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

LAr Instrumentation Scope Review CERN, 19/06/2019 Temperature monitors overview (with focus on Static T-gradient thermometers) A. Cervera M.A. García Peris (IFIC-Valencia) With assistance from C.Streff (SDSU), E. Voirin (FNAL), A. Hahn (FNAL), S. Pordes (FNAL)

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 su ffi ciently 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 Cryostat Cryogenics system LAr impurities Insulation LAr recirculation Vacuum tightness Physics Drift velocity, flow (ions), electron lifetime, energy calibration Electron lifetime prediction far from PrMs, using CFD The system will help in mitigating risks at reasonable cost (< 300 K$/10-kt) simulations validated with temperature measurements 3 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? Even if the problem cannot be solved, a 3D temperature map gives us a handle for o ffl ine corrections (next slide) 4 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 Charge Question 3 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) ? Yes. CFD simulations, tuned with temperature data, will be used to predict the electron lifetime everywhere in the TPC and make offline corrections 5 Anselmo Cervera Villanueva, IFIC-Valencia

CFD simulations 35t prototype Purity Stratification study actually triggered by precise temperature measurements • Those simulations also predict • CFD successfully used in the the 25-35 mK gradient observed 35t prototype to reproduce between RTDs in PrM1-3 the observed e- lifetime Erik Voirin Erik Voirin Produced yesterday (thanks!!!) Dune DocDB 1156 PrM1 PrM1 T1 T2 ~30 mk PrM2 T3 PrM2 PrM3 PrM3 PrM4 PrM4 e - lifetime temperature 35t prototype showed direct relation between Temperature and e-lifetime 6 Anselmo Cervera Villanueva, IFIC-Valencia

CFD simulations 35t prototype Purity Stratification study actually triggered by precise temperature measurements • Those simulations also predict • CFD successfully used in the the 25-35 mK gradient observed 35t prototype to reproduce between RTDs in PrM1-3 the observed e- lifetime Erik Voirin Erik Voirin Produced yesterday (thanks!!!) Dune DocDB 1156 PrM1 PrM1 T1 T2 ~30 mk PrM2 T3 PrM2 PrM3 PrM3 PrM4 PrM4 e - lifetime temperature 35t prototype showed direct relation between Temperature and e-lifetime 6 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 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 7 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 ~5 cm 100 $ <5 $ 1.4 cm precision sensor standard sensor in static T-gradient on wall 15 mk 1.5 mk evolution during one day 8 Anselmo Cervera Villanueva, IFIC-Valencia

Motivation of each system System Precision Motivation Vertical array Top/bottom •check cryogenics system behaviour ~mK Horizontal •constrain CFD simulations for e-lifetime prediction 2D grid RTDs at inlets and pumps Floor 20-50 mK monitor the presence of LAr when filling starts Monitor cryostat membrane temperature during cool- Walls & roof 20-50 mK down and filling Vertical array •constrain & validate CFD simulations in ullage 20-50 mK in gas •understand behaviour of mechanical structures (DP) 9 Anselmo Cervera Villanueva, IFIC-Valencia

Description of ProtoDUNE-SP systems

Standard sensors 5 sensors 12 sensors

Floor/wall sensors Wall sensors Floor sensors standard RTD mounted on PCB support with IDC-4 connector epoxied to wall epoxied to floor before removing temporary tape Aug 10th temperature Temperature evolution when LAr filling started 2 hours time 12 Anselmo Cervera Villanueva, IFIC-Valencia

Precision sensors 8 sensors 48 sensors 24 sensors dynamic static 12 sensors

Crucial to obtain mK precision Multiplexing board for 24 sensors (with 1 mA current source) X. Pons (CERN) 14 Anselmo Cervera Villanueva, IFIC-Valencia

Top/bottom sensors Top sensors Bottom sensors attached to ground planes attached to LAr pipes LAr pump dynamic LAr inlets static 15 Anselmo Cervera Villanueva, IFIC-Valencia

T-gradient monitors static dynamic # sensors 48 24 Δ z top/bottom 12 cm 10 cm Δ z middle 24 cm 50 cm calibration method laboratory in-situ (movable system) calibration precision 3 mK < 1 mK dynamic static Dedicated talk by Jelena this afternoon 16 Anselmo Cervera Villanueva, IFIC-Valencia

Calibration https://indico.fnal.gov/event/16764/session/17/contribution/157/material/slides/0.pdf • 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 Anselmo Cervera Villanueva, IFIC-Valencia

Temperature fluctuations • <2 mK below 6.5 m • Increase dramatically near the LAr surface May, 10 minutes static dynamic GPs GPs 18 Anselmo Cervera Villanueva, IFIC-Valencia

Vertical profiles • static and dynamic cross-calibrated to each other with pumps-o ff • Gradients below 15 mK • Larger gradient in static profiler: colder at the bottom, warmer at the top Vertical gradient depends on location dynamic static movable calibration pumps-o ff calibration 19 Anselmo Cervera Villanueva, IFIC-Valencia

Vertical profiles • static and dynamic cross-calibrated to each other with pumps-o ff • Gradients below 15 mK • Larger gradient in static profiler: colder at the bottom, warmer at the top Vertical gradient depends on location dynamic static movable calibration pumps-o ff calibration 15 mk 8 mk 19 Anselmo Cervera Villanueva, IFIC-Valencia

CFD predictions SDSU simulations • We have tested a couple of configurations: • T inlet = T outlet + 0.2 K • T inlet = T LAr surface + 0.2 K (closer to reality) • Still work to do dynamic static movable calibration pumps-o ff calibration 20 Anselmo Cervera Villanueva, IFIC-Valencia

Stability of vertical profiles • ΔΤ between first sensor and other sensors in the same profiler over one month Shape of profiles stable within 2-3 mK dynamic static 6,205 m 7,267 m 5,615 m 6,159 m 4,671 m 5,651 m 3,111 m 3,019 m 4,635 m 1.587 m 1,603 m 21 Anselmo Cervera Villanueva, IFIC-Valencia

Bottom sensors March May Pumps-o ff cathode LA inlets dynamic LAr pump static • Δ T between all pipe sensors at 2nd static sensor, all at ~40 cm dynamic • Temperature decreases towards the LAr pump • Along the pipes (X) temperature is LAr higher in the middle of the TPC pump static 22 Anselmo Cervera Villanueva, IFIC-Valencia

CFD predictions • CFD predicts decreasing temperature towards pump, but the e ff ect is not as large as in data • CFD does not predict higher temperature in the cathode region CFD at 40 cm (SDSU) data dynamic 4 mk ~20 mK FC cathode pump static 23 Anselmo Cervera Villanueva, IFIC-Valencia

CFD predictions E. Vorin’s simulations DocDB-928-v3 • This simulation does predict larger temperature below the cathode • It also shows: • lower temperature CFD at 2nd LAr pipe plane dynamic static towards pump • higher gradient at static profiler NOTE : map exists outside TPC but color range not appropriate pump pipe sensors 24 Anselmo Cervera Villanueva, IFIC-Valencia