Cart3D free-air simulation of D8
Transcription
Cart3D free-air simulation of D8
Aerodynamic Evalua1on of the D8 "Double-‐Bubble" Aircra; Nacelle Design Shishir Pandya, NASA Ames Research Center Overset Grid Symposium, Dayton, OH, Oct 16, 2012 1 Outline • The D8 – Concept – Goals • Modeling and Mesh Topology – Use Cart3D inviscid mesh adapta1on to guide overset viscous mesh – O-‐mesh with wake boxes – Mesh sensi1vity • Clean Configura1on – Comparison with experiment – Independent verifica1on of D8 concepts • Podded Nacelle – Geometry – Results 2 MIT D8 “Double-‐Bubble” Aircra; • MIT team concept of N+3 advanced vehicle configura1on –Lower fuel burn –Lower noise –Reduce emissions • 180 passengers • 3000 nmi range • 118 ; span • ~Boeing 737 size 3 MIT/Pra] & Whitney/Aurora D Series Airframe & Propulsion Technology Overview Novel configura1on plus suite of airframe and propulsion technologies, and opera1ons modifica1ons Reduced Secondary Structure weight Active Load Alleviation High Bypass Ratio Engines (BPR 20) with High-Efficiency Small Cores Boundary Layer Ingestion Health and Usage Monitoring Distortion Tolerant Fans Tt4 Materials and advanced cooling Natural Laminar Flow on Wing Bottom Variable Area Nozzle Lifting Body Advanced Structural Materials Faired Undercarriage LDI Advanced Combustor Advanced Engine Materials Fuselage Advantages • “Double-‐bubble” fuselage provides more li; – Gives par1al span-‐wise loading / smaller wing • Shorter cabin (wider body) –Results in lighter landing gear support structure • Provides a nose-‐up pitching moment –Shrinks horizontal tail –Lighter horizontal tail 5 Lower Cruise Speed • M=0.72 –Lower sweep wing • Reduced structural load => Lower weight • Increased CL • Can eliminate high-‐li; devices M. Drela, MIT –Proper speed at engine fan face (M=0.6) • Reduces nacelle, inlet size • Reduced nacelle drag –Nacelles embedded in the π-‐tail and fuselage –Reduced size, weight 6 Embedded Rear-‐Mounted Engines • Boundary Layer Inges1ng (BLI) engines for Fans like flow at ~ M=0.6 propulsive efficiency –Thicker boundary layer in the rear –Designed for M=0.6 flow around engine inlet area –Distor1on tolerant fan –High bypass ra1o (~20) • Lower engine-‐out yaw –Reduced ver1cal tail size Fan • Noise shield Nacelle 7 Wind Tunnel Tests • Wright brothers wind tunnel at MIT The Mach and Re don’t match the flight, but the BLI effect is not expected to be very different –Low speed (100 and 120 mph) ≈ M=0.15 –Clean configura1on: α=0, 1.9, 4.4, 6.6, 8.8, 11, 13.3 –Rec≈0.5 Million –1:20 scale model • Tests complete –1:11 scale model • Being constructed • w/, w/o empennage • Podded nacelle • Blended nacelle A. Uranga, MIT 8 Goals of This Work • Independent assessment of the D8 design assump1ons – Is the “double-‐bubble” fuselage advantageous? • Does it provide li; in the range of 10 to 20%? • Does it provide a nose-‐up pitching moment? – Boundary Layer Inges1on (BLI) • Quan1fy flow diffusion from fuselage geometry resul1ng in lower speed at the fan face • Validate Overflow against the wind tunnel tests – Use best prac1ces for nacelle assessment • Assess nacelle performance using Overflow – Clean vs. Podded vs. Embedded 9 Simula1ons • 120 mph (M=0.16) • Alpha sweep: -‐2° to 14° in 2° increments –Half-‐body cases (symmetry assumed) • Steady, turbulent flow • Clean configura1on –in Free-‐air –w/ WT –w/ WT+mount/fairing • Podded • Embedded (geometry, mesh genera1on in progress) 10 Outline • The D8 • Modeling and Mesh Topology – Use Cart3D inviscid mesh adapta1on to guide overset viscous mesh – O-‐mesh with wake boxes – Mesh sensi1vity • Clean Configura1on • Podded Nacelle 11 Cart3D/Aero Solu1on Adap1ve Mesh Original, 70K • Component based inviscid solver –Automated mesh genera1on, complex geometries • Adjoint-‐based adap1ve mesh refinement used –Takes place of a lot of user 1me and exper1se 3 adapts, 400K 6 adapts, 7M 9 adapts, 30M 12 Overflow • Overset mesh genera1on follows –Overset best prac1ces –Guidance from Cart3D adap1ve refinement • Finite difference –Beam-‐Warming (approximate factoriza1on/central difference) –Scalar dissipa1on • + Turbulence model (SA, SST) with target y ≈1 13 Wake Grid with O-‐Grid Topology • Eliminate C-‐ mesh • Wake capture • Tip vortex resolu1on 14 Rear WT core Volume Mesh WT mesh 19 near-‐body grids 1 WT grid 4 wake grids 9 box grids 2 core grids 35 total grids Clean: ~80 Million Pts. Podded: ~87 Million Pts. Wake grids Body-‐fi]ed grids Front WT core Box grids to cover test sec1on 15 Mesh Sensi1vity • Test each parameter independently –Wall spacing (y+) –Near-‐wall stretching ra1o –Surface spacing • LE, TE, ... –Off-‐body spacing • Study done at α=0° –SA turb. model 16 + CL, CD Sensi1vity to y D8 with WT walls and Strut, M=0.16, α=0 SR=1.15, DSs=1, DSo=0.15 0.45 D8, 1:20 scale model in WBWT with Mount, M=0.1578, α=0 D8, 1:20 scale model in WBWT with Mount, M=0.1578, α=0 15 10 CL %CL 0.4 5 0.35 0.3 0 0 1 2 Y 0.040 + 3 4 5 74 M D8, 1:20 scale model in WBWT with Mount, M=0.1578, α=0 78 M 80 M 82 M Number of Points 84 M 86 M D8, 1:20 scale model in WBWT with Mount, M=0.1578, α=0 35 0.038 76 M 30 0.036 25 0.034 CD %CD 20 0.032 15 0.030 10 0.028 5 0.026 0.024 0 1 2 Y + 3 4 5 0 74 M 76 M 78 M 80 M 82 M Number of Points 84 M 86 M 17 CL, CD Sensi1vity to Surface Spacing D8 with WT walls and Strut, M=0.16, α=0 0.45 D8, 1:20 scale model in WBWT with Mount, M=0.1578, α=0 15 y+=1, SR=1.15, DSo=0.15 D8, 1:20 scale model in WBWT with Mount, M=0.1578, α=0 10 CL %CL 0.4 5 0.35 0.3 0 0 1 ΔSsurface 2 3 40 M D8, 1:20 scale model in WBWT with Mount, M=0.1578, α=0 80 M 100 M Number of Points 120 M D8, 1:20 scale model in WBWT with Mount, M=0.1578, α=0 35 0.04 60 M 30 0.038 25 0.034 20 CD %CD 0.036 0.032 0.03 15 10 0.028 5 0.026 0.024 0 1 ΔSsurface 2 3 0 40 M 60 M 80 M 100 M Number of Points 120 M 18 CL, CD Sensi1vity to Stretching Ra1o D8 with WT walls and Strut, M=0.16, α=0 0.45 D8, 1:20 scale model in WBWT with Mount, M=0.1578, α=0 15 y+=1, DSs=1, DSo=0.15 D8, 1:20 scale model in WBWT with Mount, M=0.1578, α=0 10 CL %CL 0.4 5 0.35 0.3 1.05 0 1.1 1.15 1.2 Stretching Ratio 1.25 1.3 70 M D8, 1:20 scale model in WBWT with Mount, M=0.1578, α=0 35 0.040 30 0.036 25 0.034 20 CD %CD 0.038 0.032 0.030 80 M Number of Points 90 M 100 M D8, 1:20 scale model in WBWT with Mount, M=0.1578, α=0 15 10 0.028 5 0.026 0.024 1.05 1.1 1.15 1.2 Stretching Ratio 1.25 1.3 0 70 M 80 M Number of Points 90 M 100 19 M CL, CD Sensi1vity to Off-‐Body Spacing 0.45 D8 with WT walls and Strut, M=0.16, α=0 D8, 1:20 scale model in WBWT with Mount, M=0.1578, α=0 15 y+=1, SR=1.15, DSs=1 D8, 1:20 scale model in WBWT with Mount, M=0.1578, α=0 10 CL %CL 0.4 5 0.35 0.3 0 0 0.1 0.2 Off-Body Grid Spacing 0.3 0.4 60 M D8, 1:20 scale model in WBWT with Mount, M=0.1578, α=0 35 0.04 30 0.036 25 0.034 20 CD %CD 0.038 0.032 0.03 80 M 100 M Number of Points 120 M 140 M D8, 1:20 scale model in WBWT with Mount, M=0.1578, α=0 15 10 0.028 5 0.026 0.024 0 0.1 0.2 Off-Body Grid Spacing 0.3 0.4 0 60 M 80 M 100 M Number of Points 120 M 140 20 M Produc1on Mesh • Tested each parameter independently (α=0°, SA) –Wall-‐normal spacing (tested 0.5 to 5) • 0.00016 => y+ ≈ 1 –Stretching ra1o (tested 1.1 to 1.25) • 1.15 –Surface spacing (tested LE spacing of 0.025 to 0.3%) • LE spacing ≈ 0.006 (0.1% chord) • TE spacing ≈ 0.003 (0.05% chord) • Inboard airfoil covered by ~550 points • Outboard airfoil covered by ~300 points –Off-‐body spacing (tested 0.1 to 0.3) • 0.15 21 Solu1on Consistency, Sensi1vity • Pegasus vs. Xrays vs. c3Lib –li; varies 2.9%, drag varies 1.1% • AF vs. Diagonal vs. Roe vs. HLLC –li; varies 3%, drag varies 1.4% • Low-‐Mach pre-‐condi1oner –No change in li; and drag • Unsteady algorithm –Flow remained steady with minor change in li; and drag • Various WT Inlet/Exit Boundary Condi1ons –Minor varia1ons in integrated forces 22 Outline • The D8 Concept/Goals • Modeling and Mesh Topology • Clean Configura1on –Comparison with experiment – Independent verifica1on of D8 concepts • Podded Nacelle 23 Convergence and Solu1on Force/Moment History Force/Moment History 1 Total Drag Coefficient Total Lift Coefficient 0.15 0.5 0 0 10000 20000 Time Step Number 30000 40000 0.10 0.05 0.00 -0.05 0 10000 20000 Time Step Number 30000 40000 24 CL Comparison 2 WT Cart3D Overflow (SA) Overflow(SST) CL 1.5 1 0.5 0 0 5 α 10 15 25 CD Comparison 0.25 0.2 WT Cart3D Overflow (SA) Overflow (SST) CD 0.15 0.1 0.05 0 0 5 α 10 15 26 Span-‐wise Loading Fuselage provides span loading Pi-‐tail has minimum nega1ve loading Y 27 Component-‐wise Break-‐down CL 0.9 CD Cm* 0.20 0.050 0.8 0.7 0.15 0.040 0.6 0.5 0.10 0.030 0.4 0.05 0.3 0.020 0 0.2 0.1 0.010 -0.05 0 -0.1 0 Total Fuse. Wing Pi-tail -0.10 Total Fuse. Wing Pi-tail α=0 α=4 Total Fuse. Wing Pi-tail *Fuselage provides a pitch-‐up moment 28 Boundary Layer Inges1on Cruise condi1on, α=4° 29 Outline • The D8 Concept/Goals • Modeling and Mesh Topology • Clean Configura1on • Podded nacelle – Geometry – Results • Embedded nacelle 30 Podded Nacelle • In free-‐stream –Under-‐wing –Between π-‐tail –Side-‐mount • Nacelle/Hub/Pylon –7 addi1onal grids • One for each component • Two collar grids –Fuselage/pylon, nacelle/pylon • Two caps for the hub –nose, base 31 Podded Nacelle 32 Podded Nacelle • Flow-‐through nacelle • CL, CD increments w/rt clean configura1on –Lower CL –Higher CD • Compare w/ embedded • Engine model 33 Outline • The D8 Concept/Goals • Modeling and Mesh Topology • Clean Configura1on • Podded nacelle 34 Conclusions • CFD predic1ons validate basic ideas behind D8 –Fuselage carries ~15% of the li; at cruise –Fuselage provides a posi1ve pitching moment at cruise –Rear of the fuselage acts as diffuser • CL, CD compare well to the experiment –Be]er agreement with SST • Podded: Results in increased drag, reduced li; as expected 35 Future Work • Placement of Podded Nacelles • Blended Nacelles –Original configura1on • Three-‐engine –Present concept M. Drela, MIT • Two-‐engine –Geometry, mesh, CL, CD increments • CGT • Blender Sub-‐D surfaces for C-‐2 con1nuity? • Fan model (pressure disk, rota1ng blades?) 36 Acknowledgements This work is funded by NASA’s Fundamental Aeronau1cs Program under the Fixed Wing Project • MIT –Mark Drela –Ed Greitzer –Bob Haimes –Alejandra Uranga • NASA Ames –William Chan –Ce1n Kiris –Nateri Madavan –Mike Olsen –Thomas Pulliam –Mike Rogers 37