Hi people. Hope 2001 has been good thus far. Cutting to chase :

A pile of solid(probably in cone shape) of varying particle size. Exit is controlled at base of pile. We know what we put into the pile. Can we model how the solid settled and what will come out ? Can we model it on fluent I mean , to an accuracy of about 1-1.5 m. I gather it will be a air,solid mixture with high solid loading. My two problems I can forsee is that the 1) the pile will get smaller as solids exit (can that be modelled?) and 2) can the flow models in fluent model the model realistically ?

I appreciate all input. Past experiences in this field would be greatly appreciated.

Thanks

This is a complex problem of bin/hopper/bulk solids storage design/distribution. Using the granular multiphase model in FLUENT for this is overkill.

Write a program to model your pile using control volumes. A collection of control volumes forms a slice Many slices form a layer of solids of unifrom composition. The mass in each discrete slice can be adjusted to refine the analysis. The smaller the mass in the slice, the more slices per slag layer and the outlet concentration is more accurately resolved The larger the slice mass, the greater degree of smoothing or averaging of the slag concentration leaving the pile.

Well there are limited choices:

a) try the granular model in Fluent 4.5. I think this can model the settling you are talking about. I haven't tried it and I think the model is more applicable to fluidised conditions, but I think it can also model settling - there are some examples in the manuals.

I am trying a different approach for modelling a settling rubble bed. In this case I am trying to implement a kinematic model in Fluent 5, but its not quite working. In this case I assume a much simpler solids flow and for the free surface which is not yet implemented intend to use an approximation. The reason I'm tackling this approach is because I also want to add chemical reaction as the cause of the rubble settling, and the granular flow model is probably more complex than needed.

Greg