Inviscid Wall-Modeled Large Eddy Simulations for Improved Efficiency

\hyphenpenalty=1000 The accuracy of an inviscid flow assumption for wall-modeled large eddy simulations (LES) is examined because of its ability to reduce simulation costs. This assumption is not generally applicable for wall-bounded flows due to the high velocity gradients found near walls. In wall-modeled LES, however, neither the viscous near-wall region or the viscous length scales in the outer flow are resolved. Therefore, the viscous terms in the Navier-Stokes equations have little impact on the resolved flowfield. Zero pressure gradient flat plate boundary layer results are presented for both viscous and inviscid simulations using a wall model developed previously.\footnote{K.~M.~Aikens, ``High-fidelity large eddy simulation for supersonic jet noise prediction,'' Ph.D.~thesis, Purdue University, 2014.} The results are very similar and compare favorably to those from another wall model methodology\footnote{S.~Kawai and J.~Larsson, Phys.~Fluids {\bf 24}, 015105 (2012).} and experimental data. Furthermore, the inviscid assumption reduces simulation costs by about 25\% and 39\% for supersonic and subsonic flows, respectively. Future research directions are discussed as are preliminary efforts to extend the wall model to include the effects of unresolved wall roughness.