Computational Fluid Dynamics:

In Room 3305 Engineering II we do numerical simulations of turbulent flow. Most of our computations are performed on vectorial Super-computers (Cray C90, Cray C98) in the various National Supercomputing Centers. Starting from this summer we will also use Parallel Super-computers (SGI Power Challenge, Intel Paragon, Cray T3D). In our Computational Lab we have a very fast workstation DEC ALPHA 3000-400, better known as viper, that we use to test our programs before running them on Super-computers, and a DEC 5000-240, anemone, that has an high-resolution graphic cards, and that we use mostly for flow visualizations.

We then have some Mac's and PC's...for fun!

We do numerical simulations at different levels of complexity. We use spectral methods (a polynomial approximation somehow similar to finite elements) to simulate turbulent flow without any closure approximations (DNS), in flow conditions and geometries that are closer and closer to real flow conditions. From these simulations we then try to understand the physics of the problems and to write closure approximations that can be used in computing flows of industrial interest. We use finite difference methods to simulate turbulent flows with L.E.S. (large eddy simulation) models or with k-e methods (L.E.S. and k-e methods are the two models currently used in the numerical design of turbines, engines, airplanes ...in industrial practics). We are also trying new simulation strategies, like Gas-Lattice Cellular Automata, and Neural Network based algorithms.

Our computational group is a mixture of Chemical Engineers, Physicists, Mathematicians and Computer Scientists, some of them are in Santa Barbara, some others elsewhere in the world, who collaborate with us and spend time here.


The people:

Assistent Professors:

  • Maria Vittoria Salvetti (Pisa, Italy): LES Turbulent Flow
  • Alfredo Soldati (Udine, Italy): DNS of Flows in Electrostatic Precipitators

    Postdoctoral:

  • Aurelio Cortesi (PSI Switzerland): DNS of Mixing Layer
  • Francois Beux (Pisa, Italy): k-e Methods and Level Set Methods for Multiphase problems.
  • Sanjeev Kumar (Santa Barbara): Neural Network Applications
  • Martin Pattison (Santa Barbara): k-e Methods, Boltzmann Formulation for Dispersed Flows.

    Ph.D Students:

  • Brian Knowlton (Santa Barbara): DNS of Flows with Moving, Deformable Internal Boundaries.
  • Valerio De Angelis (Santa Barbara): DNS of Flows in Complex Geometries, Gas-Lattice Models.
  • Dr. Yu Pan (he just graduated): DNS of Flows with Particles