reversed flow at velocity inlet / mass flow inlet
 

hi everyone, i have a problem and am stuck on this, can't find a solution.

im modeling the flow of water entering a cylindrical chamber through 2 pipes of crossectional areas of 1mm2 and 2mm2. The Chamber has a diameter of 12 mm and height of 2 mm. Circular outlet is on top of chamber and has a cross section of 3mm2. i want to model the filling proces of the chamber, i.e. its first empty (= full with air) water flows through the two inlets at mass flow rate of 1ul/s into the chamber and fills it up, thus im using the VOF model. I've tried many BCs, mostly velocity inlets and mass inlets, but at this mass flow rate the water doesnt enter, but exits through the inlets which, of course, is wrong. Setting mass flow rates of about 20 ul or higher is accepted by fluent and water enters the domain.

Here are my model's definitions: unsteady, implicit, VOF-implicit, 2phase (i have also tried open channel flow. the result was the same) low Re-ke-turbulence, wall adhesion, surface tension, gravity & co, vel-inlet (tried mass inlet too), outflow (tried pressure outlet too), URFs are all 1, (doesnt diverge unless time steps are >0.05/sec), PISO, modified HRIC. after initialization i patched volume fraction of 1 for water-phase to inlet pipes.

Can someone help me? i cant find the reason why fluent doesnt obey the inlet boundary conditions. further keywords: microfluidics - creeping flow - stokes flow - VOF greetings, ib

Re: reversed flow at velocity inlet / mass flow in
 

hello again!

since i had used iso-clips and iso_surfaces i didnt notice that the interface between air and water is smeared over a large distance. looks like there is no way to get a true clean interface of small thickness.