NAFEMS Americas Elements presentation
Download as PPTX, PDF•3 likes•954 views
The document discusses the validation of a new CAE software suite developed by Streamline Solutions for automotive aerodynamics, utilizing extensive data from various vehicle types and wind tunnel tests. It details the software's capabilities in CFD modeling, meshing, and post-processing, aimed at achieving minimal drag coefficient errors. The project demonstrates consistent accuracy in simulations, validating drag coefficient errors below 2% for passenger cars and 4% for heavy trucks through rigorous testing and optimized computational procedures.
NAFEMS Americas Elements presentation
- 1. VALIDATION OF A NOVEL CAE SOFTWARE SUITE AGAINST AN EXTENSIVE DATA SET OF AUTOMOTIVE AERODYNAMICS TEST CASES Angus Lock Head of Aerodynamics Auto Research Center
- 2. Streamline Solutions | Who We Are AUTO RESEARCH CENTER Automotive Engineering R&D, Wind Tunnel, 7-Post & Gearbox Rig ENGYS CFD Engineering, MDO, Open Source Software Development, User Support STREAMLINE SOLUTIONS Joint Venture www.arcindy.com www.engys.com
- 3. Streamline Solutions | Open Source • Streamline Solutions have built an end to end open source CFD process • Scripting, geometry clean-up and creation, morphing, meshing, solving, optimization and post- processing are all conducted using open source tools
- 4. What Is ELEMENTS? • CAE tool with embedded best practices for automotive applications • Based on HelyX Core • Extension of ESI OpenFOAM® library • Beta repository available within 3-6 months • Up-to-date upgraded within 6-12 months • Fully supported via GUI and setup tool • 2 major releases per year • Up to 4 minor releases (issue resolution) • Currently based on latest OpenFOAM-2.1.x • 2.2.x based release scheduled for Q3 2014 © 2013 STREAMLINE SOLUTIONS. All rights reserved.
- 5. CFD Modelling | Mesh Setup • Base mesh scaled by overall vehicle dimensions • Leak/crack detection and fixing • Hex dominant mesh • Automatic feature edge capturing • Improved surface snapping • Full projection of near-wall layers to inlets, outlet, symmetry planes • Surface-based cell-spacing • Volumetric refinement boxes and buffer layer refinement • Stagnation and wake zone refinement tailored to vehicle type • Mesh size 50-80 million cells
- 6. CFD Modelling | Solver Setup • Transient DDES formulation • Spalart-Allmaras model with wall functions • Robust first/second-order blending for advection scheme based on CFL and near-wall distance • Run time is scaled by vehicle length and free- stream velocity • Multiple time-steps (coarse and fine) • Reduced overall memory usage • Improved scaling for parallel processors © 2013 STREAMLINE SOLUTIONS. All rights reserved.
- 7. CFD Modelling | Post-Processing • Automated, batch driven co- and post-processing: – Contour plots of U, p, tauw, Unw fields – Isosurfaces of Ptot – Averaged CD and CL, Cp – Force development plots – HTML / WEBGL reporting C:/Users/alock/Desktop/motorBike/test.html
- 8. Validation | Vehicle Types • 100+ different vehicle types: – Sedan – Hatchback – Estate – SUV – Sportscar – Streamliner – Nascar – Indycar – Light Duty Truck – Heavy Duty Truck © 2013 STREAMLINE SOLUTIONS. All rights reserved.
- 9. Validation | Parametric Changes • Fixed Ground • 5-belt Moving Ground • Single Belt Moving Ground • Yaw Angle • Ride Height • Test Speed • Vehicle Modifications (e.g. Open or Closed Cooling, Spoiler on/off, Underbody Panels on/off)
- 10. Validation | ARC Wind Tunnel • Max Wind Speed : 50 m/s • Max Road Speed: 50 m/s • Nozzle Size: 2.3 m x 2.1 m • Contraction Ratio: 4.8 : 1.0 • Moving Belt Size: 3.4 m x 1.7 m • Boundary Layer Thickness: 1mm • Main Fan Motor Power: 320 kW • Primary BL Motor Power: 80 kW • Secondary BL Motor Power: 19 kW • Rolling Road Motor Power: 120 kW Rapid prototyping, workshops & office space Moving ground, scale-model wind tunnel Vehicle dynamics 7-post rig
- 11. Validation | Aerodynamic Balance • Models are mounted in the tunnel by means of an overhead sting • The 6-component balance is contained within the model • Secondary balance for tractor-trailer
- 12. Validation | Instrumentation • 128 pressure channels to allow for evaluation of surface pressures from around the model • Cooling mass-flow measurement by a grid of up to 14 vane type anemometers
- 13. Validation | Flowfield Comparison • Extensive work was conducted to compare flow field structure • The figures below illustrate the correlation between the CFD predictions and wind tunnel data for the DrivAer models (SAE 2012-01-1068) 13
- 14. Validation | Flowfield Comparison 14 Upper surfaces Lower surfaces • Comparison of centerline pressures on fastback
- 15. Validation | Flowfield Comparison 15 Experiment CFD • Comparison of surface pressures on fastback windshield
- 16. Validation | Flowfield Comparison 16 Experiment CFD • Comparison of surface pressures on fastback windshield
- 17. Validation | Flowfield Comparison 17 Experiment CFD • Comparison of surface pressures on fastback windshield
- 18. Validation | Flowfield Comparison 18 Experiment CFD • Comparison of surface pressures on fastback windshield
- 19. Validation | Flowfield Comparison 19 Experiment CFD • Comparison of surface pressures on fastback windshield
- 20. Validation | Results CD CLF CLR 1 DRIVAER Estate n/a 40% SingleBelt 2.40% 2 DRIVAER Fast n/a 40% SingleBelt -0.41% 3 DRIVAER Notch n/a 40% SingleBelt 0.41% 4 Sedan 1 open 100% 5 Belt 0.67% -0.67% -6.56% open 40% SingleBelt 0.00% -7.19% 4.45% closed 40% SingleBelt 1.74% -2.48% 4.61% 5 Sedan 2 open 100% 5 Belt 0.00% -1.87% -0.37% blanked 100% Fixed 1.57% -26.38% 9.84% 6 Sedan 3 closed 100% Fixed 2.35% open 40% SingleBelt 0.32% -2.27% -2.27% closed 40% SingleBelt 2.03% -1.35% 2.03% 7 Estate1 open 40% SingleBelt -0.32% 11.04% 26.62% 8 Estate2 open 100% 5 Belt -0.95% -3.81% -3.17% 9 Hatchback1 open 40% SingleBelt 3.09% 7.21% 19.75% 10 Hatchback2 open 100% 5 Belt 2.18% -22.55% 12.00% 11 SUV 1 open 40% SingleBelt 0.81% 6.59% -16.76% 12 NASCAR 1 40% SingleBelt 2.22% 13 NASCAR 2 open 40% SingleBelt -1.25% -32.67% -10.47% 14 Sem i-Truck1 open 12.5% SingleBelt 0.19% 15 LightTruck1 open 20%% SingleBelt -0.38% -5.09% -10.38% 1.2% 9.37% 9.24%AverageErrorM agnitude Scale Ground Sim ulation CoefficientsVehicleNo. VehicleM odel Grille (open, closed, blanked) W ind TunnelData Elem ents CD CLF CLR 21 Sem i-Truck2 open 0.000 0.125 SingleBelt 0.19% open 0.000 0.125 SingleBelt -1.13% open 6.000 0.125 SingleBelt 2.63% open 6.000 0.125 SingleBelt 3.45% open 6.000 0.125 SingleBelt 3.45% open 0.000 0.125 SingleBelt 2.25% open -6.000 0.125 SingleBelt 4.85% open 9.000 0.125 SingleBelt 2.50% open -9.000 0.125 SingleBelt 3.13% open 9.000 0.125 SingleBelt 0.00% open 9.000 0.125 SingleBelt 2.06% 22 LightTruck2 open 0.000 0.125 SingleBelt -0.38% -5.09% -10.38% open 6.000 0.125 SingleBelt 2.69% -0.36% -17.95% open 3.000 0.125 SingleBelt 1.47% -3.87% 14.73% 23 Sem i-Truck3 open 0.000 0.125 SingleBelt 0.57% open 3.000 0.125 SingleBelt 4.13% open 6.000 0.125 SingleBelt 4.98% open 9.000 0.125 SingleBelt 1.11% 24 Sem i-Truck4 open 0.000 0.125 SingleBelt 5.86% 2.5% 3.11% 14.35% Yaw AverageErrorM agnitude Scale Ground Sim ulation CoefficientsVehicleNo. VehicleM odel Grille (open, closed, blanked) W ind TunnelData Elem ents • Target is a best practice that delivers drag coefficient errors less than 2% for passenger cars, and 4% for heavy trucks.
- 21. Conclusions • The result of this project is a set of best practices that consistently delivers drag coefficient error magnitudes less than 2% for passenger cars, and 4% for heavy trucks. • The CFD results have been heavily validated in a range of wind tunnels, and for a variety of different vehicle shapes. • For typical mesh sizes of 50-80 million cells, running on 144 cores, turnarounds of 16-24 hours are the median. This is considered to be extremely competitive when compared to other commercial CFD codes.
Editor's Notes
- #3: Notes: I decided to add this slide here, because the customer list is very important to show to the people which do not know us.