Develop Radiation Hard Beam Monitor and Muon Spectroscopy by using Machine Learning for Intense Neutrino Target System Katsuya Yonehara CPAD Workshop 12/08/2019
Fermilab Intensity Upgrade Plan • NuMI-AIP (Neutrinos at the Main Injector – Accelerator Improvement Plan) – Upgrade existing Fermilab accelerator complex with the same footprint to increase proton beam intensity on the NuMI target from 780 kW to > 900 kW – Machine operation starts from 2020 • LBNF (Long Baseline Neutrino Facility) – Apply PIP-II SRF Linear Accelerator to send 1.2 MW beam to the LBNF target – Machine operation will start from 2029 – Extend to PIP-III SRF Linac to reach 2.4 MW beam power – Operation year TBD 2 12/8/19 CPAD, Yonehara
Fermilab Accelerator Complex 3 12/8/19 CPAD, Yonehara
Beam Monitor for multi-MW Target System • Tolerance of the target parameter at LBNF – Tighter than NuMI • Beam monitor is a real-time (spill-by-spill) detector to check quality of multi-MW target system – High reliability and long lifetime (rad hard) required 4 12/8/19 CPAD, Yonehara
Develop Rad-Hard Beam Monitor System NuMI Target system Hadron monitor (0.8x0.8 m 2 , 7x7 pixels) Thermocouple detector (3+3 Be wires) Muon monitor Multi-pixel ionization chamber (2x2 m 2 , 9x9 pixels) 5 12/8/19 CPAD, Yonehara
Beam Monitor for Beam Based Alignment • Target beam elements were occasionally displaced or broken by various incidents – Radiation damage, thermal expansion, thermal shock, water leak, Helium gas leak, etc • Beam based alignment permits us to find baffle, target and horn positions w.r.t. the BPM coordinate by using beam monitors • Position resolution less than 0.2 mm is achieved Layout of beam based alignment Hadron Monitor Horn 2 Horn 1 Beam position and angle are measured by two bpms • Beam position on target is observed by thermocouple sensor • Scan requires a special beam condition, but it takes less than hour • 6 12/8/19 CPAD, Yonehara
Upgrade Beam Monitor for 1-MW operation • Develop rad-hard ionization chamber • Observed signal gain change by varying He gas quality – Calibration chamber can calibrate the gain change due to gas quality, but this is not the perfect solution – Apply a new gas system • Density flow control by using PLC • Add bubbler on the outlet of HM to avoid backflow • Use a radiation hard material – Apply radiation hard ceramics for insulator and cable • Optimize the dimension of monitor system – Beam profile simulation – Space charge simulation 7 12/8/19 CPAD, Yonehara
Particle Tracking in Simulation Horn 1 Horn 2 Target Proton beam spot size 1.5 mm Horn 2 Horn 1 Beam profile on hadron monitor Shows Aberration of horns 8 12/8/19 CPAD, Yonehara
Cavity body Alternate Hadron Monitor Waveguide • RF beam detector • Conceptually new rad-hard beam detector Charged particles • Apply RF field to measure the amount of ionization gas plasma which is proportional to the intensity of charged particles passing through a RF cavity by measuring gas permittivity change 𝜁 = 𝜁 ! + 𝑗𝜁 " • Proof-of-principle test was carried out by using the Main Injector 120 GeV • Beam intensity = 1.3e13 proton beam • Detector filled with ambient air 120 100 FY19 Voltage in cavity ( V ) 80 60 40 2.4 GHz RF test cavity 20 Fabricated in MI-40 Beam is turned off Beam is turned on 0 abort room 0.000000 5. × 10 - 6 0.000010 0.000015 Time ( s ) Five peaks during the beam on shows the gap of six MI beam batches 9 12/8/19 CPAD, Yonehara
Linearity of RF beam detector 0.020 1-atm N 2 , V 0 = 120 V FY19 𝑞 = 𝑊 # 𝑊 # − 𝑊 𝑢 𝑆 RF power consumption ( J ) 0.015 V 0 2 1 = V 0 r , a i y d r m t a - 2 0.010 V 0 2 1 = V 0 , e H m t a - 1 0.005 2-atm N 2 , V 0 = 60 V 0.000 1 2 3 4 Beam Intensity ( 1e12 protons / spill ) 10 12/8/19 CPAD, Yonehara
Muon Monitor • Three monitor receive different energy muons • Similar structure as Hadron monitor Muon Monitor 1 signal 11 12/8/19 CPAD, Yonehara
Systematic measurement Strong linear correlation between Horizontal scan primary proton beam and muon beam centroid on Muon Monitors MM1 FY19 MC simulation MM2 shows opposite slope from MM1 due to Aberration of horns MM2 FY19 MC simulation 12 12/8/19 CPAD, Yonehara
Pion/Muon Spectroscopy Magnetic horns have an analyzing power MC simulation X1 X9 Select pixel on X-axis (y = 0) • Individual pixel sees different muon spectrum • X1 & X9, X2 & X8, X3 & X7, X4 & X6 shows similar shape as expected 13 12/8/19 CPAD, Yonehara
Predicted Horn Current by using Machine Leaning 𝑆 !! = 𝑔 ⃗ 𝑠 "#$% , ⃗ 𝜏 & !"#$ , 𝐽 '(&) , 𝑏𝑡 𝑞𝑏𝑠𝑏𝑛𝑓𝑢𝑓𝑠 Training data Horn Current Error RMS = 0.152 kA Apply for real data 14 12/8/19 CPAD, Yonehara
Predicted Beam centroid on Muon Monitor with ML 𝑆 !! = 𝑔 ⃗ 𝑠 "#$% , ⃗ 𝜏 & !"#$ , 𝐽 '(&) , 𝑏𝑡 𝑞𝑏𝑠𝑏𝑛𝑓𝑢𝑓𝑠 Hor Error RMS = 0.15 mm Horizontal analysis Vertical analysis 195 190 180 𝐽 !"#$ = 200 𝑙𝐵 Ver Error RMS = 0.58 mm 15 12/8/19 CPAD, Yonehara
Summary • Study three beam monitors – Demonstrate that beam monitors is capable to operate the target system within the design tolerance – Introduce Machine Learning to make an automatic monitor system – Study Pion/Muon spectroscopy by using aberration of horns • Develop rad hard ion chamber for multi-MW target – New gas system to prevent gas contamination – Plan to simulation study to minimize space charge effect – Develop RF beam detector • Plan more R&D to make a practical detector 16 12/8/19 CPAD, Yonehara
Acknowledgement • Hadron Monitor – Karol Lang, Marek Proga from U of Texas Austin – Joe Beleski, Jodan Bohn from Fermilab • RF beam detector – Rol Johnson, Mary Anne, Grigory Kazakevic from Muons Inc – Al Moretti, Dave Peterson, Adam, Dent, Kyle from Fermilab • Muon Monitor – Pavel Snopok, Yiding Yu from IIT – Amit Bashyal from Oregon U – Athula Wickremasinghe from Fermilab • TSD – Bob Zwaska, Jim Hylen, Cory Crowley, Yun He, Keith Gollwitzer, Kris Anderson, Patrick Hurh 17 12/8/19 CPAD, Yonehara
Recommend
More recommend
