FLOW CONDITIONING FLOW CONDITIONING DESIGN IN TURBULENT DESIGN IN TURBULENT LIQUID SHEETS LIQUID SHEETS S.G. DURBIN, M. YODA, and S.I. ABDEL-KHALIK G. W. Woodruff School of Mechanical Engineering Atlanta, GA 30332-0405 USA
Thick Liquid Protection • Protect IFE reactor chamber first walls with liquid “curtain” to absorb radiation from fusion events • Increase chamber lifetime and decrease chamber radius HYLIFE-II (High-Yield Lithium-Injection Fusion Energy) • Oscillating slab jets, or liquid sheets, create protective pocket to shield chamber side walls • Lattice of stationary sheets shield front/back walls while allowing beam propagation and target injection 2
Motivation • Effective protection ⇒ Minimize clearance between edge of liquid sheet and driver beams � Minimize interference with target injection, beam propagation � How are velocity fluctuations near the nozzle exit influenced by different flow conditioner ( vs . nozzle) designs? � How are velocity fluctuations related to free-surface fluctuations downstream of the nozzle exit? � Are fine screens required in the flow conditioner? � Will more screens reduce free-surface fluctuations? 3
Objectives • Quantify effect of flow conditioner designs in terms of mean velocity and turbulence intensity just upstream of nozzle exit • Quantify surface ripple in terms of free- surface fluctuations within range of interest for HYLIFE-II • Measure loss coefficient across the flow conditioner / nozzle assembly for different flow conditioner configurations 4
Flow Loop • Pump-driven D recirculating flow loop E E • Test section height ~ 1 m F F • Overall height ~ 5.5 m G G A Pump H 400 gal tank C B Bypass line I Butterfly valve C Flow meter J 700 gal tank H B D Pressure gage K 20 kW chiller I E Flow conditioner E Flow conditioner F Nozzle F Nozzle K J A G Liquid sheet G Liquid sheet 5
Experimental Parameters z • δ = 1 cm; aspect ratio AR = 10 y δ x • Near-field: x / δ ≤ 25 • Reynolds number Re = 0.5 – 1.2 × 10 5 � [ Re = U o δ / ν ; U o average speed; ν liquid kinematic viscosity] • Fluid density ratio ρ L / ρ G = 850 [ ρ G gas density] • Velocity and rms fluctuations � u and u ′ : Streamwise ( x -component) � w and w ′ : Transverse ( z -component) • σ z standard deviation of free surface z -location 6
Flow Conditioning Elements • Rectangular flow cross-section 10 cm × 3 cm ( y × z ) • Perforated plate (PP) � Open area ratio 50% with staggered 4.8 mm dia. holes • Honeycomb (HC) � 3.2 mm dia. × 25.4 mm staggered circular cells • Fine screen (FS-1) � Open area ratio 37.1% � 0.33 mm dia. wires woven w/ open cell width of 0.51 mm (mesh size 30 × 30) • Fine screen (FS-2) � Open area ratio 36.0% � 0.25 mm dia. wires woven w/ open cell width of 0.38 mm (mesh size 40 × 40) • Nozzle � 5 th order polynomial contour with contraction ratio = 3 7
Flow Conditioning Configurations No Screen / (One Screen * ) Two Screens PP 3.9 cm PP HC 3.9 cm 3.0 cm HC FS-1 3.0 cm 3.3 cm (FS-1) FS-2 14.7 cm 14.7 cm LDV z z window x y x y * Standard design 8
Laser-Doppler Velocimetry (LDV) • Single component backscatter operation • Frequency shift of 1.3 MHz for y w measurements z • Probe volume 230 × 50 × 1250 µ m ( x, y, z ) at FWHM x • Positioning controlled by two g linear stages Probe shown in • Water seeded with TiO 2 streamwise particles (typical dia. 0.3 µ m) b configuration • Velocity profiles at x = -6 mm z y � b = 10.95 mm x 9
Planar-Laser Induced Fluorescence (PLIF) • y Water dyed w/disodium fluorescein (26 mg/L) • Free surface = interface between fluorescing Nozzle z (bright) water and (dark) air x • Image obliquely with B/W CCD camera Light � Exposure one convective time scale τ = δ / U o = sheet 0.9 – 2.2 msec • Surface ripple measurements span > 2000 τ CCD • Threshold individual images � Threshold value from image grayscale histogram g � Grayscale > threshold ⇒ water ≤ threshold ⇒ air Original image Thresholded image Free surface 1 cm z y x 10
Velocity Profiles: Re = 120,000 (One Screen) 1.00 0.15 0.10 b 0.75 z y 0.05 w / U o u / U o x 0.50 0.00 b = 10.95 mm 0.25 z -0.05 x b 0.00 -0.10 y -0.50 -0.25 0.00 0.25 0.50 Profiles along dotted lines z / t(x) z / b y / δ � w / U o � u / U o � 4.838 � 4.438 � 4.163 � 3.688 + 1.188 11
Velocity Profiles: Re = 120,000 (No Screen and Two Screens) No Screen Two Screens 1.00 0.15 1.00 0.15 0.10 0.10 0.75 0.75 0.05 0.05 w / U o w / U o u / U o u / U o 0.50 0.50 0.00 0.00 0.25 0.25 -0.05 -0.05 0.00 -0.10 0.00 -0.10 -0.50 -0.25 0.00 0.25 0.50 -0.50 -0.25 0.00 0.25 0.50 z / t(x) z / b z / t(x) z / b � w / U o u / U o � y / δ � 4.838 � 4.438 � 4.163 � 3.688 + 1.188 12
RMS Profiles: Re = 120,000 (One Screen) 5 • Non-homogeneous turbulence 4 � u ′ / w ′ ≈ 2 rms / U o (%) RMS / U o (%) • Nearly constant 3 fluctuations for central 2 75% of b • Turbulent boundary 1 layer indicated by marked increase in u ′ , 0 w ′ near nozzle walls -0.50 -0.25 0.00 0.25 0.50 z / t(x) z / b � w ′ / U o y / δ � u ′ / U o � 4.838 � 4.438 � 4.163 � 3.688 + 1.188 13
RMS Profiles: Re = 120,000 (No Screen and Two Screens) No Screen Two Screens 5 5 4 4 rms / U o (%) rms / U o (%) RMS / U o (%) RMS / U o (%) 3 3 2 2 1 1 0 0 -0.50 -0.25 0.00 0.25 0.50 -0.50 -0.25 0.00 0.25 0.50 z / t(x) z / t(x) z / b z / b w ′ / U o � u ′ / U o � y / δ � 4.838 � 4.438 � 4.163 � 3.688 + 1.188 14
Average Streamwise RMS ( Re = 120,000) u' / U o (%) • Averaged over | z| / b No One Two ≤ 0.375 y / δ Screens Screen Screens • 95% confidence 4.838 2.6 2.6 2.0 interval for all data 4.638 2.6 2.4 2.0 4.438 2.6 2.5 2.0 ~0.1 % 4.163 2.6 2.6 1.7 • Two screens has 3.688 2.6 2.3 1.8 less streamwise 2.438 2.5 2.4 fluctuation Average 2.6 2.5 1.9 15
Average Transverse RMS ( Re = 120,000) • Small decrease in w ′ with w' / U o (%) one screen No One Two y / δ Screens Screen Screens • No change in w ′ between 4.838 1.7 1.7 1.9 one and two screens 4.638 1.5 1.3 1.2 • Significant central 4.438 1.4 1.0 1.0 4.163 1.4 1.1 1.0 disturbance without 3.688 1.3 1.1 1.0 screen 2.438 1.1 1.0 1.0 � w ′ / U o = 1.6 – 2.3 % 1.188 1.9 1.0 1.0 at y / δ = 1.638 – 0.338, Average 1.5 1.2 1.2 respectively 16
PLIF Results (Effect of Initial Conditions) • x / δ = 25 0.15 • Re = 120,000 0.10 • Large central σ z / δ fluctuation without 0.05 fine screen � Also observed in 0.00 transverse velocity -5.0 -2.5 0.0 2.5 5.0 fluctuations y / δ z y x No Screen One Screen 17
Average PLIF Results ( Re = 120,000) 0.08 • Fluctuations ~ 1.5 × for flows without fine screen 0.06 � Due to central σ z / δ σ z / δ 0.04 disturbance • Flows similar for one 0.02 and two screen flow 0.00 conditioning 0 10 20 30 x / δ � No Screen � One Screen � Two Screens 18
Average PLIF Results ( Re = 50,000) 0.05 • One screen 0.04 configuration produces smoother jet 0.03 σ z / δ σ z / δ • Streamwise and 0.02 transverse rms identical 0.01 within experimental 0.00 error 0 10 20 30 x / δ � One Screen � Two Screens 19
Loss Coefficient ( Re = 120,000) 12 ∆ + 2 P ρ U L in = • K L 12 2 ρ U ∆ P U in U o L o K L (kPa) (m/s) (m/s) where ∆ P = Press. drop across No flow conditioner assembly, ρ L = Screens 69 2.35 10.72 1.25 fluid density, U in = inlet velocity, One and U o = exit velocity Screen 103 2.36 10.75 1.84 • Addition of screens Two increases loss coefficient Screens 121 2.38 10.84 2.11 � Pumping power ↑ as K L ↑ for given flowrate 20
Conclusions (Initial Conditions) Characterized turbulent liquid sheets from three flow conditioning configurations for Re = 50,000 and 120,000 • Quantified velocities / turbulence intensities near nozzle exit � Streamwise velocity measurements indicate uniform flow for all flow conditioner configurations � Second screen decreases streamwise velocity fluctuations � Elevated levels of transverse velocity fluctuations in center of jet for conditioning without a fine screen • Quantified loss coefficient across flow conditioner � Addition of fine screens increases K L – Requires higher pumping power – Increases likelihood of flow blockage due to trapped debris 21
Conclusions (Surface Ripple) • Quantified surface ripple for flows of interest in HYLIFE-II � One screen configuration produced smoothest flows for Re = 50,000 and 120,000 – σ z / δ < 0.04 in near-field – Second screen increases surface ripple at Re = 50,000 � Central disturbance observed in transverse velocity fluctuations and free-surface fluctuations for conditioning with no screen 22
Recommend
More recommend