11 th International ERCOFTAC Symposium on Engineering Turbulence Modelling and Measurement Palermo, Sept. 21-23, 2016 Direct Numerical Simulation of Drag Reduction with Uniform Blowing over a Rough Wall Eisuke Mori 1,2 , Maurizio Quadrio 2 and Koji Fukagata 1 1 Keio University, Japan 2 Politecnico di Milano, Italy
Background Turbulence - Huge drag - Environmental problems - High operation cost - How to control ? E.Mori, Fukagata lab. DNS/Drag Reduction/Uniform blowing(UB)/Rough wall 2/15
Flow control classification (M. Gad-el-Hak, J. Aircraft , 2001) Flow control strategies Active Passive Feedback Feedforward - Uniform blowing E.Mori, Fukagata lab. DNS/Drag Reduction/Uniform blowing(UB)/Rough wall 3/15
Uniform blowing (UB) (Sumitani & Kasagi, AIAA J. , 1995 Kametani & Fukagata, J. Fluid Mech. , 2011) β’ Drag contribution in a channel flow with UB(/US) π π πΎ π : Blowing velocity π« π = ππ βπ β² π β² ππ + ππ ΰΆ± π β π π β π ΰ΄₯ βπππΎ π ΰΆ± πππ ππ π π π Viscous Turbulent Convective (=UB/US) Contribution contribution contribution (= laminar drag, const. ) (Fukagata et al., Phys. Fluids , 2002) β’ Excellent performance (about 45% by πΎ π = π. π%π½ β ) β’ Unknown over a rough wall On a boundary layer, White: vortex core, Colors: wall shear stress E.Mori, Fukagata lab. DNS/Drag Reduction/Uniform blowing(UB)/Rough wall 4/15
Goal Investigate the interaction between roughness and UB for drag reduction - DNS of turbulent channel flow - Focus on drag reduction performance and mechanism E.Mori, Fukagata lab. DNS/Drag Reduction/Uniform blowing(UB)/Rough wall 5/15
Numerical procedure β’ Based on FD code (for wall deformation) (Nakanishi et al., Int. J. Heat Fluid Fl. , 2012) β’ Constant flow rate, ππ π = ππ½ π πΊ/π = ππππ - so that ππ π β πππ in a plane channel (K.M.M.) β’ βπ + = π. π, π. ππ < βπ + < π, βπ + = π. π β’ UB magnitude: π = 0, π. πππ, π. πππ ROUGH CASE SMOOTH CASE E.Mori, Fukagata lab. DNS/Drag Reduction/Uniform blowing(UB)/Rough wall 6/15
Model of rough wall (E. Napoli et al., J. Fluid Mech. , 2008) Roughness displacement π π© π ππ£π¨ π π ππ π π = πΊ ΰ· Ξ€ π΄ π π π=π πΊ : channel half height π΄ π : Channel length, πππΊ π π π§ππ² = π. πππΊ π© π : Amplitude of each sinusoid π© π = α π, π π©π¬ π = π π = π π, π , π π©π¬ π β π (Defined randomly) with rescaling so that π π = π. πππΊ E.Mori, Fukagata lab. DNS/Drag Reduction/Uniform blowing(UB)/Rough wall 7/15
Coordinate transformation (S. Kang & H. Choi, Phys. Fluids , 2000) Calculation grids: π π (Cartesian with extra force) π = π π π = π π π + π + π π α Actual grid points allocation π = π π ( π¦, π§, π¨ : physical coordinate) π β‘ π π― β π π Ξ€ π = β π π² Ξ€ π π π = π¬ π² , π π = π π π , π π― : displacement of wall lower/upper wall E.Mori, Fukagata lab. DNS/Drag Reduction/Uniform blowing(UB)/Rough wall 8/15
Post processing β’ Drag coefficient decomposition for rough case 8 πΰ΄€ π£ π· πΈπ£π = α€ Re π ππ§ π 2 =2 π· πΈππ = 8 ππ£ ππ€ α€ + α€ (Friction component) Re π ππ§ π 2 =0 ππ¦ π 2 =0 π· πΈππ = β16 ππ β π· πΈππ + π· πΈπ£π (Pressure component) ππ 1 β’ Drag reduction rate βπ· πΈ π πΈπ = Γ 100 [%] π· πΈ,π=0 Only focusing on lower side , subscript β π β omitted hereafter E.Mori, Fukagata lab. DNS/Drag Reduction/Uniform blowing(UB)/Rough wall 9/15
Drag reduction rate, πΊ π¬ SMOOTH CASE ROUGH CASE β ππ% β ππ% β π% β ππ% Total π πΈ β ππ% β ππ% β π% β ππ% Friction π πΈ,πΊ β π% β ππ% Pressure π πΈ,π - - E.Mori, Fukagata lab. DNS/Drag Reduction/Uniform blowing(UB)/Rough wall 10/15
How does friction drag decrease? Bulk mean streamwise velocity Smooth Rough Black: π = 0 Green: π = 0.001 Red: π = 0.005 Normalization based on π = 0 E.Mori, Fukagata lab. DNS/Drag Reduction/Uniform blowing(UB)/Rough wall 11/15
How does pressure drag decrease? averaged in the spanwise and time Pressure contours dashed lines: zero contour π΅ = π π΅ = π. πππ π + π π π§ππ² π π π§π£π¨ E.Mori, Fukagata lab. DNS/Drag Reduction/Uniform blowing(UB)/Rough wall 12/15
Stream function Solid lines: π = 0 Dashed lines: π = 0.005 Uniform Blowing E.Mori, Fukagata lab. DNS/Drag Reduction/Uniform blowing(UB)/Rough wall 13/15
In practical applications Drag reduction amount, βπ« π¬ = π« π¬,π΅=π β π« π¬,π΅=π.π,π.π [%] E.Mori, Fukagata lab. DNS/Drag Reduction/Uniform blowing(UB)/Rough wall 14/15
Concluding remarks DNS of turbulent channel flow over a rough wall with UB β’ UB is effective over rough walls - Lower drag reduction rate ( 7%, ππ% / 11%, ππ% in rough / smooth case, with π = 0.001, 0.005 ) β’ Drag reduction mechanism - Friction drag by wall-normal convection (=conventional) - Pressure drag by prevention of stagnant flow β’ Outlook toward practical applications - More saving opportunity over rough walls E.Mori, Fukagata lab. DNS/Drag Reduction/Uniform blowing(UB)/Rough wall 15/15
Future plans β’ Another drag reduction technique on rough surface - Spanwise oscillation (ongoing) β’ Assessment of net energy? (should be external flow) β’ Calculation at higher Reynolds number? β’ Other types of rough surfaces (e.g., 3D structure)? E.Mori, Fukagata lab. DNS/Drag Reduction/Uniform blowing(UB)/Rough wall 17/15
Uniform blowing (UB) (Sumitani & Kasagi, AIAA J. , 1995 Kametani & Fukagata, J. Fluid Mech. , 2011) β’ Drag contribution in a channel flow with UB(/US) π π πΎ π : Blowing velocity π« π = ππ βπ β² π β² ππ β πππΎ π ΰΆ± π β π ΰ΄₯ + ππ ΰΆ± π β π πππ πΊπ π Blowing π π Viscous Turbulent Convective (=UB/US) side Contribution contribution contribution (= laminar drag, const. ) (Fukagata et al., Phys. Fluids , 2002) β’ Excellent performance (about 45% by πΎ π = π. π%π½ β ) β’ Unknown over a rough wall White: vortex core, Colors: wall shear stress E.Mori, Fukagata lab. DNS/Drag Reduction/Uniform blowing(UB)/Rough wall 18/15
Governing equations (S. Kang & H. Choi, Phys. Fluids , 2000) Incompressible Continuity and Navier-Stokes in π π coordinate ππ π = βπ» ππ π π π π π π π π π π ππ π β ππ + π β ππΈ ππ = β πΊ ππ + π» π ππ π ππ π ππ π ππ π π π ππ π where π 2 π£ π π 2 π£ π π π£ π π£ π π π π π π ππ£ π ππ π ππ + 1 2 + 1 ππ£ π π π = βπ π’ β π π β π π ππ 2π π + π π π π ππ 2 ππ 2 ππ 2 ππ π π 2 2 ππ 2 ππ 2 ππ 2 ππ£ π 1 ππ + ππ 0 π = π π β 1 + π π 2 , for j = 1,3 π π = π π β π π2 ππ 2 ππ π ππ π π π = 1 1 + π , for j = 2 E.Mori, Fukagata lab. DNS/Drag Reduction/Uniform blowing(UB)/Rough wall 19/15
Discretization methods β’ Energy-conservative second-order finite difference schemes (In space) β’ Low-storage third-order Runge-Kutta / Crank- Nicolson scheme (In time) + SMAC method for pressure correction Discretized in the staggered grid system E.Mori, Fukagata lab. DNS/Drag Reduction/Uniform blowing(UB)/Rough wall 20/15
Validation & Verification (B. Milici et al., J. Fluid Mech. , 2014) Time trace of instantaneous C D,r Bulk mean streamwise velocity Less than 2% of difference with most resolved one E.Mori, Fukagata lab. DNS/Drag Reduction/Uniform blowing(UB)/Rough wall 21/15
How does friction drag decreases? *averaged in the spanwise and time Reynolds shear stress contour dashed lines: zero contour π΅ = π π΅ = π. πππ π£ β²+ π€ β²+ π π π§ππ² E.Mori, Fukagata lab. DNS/Drag Reduction/Uniform blowing(UB)/Rough wall 22/15
π distribution (2D contour) ΰ΄₯ Based on π£ π in w/o control case UB 0.5% case No control case E.Mori, Fukagata lab. DNS/Drag Reduction/Uniform blowing(UB)/Rough wall 23/15
π distribution (2D contour) ΰ΄₯ Based on π£ π in w/o control case UB 0.5% case No control case E.Mori, Fukagata lab. DNS/Drag Reduction/Uniform blowing(UB)/Rough wall 24/15
Stream function (detailed) E.Mori, Fukagata lab. DNS/Drag Reduction/Uniform blowing(UB)/Rough wall 25/15
π π¬π§π distribution Black: w/o control Green: UB 0.1% Red: UB 0.5% Normalized by π£ π in w/o control case E.Mori, Fukagata lab. DNS/Drag Reduction/Uniform blowing(UB)/Rough wall 26/15
Recommend
More recommend