Hi Everyone,

I'm modelling two-phase flows with fluent, and I want to try out some of the non-Newtonian viscosity models. These models only work with the laminar viscous model, so I need to make sure that the respective phases are in the laminar regime before applying the models. Does anyone know how to determine via hand calculation the regime? I know its easy to calculate the phase superficial velocity and subsequent reynolds number, but I feel this isn't an ideal method. Does anyone know of another method?

Thanks in advance,

Atholl

can you be more specific? geometry, fluids used... also do you mean regime as in turbulent or laminar or regime from a multiphase perspective (stratified, dispersed, annular...) ?

Hi, The geometry is an annulus, fluids used up to now are air and water. When I say flow regime I mean laminar, transition or turbulent. Flow pattern (for me) refers to the spatial configuration of the phases, i.e. stratified, slug, etc.

if you mean annulus as in annular flow you can usually assume the surrounding water flow in the ring to be laminar. especially if it's thin which is usually the case for annular flow. typically if any of the phases goes turbulent then you'll have increased mixing at the interface and annular flow is not likely anymore (depending of course on the stabilizing effects of buoyancy, surface tension and gravity). transition to dispersed regime would occur. now for the core flow (air) you can just consider it as a flow in a cylinder where the turbulent transition occurs at Re = 2100 roughly (use relative phase velocity).

I'm not sure the inversed situation of water core in an air annulus is possible but I might be wrong. good books on the subject are Govier & Aziz (flow of complex mixtures) and Wallis (one-dimensional two-phase flow) for a good global perspective.

hope this helps...

Hi, and thanks for the reply

My geometry is an annulus, so the flow patterns can be thought of as the same encountered in horizontal two phase pipe flow.

I need to find out which regime (turbulent or laminar) one or both of the phases are in. If I simly produce a two phase flow pattern map (superficial gas velocity vs superficial liquid velocity) and overlay the four Reynolds number zones (laminar gas & liquid, laminar liquid & turbulent gas, turbulent liquid and laminar gas and turbulent liquid and gas) this shows that when both gas & liquid are laminar, the stratified flow pattern is present.

I think this is too simplified, as I've simply calculated the reynold numbers of the phases using superficial velocity (phase flow rate divided by total flow area).

I need to know of any other ways to find the regime of each phase? I've got the Wallis book, so I'll have a scan through that in the meantime..

Cheers, Atholl

If your geometry is an annulus then you cannot apply the conventional flow regime maps for circular pipe flow. even the Re=2100 turbulent transition value doesn't apply anymore. you better try to find some papers on experimental transition work. something specific to your geometry.

But let's say this is a pipe flow problem, in which case what is the method to find whether either (or both) phases are in laminar, transition or turbulent regime?

for pipe flow you can always calculate the Re of each phase based on the superficial velocity and pipe diameter. use the 2100 limit. it is a good order of magnitude approximation especially for stratified flow. for dispersed, depending on the particle size and spacing with respect to the pipe diameter and turbulence inertial range you might have production or suppression of turbulence (Crowe did some work to that effect. published in the int. journal of multiphase flow). however, the particle turbulence suppression is not likely to eliminate turbulence all together so the 2100 is still a good rule of thumb.

the real way to do it is linear or nonlinear stability analysis (Orr-Sommerfeld and the likes) which is a phd thesis in itself and probably beyond the scope of your work.

Hi Ziad,

I was thinking along the lines of plotting pressure drop against flow rate (or possibly reynolds number) from an existing mechanistic model I've implemented, and looking for the point in the curve where the rate of pressure drop changes. I read somewhere that this is another indication of the onset of turbulence. Trouble is, the pressure drop expressions in two-phase mechanistic models vary according to assumed flow pattern, and I'm not sure how this will effect things. Have you any thoughts on this?

as you said, pressure drop models vary according to the flow regime. besides, sudden peaks pressure drop variation could be due to flow inversion (switch in the continuous phase). they usually are anyway. in the absence of transition analysis, I don't think you can rely on much more than empirical values or order od magnitude approximations for turbulence transition evaluation.

Ziad,

Thanks for your input on this,

Atholl

you're welcome. let me know what you end up with eventually.

cheers, Ziad

how different laminar or turbulent two-phase flow with other tow-phase flow.

student