I have a problem that I am looking at, basically I am looking at flow through regenerators (Heat Exchangers). This basically consist of 500 layers of fine steel mesh packed into a tube 40mm long. A lot of the work I have seen assumes that due to the 'flow straightening' characteristics, I can assume the flow through this mesh is Laminar. I'm not so sure? The query is with respect to looking at pressure drops. Could I analyse the pressure drop of say 1/10th of the length and by analogy just sum up to get the full length pressure drop. I know that Fluent has porous material analysis, does anyone know how the analysis works?

You are on sth which I am working on now.I assume flow though regenerator laminar because the reynolds based on hydraulic diameter is 800 or so. There do exsist some debate about the flow pattern and which model should be used under certain reynolds number. I did CFD modeling and experiment both , the pressure drop agrees well as laminar model applies. I use CFDRC.

You can treat with 1/10 then by analogy sum up to the full length.

No clue about the "flow straightening".

The real trick lies in heat transfer in porous media. I am really getting pain with CFD commercial packages when heat transfer is introduced.

Hope above all can help you

Most HT models use the form Nu=c * Re^m * Pr^n, where Nu, Re and Pr are Nusselt, Reynolds and Pradtl numbers, and c, m and n are the corellation coefficients, which differ depending on the porous medium microstructure and the flow regime.

The closest case to yours I had simulated was flow through a honeycomb (which indeed straightens the flow along the honeycomb cells). Therefore, the case is similar to flow and HT in pipes, and correlations may be found in standard textbooks (e.g., M.N. Ozisik, Heat Transfer: A Basic Approach, McGraw-Hill, 1985).

For turbulent flow (Re>10,000, 0.7<Pr<160, L/d>60) I applied Dittus-Boetler correlation (c=0.023, m=0.4, n=0.3 for cooling or 0.4 for heating).

When the flow is laminar, Nu is constant in my case. For square-cell honeycomb it is Nu=2.976 (see Ozisik).

I hope this helps, Rami

It is interesting topic. Do you have some publishes can be offered in here?

Steve,

If you meant whether we published this (in the context of flow through a honeycomb, which - in my case - is a part of a solar receiver) - unfortunately not yet (only in internal reports). However, this is rather standard treatment of pipe flow, as mentioned in my former posting.