Turbulence Measurements in Flows with Particles

Abstract

Despite the ubiquity of turbulent flows in nature, a general solution to the governing equations of motion remain elusive. These equations are nonlinear and nonlocal, and they describe a chaotic system with multi-scale characteristics, making its theoretical analysis difficult. One viable way of investigating high Reynolds number turbulence is by experimental methods, either in laboratory conditions or environmental measurements. The latter, however, are difficult. In one hand, environmental flows have a high Taylor scale Reynolds number ( > 10000) with a high mean velocity flow. This demands a high temporal and spatial resolution anemometer in order to adequately observe the smaller scales. High resolution anemometers, however, require very specific conditions to operate or are very delicate, being able to break even if a sand particle collides with them. These characteristics are in direct opposition to the uncontrolled, particle laden environmental conditions usually found in several types of flows of interest (e.g. inside ice clouds and sandstorms). Laboratory setups offer a viable alternative, albeit at a reduced turbulence intensity. For example, the Max Planck Variable Density Turbulence Tunnel (VDTT) allows for turbulent flows of Taylor scale Reynolds number Reλ ~ 6000. This Reλ is generated by an active grid and kinematic viscosity tuning of the working gas achieved by changing the operating pressure. This allows great flexibility at modulating the turbulence over a wide and continuous range of Reλ values. This setup, in particular, offers the possibility of not only obtaining single point Eulerian statistics, but also to obtain high Re Lagrangian statistics via particle tracking, which cannot be easily performed on atmospheric conditions. This allows not only to complement and corroborate atmospheric Eulerian statistics with more controlled experiments in laboratory, but also to study Lagrangian statistics (which is a field with much less experimental data available) and to analyze certain properties of the VDTT, such as its isotropy. In this thesis, a novel hot-wire sensor which uses a carbon nanotube based material as its sensing element is developed and tested. This device shows a frequency response larger than 20kHz, as well as superior mechanical resistance when compared to conventional Pt-tungsten anemometers of similar dimensions. This would allow the acquisition of high resolution Eulerian statistics on a wide range of environmental flows previously inaccessible due to technological constraints. This thesis also shows, as a separate chapter, the improvements performed on the already existing Lagrangian particle tracking (LPT) setup on the VDTT. The system has been improved to have a Stokes number 4 times smaller than the previous iteration. This thesis also demonstrates, by two independent but complementary methods, that the particles in the system are not electrostatically charged, but inertial. By comparing the Eulerian statistics obtained from both LPT and a pre-existing hot-wire statistics database, it was found that the turbulence intensity from LPT is consistently lower than before, pointing to the existence of sampling biases. A measure of the anisotropy of the system is measured from the LPT statistics, and it is found that the flow is mostly isotropic, even at the inertial range.