Actuated turbulent boundary layer flows
This data collection contains time-series of three-dimensional snapshots for a parametric study of flat plate turbulent boundary layer flow actuated by spanwise traveling transversal surface waves with the goal of friction drag reduction. In the following some details regarding the numerical setup and the structure of the data is given.
A full description of the numerical setup and the parametric study is presented in .
The non-actuated turbulent boundary layer flow has a momentum thickness Reynolds number of around Reθ=1000.
The numerical method discretizes the compressible Navier-Stokes equations at a Mach number of M = 0.1, i.e., the flows can be regarded practically incompressible.
A total of 79 variations of the actuation parameters wavelength, period, and amplitude were considered. To save computational effort shorter wavelengths were partially simulated in domains which are narrower than the maximum wavelength. That is, six different domain widths were considered, therefore there are six different grids, one for each domain width.
Figure 1: Comparison of the flow structures between (left) the non-actuated reference case and (right) a case with high drag reduction N80; the contours of the λ2-criterion are shown.
Figure 2: Overview of the computational domain of the actuated flat plate. The incoming flow (blue) is created by an synthetic turbulence generation method (STG) .
Wall actuation function:
Transition function at the start (flat wall to actuated wall):
Transition function at the end (actuated wall to flat wall):
Figure 3: Sideview of the actuated wall showing the two transition zones at the start and at the end.
Figure 4: Overview of the sinusoidal wall function with all used wavelengths and the corresponding maximum amplitude. The spanwise extent is varied to fit an integer number of wavelengths into the domain; each of the sections defined by the dashed lines represents one of the six different mesh widths.
Description of the dataset
For every case an HDF5 file containing [u, v, w, ρ, p] is written out every ∆n = 300 time steps. The file includes only cells that are in the actuated region. Each snapshot is containted within a single boxOutputXYZ.hdf5 file, where XYZ is the time step n at which the snapshot was written out. Each boxOutput-file contains the following structure:
−> time [ attribute of solver time ]
−> timeStep [ attribute of solver time step ]
−> physicalTime [ attribute of convection time ]
−> physicalTimeStep [ attribute of convection time step ]
−> u [ structured dataset with (i,j,k)−indexing ]
−> v [ structured dataset with (i,j,k)−indexing ]
−> w [ structured dataset with (i,j,k)−indexing ]
−> rho [ structured dataset with (i,j,k)−indexing ]
−> p [ structured dataset with (i,j,k)−indexing ]
−> offseti [ attribute of i-offset of subvolume ]
−> offsetj [ attribute of j-offset of subvolume ]
−> offsetk [ attribute of k-offset of subvolume ]
−> sizei [ attribute of i-offset of subvolume ]
−> sizej [ attribute of j-offset of subvolume ]
−> sizek [ attribute of i-offset of subvolume ]
Works using this datset need to cite this manuscript:
 Albers, M., Meysonnat, P. S., Fernex, D., Semaan, R., Noack, B. R., & Schröder, W. (2020). Drag Reduction and Energy
Saving by Spanwise Traveling Transversal Surface Waves for Flat Plate Flow. Flow, Turbulence and Combustion, 105(1),
 Roidl, B., Meinke, M., & Schröder, W. (2013). A reformulated synthetic turbulence generation method for a zonal RANS–LES
method and its application to zero-pressure gradient boundary layers. International Journal of Heat and Fluid Flow, 44, 28–