DLR-F4 Results using Bombardier AerospaceDLR-F4 Results using Bombardier AerospaceFull-Aircraft Full-Aircraft NavierNavier-Stokes Code FANSC-Stokes Code FANSC

Eric Eric LaurendeauLaurendeauEngineering SpecialistEngineering SpecialistAdvanced Aerodynamics, MontrealAdvanced Aerodynamics, Montreal

AIAA CFD Drag Prediction Workshop, Anaheim, CA June 9-10 2001

Josee BoudreauJosee BoudreauAerodynamicistAerodynamicistAdvanced Aerodynamics, MontrealAdvanced Aerodynamics, Montreal

Outline

• Description of FANSC

• DLRF4 Results:

•Convergence•Pressure Distribution•Forces

• Conclusions

FANSCFANSC

Developed in House under a multi-year Defense/Industry Research Program which included:

- Development of a grid-generation package (MBGRID)

- Development of structured and unstructured Navier-Stokes methods for full-aircraft configurations

FANSCFANSC

Flow Solver

- Multi-block structured domain decomposition

- Finite-volume discretization (cell-centered)

- Explicit artificial dissipation (JST, AUSM, CUSP)

- Explicit Runge-Kutta scheme

FANSCFANSC

Acceleration techniques

- Full-multigrid algorithm

- Local time stepping

- Implicit residual smoothing, directional scaling for high aspect ratio meshes

- Coarse-grain parallelization on blocks (3.6/4 CPU)

- Vectorization (94% efficient)

FANSCFANSC

Boundary Conditions

- No-Slip/Slip Wall

- Transpiration Wall (Boundary-Layer Coupling)

- Symmetry, Degenerate line/point

- Riemann, Engine Inlet/Outlet

- Multiple boundary condition per block face

FANSCFANSC

Turbulence Model

- Spalart-Allmaras one-equation with/without wall functions

- Discretized and solved as described in their 1992 paper

- Loosely coupled

- Computed on fine grid only (weighted-average and frozen on coarse grids)

- Zero turbulent eddy viscosity at solid walls on all grid levels

HardwareHardware

Run-time (for provided mesh)

- 9.0 hours on 4 CPU

- 2.4 Gbytes Memory Requirements

Hardware

- 8-CPU Cray SV1

- 8 Gbytes RAM

- 1.2 Gflops/CPU

-

-

-

DLR-F4 Results: Convergence (DLR-F4 Results: Convergence (DPW gridDPW grid))

Iterations

log(

Res

ρ)

0 250 500 750 1000-6

-5

-4

-3

-2

-1

0

Constant αConstant CL

Iterations

CL

0 250 500 750 10000

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Constant αConstant CL

Iterations

CD

0 250 500 750 10000

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

0.09

0.1

Constant αConstant CL

Iterations

log(

nut)

0 250 500 750 1000-5

-4

-3

-2

-1

0

1

Constant αConstant CL

DLR-F4 Results (M 0.60 CL 0.50)DLR-F4 Results (M 0.60 CL 0.50)

x/c

Cp

0 0.25 0.5 0.75 1

-1.50

-1.00

-0.50

0.00

0.50

1.00

EXPFANSC

η = 18.5%

x/c

Cp

0 0.25 0.5 0.75 1

-1.50

-1.00

-0.50

0.00

0.50

1.00

EXPFANSC

η = 23.8%

x/c

Cp

0 0.25 0.5 0.75 1

-1.50

-1.00

-0.50

0.00

0.50

1.00

EXPFANSC

η = 33.1%

x/c

Cp

0 0.25 0.5 0.75 1

-1.50

-1.00

-0.50

0.00

0.50

1.00

EXPFANSC

η = 40.9%

x/c

Cp

0 0.25 0.5 0.75 1

-1.50

-1.00

-0.50

0.00

0.50

1.00

EXPFANSC

η = 51.2%

x/c

Cp

0 0.25 0.5 0.75 1

-1.50

-1.00

-0.50

0.00

0.50

1.00

EXPFANSC

η = 63.6%

x/c

Cp

0 0.25 0.5 0.75 1

-1.50

-1.00

-0.50

0.00

0.50

1.00

EXPFANSC

η = 84.4%

DLR-F4 Results (M 0.80 CL 0.50)DLR-F4 Results (M 0.80 CL 0.50)

x/c

Cp

0 0.25 0.5 0.75 1

-1.50

-1.00

-0.50

0.00

0.50

1.00

EXPFANSC

η = 18.5%

x/c

Cp

0 0.25 0.5 0.75 1

-1.50

-1.00

-0.50

0.00

0.50

1.00

EXPFANSC

η = 23.8%

x/c

Cp

0 0.25 0.5 0.75 1

-1.50

-1.00

-0.50

0.00

0.50

1.00

EXPFANSC

η = 33.1%

x/c

Cp

0 0.25 0.5 0.75 1

-1.50

-1.00

-0.50

0.00

0.50

1.00

EXPFANSC

η = 40.9%

x/c

Cp

0 0.25 0.5 0.75 1

-1.50

-1.00

-0.50

0.00

0.50

1.00

EXPFANSC

η = 51.2%

x/c

Cp

0 0.25 0.5 0.75 1

-1.50

-1.00

-0.50

0.00

0.50

1.00

EXPFANSC

η = 63.6%

x/c

Cp

0 0.25 0.5 0.75 1

-1.50

-1.00

-0.50

0.00

0.50

1.00

EXPFANSC

η = 84.4%

DLR-F4 Results (M 0.75 DLR-F4 Results (M 0.75 α α = 1.0 = 1.0 degdeg.).)

Oil flowTrailing edge separation and corner vortex

MBGRID meshMBGRID mesh

- An internally generated mesh of the DLRF4 configuration using Bombardier Aerospace MBGRID grid-generation package was produced:

Orthogonality of the mesh on the body surface

10e-6 wall spacing

3 Million mesh points, open wing-tip, blunt trailing edge

- Similar convergence as when the DPW mesh is used

DLR-F4 Results (M 0.75)DLR-F4 Results (M 0.75)

αo

CL

-4.0 -3.0 -2.0 -1.0 0.0 1.0 2.0 3.0 4.00

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Exp NLRExp ONERAExp DRAFANSC-DPW gridFANSC-Bombardier grid

The solid line prediction lies in line with the most accurate unstructured CFD results presented at the DPW.

DLR-F4 Results (M 0.75)DLR-F4 Results (M 0.75)

CL

Cm

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0-0.2

-0.19

-0.18

-0.17

-0.16

-0.15

-0.14

-0.13

-0.12

-0.11

-0.1

-0.09

-0.08

-0.07

-0.06

Exp NLRExp ONERAExp DRAFANSC-DPW gridFANSC-Bombardier grid

The solid line prediction lies in line with the most accurate unstructured CFD results presented at the DPW.

DLR-F4 Results (M 0.75)DLR-F4 Results (M 0.75)

CD

CL

0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.100

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

Exp NLRExp ONERAExp DRAFANSC-DPW gridFANSC-Bombardier grid

The solid line prediction lies in line with the most accurate unstructured CFD results presented at the DPW.

CONCLUSIONSCONCLUSIONS

- FANSC showed excellent convergence characteristics (density and turbulent viscosity) on the required grid, despite its excessive skewness

- Integrated lift and pitching moment predictions on the required grid were accurate, but mesh skewness introduced discretization errors on the skin-friction evaluations (Pressure drag was correctly predicted)

CONCLUSIONSCONCLUSIONS

- Results obtained with FANSC on an internally generated mesh using the MBGRID grid-generation package resulted in accurate lift, pitching moment and drag prediction, as was expected from our

industrial experience.

- The overall performance of FANSC in terms of robustness, convergence and accuracy at this drag workshop ranks it as one of the best Navier-Stokes code available.

CONCLUSIONSCONCLUSIONS

- It was unfortunate that the DPW structured mesh was not better suited for the problem in hand.

However, it did provide a basis for evaluating the robustness of the structured codes.

- The enthusiasm present at the workshop was very encouraging and demonstrated the need to pursue further research in this area.

CONCLUSIONSCONCLUSIONS

- Bombardier Aerospace whishes to thank the DPW committee for their efforts in organizing the event.