREMENTS ON:

• THE GAS FLOW PURITY • THE BOX AIR –TIGHTNESS

• PROPOSAL: USE NITROGEN FLOW IN A CLOSED LOOP WITH AN EXTERNAL COOLING, MEASURING BOX INTERNAL DEW-POINT

• HOT SPOTS (THE ASICs) WOULD DISSIPATE SOME THERMAL POWER TOWARDS THE GAS (MITIGATING TEMPERATURE PEAKS

Page 22: First look at the  cooling  requirements for the UT  detector

22

• THERE IS ALSO A UNAVOIDABLE RADIATIVE THERMAL TRANSFER (INFRARED EMISSION)• HELPS HOT SPOTS (THE ASICs) TO DISSIPATE SOME THERMAL POWER TOWARD COLDER SURROUNDING

AND MITIGATING TEMPERATURE PEAKS• DETECTOR THERMAL EXCHANGE TAKES PLACE WITH THE SURFACES INCLOSING THE DETECTOR• DEPENDING MAINLY ON THE TEMPERATURE DIFFERENCE BETWEEN THE EXTERNAL DETECTOR PLANESAND THE CLOSING BOX SURFACES (THAT COULD BE NEAR TO THE AMBIENT TEMPERATURE)• RADIATION HEAT TRANSFER CALCULATION WOULD NEED A KNOWLEDGE OF MANY PARAMETERS (EXACT

GEOMETRY AND TEMP. DISTRIBUTION, VIEW FACTORS AND EMISSIVITY FACTORS)=> CALCULATION WILL BE DONE,=> A SAFETY MARGIN IN THE COOLING DESIGN SHOULD BE TAKEN

• FOLLOWS A SUM-UP OF THE RESULTS FOR THE TWO DESIGN :• THE SNAKE PIPE DESIGN• THE TWO STRAIGHT PIPES DESIGN

• TEMPERATURES OF:• SENSORS• ASICs

Page 23: First look at the  cooling  requirements for the UT  detector

23

• SNAKE PIPE DESIGN

• STRAIGHT PIPES DESIGN

SENSOR TEMP. ASICs TEMP.

SENSOR TEMP.

INTERNAL STAVE GEOMETRY

INTERNAL STAVE GEOMETRY

ASICs TEMP.

TEMP. CONDUCTIVE GRADIENT RELATIVE TO THE COOLING PIPE INTERNAL SURFACE

Max SENSOR T ~ 11 °C

Max ASIC T ~ 30°C

Max ASIC T ~ 40°C

Max SENSOR T ~ 3 °C

Page 24: First look at the  cooling  requirements for the UT  detector

24

NOTES ON THE RESULTS:

• COMMON TO THE TWO DESIGN OPTIONS

• SENSORS MAX TEMP. IS AFFECTED BY THE HOTTEST ASIC THAT TRANSMITS HEAT TO THE NEAR SENSOR GIVING RISE TO THE SENSOR HOT-SPOT

• ASIC-SENSOR DISTANCE AND FLEXBUS CONDUCTIVITY DRIVING THE EFFECT• ASIC HEAT IS TRANSFERRED ALSO TO THE OPPOSITE FACE SENSOR

• SNAKE PIPE DESIGN (COMPLICATION: PIPE BENDING)• ASICs TEMP. ARE RELATIVELY UNIFORM• DUE O THE SNAKE PIPE CONFIGURATION THAT IS DESIGNED TO PASS UNDERNEAT

EACH ASIC• ASICs THERMAL FIGURE OF MERIT VARYING FROM 10-20 °C/W/cm2• WORST ΔT over the sensor AROUND 3 °C

• TWO STRAIGHT PIPES DESIGN (COMPLICATION: INSERTS IMPLEMENTATION, DOUBLE PIPE CONNECTIONS)• TPG INSERTS (ARE USED WITH K = 900-900-10 UNDER THE ASICs• ASIC DISTANCE FROM THE PIPE DRIVING THE TEMP. PEAK• ASICs THERMAL FIGURE OF MERIT VARYING FROM 5-30 °C/W/cm2• WORST ΔT over the sensor AROUND 8 °C

=> GENERAL CONCLUSION: THE SNAKE PIPE DESIGN IS THE BEST TO BE PURSUED

THERMAL FIGURE OF MERIT ISBASED ON: • ASIC POWER ~ 0.768 W• ASIC SURFACE ~ 0.52 cm2

Page 25: First look at the  cooling  requirements for the UT  detector

25

STARTING FROM THE THERMAL SIMULATION RESULTS:=> COOLING PIPE TEMPERATURE COULD BE SET TO STATISFY THE REQUIREMENTS

SENSOR OPERATING TEMPERATURE ALWAYS UNDER - 5 °CASICs OPERATING TEMPERATURE NOT EXCEEDING 40 °C

• SNAKE PIPE DESIGN => COOLING PIPE T ~ - 8 °C• TWO STRAIGHT PIPES DESIGN => COOLING PIPE T ~ - 16 °C• NOTE THIS IS WITHOUT MARGIN

REQUIRED INLET CO2 FLUID TEMPERATURE COULD THEN BE CALCULATEDTHE EFFECTS RELATED TO THE USE OF A TWO-PHASE COOLANT MUST BE TAKEN INTO ACCOUNT :

• PRESSURE DROP IN THE COOLING PIPE IS CAUSING A TEMPERATURE DROP (SATURATE FLUID)Þ INLET TEMPERATURE IS HIGHER THAN OUTLET (THE FIGURE DEPENDING ON THE PIPE AND ON

THEFLUID PARAMETERS )• INTERNAL HEAT TRANSFER COEFFICIENT (HTC) BETWEEN THE FLUID AND THE PIPE INTERNAL WALL ADDING A TEMPERATURE GRADIENT TO BE SUPERIMPOSED ON THE LOCAL FLUID TEMP.• ΔT convective contribution: temperature difference between fluid (bulk) and inner pipe wall• ΔT conductive contribution: temperature difference across the stave simulated by Finite Element Analysis

Page 26: First look at the  cooling  requirements for the UT  detector

26

TO START THE DESIGN SOME HYPOTHESIS ON THESE THERMAL GRADIENTCAN BE MADEFIGURES BASED ON PREVIOUS EXPERIENCES AND SIMILAR DESIGN ARE AVAILABLEANY SUCH DESIGN NEEDS AN ITERATIVE PROCESS BECAUSE THE COOLING GEOMETRY AND THE THERMO-HYDRAULIC PROCESS ARE INTER-RELATEDDISCLAIMER:THESE CALCULATION NEED TO BE REFINED AND FOLLOWED-UP DURING THE COOLING SYSTEM DESIGN=> REAL SCALE TEST NEED TO BE MADE TO VALIDATE THE CALCULATIONS

NOTEGIVEN THE RESPECT OF THE CITED REQUIREMENTSTHE UT DETECTOR IS NOT VERY SENSITIVE TO A SMALL CHANGE (FEW DEGREES) IN THE OPERATIVE TEMPERATURE OF A SENSOR LOCATED ON THE INLET OR OUTLET PART OF THE STAVE

=> USING SOME MARGIN THE OPERATIVE INLET CO2 COOLANT TEMPERATURE COULD BE SUPPOSED TO BE FROM - 20 °C TO - 30 °CDEPENDING ON THE DESIGN CHOICE (SNAKE OR STRAIGHT PIPE)

THE FOLLOWING CONSIDERATIONS FOR THE COOLING SYSTEM ARE VALID IN GENERAL FOR BOTH THE DESIGN CASES

HALF STAVECENTRAL GEOMETRY

Page 27: First look at the  cooling  requirements for the UT  detector

27

THE PURE CO2 SATURATION CURVEHEREAFTER CORRELATES TEMPERATURE AND PRESSURE INSIDE THE EVAPORATION CHANNEL

- 20 °C TO - 30 °C COOLING FLUID OPER. TEMP.

=> 10 TO 20 bar COOLING FLUIDOPER. PRESSURE

Page 28: First look at the  cooling  requirements for the UT  detector

28

IN THE RANGE OF INTEREST DELTA H liq.=> vap. = 280 kJ/kg

THE LATENT HEAT OF VAPORIZATION FOR CO2 CAN BE KNOWN FROM THE CO2 PRESSURE-HENTALPY DIAGRAM

Page 29: First look at the  cooling  requirements for the UT  detector

29

CONSIDERING FIRST WHAT HAPPENS IN A SINGLE STAVETHE CENTRAL STAVE TOTAL POWER TH THE COOLING PIPE HAS TO EXTRACT IS AROUND 85 W(TAKING INTO ACCOUNT SOME DISSIPATION IN THE FLEXBUS)

DESIGN OF THE COOLING SYTEMREQUIRES:• INLET CO2 LIQUID SUBCOOLED• BUT NEAR TO SATURATION• OUTLET VAPOUR FRACTION X OUT= 0.5

TO 0.6• TO AVOID THE DRY-OUT

MASS FLOWRATE

MASS FLOWRATE

TAKING X OUT = 50% MASS FLOWERATE NECESSARY CAN THEN BE EXTIMATED:

MASS FLOWRATE= 2 * POWER /DELTA HLIQ-VAP

= 2* 85 W / 280 kJ/kg = 0.6 g/s

LESS FLOWRATE IS TO BE AVOIDEDBECAUSE WILL BE GIVING A RISK OF DRYING-OUT

Page 30: First look at the  cooling  requirements for the UT  detector

30

TO LIMIT THE TEMP.DECREASE => DESIGN CHANNEL PRESSURE DROP HAS TO BE LIMITEDACCEPTING FEW DEGREES MEANS ACCEPTABLE PRESSURE DROP AROUND 1 TO 2 BAR

MASS FLOWRATE

MASS FLOWRATE

CHANNEL

PRESSURE DROP IS DRIVEN MAINLY BY THE FLOWRATE, GEOMETRY (TOTAL LENGHT AND BENDS), INTERNAL ROUGHNESS, CHANNEL ORIENTATION

REAL CONDITION TEST IS NEEDED MEASURING ALL THE HYDRAULIC PARAMETERSFLOWRATEP, T INLET AND OUTLET

Page 31: First look at the  cooling  requirements for the UT  detector

31

COURTESY OF BART VERLAAT (NIKHEF - CERN)

WHAT HAPPENS INSIDE THE CHANNEL

ALONG EVAPORATING CHANNEL:• DECREASING P• DECREASING T

TARGET DESIGN FOR THE THERMO-HYDRAULIC CONDITIONS IN THE DETECTOR COOLING CHANNEL

AVOID WITH, A SAFETY MARGIN, THE DRY-OUTX OUT < MAX

=> REQUIRES SUFFICIENT FLOWRATE

Page 32: First look at the  cooling  requirements for the UT  detector

32

DETECTOR COOLING CO2 SYSTEM2PACL (2-Phase Accumulator Controlled Loop) (Verlaat et al. 2011)

UT DETECTOR COOLING CONFIGURATION SHOULD LOOK LIKE THIS:

This figure shows a simplified schematic of the circulation system which is used often in particle physics detector cooling.This method controls the system pressure and hence evaporation temperatures by a 2-phase accumulator which is heated or cooled. A heat exchanging concentric transfer line (2-3) brings the liquid in the inlet of the detector cooling pipes into saturation. Like this the evaporation in the detector cooling pipes is always happening at the pressure regulated in the accumulator. Experimental caverns are inaccessible during running of the LHC beam. All the active hardware of a 2PACL can be located in a safe zone (left side of figure 2). This area is always accessible for maintenance. In the experimental cavern preferably only passive piping is present.

Page 33: First look at the  cooling  requirements for the UT  detector

33

Þ «CENTRAL» stave power ~ 90 W Þ «HALF PLANE»power ~ 500 W

To start thinking on the connectivity of the cooling system exploiting CO2 evaporation system

Proposal: use for each «half plane» • 1 lower inlet manifold,

distributing liquid CO2 to the staves

• 1 upper manifold, collecting hexaust CO2 (partially evaporated) from the staves

«right half plane»«left half plane»

CO2 (X = 0)

CO2 (~ 50%)

X := thermodynamic title Saturated liquid = 0%Saturated vapour =100%

Half planes areSupposed to move to open like in the actual tracker

DETECTOR COOLING LAY-OUT supposing to have a modularity with the four UT detector planes divided in:• 1 right half box (composed of 4 half planes)• 1 left half box (composed of 4 half planes)

Page 34: First look at the  cooling  requirements for the UT  detector

34

The CO2 cooling plant should be a 2PACLsystem with cooling capacity: 4000 Watt@-30 °CÞ Need a specific plant

design Þ Similar to VELO UpgradeActual LHCb- VELOCooling capacity: 1500 W@-30°C

DETECTOR COOLING LAY-OUT

CONCEPTUAL BRANCHES LAY-OUT

Page 35: First look at the  cooling  requirements for the UT  detector

35

ANOTHER POINT IS THE PRESSURE DROP OF THE OUTLETLINE Þ LIFTS UP THE TEMPERATURE PROFILE OF THE COOLING CHANNEL

For the study of long length cooling branches with CO2 a program is developed called CoBra Calculator (CO2 BRAnch Calculator)

DESIGN OF THE COOLING SYSTEM