[josephmp]

Hi everybody,

I am trying to model Liquid flow in a microchannel over air filled cavities using transient VOF in 2D. At low Reynolds numbers(Re<10) I m facing difficulties for convergence and often reversed flow appears. I m using a velocity inlet and pressure outlet condition. even after large number of iterations i am not getting expected pressure variation along the channel length. But at higher Re values i am getting fair results. what can be the reason for this instability at low Re values?. How can i rectify it?

[sadjad.s]

hi mate,
if you're using Patch tool, try using "Patch reconstructed interface" method which introduced in FluentV16.
Also structured mesh always gives better convergence.
And try switching between "Body face weighted" and "PRESTO!" for pressure interpolation. i don't really remember, but one of them gave fast convergence for my case.

[josephmp]

thank you sadjad for your suggestions......i m trying it.

[josephmp]

hi sadjad,
i tried the structured mesh case and it gives better convergence but the streamlines obtained from this is slightly different from that of the unstructured case and is wrong. the problem is at sharp edges. but the streamlines from unstruct-mesh looks good eventhough converegence is slow. why two meshes with same BCs give different results?

[josephmp]

still my low Re problem persists......can i use Body force weighted scheme when
body forces are negligible in my case?

[sadjad.s]

Originally Posted by josephmp

hi sadjad,
i tried the structured mesh case and it gives better convergence but the streamlines obtained from this is slightly different from that of the unstructured case and is wrong. the problem is at sharp edges. but the streamlines from unstruct-mesh looks good eventhough converegence is slow.

I don't know the exact reason, but in mechanical problems, struct. always converge faster and mesh independency is gained with lot less number of elements.
Also in drag/lift cases and wall bounded problems (where boundary layer is of great importance), struct. gave me relative better results.
what were your elements' aspect ratio.
also try refining mesh in interfaces.
for sharp interface, try using ANSYS16 introduced sharp interface capturing.
it adds artificial compression for volume fraction equation. it has 3 methods (sharp, dispersed, sharp/dispersed).

Originally Posted by josephmp

. why two meshes with same BCs give different results?

in theory, if you refine two meshes at a very low level, results should be same.

[sadjad.s]

Originally Posted by josephmp

still my low Re problem persists......can i use Body force weighted scheme when
body forces are negligible in my case?

did you consider gravity?

[josephmp]

thank you sadjad,

i didnt activate gravity, since i m studying microchannel flow with very small length scales the effect of gravity is negligible......so only forces are viscous force, pressure force and surface tension.

[sadjad.s]

So the base choice would be PRESTO!.

[josephmp]

thank you,
i shall update developments....