[JPBodner]

I'm working on a model of the dispersion of a solute injected into a fluid that is driven in an oscillatory manner at the boundaries. I want to model the dispersion over several hours, but the oscillation is at the 1 Hz scale, so resolving each oscillation is very expensive.

I am thinking about a simplified modeling approach where I take larger time steps (say 1 min), and use a modified diffusion coefficient to try to include the effect of the individual small oscillations. Can anyone tell me if this is a common approach and whether you know of literature or problem-types where this sort of scheme is used? The questions I have are around how to derive a suitable diffusion coefficient from a short-run model that includes the oscillations.

Thanks in advance
Jeff