SEMINAR: A Variational Multiscale Approach for Dynamic and Adaptive Simulation of Complex Flows

Abstract:

In this talk we will present a variational multiscale (VMS) approach for performing high-fidelity simulation of complex flows. The VMS paradigm relies on the idea of decomposing the solution space into coarse-scale and fine-scale subspaces. The coarse-scale solution space depends on our choice of discretization, whereas the fine-scale solution space encompasses the remainder of the space that our discretization cannot represent. The coarse-scale solution is resolved, while the fine-scale solution is modeled using the fine-scale Green’s function. The fine-scale solution is used as a subgrid-scale model in computing the coarse-scale solution. It is also used to estimate the discretization error due to the coarse-scale solution. Specifically, we will focus on two aspects: a dynamic procedure to evaluate the subgrid-scale model coefficients and an adaptive approach based on the discretization error estimate. The current dynamic procedure takes in the given structure/form of the subgrid-scale model with unknown coefficients and computes them dynamically in a local fashion. A local version of the variational Germano identity that is suitable for unstructured meshes and high orders is developed to perform the dynamic computation of the coefficients. Further, an explicit error estimator is developed and used to control the spatial discretization error through mesh adaptation. We will demonstrate the utility of the current approach on a variety of complex flow problems including active flow control (e.g., synthetic jets), and moving and deforming geometries (e.g., a surging airfoil with dynamic cambering or a projectile fired from a cannon).

Bio:

Dr. Onkar Sahni is an Associate Professor of Aerospace Engineering at Rensselaer Polytechnic Institute. Prior to joining Rensselaer, he was a research scientist at the Center for Predictive Engineering and Computational Science (PECOS) at the University of Texas at Austin. He received his Ph.D. (2007) in Mechanical Engineering and M.S. (2004) in Aeronautical Engineering from Rensselaer. He completed his B.S. (2002) in Aerospace Engineering from IIT-Bombay, India. Dr. Sahni is a recipient of the NSF CAREER Award and Rensselaer’s Outstanding Team Award.

Dr. Sahni has expertise in large-scale coupled fluid flow problems with a focus on turbulence and flow control, specifically for flow problems that exhibit multiscale phenomena (i.e., turbulent flows) and uncertainty (i.e., due to random inputs). His research is focused on formulating advanced models and methods. His research places emphasis on multiscale methods, unstructured meshes, parallel and anisotropic adaptive procedures, and uncertainty quantification techniques. He has designed dynamic and adaptive procedures that are applicable on unstructured grids and demonstrated it on complex problems involving turbulent flows and uncertain inputs. He has published over 90 peer-reviewed journal papers and conference proceedings.