In simulations of unsteady turbulent flow by Reynolds-averaged approaches (URANS), turbulent time-scales always require special attention. The resolved time-scales of the large-scale motions have to be significantly larger than the high-frequency small-scale motions that are captured by the turbulence model. Problems can occur if parts of the spectrum are modeled as well as resolved. In the present investigations, the entire flow field is checked for the smallest occurring time-scales. Particularly in the near wake behind the Gurney-flap, the turbulent frequencies cover the same range of the spectrum as the resolved time-scales. Consequently, the application of the Reynolds-averaged approach is questionable. It should also be remarked that the turbulent frequencies depend on the time stepping itself [11]. In order to prove the reliability of the simulation, additional DES investigations are performed which should avoid the time-scale problem.
The DES results are obtained on the same numerical meshes and using the same
time step like for the URANS results.
Fig. gives an impression from where the hybrid method
switches between URANS (white) and LES (dark) depending on equation
(
). The LES mode is clearly limited to the near wake up to
about
downstream of the Gurney-flap, covering the region
of strongest unsteady effects and of greatest importance for the flow.
DES predicts similar mean aerodynamic forces as URANS
(table ). The
unsteady features of the flow can be compared in a spectrum of predicted lift
in fig.
.
One dominant peak corresponding to roughly the
same shedding frequency can be identified in the results of all methods.
However, the level of noise in the hybrid URANS/LES simulation is
higher than in the Reynolds-averaged method. This typical behavior of DES is caused by the resolution of turbulent
fluctuations in the LES region. Such very complex structures appear in the
near wake (see fig.
, lower figure). Compared to URANS,
stronger vortices occur but these are no longer concentrated in a single
dominating structure. The prediction of characteristic frequencies compared to
experiments can be improved by DES.