ABSTRACT
Th e eddy viscosity models (EVM) which have become the workhorses of engineering turbulence analyses do not describe the eddy structure of the turbulent fi elds, anisotropic turbulence, turbulence/chemistry interactions, or transition from laminar to turbulent fl ow. Much research to address such issues has been and is being pursued to understand and model these eff ects. Solving more complete equations for the Reynolds stresses by diff erential second-moment (DSM) closure is expected to remove some of the shortcomings of the EVM analyses. Using probability density functions (PDF) to address more of the statistical properties of turbulent fi elds, including turbulence/chemistry interactions, is being actively researched. Th e Navier-Stokes equations are being integrated directly to provide direct numerical simulations (DNS) of turbulence to represent all spatial scales of the turbulent fi eld. Large eddy simulations (LES) treat large and small eddy scales separately to provide a model intermediate between DNS and PDF models in computational effi ciency. Models to describe transitional fl ow from laminar to turbulent are being researched. All of these models are being studied with the goal of providing a better description of turbulent fl ows. Most have provided some interesting insight. None have been developed to the point of being used by nonspecialists to analyze transport phenomena. All are extremely computationally ineffi cient and have no demonstrated performance on other than very simple fl ows. Th ese models are not expected to be practical for simulations on personal computer systems. Nevertheless, they are tools which will be utilized in
future engineering work, as indicated by the active research surveyed by Launder and Sandham (2002).
