[PPT] - NuMI-NOvA Target & Window K. Ammigan 10 th International PowerPoint Presentation

SLIDE 1

K. Ammigan

10th International Workshop on Neutrino Beams and Instrumentation 20 September 2017

NuMI-NOvA Target & Window

SLIDE 2

Outline

9/21/17

K. Ammigan | NuMI – NOvA Target and Window, NBI 2017

2

§ Introduction § Target design history

§ IHEP § FNAL & STFC/RAL

§ Target analysis, testing and QA

§ Target fins, windows, baffle

§ Operational experience

§ Protons on target & DPA § Window failure

§ Target autopsy and R&D

SLIDE 3

ME target for NOvA experiment

9/21/17

K. Ammigan | NuMI – NOvA Target and Window, NBI 2017

3

Proton-graphite interactions to produce pions, which are focused before decaying to

produce neutrinos

Long/narrow target to allow pions to escape out sides and reduce pion re-interactions
50 graphite fins: 24 mm long and 7.4 mm wide
Beam sigma: 1.3 mm
Helium atmosphere enclosed by US/DS beryllium windows
Heat removal: Al pressing plates and water cooled outer can

Design parameters (**center of peak fin) Beam energy (per proton) 120 GeV POT/10 µs spill 4.9e13 Repetition time 1.33 sec Proton beam power 700 kW Peak max. Edep. per spill ** 310 J/g Peak max. power deposition ** 235 W/g Instantaneous power during spill ** 30 MW/g

J. Hylen (FNAL)

SLIDE 4

IHEP target design I

IHEP 2006 report 9/21/17

K. Ammigan | NuMI – NOvA Target and Window, NBI 2017

4

Initial design and study of the Medium Energy Target (MET) was performed by

IHEP (Protvino, Russia) in 2006

Designed for 440 kW beam: 120 GeV beam, 4 x 1013 ppp, repetition time: 1.9 s
12 x 100 mm long, 3.2 mm wide graphite plates

SLIDE 5

IHEP target design II

IHEP 2008 report 9/21/17

K. Ammigan | NuMI – NOvA Target and Window, NBI 2017

5

Design revised by IHEP in 2008 for higher proton beam power
Designed for 880 kW & 1440 kW beam: 120 GeV beam, 5.5 x 1013, 9.0 x 1013 ppp

every 1.2 s Main changes

Target core with 48 graphite fins (24.5 mm L and 6.4 mm W): decrease quasi-static/dynamic

stresses and improve fin manufacturability

Charge read-out (Budal) target position monitoring fins added upstream of the target
Water cooling system developed for target casing to reduce heating from secondary particles

SLIDE 6

FNAL/RAL target design tweaks

9/21/17

K. Ammigan | NuMI – NOvA Target and Window, NBI 2017

6

Design finalized by FNAL (M. McGee) and STFC/RAL (C. Densham) target groups
Mainly to improve fabrication/operation of components

48 POCO ZXF-5Q fins DS Be window Cooling water inlet/outlet Budal monitors Water cooled clamping plates US Be window He filled target canister Canister cooling water inlet/outlet

SLIDE 7

NOvA targets

9/21/17

K. Ammigan | NuMI – NOvA Target and Window, NBI 2017

7

MET-01 manufactured by STFC/RAL (used until failure) MET-02 manufactured by Fermilab (currently installed) MET-03 manufactured by STFC/RAL, modifications, 1) internal welds of outer vessel removed, 2) outer vessel anodized MET-04 manufactured by Fermilab

M. Fitton, (STFC/RAL)

SLIDE 8

Fins heat transfer performance

M. Fitton (STFC/RAL)

9/21/17

K. Ammigan | NuMI – NOvA Target and Window, NBI 2017

8

Experimental tests using thermal camera imaging confirmed consistent thermal contact across all

graphite fins

Cooling issue with horizontal Budal monitor fin was identified

310°C 152°C

Horizontal Budal bracket redesigned by thickening material and shortening conduction path. Simulation shows operating temperature reduced by ~50%

SLIDE 9

Target temperature simulations

T. Davenne (STFC/RAL)

9/21/17

K. Ammigan | NuMI – NOvA Target and Window, NBI 2017

9

Estimate of target fin operating temperatures and heat to outer vessel ANSYS CFX simulations of natural convection and thermal radiation inside target vessel

Beam Energy: 120 GeV Gaussian beam sigma: 1.3 mm in x & y Position: 4mm from top of fins. Protons per bunch: 4.9e13 Rep rate: 1.33s

Total heat load = 7482 W
Heat conduction from fins to

cooling rail = 6867 W

Heat radiation from fins to
uter vessel = 476 W
Convection and conduction

through helium from fins to

uter vessel = 139 W

Heat balance

SLIDE 10

Target fin stress calculations

T. Davenne (STFC/RAL)

9/21/17

K. Ammigan | NuMI – NOvA Target and Window, NBI 2017

10

POCO ZXF-5Q steady state temperature and stress

Tpeak= 672 °C Peak stress = 4.7 MPa

Largest temperature gradient near cooling block

Peak dynamic stress during beam pulse: ~26 MPa ZXF-5Q tensile strength: 79 MPa

SLIDE 11

Welding development

STFC/RAL 9/21/17

K. Ammigan | NuMI – NOvA Target and Window, NBI 2017

11

Water cooled vessel of NOvA target

Design modified after MET-01 so there are no internal welds

MET-03, only 2 external welds

SLIDE 12

Welding development

STFC/RAL 9/21/17

K. Ammigan | NuMI – NOvA Target and Window, NBI 2017

12

High quality welds (NAS 1514 Class II or better) in Al 6061 proved very difficult to produce consistently for plug weld

Weld looks very good by visual inspection Radiography reveals significant porosity

Parameters explored to optimize weld

§ Cleaning (including etching) § Filler wire selection § Pre-heating § AC current/balance § Gas (Helium/Argon blend)

SLIDE 13

Welding development

STFC/RAL 9/21/17

K. Ammigan | NuMI – NOvA Target and Window, NBI 2017

13

Original design – 9 Aluminum welds Modified design – 3 Aluminum welds

Remove as many of the Aluminum welds as possible from the water cooling circuit

AL-SS roll bonded plate replaced with friction welded blanks Blind gun drilled holes, with single cross drill Pipe stubs machined into rail

SLIDE 14

Welding radiography QA

STFC/RAL 9/21/17

K. Ammigan | NuMI – NOvA Target and Window, NBI 2017

14

Achieved NAS-1514, Class 1 certification for cooling rail

SLIDE 15

Hydrostatic pressure testing

STFC/RAL 9/21/17

K. Ammigan | NuMI – NOvA Target and Window, NBI 2017

15

Testing for leaks and weakness at much

higher pressures than operational pressure

Cooling rail tested to 200 psig
Outer vessel tested to 90 psig

SLIDE 16

Helium leak check

STFC/RAL 9/21/17

K. Ammigan | NuMI – NOvA Target and Window, NBI 2017

16

All components and complete assembly helium leak checked Specification: total leak rate <1 x 10-9 atm.cm3/s

SLIDE 17

Target windows

M. McGee (FNAL)

9/21/17

K. Ammigan | NuMI – NOvA Target and Window, NBI 2017

17

PF-60 Beryllium US and DS windows
US window: Ø 44.45 mm x 0.254 mm thick
DS window: Ø 135 mm x 1.25 mm thick
Designed to withstand 15 psig without beam and 3 psig with beam (during operation)

US window

US window reaches thermal equilibrium after 30 s (23 pulses) with Tmax = 66 °C

Figure (a) US 0.25 mm thick window structural results with 15 psig load (max equiv. stress: 223 MPa), (b) With 3 psig He (maximum operational) load (max equiv. stress: 87.8 MPa)

(a) (b)

SLIDE 18

Target windows

M. McGee (FNAL)

9/21/17

K. Ammigan | NuMI – NOvA Target and Window, NBI 2017

18 DS window

DS window reaches thermal equilibrium after 380 s (286 pulses) with Tmax = 67.3 °C (a) (b)

Figure (a) DS 1.25 mm thick window structural results with 15 psig load (max equiv. stress: 211 MPa), (b) with 3 psig He (maximum operational) load (max equiv. stress: 100 MPa)

Both US and DS window maximum Von Mises stress occurs at the edge of the window,

< 224.1 MPa defined by ½ the UTS (FS: 2) and fatigue limit (> 107 cycles) of 268 MPa.

Expected maximum load is 3 psig (not 15 psig), after considering fluctuations in

external barometric pressure conditions, internal pressure control and gas heating from beam

Max. Von Mises stress of 87.8 MPa and 100 MPa for US and DS window, respectively
Be yield strength ~250 MPa , tensile strength ~450 MPa

SLIDE 19

Target baffle (IHEP design)

IHEP technical design report (2002) 9/21/17

K. Ammigan | NuMI – NOvA Target and Window, NBI 2017

19

Requirement: operate with continuous 3% beam scraping
Consisted of graphite slugs inserted in Al sleeve with clamped pin radiators

NuMI 400kW baffle design (IHEP) Thermal analysis with NOvA beam parameters

Peak temperature of 190 °C in Aluminum sleeve (3% beam scraping)
Need to keep below 160 °C to avoid Al ageing

SLIDE 20

Target baffle (new design)

FNAL 9/21/17

K. Ammigan | NuMI – NOvA Target and Window, NBI 2017

20

Re-designed Al tube with 4 mm thick wall with cooling fins machined into tube
Eliminate Grafoil interface and hose clamps
Shrink fit at 150 °C with interference fit of 0.003”
Al ID: 57 mm, OD: 65 mm
18 fins at 20° spacing
Fin height: 20 mm, fin width: 2 mm
Graphite slug baffle hole diameter: 13 mm

Temperature distribution in baffle with 3% beam scraping

Graphite Tmax: 157 °C
Aluminum Tmax: 141 °C

Very conservative analysis

3% beam scraping
Low air velocity over baffle and convective heat

coefficient assumed Tmax: 157 °C

SLIDE 21

Target baffle operation

K.E. Anderson (FNAL) 9/21/17

K. Ammigan | NuMI – NOvA Target and Window, NBI 2017

21

TA-02 baffle (IHEP design) thermocouple readings DS end of Al containment tube

Beam power: ~680 kW
Al tube temperature: 60 – 65 °C
Much lower than predicted temperatures
Beam scraping < 1%, higher convective cooling

Target assemblies starting with TA-04 will use new fin-style baffle

SLIDE 22

Target operation history

9/21/17

K. Ammigan | NuMI – NOvA Target and Window, NBI 2017

22

MET-01 lasted 3 years and received the most POT than any other target, with no

noticeable degradation in neutrino yield

Failure due to helium leak developed in DS Be window
MET-02 currently in service
3 of the 50 graphite target fins were replaced by Beryllium as part of target R&D
J. Hylen (FNAL)

SLIDE 23

DS Be window leak (MET-01)

9/21/17

K. Ammigan | NuMI – NOvA Target and Window, NBI 2017

23

§ MET-01 DS Be window helium leak developed during operation § Located at the edge of window near electron beam welded joint

Gap

CT scan images of spare Be window of same design confirmed that there is a gap between the Be foil and Al 6061 flange on some sections

Dissimilar thickness at the EBW joint

SLIDE 24

DS Be window redesign

Y. He (FNAL)

9/21/17

K. Ammigan | NuMI – NOvA Target and Window, NBI 2017

24

Involves 2-step EB-welding

Now using 5000 series aluminum (more compatible than Al 6061-T6 for Be-Al EBW)
Improved weld prep: welding grooves allows material to fully melt before the heat is transferred

away in the flange

Sandwich structure provides protection/support of the Be-Al EBW joint for both vacuum and

internal pressure conditions 1st EB-welding (Be to Al) 2nd EB-welding (Al to Al) Be foil, PF-60

1.25 mm thick

Al 5052

SLIDE 25

Beryllium fins in MET-02

FNAL 9/21/17

K. Ammigan | NuMI – NOvA Target and Window, NBI 2017

25

Be fins

Target currently in operation and will run

until failure (5e20 POT so far)

Plan to autopsy target and perform PIE to

evaluate Be fin conditions relative to graphite

Microstructural analyses and possible

macro and micro-mechanical tests

R&D to evaluate Beryllium (S-65F) as a potential target material to increase target lifetime

Beryllium known to have better radiation damage resistance than graphite
However, fatigue life may be an issue due to higher dynamic stresses in Be

Beryllium and graphite dynamic stress comparison

Parasitic test with three Be fins in target
1 fin at location of highest stress
2 control fins at lower stress regions
T. Davenne (STFC/RAL)

ZXF-5Q, UTS:79 MPa, E:14.5 GPa IG-430, UTS: 37 MPa, E: 10.6 Gpa Be, UTS: ~450 MPa, E: 303 GPa

SLIDE 26

NT-02 target showed neutrino yield degradation (10-15%) towards end of its life
POT: 6.5e20 , DPA ~ 0.7
But MET-01 target showed no significant degradation
POT: 1.1e21, DPA ~ 1.15

NT-02 vs. MET-01 performance comparison

9/21/17

K. Ammigan | NuMI – NOvA Target and Window, NBI 2017

26 Fracture surface EDS (PNNL) Fracture surface fractography (PNNL) Cracks along centerline

NT-02 autopsy & PIE

PIE work at PNNL showed bulk swelling and suggested crack initiated from center of fins along beam axis

Higher operating temperature of MET-01 possibly provided annealing of

radiation damage

60°C vs. 600 °C
Fins in MET-01 are only constrained at the bottom end faces and region

around beam spot are more free to expand

Failure or cracks more likely to develop in top corners of fins, away from

beam axis

Plan to autopsy MET-01 and perform PIE next fiscal year

SLIDE 27

Thank you for your attention

Questions?

9/21/17

K. Ammigan | NuMI – NOvA Target and Window, NBI 2017

27

SLIDE 28

Back-up Slides

9/21/17

K. Ammigan | NuMI – NOvA Target and Window, NBI 2017

28

SLIDE 29

Metrology

STFC/RAL 9/21/17

K. Ammigan | NuMI – NOvA Target and Window, NBI 2017

29

All components and assemblies carefully measured using Coordinate Measuring Machine (CCM)

Pressing plate (4.5 µm flatness)

SLIDE 30

Components alignment

STFC/RAL 9/21/17

K. Ammigan | NuMI – NOvA Target and Window, NBI 2017

30

Clocks used to ensure outer vessel has not moved during adjustment

Target rail aligned in outer vessel using large CMM and long probe

SLIDE 31

NOvA welding procedure

9/21/17

K. Ammigan | NuMI – NOvA Target and Window, NBI 2017

31

CLEANING

Applied to all parts and welding rods

Degrease using Trichloroethylene (TCE) a vapor solvent for organic materials
Alkaline bath w/ pH 11-11.5 (for 5 min)
Tap Water Rinse (ambient) to remove residual alkaline solution (for 5 min)
Nitric Acid Etch 38-48% Nitric, 0.7-0.9% HF-3 (for 2.5 min)
DI water Rinse (for 5 min) at ~ 90 F
Oven dry at 130 F (for 30 min)

WELDING

AC TIG (HF Impulse) Welding
Plug Weld Parameters:

Pos electro +170/ Neg. electro -180, balance 80% & 130 Hz frequency

Socket Weld Parameters:

Pos electro +100/ Neg. electro -120, balance 80% & 130 Hz frequency

Tungsten cup #5, flow 25 – 30 scfm (75% He/25% Ar) gas purge

SLIDE 32

9/21/17 Kavin Ammigan | Preliminary Results of the NT-02 graphite fin PIE at PNNL 32

Operation between 2006 to 2009, and

again in 2011

Subjected to 120 GeV protons

– Integrated POT ~ 6.1 x 1020 – Gaussian beam spot size (1σ): 1.1 mm – Peak fluence: 2.5 x 1022 p/cm2 – Estimated DPA ~ 0.63

Peak temperature ~ 330°C

– Heat to 330 C in 10 µs, cool to 60 C before next pulse (1.85 s cycle time)

Neutrino yield declined 10-15% during life

– Yield reduction not observed in other NT targets – NT-02 lifetime significantly longer than any

ther NT targets (2x or more)

NuMI NT-02 target

SLIDE 33

Swelling data

9/21/17 Kavin Ammigan | Preliminary Results of the NT-02 graphite fin PIE at PNNL 33

US Half Fin US Full Fin DS Half Fin DS Full Fin

Avg. End Thickness (mm)

6.54 6.57 6.55 6.55

Avg. Middle Thickness (mm)

6.67 6.64 6.60 6.57 Relative Swelling (%) (Middle-to-end) 2.0 1.1 0.7 0.2 Absolute Swelling (%) (Middle-to-Reference*) 4.3 3.8 3.1 2.6

*Reference Thickness = 6.4 mm

Amount of swelling consistent