Numerical investigation around notchback SAE model Franois - PowerPoint PPT Presentation

Numerical investigation around notchback SAE model François DELASSAUX 11/12/2019

Numerical setup - Boundary conditions Inlet: uniform 𝑊 " = 40 𝑛/𝑡 (notchback) and 16m/s (fastback/estate) and experimental turbulent intensity ~0,2% - - Outlet: pressure outlet condition with gauge pressure = 0 Pa - No slip wall condition on the body - Software: ANSYS Fluent - Grids: HexaPoly provided for the conference SST DDES RKE RANS - Numerical schemes – Finite Volume Method - Advections terms: Bounded Central Differencing (BCD) scheme - Numerical schemes – Finite Volume Method - k, ω and pressure: 2nd order Upwind scheme 2 nd order upwind - - Temporal derivative: Bounded 2nd order implicit - Coupled algorithm scheme - Segregated SIMPLEC algorithm Computational time - Δ𝑢 = 5.10 ./ s - Steady computation using pseudo-transient method - to increase flow convergence - Total physical time = 2 s - 1000 iterations - Convergence = 1 s - Averaging = 1 s - From 20 to 4 inner iterations 2

Notchback SAE – Aerodynamics coefficients and Cp centerline -Cd ΔCd -Cl ΔCl - Good agreement on drag prediction Exp. 0.2071 - -0.0548 - - Bad prediction on lift coefficient RKE 0.1925 -7.0% -0.0698 27.4% - Not meaningful due to very small SST DDES 0.2121 2.4% -0.0778 42.0% lift values Cp centerline -450 -410 -370 -330 -290 -250 -210 -170 -130 -90 -50 -10 30 70 110 150 190 230 270 310 350 390 430 0.5 0 - Good agreement on Cp coefficient along the centerline for both turbulence models -0.5 Exp. Cp [-] -1 RANS DDES -1.5 -2 -2.5 3 X [mm]

Notchback SAE – Cp rear slant surface Exp. RKE SST DDES • Good pressure recovery over the rear slant with RANS => no separation • More energetic C-pillar vortices with DDES compared to RANS • Not enough Cp probes in experiments to capture the footprint of C- pillar vortices 4

Notchback SAE – iso Vx<0 / streamlines RKE RKE No separation L/H=1,53 SST DDES SST DDES Small separation L/H=1,08 5

Notchback SAE – Y=0 mm backlight – Vx Exp. RKE SST DDES Separation Δ 123 = max Δ𝑦, Δ𝑧, Δ𝑨 = 2,6 𝑛𝑛 Vx and fd evolution at X=-300 mm 𝜀 ≈ 20 − 30 𝑛𝑛 0.06 0.05 𝟏, 𝟏𝟗 < 𝒔 = 𝜠 𝒏𝒃𝒚 < 𝟏, 𝟐𝟒 Z from surface 0.04 𝜺 0.03 Vx • XX 0.02 fd r should be at least >0,2 [Menter] => 5 mm 0.01 0 -0.1 0.1 0.3 0.5 0.7 0.9 1.1 6 Vx and fd

Numerical investigation around DrivAer models François DELASSAUX 11/12/2019

Fastback – Aerodynamics coefficients Without wheels Full body Cd Cl Cl Front Cl Rear Cd Cl Cl Front Cl Rear 0. RKE 0.145 0.031 -0.062 0.093 RKE 0.200 0.018 -0.065 0.083 Coarse Coarse 0. DDES SST 0.160 -0.004 -0.095 0.091 DDES SST 0.234 -0.001 -0.083 0.082 RKE 0.198 0.023 -0.067 0.090 RKE 0.146 0.036 -0.062 0.099 Medium Medium 0. DDES SST 0.234 -0.025 -0.090 0.066 DDES SST 0.156 -0.022 -0.100 0.078 RKE 0.197 0.031 -0.065 0.095 RKE 0.145 0.042 -0.062 0.104 0. Fine Fine DDES SST 0.234 -0.032 -0.101 0.069 DDES SST 0.161 -0.032 -0.111 0.079 Cl Cd 0.240 0.040 0.030 0.230 0.020 0.220 0.010 0.210 0.000 0.200 -0.010 0.190 -0.020 0.180 -0.030 0.170 -0.040 Coarse Medium Fine Coarse Medium Fine 8

Fastback – CFL Y0 Coarse Fine Y300 Coarse Fine Time step = 5.10 -4 s is enough regarding CFL and using implicit temporal scheme 9

Fastback – Iso surface Vx=0 RKE Coarse Medium Fine DDES Coarse Medium Fine 10

Fastback – Cp side RKE Coarse Medium Fine DDES Coarse Medium Fine 11

Fastback – Cp underbody RKE Coarse Medium Fine DDES Coarse Medium Fine 12

Fastback – Y0 – Vx and Vz RKE Coarse Fine DDES Coarse Fine 13

Fastback – Cp top RKE Coarse Medium Fine DDES Coarse Medium Fine 14

Fastback – Streamlines Y0 RKE Coarse Medium Fine DDES Coarse Medium Fine 16

Fastback – Streamlines Y3125 RKE Coarse Medium Fine DDES Coarse Medium Fine 17

Estate – Aerodynamics coefficients on full body Without wheels Full body Cd Cl Cl Front Cl Rear Cd Cl Cl Front Cl Rear RKE 0.170 -0.097 -0.066 -0.031 RKE 0.230 -0.112 -0.072 -0.040 Coarse Coarse DDES SST 0.198 -0.139 -0.119 -0.020 DDES SST 0.276 -0.138 -0.107 -0.031 RKE 0.171 -0.090 -0.064 -0.026 RKE 0.226 -0.104 -0.070 -0.034 Medium Medium DDES SST 0.280 -0.171 -0.094 -0.076 DDES SST 0.199 -0.170 -0.098 -0.072 RKE 0.171 -0.090 -0.065 -0.025 RKE 0.226 -0.104 -0.071 -0.033 Fine Fine DDES SST 0.205 -0.173 -0.100 -0.072 DDES SST 0.285 -0.171 -0.097 -0.075 Cl Cd 0.000 0.300 -0.020 0.250 -0.040 -0.060 0.200 -0.080 0.150 -0.100 -0.120 0.100 -0.140 -0.160 0.050 -0.180 0.000 -0.200 Coarse Medium Fine Coarse Medium Fine 18

Estate – Iso surface Vx=0 RKE Coarse Medium Fine DDES Coarse Medium Fine 19

Estate – Cp side RKE Coarse Medium Fine DDES Coarse Medium Fine 20

Estate – Cp underbody RKE Coarse Medium Fine DDES Coarse Medium Fine 21

Estate – Y0 – Vx and Vz RKE Coarse Fine DDES Coarse Fine 22

Estate – Cp top RKE Coarse Medium Fine DDES Coarse Medium Fine 23

Estate – Cp rear RKE Coarse Medium Fine Peugeot 308 SW DDES Coarse Medium Fine 24

Estate – Streamlines Y0 RKE Coarse Medium Fine DDES Coarse Medium Fine 25

Exp. 308 Estate – Streamlines Y0 Y0 RKE • RANS: 2 massive separations • DDES: very close to exp. DDES => accurate flow topology 26

Estate – Streamlines Y3125 RKE Coarse Medium Fine DDES Coarse Medium Fine 27

Conclusions - Notchback SAE - Good prediction drag coefficient - Discrepancies on lift coefficient prediction - Improve DDES flow on the rear slant surface - DrivAer models - Higher drag and lift values with DDES compared to RANS - Both RANS and DDES are insensitive to the tested grid refinements for drag prediction - Estate shape: good agreement for DDES model compared to PSA work on Peugeot 308 SW 28