Turbulence Modeling for Computational Aerodynamics

摘要

Computational aerodynamics will have a significant role in the development of future aerospace vehicles. Several key pacing items must be addressed, however, before its full potential can be realized. Important among these is turbulence modeling, because the exact simulation of the dynamics of turbulence for practical applications is unattainable. Although new modeling concepts are being investigated, today's practical computations, and probably most of those that will be made in this decade, use the traditional modeling approach with some form of closure for the Reynolds-averaged Navier-Stokes equations. This approach is taken with the expectation that studies of the deterministic structures of turbulence and of numerical simulations of turbulence will provide more understanding and eventually will guide the development of models that include more physical information. Modeling in which traditional closure concepts are used eliminates important information about the dynamics of turbulence, and the universality of the approach has not been established. Therefore, limits of application must be determined by careful comparison with well-conceived experiments. Progess in that direction was reported for incompressible boundary-layer flows in the late 1960's. During the last decade, however, advances in computational aerodynamics have increased our ability to compute more complex flows markedly, and the burden of establishing limits of applicability for turbulence modeling has increased proportionately. The recent AFOSR/HTTM-Stanford Conference is witness to the breadth of that task. The scope of that conference was broad and there was no particular emphasis on external aerodynamic flows, as there is in this paper. For those unfamiliar with the progress and direction of the modeling of external flows, this paper will provide a perspective from which to view further developments certain to come in the next decade. For those entering the field of computational aerodynamics and who are interested in using various models in their computations, it will provide a basis of reference and, it is hoped, help them avoid unnecessary difficulties. For those active in the field, it will simply reinforce what they already know about the problems of modeling, although it may also suggest some new ways to view combined computational and experimental investigations. A complete review of modeling, even within the closure framework for the Reynolds-averaged Navier-Stokes equations, is beyond the scope of this paper, and the reader is referred to several excellent books that review the development of modeling and give details on the formulations of various models. This discussion is limited to a few important external aerodynamic flows of practical interest. Many of the examples, familiar to the author and to his colleagues, were chosen because the author had first-hand access to information about them, and because it is believed that they are a fair representation of the broader spectrum of today's modeling activities.