Project reference: 2131
CFD is solving time dependent partial differential equations (PDE) and PDE is solved by numerical approximation. Which brings the system of equations (matrices and vectors) that need to be solved either by direct or indirect solvers. These solvers need a huge computational power in terms of the chosen problem. Pre-processing and post-processing (time step solutions) also require lot of computational power to visualize or to make a movie.
Efficient car modelling is essential for fuel efficiency, leading to environmental friendly, particularly with CO2 emission and electricity consumption. A car can be a passenger car or race car, but lift and drag are the main parameters that need to be studied for efficient car modelling.
To understand the cars’ aerodynamics, most of the time, the simple models have been studied a lot and validated against experimental and numerical simulations, for example, Ahmed Body or the SAE body. But in reality, to understand the flow phenomena around the car, a simple car model should be investigated. DrivAer model is a simple car model, similar to a standard vehicle in the present production model. Understanding the aerodynamics of the DrivAer model will open up lots of challenges in the design and understanding of the standard car’s aerodynamical forces.
Project Mentor: Dr. Ezhilmathi Krishnasamy
Project Co-mentor: Dr. Sebastien Varrette
Site Co-ordinator: Dr. Ezhilmathi Krishnasamy
Participants: Benet Eiximeno Franch, Paolo Scuderi
* Computational fluid dynamics using either open source or commercial tools to solve the given problem on a HPC setting and more insight into Aerodynamics.
* Pre-processing, computation and post processing techniques (data analysis).
* Parallel visualization
* Optimization (load balancing on HPC setting and design parameters for DrivAer model)
Student Prerequisites (compulsory):
* Fluid mechanics and basic programming skills.
Student Prerequisites (desirable):
* Familier with any of the open source or commercial CFD software and computational mathematics.
* Paraview : https://www.paraview.org/hpc/
* VisIt : https://wci.llnl.gov/simulation/computer-codes/visit/
* OpenFOAM : https://www.openfoam.com/
* Geometry detail : https://www.mw.tum.de/en/aer/research-groups/automotive/drivaer/geometry/
* Experimental results : https://www.mw.tum.de/en/aer/research-groups/automotive/drivaer/
Week 1: HPC training
Week 2: Project preparation
Week 3: Preprocessing
Week 4: Simulation
Week 5: Simulation
Week 6: Post processing
Week 7: Results analysis
Week 8: Report writing
Final Product Description:
* Compare the simulation results against any one of the model with its wind tunnel experimental results.
Adapting the Project: Increasing the Difficulty:
* DrivAer model has up to 18 configurations, so it is better to study only 1 or 2 configurations. Considering more models would make the project more difficult.
Adapting the Project: Decreasing the Difficulty:
* If we do not need to compare the results with experimental results that could make the project just easier.
* Considering just few parameters with simulation makes it even simpler.
ANSYS and OpenFOAM are available. Other open source tools (both for simulation and pre-processing) can be installed upon student request.
Paraview and VisIt (data processing and visualization) are also available.
ULux-University of Luxembourg
Leave a Reply