Title: Predictive Simulations of Fluid Flows in Engineering: Challenges and Opportunities

Speaker: Dr. Onkar Sahni, Rensselaer Polytechnic Institute, Troy, NY

Date/Time: Wednesday, January 27, 2010, 10:00 am        

Location: CSRI/90 in NM and 915/S101 in CA

Brief Abstract: Engineering flows often involve adverse conditions and 3D platforms, e.g., massively separated flow in short inlet ducts, where the major design goal is to improve the system performance. To achieve this goal, fluidic actuation has been applied as a promising active flow control technology in various forms including steady and unsteady blowing as well as periodic excitation (e.g., via synthetic jets) but has been mostly limited to 2D and steady conditions. This talk will present numerical analysis of fluidic actuation on 3D platforms based on two examples involving aerodynamic flows. One where a tangential control jet is considered in a short inlet duct with low aspect ratio and other with a finite-span synthetic jet surface mounted on a unit wing section. In complement to these simulations, closely coordinated experimental investigations are also performed and therefore, comparisons between numerical calculations and experimental measurements will be provided.

To make such simulations practically feasible and yield insightful solution in a relevant time frame, we employ scalable implicit solution techniques (shown to scale up to 288k processor-cores) along with novel adaptive boundary layer meshing procedures (suitable for viscous flows).

This talk will also discuss a number of recent observations and progress made during the support of these simulations, such as high-order (curved) boundary layer meshing, control of unstructured mesh partitions, operating system interference/jitter, etc.; other examples involving cardiovascular and two phase flows will be included. It will be shown that even with recent advances in the computational approach, the vastness and unknowns associated with natural and engineered systems limit the ability of current computational tools to be used in a comprehensive and predictive manner.

CSRI POC: Patty Hough, (925) 294-1518