SOUTH – Shape Optimization in HPC-Clouds

Within the SOUTH (Shape Optimization under Uncertainity THrough HPC Clouds) experiment of the EC FP7 FORTISSIMO project, Optimad Engineering developed a cost-effective platform for automatic shape optimisation in collaboration with University of Strathclyde, CINECA SCAI and Automobili Lamborghini spa. SOUTH is designed to help simulation experts in setting up a shape optimisation workflow in a few User-intuitive steps.
For more details, please visit the official page of the experiment.

The SOUTH platform integrates mimic as the geometrical manipulation engine and Dakota as the primary optimization tool.
OpenFoam is natively supported in SOUTH for fluid dynamic simulations and other commercial solvers will be supported in the next future.

SOUTH is in beta-testing and will be deployed on the market as a HPC backed cloud service soon.
Please contact us if you’re interested in testing it.

HPC Cloud Shape Optimization with SOUTH

SOUTH: Optimization of the front bumper of the DrivAer model (courtesy of the Technical University of Munich), in order to minimize the overall drag coefficient Cd of the car. Optimization tries to avoid the rise of the upforce coefficient CL(f) on the front of the car, with respect to its baseline configuration. The presence of a geometrical constraint (a simple plane on the front of the car) let to discard unfeasible bumper configurations when a collision against it occurs. Aerodynamic coefficients in the figure refers to the half-symmetric model. Budget of the optimization was 25K cpu-hours with a number of 24 degrees of freedom for geometry manipulation.