Temperature monitors overview (with focus on Static T-gradient - - PowerPoint PPT Presentation

Temperature monitors overview

(with focus on Static T-gradient thermometers)

• A. Cervera

M.A. García Peris (IFIC-Valencia)

LAr Instrumentation Scope Review CERN, 19/06/2019

With assistance from C.Streff (SDSU), E. Voirin (FNAL), A. Hahn (FNAL), S. Pordes (FNAL)

Anselmo Cervera Villanueva, IFIC-Valencia

Overview

• Motivation of precise temperature measurements
• Description of ProtoDUNE-SP systems and results
• From ProtoDUNE to DUNE

2

Anselmo Cervera Villanueva, IFIC-Valencia

Introduction

• The first DUNE far detector module will use the largest cryostat ever

built (20xProtoDUNE) for a TPC

• At least for this module a sufficiently dense 3D grid of temperature

sensors should allow us to understand the behaviour of the system and ensure that lack of understanding of LAr velocity, density & purity does not compromise physics

3

Cryostat

Insulation Vacuum tightness

Electron lifetime prediction far from PrMs, using CFD simulations validated with temperature measurements

Physics LAr impurities Cryogenics system

LAr recirculation

The system will help in mitigating risks at reasonable cost (< 300 K$/10-kt)

Drift velocity, flow (ions), electron lifetime, energy calibration

Anselmo Cervera Villanueva, IFIC-Valencia

Cryogenics system

• A precision 3D temperature map will demonstrate that the cryostat

and the cryogenics system are working well:

• Incoming LAr temperature as expected in all inlets
• LAr mixing as expected: gradients as expected everywhere
• What do we do if data contradicts one of the above ?
• Can the problem be solved externally ?
• Can inlet temperature be controlled ? How much ?
• Can we live without this ? How do we diagnose a potential problem?

4

Even if the problem cannot be solved, a 3D temperature map gives us a handle for offline corrections (next slide)

Anselmo Cervera Villanueva, IFIC-Valencia

Physics

• Energy calibration depends very much on the local electron lifetime
• e- lifetime can only be measured in few locations near the cryostat

corners by PrMs. This is an instantaneous measurement

• e- lifetime can be

computed using cosmic rays, but those are rare 1500 m underground

5

Is it important from a physics standpoint to validate theoretical calculations of fluid flow inside the DUNE far detector modules (no longer a design issue) ?

Charge Question 3

• Yes. CFD simulations, tuned with temperature data, will be used to predict the

electron lifetime everywhere in the TPC and make offline corrections

Anselmo Cervera Villanueva, IFIC-Valencia

CFD simulations

• CFD successfully used in the

35t prototype to reproduce the observed e- lifetime

6 Erik Voirin Dune DocDB 1156

PrM1 PrM2 PrM3 PrM4

~30 mk

PrM1 PrM2 PrM3 PrM4

Erik Voirin Produced yesterday (thanks!!!)

• Those simulations also predict

the 25-35 mK gradient observed between RTDs in PrM1-3

35t prototype showed direct relation between Temperature and e-lifetime

T1 T2 T3

Purity Stratification study actually triggered by precise temperature measurements

35t prototype

e- lifetime temperature

Anselmo Cervera Villanueva, IFIC-Valencia

CFD simulations

• CFD successfully used in the

35t prototype to reproduce the observed e- lifetime

6 Erik Voirin Dune DocDB 1156

PrM1 PrM2 PrM3 PrM4

~30 mk

PrM1 PrM2 PrM3 PrM4

Erik Voirin Produced yesterday (thanks!!!)

• Those simulations also predict

the 25-35 mK gradient observed between RTDs in PrM1-3

35t prototype showed direct relation between Temperature and e-lifetime

T1 T2 T3

Purity Stratification study actually triggered by precise temperature measurements

35t prototype

e- lifetime temperature

Anselmo Cervera Villanueva, IFIC-Valencia

CFD simulations

• Used to design the cryogenic system in ProtoDUNE and DUNE:

number/distribution of LAr inlets/outlets, LAr flow-rate/temperature

• In DUNE (and ProtoDUNE) we will use it to:
• Validate the cryogenics system with temperature and purity data. Get

confidence on the full system and detect malfunctioning

• Predict LAr properties everywhere in the TPC after constraining with

temperature (and purity) data

• CISC working primarily with SDSU with some advice from E. Voirin

7

Temperature maps + CFD simulations is probably the only practical way of quickly predicting everywhere in the TPC the LAr properties that are relevant for Physics

Anselmo Cervera Villanueva, IFIC-Valencia

Precision and standard RTDs

• Using same cables and readout, precision RTDs have a factor 10

lower RMS

• Reproducibility for precision sensors is 2.5 mK, it is expected much

worst for standard sensors

• Longevity should be also worst

8

15 mk 1.5 mk

standard sensor

• n wall

precision sensor in static T-gradient

evolution during one day

<5 $

100 $

1.4 cm ~5 cm

Anselmo Cervera Villanueva, IFIC-Valencia

Motivation of each system

9

System Precision Motivation

Vertical array ~mK

• check cryogenics system behaviour
• constrain CFD simulations for e-lifetime prediction

Top/bottom Horizontal 2D grid RTDs at inlets and pumps Floor 20-50 mK monitor the presence of LAr when filling starts Walls & roof 20-50 mK Monitor cryostat membrane temperature during cool- down and filling Vertical array in gas 20-50 mK

• constrain & validate CFD simulations in ullage
• understand behaviour of mechanical structures (DP)

Description of ProtoDUNE-SP systems

5 sensors 12 sensors

Standard sensors

Anselmo Cervera Villanueva, IFIC-Valencia

Floor/wall sensors

12

2 hours

Temperature evolution when LAr filling started

temperature

time

epoxied to floor before removing temporary tape

Floor sensors Wall sensors

Aug 10th

epoxied to wall

standard RTD mounted on PCB support with IDC-4 connector

24 sensors 48 sensors 8 sensors 12 sensors

Precision sensors

dynamic static

Anselmo Cervera Villanueva, IFIC-Valencia

Crucial to obtain mK precision

14 Multiplexing board for 24 sensors (with 1 mA current source)

• X. Pons (CERN)

Anselmo Cervera Villanueva, IFIC-Valencia

Top/bottom sensors

15

Top sensors

attached to ground planes

Bottom sensors

attached to LAr pipes

LAr pump LAr inlets

static dynamic

Anselmo Cervera Villanueva, IFIC-Valencia

T-gradient monitors

16

Dedicated talk by Jelena this afternoon

static dynamic

# sensors 48 24 Δz top/bottom 12 cm 10 cm Δz middle 24 cm 50 cm calibration method laboratory in-situ (movable system) calibration precision 3 mK < 1 mK

static dynamic

Anselmo Cervera Villanueva, IFIC-Valencia

Calibration

• Static profiler and top/bottom

sensors calibrated in the lab (3 mK)

• The dynamic T-gradient monitor in-

situ calibration is used to test the hypothesis of homogeneous temperature for no recirculation

• Static T-Gradient can be cross-calibrated using that assumption
• This matches with the laboratory calibration

17

https://indico.fnal.gov/event/16764/session/17/contribution/157/material/slides/0.pdf

Anselmo Cervera Villanueva, IFIC-Valencia

Temperature fluctuations

• <2 mK below 6.5 m
• Increase dramatically near the LAr surface

18

static dynamic GPs May, 10 minutes GPs

Anselmo Cervera Villanueva, IFIC-Valencia

Vertical profiles

• static and dynamic cross-calibrated to each other with pumps-off
• Gradients below 15 mK
• Larger gradient in static profiler: colder at the bottom, warmer at

the top

19

dynamic static

Vertical gradient depends on location

pumps-off calibration movable calibration

Anselmo Cervera Villanueva, IFIC-Valencia

Vertical profiles

• static and dynamic cross-calibrated to each other with pumps-off
• Gradients below 15 mK
• Larger gradient in static profiler: colder at the bottom, warmer at

the top

19

dynamic static

8 mk 15 mk

Vertical gradient depends on location

pumps-off calibration movable calibration

Anselmo Cervera Villanueva, IFIC-Valencia

CFD predictions

• We have tested a couple of configurations:
• Tinlet = Toutlet + 0.2 K
• Tinlet = TLAr surface + 0.2 K (closer to reality)
• Still work to do

20

static SDSU simulations

movable calibration pumps-off calibration

dynamic

Anselmo Cervera Villanueva, IFIC-Valencia

Stability of vertical profiles

• ΔΤ between first sensor and other sensors in the same profiler over
• ne month

21

1,603 m 3,019 m 4,671 m 5,615 m 6,205 m 1.587 m 3,111 m 4,635 m 5,651 m 6,159 m 7,267 m

dynamic static

Shape of profiles stable within 2-3 mK

Anselmo Cervera Villanueva, IFIC-Valencia

Bottom sensors

• ΔT between all pipe sensors at

2nd static sensor, all at ~40 cm

• Temperature decreases towards

the LAr pump

• Along the pipes (X) temperature is

higher in the middle of the TPC

22

March May Pumps-off

LAr pump static dynamic LAr pump dynamic static LA inlets cathode

Anselmo Cervera Villanueva, IFIC-Valencia

CFD predictions

• CFD predicts decreasing temperature towards pump, but the effect

is not as large as in data

• CFD does not predict higher temperature in the cathode region

23

4 mk ~20 mK

CFD at 40 cm (SDSU) data

static

cathode

dynamic

FC pump

Anselmo Cervera Villanueva, IFIC-Valencia

CFD predictions

• This simulation does predict larger temperature below the cathode
• It also shows:
• lower temperature

towards pump

• higher gradient at static

profiler

24

CFD at 2nd LAr pipe plane

pipe sensors

static dynamic

DocDB-928-v3

NOTE: map exists outside TPC but color range not appropriate

pump

• E. Vorin’s simulations

Anselmo Cervera Villanueva, IFIC-Valencia

Charge questions

25

How did the three types of temperature sensors employed in the ProtoDUNE detectors (Static T-gradient thermometer, Dynamic T-gradient thermometer, and individual temperature sensors) perform relative to one another?

Charge Question 1

Shown in previous slides. Many studies still to come Are all three different types of temperature monitors employed in the ProtoDUNE-SP cryostat necessary for the DUNE far detector cryostats in that they have unique features necessary for meeting monitoring requirements?

Charge Question 4

Dynamic T-gradient monitor: can be calibrated in-situ at any time, ensuring that we have a reliable understanding of the gradient regardless of ageing, calibration or electronics problems. Gives a reference for all other systems. Static T-gradient monitor: It has much simpler mechanics. It can be installed in places were space is restricted. In DUNE-SP, where there is no space behind APAs, this is the only viable solution for the vertical gradient

• measurement. Moreover, lab calibration works since the first day.

Individual sensors: They complement the vertical measurement with a 2D horizontal map. In ProtoDUNE-SP it has been observed that the horizontal behaviour is far from trivial.

Anselmo Cervera Villanueva, IFIC-Valencia

CFD simulations: conclusions

26

Were the employed temperature sensors able to provide the data necessary to validate theoretical calculations of fluid flow inside the ProtoDUNE cryostat?

Charge Question 2

We are still in the process of understanding CFD simulations and tuning them using ProtoDUNE-SP data Current simulations predict ~10 mK gradients, which agree with data to first

• rder, but we need to go into the details

We should demonstrate that:

• We know how to deal with CFD input parameters to improve agreement with data
• e- lifetime can be predicted in ProtoDUNE-SP as was done in the 35t prototype

Anselmo Cervera Villanueva, IFIC-Valencia

CFD simulations: next steps

• 1. Identify the relevant CFD input parameters affecting vertical and

horizontal patterns

• 2. Get a feeling of how gradients depend on those parameters
• 3. Tune parameters to match data
• 4. If no agreement revise geometry implementation

27

Parameter Input or Output of CFD Comments Nominal values Cryostat height Input Fixed 7.878 m LAr surface height Input Parameterised 7.406 m LAr surface temperature Input Parameterised 87.596 K ullage pressure Input Does not affect liquid (overwritten by surface temperature) 1.045 bar LAr inlet temperature Input Varied bulk LAr temp + 0.2 K LAr flow rate per pipe Input Varied 0.4170025 kg/s Heat flux Output 5.76 W/m^2 vapor being drawn from the chimneys Output 5-7 gr/sec

Anselmo Cervera Villanueva, IFIC-Valencia

ProtoDUNE-DP

28

Are the temperature monitor strings currently being implemented in the ProtoDUNE-DP cryostat (and hence not yet tested) potentially a simpler and less expensive option to the three types of devices implemented in the ProtoDUNE-SP cryostat?

Charge Question 5

The ProtoDUNE-DP system has a different purpose. It uses one string in one of the corners with standard uncalibrated

• sensors. In principle this can only be used for level

measurements and during cool-down and filling. Those sensors don’t have sufficient precision for:

• Detecting cryogenic system problems: < 20 mK (in the 35t

stratification was observed with 20 mK gradient)

• Check the temperature uniformity: in ProtoDUNE-SP <20 mK
• Predict the electron lifetime based on CFD simulations

This system, as it is, can only be installed in the corners where we expect a special regime, not representative of most

• f the cryostat

Dedicated talk by Filippo

From ProtoDUNE to DUNE

Anselmo Cervera Villanueva, IFIC-Valencia

Lessons learned I

• Sensors at pipe’s edges see (partially) the incoming LAr
• Difficult to use those temperatures since they depend very much on exact

sensor position and incoming LAr direction

• In DUNE we can have few sensors at inlets to check the incoming LAr

temperature, but most of them should be farder from inlets

• Also locate some sensors at other special locations: near pumps, …

30

Anselmo Cervera Villanueva, IFIC-Valencia

Lessons learned II

• In ProtoDUNE-SP cables for individual sensors were routed

independently of each other to avoid grounding loops in the case the teflon jacket was broken

• Also, 1% slack was used to account for shrinkage in cold
• In DUNE we plan to use compact cable bundles with less slack

31

Small shrinkage of cable Cables routed individually

Anselmo Cervera Villanueva, IFIC-Valencia

Internal cryogenics

• Final distribution of sensors will depend on CFD simulations

(ongoing) and distribution of LAr inlets & outlets

32

64 inlets in each blue pipe

LAr return pipes GAr pipes

• Baseline is 6 static T-gradient monitors, with 48 sensors each
• Behind APAs, with ~30 cm of clearance
• Two of them in LAr inlet planes (to be decided which)
• No need of dedicated ports. Use DSS ports to extract cables

Anselmo Cervera Villanueva, IFIC-Valencia

T-gradient monitors

33

S S

24 24 24 24 24 24

S S S

24 24 24 24 24 24 24 S

Static T-gradient monitor with 48 sensors Dynamic T-Gradient monitor with 72 sensors Vertical string Number of cables in each port

D

D

S

72

34

APA side cryostat wall 14 m 95 mm

?

sensor SS string M10 bolts

34 cm

cryostat top membrane < 30 cm Intermediate guiding ring to reduce swinging 32 cm cryostat bottom membrane 95 mm

side view front view

cable bundles to DSS port (24 cables each) DSS port DSS port rigid SS rod to prevent string torsion

Conceptual design (SP)

M10 bolts

340 mm

95 mm

Use M10 bolts at top and bottom corners to attach two SS strings, one for cables and another one for sensors

• It is important to have sensors “in all 4 drift volumes” to understand

the effect of the non-transparent cathodes

• Assume many sensors to understand feasibility of a complex system
• Larger density at the position of the static T-Gradients to have good

sampling in a complete ring surrounding the TPC

• Similar distribution at top and bottom

Anselmo Cervera Villanueva, IFIC-Valencia

Horizontal 2D grid

35

S

Top/bottom individual temperature sensor Static T-gradient monitor with 48 sensors Vertical string Gas purge pipes Liquid argon return pipes

S S S S S S

134 individual sensors (top +bottom)

Anselmo Cervera Villanueva, IFIC-Valencia

Additional sensors

• Sensors at LAr inlets:
• We have learned that sensors directly exposed to the incoming LAr

could be very useful both to validate the cryogenics system and as input for CFD simulations

• Successfully done in 35t prototype
• Sensors at LAr pumps

36

Anselmo Cervera Villanueva, IFIC-Valencia

37

Is it important from a physics standpoint to validate theoretical calculations of fluid flow inside the DUNE far detector modules (no longer a design issue) and if so what level of instrumentation will be required to accomplish this?

Charge Question 2

In principle the more sensors the better constraint, which implies a lower systematic error on physics. There is obviously saturation here. To understand the optimal number of sensors and their locations we need to implement the full correction chain:

• several CFD simulations varying relevant parameters
• mock temperature data with varying number of sensors to constraint the CFD

simulations

• Estimate the error on the predicted electron lifetime, which will depend on

the number of sensors and their location

Anselmo Cervera Villanueva, IFIC-Valencia

Dual-Phase detector

• DUNE-DP will use the same cryogenics system and the same LAr

recirculation scheme. ¿ Can we assume T-maps will be the same ?

• Not obvious because detector segmentation is different
• We can probably have less sensors but

we think a 3D temperature map is

• important. No quantitative answer yet

38

S

Bottom individual temperature sensor Floor individual temperature sensor Wall sensor array with 13 sensors (one every three corrugations) Static T-gradient monitor with 24 sensors Sensor array in gas phase with 8 sensors Gas purge pipes Liquid argon return pipes

G W

S S S S S S W W

G G G G G G G G G G G G G G G G G G G G

TDR distribution

39

Conceptual design (DP)

Comsol simulation

• f Faraday Cage

Anselmo Cervera Villanueva, IFIC-Valencia

Question 6

40

Are the proposed mechanisms for supporting the monitors within the cryostat and connecting them to the outside of cryostat mechanically sound and cost effective?

Charge Question 6

Static T-Gradient monitors:

• For the SP detector the proposed mechanics is much cheaper that it was in

ProtoDUNE-SP since no Faraday Cage is needed

• For DP we don’t expect larger cost since the use of faraday cage is compensated

by lower number of sensors

• The installation is also simpler since no rigid elements are used and can be

transported in a roll to the cryostat Individual sensors:

• Mechanics will be simpler than in ProtoDUNE-SP. Cables will be routed in bundles

In both cases cost is driven by sensors & cables, not by the supporting system

Anselmo Cervera Villanueva, IFIC-Valencia

Conclusions

• The first DUNE far detector module will use the largest cryostat ever

built (20xProtoDUNE) for a TPC

• At least for this module a sufficiently dense 3D grid of temperature

sensors should allow us to understand the behaviour of the system and ensure that lack of understanding of LAr velocity, density & purity does not compromise physics

41

The proposed system will help in mitigating risks at reasonable cost

Temperature maps + CFD simulations is probably the only practical way of quickly predicting everywhere in the TPC the LAr properties that are relevant for Physics

Risks introduced by the system are small ProtoDUNE-SP data analysis looks promesing

backup

42

Calibration of temperature sensors in the Lab

Anselmo Cervera Villanueva, IFIC-Valencia

Calibration setup

44

Aluminum capsule with 4 sensors inside Aluminum capsule inside PLA Box

5 cm 12 cm

4 sensors are put inside a cylindrical aluminum capsule to minimise convection and avoid thermal shocks

PT102 with IDC-4 connector

1.4 cm ~5 cm

4-wires readout to avoid lead resistance and effect

• f connectors

PLA support to hold the sensors and keep them at a fix distance in the middle of the capsule

35 cm 10 cm

Home made cryostat: polystyrene box with 12 cm thick walls and a 3D printed PLA box with two independent volumes

4 sensors together

Anselmo Cervera Villanueva, IFIC-Valencia

Results

• Dispersion of the measurements is below 2

mk for the middle sensor (closer to reference) and below 4 mk for outer sensors

• Reproducibility, defined as the RMS of the

mean offset in the stable region (1000-2000 s) is below 1 mk for most sensors

45

stable region 1000-2000 s

1 2 3 (ref) 4 1 2 4

Anselmo Cervera Villanueva, IFIC-Valencia

Calibration strategy

• Use a sensor as reference. Keep that sensor in all sets
• Thus, 3 sensors are calibrated in each set
• Two calibration methods:

46

1st round 2nd round 3rd round

Anselmo Cervera Villanueva, IFIC-Valencia

Stability over time

• Check the same sensor (middle one) after a month
• NO changes (< 1 mk) after a month
• Need more checks like this

47 March 22nd April 17th March 22nd + April 17th

Anselmo Cervera Villanueva, IFIC-Valencia

Conclusions on calibration

• Dynamic T-gradient is essential to have a reliable reference
• All other systems can be referenced to this one, calibrated to better

than 1mK, at any time (no affected by ageing)

• Laboratory calibration works well, except for few outliers its

precision is ~ 3 mK

• Pumps-off calibration is an important tool, useful to understand in-

situ the reliability of the laboratory calibration

• We know the lab. calibration can be improved since when we did it

we were still converging and improving the method

• The new calibration system will be prototyped and tested in 2020

48

Anselmo Cervera Villanueva, IFIC-Valencia

LAb calibration

• We believe ~1 mk reproducibility can be achieved
• Steel need to understand variations of behaviour from one set to

another

• Improve some elements of the calibration setup, with better insulation
• f the sensors
• Monitor pressure and ambient temperature
• There is still room for improvement:
• Improve cryostat
• Modify the sensor PCB such that all 4 sensors can be put even closer
• Amplify the voltage signal such that we use the full range of the ADC.

Currently using 20 mV, having a range of 1 V

49

Interface with APA

Anselmo Cervera Villanueva, IFIC (UV-CSIC)—Valencia

Cable trays

• Single cable tray sections are attached to the APAs
• The design is not final but they will certainly reduce the available space

between APA and cryostat membrane

51

Anselmo Cervera Villanueva, IFIC (UV-CSIC)—Valencia

Conflict with our system

52

• ur strings

conflict

Anselmo Cervera Villanueva, IFIC (UV-CSIC)—Valencia

• 44 mm mm between cable tray and corrugation

53

corrugation cable tray

• Sensor’s string is passed through the space between the cable trays in

two adjacent APAs

• Adjustable string position to match the APA gap
• Cable’s string behind cable trays
• Strings mounted prior to APA installation

Anselmo Cervera Villanueva, IFIC (UV-CSIC)—Valencia

Option 1

54

Front view side view

M10 bolts in cryostat corner M10 bolts in cryostat corner

cable’s string sensor’s string cable’s string sensor’s string

Anselmo Cervera Villanueva, IFIC (UV-CSIC)—Valencia

Option 2

• String are attached to the cable tray and mounted

in the clean room together with all cables and sensors

• Significant interfaces with APA
• Would need to transfer the load to the DSS once

moved to its final position

• No DSS at the bottom
• A bad idea !!!!

55

35t

Anselmo Cervera Villanueva, IFIC-Valencia

Observations in 35t prototype

57

Anselmo Cervera Villanueva, IFIC-Valencia

• the PrM data, showed that indeed there was stratification, and that it

was somehow related to the LAr recirculation pump being on. In

• rder to match the data, Erik V. had to make the inlet LAr be colder

than the bulk LAr. This was then “verified” by the trick we made with the three RTD’s on the PrM top plates.

• It was at that point that we realized that the CFD is no better than

getting the boundary conditions correct. In the next 35T run (Sarah’s first HV run), we did install an RTD in the LAr inlet, as well as altering where the inlet was placed and verified the inlet temperature was lower than the bulk LAr, and that with the inlet positioned higher in the cryostat, that the purity stratification was greatly minimized over the previous run.

58

Anselmo Cervera Villanueva, IFIC-Valencia

RTDs on PrMs

59

CFD parameters

Anselmo Cervera Villanueva, IFIC-Valencia

LAr surface temperature

• It changes the overall absolute scale but not the shape
• We can estimate it for a given pressure
• Compatible to sensor closest to LAr surface (sensor not absolutely

calibrated)

61

87.596 K 1.045 bar

https://lar.bnl.gov/properties/basic.html#phase

measured in PD estimated

87.645 K

measured in PD by sensor closest to surface (7.385 m)

Anselmo Cervera Villanueva, IFIC-Valencia

GAr pressure and LAr level

• Since CFD simulations are computationally expensive we should

minimise the number of parameters to vary:

• Gas Pressure: it only affects the ullage, since overall LAr temperature

scale is given by LAr surface temperature

• LAr level: it just produces a shift on the last sensors. Should be easy to

parameterise

• Those two parameters change with time, so if we want to compare

different periods and avoid redoing the simulation it is better to parameterise their effects

62

Anselmo Cervera Villanueva, IFIC-Valencia

Heat flux

• Current nominal value 5.76 w/m2 has been estimated by

M.Chalifour. We need to understand how good this estimation is

• Not an input to simulation, it could be extracted from the

simulation and compared with the estimation (5.76 w/m2)

• The heat flux in the simulation is the result of simulating the cryostat

insulation and membrane

• CFD takes into account ambient temperature but it depends very little
• The insulation properties can be varied to control the heat flux

63

Anselmo Cervera Villanueva, IFIC-Valencia

Vapor being drawn

• Current nominal value 5-7 gr/sec has been estimated by

M.Chalifour. We need to understand how good this estimation is

• As far as I understood this affects only the gas phase
• Is this an input parameter of an output of the simulation ?

64

Anselmo Cervera Villanueva, IFIC-Valencia

Next steps

• The plan is to identify the key parameters and make several

simulations varying one parameter at the time:

• LAr flow rate
• LAr inlet temperature

65

Parameter Input or Output of CFD Comments Nominal values Cryostat height Input Fixed 7.878 m LAr surface height Input Parameterised 7.406 m LAr surface temperature Input Parameterised 87.596 K ullage pressure Input Does not affect liquid (overwritten by surface temperature) 1.045 bar LAr inlet temperature Input Varied

• utlet temp + 0.2 K

LAr flow rate per pipe Input Varied 0.4170025 kg/s Heat flux Output 5.76 W/m^2 vapor being drawn from the chimneys Output 5-7 gr/sec

table not up to date current inputs

inner-outer

Anselmo Cervera Villanueva, IFIC-Valencia

Inner-outer temperature

67

5 K 12 mk

LAr temperature ambient temperature Inner GAr pressure ambient pressure

3 mbar 20 mbar 9.1 mK / mbar 4 mK/mbar 2.75 mK/K

Anselmo Cervera Villanueva, IFIC-Valencia

68

Anselmo Cervera Villanueva, IFIC-Valencia

69

Anselmo Cervera Villanueva, IFIC-Valencia

70

CFD simulations

Anselmo Cervera Villanueva, IFIC-Valencia

72

• lab. calibration

static

Anselmo Cervera Villanueva, IFIC-Valencia

CFD predictions

73

Static Dynamic

pumps off pumps off pumps on pumps on

OLD SOLUTION (INLET TEMP = OUTLET TEMP + 0.2K) NEW SOLUTION (INLET TEMP = SURFACE TEMP + 0.2K) @ 30,000 iterations NEW SOLUTION (INLET TEMP = SURFACE TEMP + 0.2K) @ 50,000 iterations

ProtoDUNE-SP data

Anselmo Cervera Villanueva, IFIC-Valencia

Static-dynamic

• Reasonably stable in time (from

2 to 10 mK from bottom to top)

• Larger gradient for static
• colder at the bottom and hotter

at the top

75

0,541 m 5,615 m 6,677 m 2,075 m 1,603 m

May March

3,727 m

T Dispersion (mK)

0,0 5,0 10,0 15,0 20,0 25,0

Height (m)

0,56 1,59 2,08 5,14 6,67 7,16 7,27

Anselmo Cervera Villanueva, IFIC-Valencia

Pipes sensors: pumps-off

• ΔΤ for two pipe sensors at same height -> Dispersion in time of 3

mK for sensors separated 3 m

76

• 2,0
• 1,5
• 1,0
• 0,5

0,0 0,5 1,0 1,5 2,0

October December February April

• 5,0
• 4,0
• 3,0
• 2,0
• 1,0

0,0 1,0 2,0 3,0 4,0 5,0 October December February April

dynamic static

1 mK dispersion 3 mK dispersion 1 mK RMS 3 mK RMS

February, 25 h

Anselmo Cervera Villanueva, IFIC-Valencia

Pipes-Profiler: pumps-off

• ΔΤ for sensors at the same height in T-gradients and pipes ->

Dispersion in time of 3 mK for sensors separated 2 m

77

• 5,0
• 4,0
• 3,0
• 2,0
• 1,0

0,0 1,0 2,0 3,0 4,0 5,0

October December February April

• 5,0
• 4,0
• 3,0
• 2,0
• 1,0

0,0 1,0 2,0 3,0 4,0 5,0 October December February April

dynamic static

6 mK dispersion 3 mK dispersion 3 mK RMS 3 mK RMS

February, 25 h February, 25 h

Anselmo Cervera Villanueva, IFIC-Valencia

Bottom horizontal line

• ΔΤ between second static sensor (0,42 m) and other sensors at

similar height in pipes and dynamic profiler

78 dynamic static

5 mk

• ne month with pumps on & off (Feb)

2.2 mk

• ne month pumps-on (May)

pumps-on

ΔT (mK)

0,0 4,0 8,0 12,0 16,0 20,0

sensor

black red green blue

reference

Anselmo Cervera Villanueva, IFIC-Valencia

Pipe middle sensors

• Temperature decreases beam downstream
• Proximity to pump ?
• Along the pipes temperature is higher in

the middle of the TPC

79 dynamic static reference

ΔT (mK)

• 25,0
• 20,0
• 15,0
• 10,0
• 5,0

0,0

x position

• 2

2 4

pumps-off pumps-off

• 5 mk
• 5 mk

Anselmo Cervera Villanueva, IFIC-Valencia

Hypothesis

• Warm LAr stream passes closer to middle sensors
• The warm LAr stream is interrupted by the pump and this is why

beam downstream sensors are colder

80

ΔT (mK)

• 25,0
• 20,0
• 15,0
• 10,0
• 5,0

0,0

x position

• 2

2 4

inlet

• utlet

(pump)

Anselmo Cervera Villanueva, IFIC-Valencia

Pumps inertia

• It takes a while to stabilise after pumps are

switched off

• Sensor indicated with an arrow (closest to

dynamic T-Gradient) quite insensitive to pumps, as the dynamic T-gradient it-self

81

pumps-on

dynamic

Anselmo Cervera Villanueva, IFIC-Valencia

All pipe sensors

82 dynamic static reference dynamic static reference compatible with temperature decreasing when approaching the pump

Anselmo Cervera Villanueva, IFIC-Valencia

Top sensors

• First serious look at top sensors
• RMS is larger in the half closer to the pump
• RMS seems to be lower near the cathode

83

May, 10 minutes

work in progress

Anselmo Cervera Villanueva, IFIC-Valencia

Top sensors

• First serious look at top sensors
• RMS is larger in the half closer to the pump
• RMS seems to be lower near the cathode

83

May, 10 minutes

work in progress

Anselmo Cervera Villanueva, IFIC-Valencia

Top sensors

• First serious look at top sensors
• RMS is larger in the half closer to the pump
• RMS seems to be lower near the cathode

83

May, 10 minutes

work in progress

Anselmo Cervera Villanueva, IFIC-Valencia

Top sensors

• First serious look at top sensors
• RMS is larger in the half closer to the pump
• RMS seems to be lower near the cathode

83

May, 10 minutes

work in progress

Anselmo Cervera Villanueva, IFIC-Valencia

Top sensors

• First serious look at top sensors
• RMS is larger in the half closer to the pump
• RMS seems to be lower near the cathode

83

May, 10 minutes

work in progress

GP-1 GP-5

Anselmo Cervera Villanueva, IFIC-Valencia

Lessons learned II

• Top sensors should be further away from LAr surface to avoid large

fluctuations and better represent the behaviour in the active volume

84

DUNE-SP

Anselmo Cervera Villanueva, IFIC-Valencia

Temperature sensor distribution

86

S S S 24 24 24 24 24 24 8 S 2 9 9 2 9 9 S S S 24 24 24 24 24 24 1 6 6 144 1 8 8 1 8 8

Top/bottom individual temperature sensor Floor individual temperature sensor Wall sensor array with 13 sensors (one every three corrugations) Static T-gradient monitor with 48 sensors Dynamic T-Gradient monitor with 48 sensors Vertical SS string DSS port used for temperature sensors Individual sensors associated to a DSS port (same for top and bottom) Number of cables in each port (black for T-Gradient, red for top, blue for bottom, green for floor/wall) Gas purge pipes Liquid argon return pipes

D D 48 W 13 2 6 6 244 277 166 W W 13

Anselmo Cervera Villanueva, IFIC-Valencia

DSS ports

• There are 120 DSS ports (in orange)
• In principle all of them available to extract cables from

instrumentation devices

87

DSS ports

Anselmo Cervera Villanueva, IFIC-Valencia

DSS ports

• DSS ports will have a side port for instrumentation cables
• Baseline is CF40 (40 mm outer diameter tube)
• DSS engineers suggest circular connector (e.g. 32 pins)
• Port could be larger CF63 or even inclined

88

Anselmo Cervera Villanueva, IFIC-Valencia

DSS ports

89

13.8 mm 6.35 mm

In principle use space between port tube and DSS tube to route cables But the available space for the two

• ptions being

discussed seems not sufficient If this is finally the case we will have to perforate the DSS tube to put cables

• inside. Is this safe

considering that DSS rod will move ?