Viscous Cartesian Flow Solver with AMR Capabilities for Automated Flow Simulation

The goal of this project is to develop and leverage the novel Normal Ray Refinement (NRR) methodology investigated in Phase I, and to apply it to an efficient and robust prototype adaptive Cartesian solver for automated simulations of viscous flows. The NRR approach will provide a substantial reduction in the number of grid points in viscous boundary layers and will employ adaptive normal ray placement strategies for normal ray location, clustering and length in 2D and 3D. The prototype flow solver will feature this and other cutting edge technologies for efficient, automated and accurate simulations of viscous flows with arbitrary complex geometries. The solver will feature the immersed boundary method which was shown to eliminate numerical fluctuations at the boundaries inherent to the cut-cell methods on Cartesian grids. The prototype solver will also use the advanced parallelization techniques with dynamic load balancing in an efficient object-oriented framework. The proposed technology will help to significantly reduce costs of viscous CFD simulations for assessing air vehicle performance, supporting air vehicle trade and design studies and for building critical aerodynamic databases and considerably decrease the design cycle-time, thus reducing the turn-around time and gaining critical advantages over competitors.