(192a) On the Trailing Vortices behind Impeller Blades in Stirred Vessels

A still increasing computational speed and power, among other things due to the use of graphics processing units (GPUs), reveals ever more details of the flow fields in, among other devices, stirred vessels. A decade or more ago, we were happy to use vorticity to show that our simulations were capable of resolving the trailing vortices behind impeller blades, at least qualitatively. These days, there is a growing tendency to try and predict these trailing vortices quantitatively, and to use then these data for forecasting particle trajectories and bubble or droplet breakup in the impeller domain. Time for a review of the current state of the art in computationally simulating particularly these trailing vortices.

Over the years, both experimental and computational techniques have developed tremendously. We now have Stereo Particle Tracking Velocity with 2 or 4 high-speed cameras on the experimental side alongside Large Eddy Simulations (based on the lattice Boltzmann technique) and Detached Eddy Simulations, often supplemented with Lagrangian particle tracking routines, which are increasingly replacing the common RANS-based simulation techniques (on the basis of k-epsilon modelling). Of course, this leaves the option open that old experimental as well as old calculated data get outdated. However, what both the novel experimental and simulation techniques have in common is that they heavily rely on numerical techniques including data processing routines which are outside the scope of most users.

In this review, older and recent data about the flow over impeller blades from both experiments and simulations are evaluated and mutually assessed. The focus is on the trailing vortices behind impeller blades of both Rushton and Pitched-Blade impellers. The intriguing question is addressed whether or not, in these trailing vortices or otherwise, velocity vector magnitudes in excess of the impeller tip speed are feasible.