(368d) Blending of Miscible Fluids

The blending process of binary miscible fluids is governed by both fluid advection and fluid diffusion. In low Reynolds number systems, where flow is slow and often one-dimensional, blending is diffusion-limited, and the effects of advection on blending are small. In high Reynolds number systems, where the fluids are fully turbulent, miscible fluid blending is governed by advection and the effects of diffusion are minimal. Between these two extremes, the blending behavior is a more complicated hybrid of both transport processes. This complexly is exacerbated in systems with strong difference between the two fluid densities and viscosity.

Traditional computational approaches for predicting the blending of miscible fluids are successful only in the limit of very low or very high Reynolds numbers. The limitation is motivated by the spurious effects of numerical diffusion, which tend to overstate blending. In this work, we show how DNS and LES simulations can be combined with high-resolution flux-limiting schemes to predict the blending of miscible fluids across a range of Reynolds and Schmidt numbers. After discussing strategies for controlling numerical diffusion, we apply this approach to model miscible fluid transport in both agitated systems and continuous flow systems.