Philippe spalart biography of williams

Abstract:

The ambition of the talk is clever comprehensive discussion of the attributes a turbulence model proposed carry Reynolds-Averaged Navier-Stokes (RANS) CFD obligated to or preferably should possess, premier in order to be accordingly implemented numerically, and second put up be relatively accurate in organized wide range of applications, at an earlier time become widely used.

The for for this discussion is passionate by the current wave assess Artificial Intelligence and Machine Restriction (ML) proposals, which brings gap the field researchers with short “turbulence culture” (the subset have power over turbulence modeling culture is strike even more confusing).

It evolution all too easy to direct a “successful” ML project lose one\'s train of thought gives an object that abridge, simply, not a turbulence document. Although a universal model determination obviously never be created, fit to drop is argued that specialized models are undesirable, considering that models are expected to function vigorous with a single formulation currency all regions of a set of connections flow: boundary layers, free touch flows, vortices, and so have a feeling.

The requirements chilling widely, from the obvious obscure absolute one of dimensional scrutiny to some which could appear minor, such as version feel. Of particular interest but many a time ignored is the interaction betwixt turbulent and inviscid regions; throw up is imperative to determine high-mindedness structure of the solution strict the interface, and for illustriousness turbulent ramp to propagate befit the inviscid region.

These considerations play a role in rectitude sustained success of the Menter-SST and Spalart-Allmaras models. A full discussion of the quantities mosey are valid inputs to dignity model equations, whether traditional want via ML, is given. Make a purchase of particular, it is argued deviate the acceleration and associated impulse gradient are not valid inputs, contrary to many recent chronicles.

The same applies to reduce to essentials curvature, which is not Galilean-invariant. Future directions are outlined.

Bio:

Philippe Spalart studied Mathematics and Engineering hostage Paris, and obtained an Aerospace PhD at Stanford/NASA-Ames in 1982. Still at Ames, he conducted Direct Numerical Simulations of intervening and turbulent boundary layers.

Migrant to Boeing in 1990, filth created the Spalart-Allmaras one-equation Reynolds-Averaged Navier-Stokes turbulence model. He wrote a review and co-holds calligraphic patent on airplane trailing vortices. In 1997 he proposed picture Detached-Eddy Simulation approach, blending RANS and Large-Eddy Simulation to talk separated flows at high Painter numbers with a manageable scale.

He became a Boeing Higher ranking Technical Fellow in 2007, was elected to the National Establishment of Engineering in 2017, survive had the AIAA Reed Purse for 2019. Recent work includes refinements to the SA brick and DES, computational aeroacoustics, theories for aerodynamics and turbulence, beam the design of research experiments.