ThermalHydraulics Studies of HeliumCooled Divertors M Yoda S

Thermal-Hydraulics Studies of Helium-Cooled Divertors M. Yoda, S. I. Abdel-Khalik, B. Zhao and S. A. Musa July 28, 2016

He-Cooled Divertors • “Divert” ~20% of energy from burning plasma away from first wall onto divertor target surfaces • Target surfaces subject to very high heat fluxes (>10 MW/m 2 steady state; much higher transient values) • Helium-cooled (solid) tungsten-alloy divertors most developed divertor concept – Concepts studied by KIT (FZK), ARIES: HEMJ, HEMP, T-tube, HCFP, integrated plate-finger design – Issues with W surfaces exposed to plasma containing He: erosion/redeposition, nanobubbles, fuzz, D retention, … – Only divertor concept experimentally shown to remove >10 MW/m 2 steady-state Fusion Matls Workshop (7/16) 2

Introduction Aim • Evaluate thermal-hydraulic performance of leading He-cooled divertor designs at prototypical conditions using dynamically similar experiments at nearly prototypical conditions and numerical simulations Research Objectives • Experimentally test modules that closely mimic current divertor designs at nearly prototypical conditions • Perform numerical simulations validated by experimental data • Estimate maximum heat flux and He pumping power requirements • Determine numerically whether divertor designs can be simplified and/or improved Fusion Matls Workshop (7/16) 3

He-Cooled Multi-Jet Divertor • HEMJ: design proposed for DEMO – Jet impingement cooling: 6. 8 g/s He at (600 °C, 10 MPa) exits from 25 (24 0. 6 mm dia. + one 1. 04 mm dia. ) holes into H = 0. 9 mm gap – ~106 modules to cool O(100 m 2) divertor • KIT/Efremov experiments Norajitra et al. 15 – Single module at prototypical conditions withstood >103 cycles at 9 14 MW/m 2 – Nine-finger module at (500 °C, 10 MPa) tested to 6 MW/m 2 – Complex geometry need to develop new fabrication methods: W Tile 18 m W-alloy shell 15 mm 18 mm m Steel cartridge deep drawing thimbles, powder injection molding (PIM) tiles Fusion Matls Workshop (7/16) 4

Experimental Approach • Test at near-prototypical conditions in GT helium loop – He mass flow rate ṁ 10 g/s, inlet temperature Ti < 420 °C (vs. 600 °C), inlet pressure pi 10 MPa – RF induction heater [INL STAR facility]: heat flux q from He energy balance (less than actual incident flux due to losses) – q 6. 6 MW/m 2: decreases as Ti due to thermal losses – Match dimensionless ṁ (Reynolds number Re) and fraction of heat removed by convection, vs. conduction (Biot number Bi thermal conductivity ratio ks / k for geometrically similar studies) – Measure cooled surface and coolant temperatures, pressure drop – Determine correlations for dimensionless HTC (Nusselt number) Nu = f (Re, ) [neglecting Prandtl number effects] and loss coefficient KL = g (Re) Fusion Matls Workshop (7/16) 5

Current GT Helium Loop Induction heater • Evacuate loop, then charge to 10 MPa Test section P T T P from 41. 3 MPa source tanks • Control mass flow rate ṁ with bypass – Filters remove >7 μm particulates • Heat He with recuperator + electric heater Fusion Matls Workshop (7/16) He source tank Compressor Venturi P Recuperator meter Buffer tank Vacuum pump T Buffer tank Heater Cooler 6

HEMJ (J 1 -c) Test Section • WL 10 (99% W, 1% La 2 O 3) outer shell + stainless jets inner cartridge with adjustable gap width H q – Thermocouples (TCs) embedded 0. 5 mm from cooled surface of outer shell average cooled surface temperature Tc • Experiments at H = 0. 47 0. 03 mm, 0. 86 0. 02 mm, 1. 28 0. 04 mm – Focus on higher inlet temperatures Ti = 300 400 C – Determine average and standard deviation from 10 measurements taken with air-dry clay at a given setting • Reynolds number Re = 1. 3 104 5. 4 104 (ṁ 3 8 g/s) – Prototypical ṁ = 6. 8 g/s Rep = Fusion Matls Workshop (7/16) 2. 14 104 at Ti = 634 °C Dimensions in mm 7

Experimental Results Ti = 300 C p 400 C Nu [-] q 4. 5 MW/m 2 – Nu independent of H – Re-examine correlation at Ti 300 C Mills et al. 15 • KL 1. 8 for H = 0. 47, 0. 86 mm H = 0. 47, 0. 86, 1. 28 mm Re [/104] Fusion Matls Workshop (7/16) • Nu less than previous results at lower Ti and correlation – KL 1. 86 for 1. 28 mm – Pressure transducer recalibrated – Simulations predict KL 1. 77 8

HEMJ Optimization • Can HEMJ design be modified to simplify manufacture while keeping similar thermal performance? Curved – 25 jets of different diameters impinging on curved surface – Consider instead fewer and larger impinging jets, all of Jets same diameter, impinging on and cooling flat surface Thimble Cartridge • Numerical simulations – Coupled CFD / thermal stress analysis: model 60° wedge using standard k turbulence model – Validate with experimental data from He loop – Investigate effect of jet exit diameter D (all jets same dia. ), spacing s on thermal performance quantified in terms of Nu, maximum h, maximum Tc, thermal stresses – Thermal stress analysis: one-way coupling, with CFD results used as loads for FEM model – Test “best” result in He loop to validate simulations Fusion Matls Workshop (7/16) Flat Fluid 9

Jet Array Parameterization • Vary number of rows and jets # Rows – Total jet area, V kept constant radii [mm] # Jets 1 D [mm] 3. 11 7 1. 18 13 19 25 0. 865 0. 715 0. 624 – Only designs with projected row (s/D) and jet (j/D) spacings within ~1 – Rows spaced evenly cartridge diameter for flat surfaces – Hexagonal array of jets: 30° 120° wedges, depending on symmetry – Up to 7 106 cells • Vary H for each geometry – H = 0. 5, 0. 75, 0. 9, 1. 25, 1. 5 mm Fusion Matls Workshop (7/16) # Jets 0 1 3. 5 2 2. 2, 6. 5 3 2. 2, 4. 1, 6. 5 4 2. 2, 3. 5, 4. 8, 6. 5 7 7 13 19 25 CR 4 H 25 25 FR 4 H 25 s j 10

Preliminary Results • Results for HEMJ at prototypical conditions suggest: – Temperature differences over cooled surface as great as ~127 °C – Maximum von Mises stress on cooled surface ~388 MPa – Change in H due to diff. thermal expansion as great as 0. 22 mm (vs. 0. 9 mm) • Optimizations can achieve similar h, lower p with fewer jet holes and/or rows in some cases CFD Analysis (HEMJ) T [°C] 997 60° Structural Analysis (HEMJ) v. Mises stress [MPa] Expansion [mm] 7724 0. 318 0 0 r 870 z Re = 2. 2 104 Ti = 600 C q = 10 MW/m 2 h = 35. 3 k. W/(m 2·K) 11

New GT He Loop • Design and build larger He loop with ṁ 100 g/s – Experimental studies of other divertor designs: HCFP, 9 -finger HEMJ module – Space for loop under renovation • Biggest challenge: creating nearly prototypical heat fluxes on test section – Current induction heater cannot provide 10 MW/m 2 on large (at least 20 cm 2) areas – Try reversed heat flux approach: reverse direction of heat transfer by heating plasma-facing surface with hot He, cooling with water: estimate heat transfer and h from energy balance of coolant, i. e. , water Ovchinnikov et al. 05 – Pros: Suitable for large areas; greatly reduces maximum T loop components and test sections can be fabricated from standard materials – Cons: Only characterizes heat transfer; no information on materials Fusion Matls Workshop (7/16) 12

Design Simulations • Initial study: numerical simulations of HEMJ test section Fusion Matls Workshop (7/16) Cooler Test section P T T P Heater – Current configuration: heat transferred to He: Ti 600 C; To 700 C – Reversed configuration: He heats water, which cools surface by jet impingement: Ti 700 C; To 600 C – Estimate impinging jet parameters required to remove q = 5 10 MW/m 2 without boiling 13

Impinging Jet “Cooler” • Simulate impinging jet of water in pressurized tank – Jet issues from 2 cm dia. pipe into tank at 2 MPa, 27 C: fully-developed turbulent pipe flow at exit at Exit Rew = 7. 5 104 3. 7 105 (ṁw = 1 5 kg/s) – Impinges upon brass disk modeling HEMJ plasma-facing surface (2 cm dia. Tank Pipe 0. 4 cm thickness): impose uniform heat flux BC over disk – Numerical model = radial slice of entire tank + pipe 2. 3 105 cells Jet 4 cm – Standard k- turbulence model Disk q – Temperature-dependent properties (NIST) Fusion Matls Workshop (7/16) 14

Initial Results – Can cool heat fluxes up to 7. 5 MW/m 2 q = 5, 7. 5, 10 MW/m 2 TL [ C] • Based on h from HEMJ simulations at prototypical conditions, T 365 C TL 335 C for Ti 700 C • Temperatures on impingement surface Tu < Tsat (2 MPa) = 212 C no boiling T U TL Fusion Matls Workshop (7/16) q Rew [/104] 15

Current Status • Helium loop upgraded to higher inlet temperatures: Ti 400 °C – Challenges in measurements at Ti 300 C: consistency in H, seals, achieving steady-state conditions – Replacing Cu with Inconel X-750 gaskets (PHENIX) • Numerical simulations of HEMJ “variants”: thermal stress analysis and CFD – How significant is differential thermal expansion at higher Ti? – Can HEMJ design be simplified (fewer jets, flat surface)? • Experimental studies on HEMJ at Ti 300 C, H = 0. 5 1. 5 mm – Acquire more data to evaluate correlation • Designing larger He loop: ṁ 100 g/s – Initial design simulations for HEMJ suggest q = 7. 5 MW/m 2 possible with reversed heat flux approach using jet-impingement cooling – Validate simulations with small-scale impinging jet experiment Fusion Matls Workshop (7/16) 16