The papers presented here describe research to improve the general understanding of the application of SAMR to practical problems, to identify issues critical to efficient and effective implementation on high performance computers, and to stimulate the development of a community code repository for software including benchmarks to assist in the evaluation of software and compiler technologies. The ten chapters have been divided into two parts: programming complexity of SAMR algorithms and the applicability and numerical challenges of SAMR methods.

A new adaptive mesh refinement strategy that is based on a coupled feature-detection and error-estimation approach is developed. The overall goal is to apply the proper degree of refinement to key vortical features in aircraft and rotorcraft wakes. The refinement paradigm is based on a two-stage process wherein the vortical regions are initially identified for refinement using feature-detection, and then the appropriate resolution is determined by the local solution error. The feature-detection scheme uses a local normalization procedure that allows it to automatically identify regions for refinement with threshold values that are not dependent upon the convective scales of the problem. An error estimator, based on the Richardson Extrapolation method, then supplies the identified features with appropriate levels of refinement. The estimator is shown to be well-behaved for steady-state and time-accurate aerodynamic flows. The above strategy is implemented within the Helios code, which features a dual-mesh paradigm of unstructured grids in the near-body domain, and adaptive Cartesian grids in the off-body domain. A main objective of this work is to control the adaption process so that high fidelity wake resolution is obtained in the off-body domain. The approach is tested on several theoretical and practical vortex-dominated flow-fields in an attempt to resolve wingtip vortices and rotor wakes. Accuracy improvements to rotorcraft performance metrics and increased wake resolution are simultaneously documented.

We present a patch-based direct Eulerian adaptive mesh refinement (AMR) algorithm for modeling real equation-of-state, multimaterial compressible flow with strength. Our approach to AMR uses a hierarchical, structured grid approach first developed by (Berger and Oliger 1984), (Berger and Oliger 1984). The grid structure is dynamic in time and is composed of nested uniform rectangular grids of varying resolution. The integration scheme on the grid hierarchy is a recursive procedure in which the coarse grids are advanced, then the fine grids are advanced multiple steps to reach the same time, and finally the coarse and fine grids are synchronized to remove conservation errors during the separate advances. The methodology presented here is based on a single grid algorithm developed for multimaterial gas dynamics by (Colella et al. 1993), refined by(Greenough et al. 1995), and extended to the solution of solid mechanics problems with significant strength by (Lomov and Rubin 2003). The single grid algorithm uses a second-order Godunov scheme with an approximate single fluid Riemann solver and a volume-of-fluid treatment of material interfaces. The method also uses a non-conservative treatment of the deformation tensor and an acoustic approximation for shear waves in the Riemann solver. This departure from a strict application of the higher-order Godunov methodology to the equation of solid mechanics is justified due to the fact that highly nonlinear behavior of shear stresses is rare. This algorithm is implemented in two codes, Geodyn and Raptor, the latter of which is a coupled rad-hydro code. The present discussion will be solely concerned with hydrodynamics modeling. Results from a number of simulations for flows with and without strength will be presented.