Turbulent drag reduction in a closed flow system: Boundary layer versus bulk effects

Abstract

We present an experimental study of drag reduction by polymer additives in a closed turbulent flow with a zero mean velocity at large Reynolds numbers. The study is performed using two different forcing schemes: a smooth forcing, for which the fluid is brought into movement through viscous boundary layers, and a very rough forcing where the fluid is stirred inertially by baffles. For a polymer solution of 30 wppm, a steady drag reduction is only observed for the smooth forcing. In the other case, absolutely no decrease of the turbulent energy dissipation is observed. When drag reduction is observed, the turbulent intensity in the bulk is unaltered upon addition of the polymer. This result demonstrates that, in our experimental system, the reduction of the energy dissipation does not only occur mainly in the boundary layers (as is well known for a long time), but takes place solely in the boundary layers. We also study heterogeneous drag reduction by local injection of a concentrated polymer solution. During and just after the injection, similar effects on the measured pressure histograms are observed for the two forcing schemes. However, these effects are much stronger and last much longer, for the “smooth” forcing than for the “inertial” one. With the use of a hot film probe as a tool to detect the homogeneity of the solution, we can show that these effects are due to the mixing of the concentrated polymer solution. The measured pressure histograms also provide information about coherent structures in the flow. We show that during the mixing of the polymer solution, coherent low-pressure filaments are partially suppressed, also when no drag reduction is present: the coherent structures are thus not directly related to the energy dissipation.