Dilute polymers in an oscillating grid
turbulence
U. Reiter, M. Guala, M. Holzner, W.Kinzelbach1 and A. Liberzon2
1 Institute of Environmental Engineering, ETH Zurich
reiter,guala,holzner,kinzelbach@ifu.ethz.ch
2 Department of Fluid Mechanics and Heat Transfer, Tel Aviv University
alex.liberzon@gmail.com
Since the drag reduction phenomenon discovery by Toms [1], it is well known
that turbulent °ows can be strongly modi¯ed by very dilute solutions of °ex-
ible polymers [2].
In spite of very extensive research, the physical mechanisms underlying
the e®ect of dilute polymers on turbulence remain poorly understood. Drag
reduction is the most well known large scale phenomena, though there is
a consensus that the direct action of the polymers is on the small scales of
turbulent °ows. It is of interest, therefore, to conduct experiments in which the
e®ects at large and small scales are easily separated and the e®ect associated
with the way turbulence is agitated could be isolated from the e®ect in the
turbulent bulk. There are several examples of turbulent °ows of this kind such
as spreading of turbulent cloud [3] or counter-rotating disk setup [4, 5], among
others. A di®erent approach is proposed in the present study. We analyze the
propagation of a turbulent front generated by an oscillating grid in an aqueous
dilute polymer solution. Using particle image velocimetry (PIV) the turbulent
°ow in a squared tank is measured. The study is based on a comparison of
the front propagation in an aqueous solution of Polyox WSR 301 (25 and 50
wppm) with the respective results in water [6].
It is found that the front propagates faster in polymer solution compared
to the distilled water case. This is due to the observation that the energy in-
put of the grid into the °ow is increased with polymer additives for the same
stroke, mesh and frequency of the grid. However, this is not the only e®ect.
When normalized with the °ow parameters, namely velocity and length scale,
estimated close to the grid, the front propagation after a certain time exhibit
a di®erent behavior in the polymer solution with respect to water. The es-
timated °ow parameters close to the grid correspond to the e®ective energy
input to the °ow. The di®erent propagation of the front in dimensionless form
directly indicates that the e®ect of polymer is signi¯cant also in the bulk of
the °ow. Apparently the strongest e®ect is at the turbulent front (where the
irrotational °uid is entrained into the turbulent °ow), the region of strong