Hybrid finite-difference methods for DNS of compressible turbulent boundary layers

High-order accurate hybrid spatial differencing methods for compressible turbulent boundary layer simulations are developed, analyzed, and tested. The hybrid schemes combine characteristic-based fluxes with explicit and implicit central differences of those fluxes to exploit the advantages of both. The extension of this method to high-order accurate shock-capturing schemes is discussed, as are issues pertaining to efficient parallel implementation. A technique which permits the use of periodic boundary conditions in the streamwise direction while simultaneously allowing spatial boundary layer growth is extended to compressible flow. Preliminary turbulent boundary layer simulations are presented.