A numerical study of self-similarity in a turbulent plane wake using large-eddy simulation

Sandip Ghosal*, Michael M. Rogers

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

44 Scopus citations


Turbulent wakes are known to develop self-similarly sufficiently far downstream from obstacles that generate them. It has long been assumed that the spreading rate of the wake in the self-similar regime is independent of the details of the body generating the wake, being dependent only on the total drag (or momentum deficit) This assumption seems to be in contradiction with some recent experiments. In this study we attempt to complement these experimental investigations through a numerical study of a time-developing wake. A numerical study has the advantage of eliminating many of the uncontrolled factors present in experiments and allowing precise control of initial conditions. Large-eddy simulations employing the recently developed dynamic localization model are used to extend previous results from direct numerical simulations. The large-eddy simulation results are compared to the direct numerical simulation database, wherever such comparisons are feasible, as a check of the method. Like the experiments, the large-eddy simulations suggest that non-unique self-similar states, characterized by different spreading rates and turbulent statistics, are possible and that they can be maintained for significant time periods. The study also demonstrates the predictive capability of the dynamic localization subgrid model.

Original languageEnglish (US)
Pages (from-to)1729-1739
Number of pages11
JournalPhysics of Fluids
Issue number6
StatePublished - Jun 1997

ASJC Scopus subject areas

  • Computational Mechanics
  • Condensed Matter Physics
  • Mechanics of Materials
  • Mechanical Engineering
  • Fluid Flow and Transfer Processes


Dive into the research topics of 'A numerical study of self-similarity in a turbulent plane wake using large-eddy simulation'. Together they form a unique fingerprint.

Cite this