ABSTRACT
Direct Numerical Simulations (DNS) are frequently used to evaluate subgrid-scale models, but for practical flows full three-dimensional (3D) DNS is often computationally infeasible. It is a common practice to employ two-dimensional (2D) DNS, but since turbulence is inherently a three-dimensional phenomenon, the validity of this approach is questionable. In this study, we compare the evolution of a turbulent fuel/air mixing layer employing 3D and 2D DNS. While there are significant differences in detail, it is shown that employing either set of results leads to the same conclusion about the superiority of one subgrid-scale model over another.
Acknowledgments
The contributions of Professor Vinicio Magi in the development of the numerical code are gratefully acknowledged. The authors would also like to thank Professor Carlo Scalo for useful discussions about homogenous isotropic turbulence spectra.