Rob
Ansys Employee

Are you expecting the bed to be entrained by the air? DDPM sounds a good fit, but the packing model isn't necessarily ideal for the static part; Eulerian granular is better for the two sections but you'd need to consider whether you need a particle size distribution as that requires additional phases or population balance models. Try both and see how they behave. Ansys Rocky coupled to Fluent is a good choice, but is lacking combustion models unless you fancy messing with the advanced parts. 

The particle volume fraction is a normal report from Fluent. If you mean the individual particle porosity following a reaction, that's rather more involved as there isn't a model for that: we assume (or calculate) a size and composition based on what's going on in the domain.