[fgommer]

Hi,

I’m trying to model the transient flow of a fluid through an array of circles in 2D. The flow is assumed to be non-turbulent and creeping. I tried to follow the recommendations given in the Fluent user’s guide and other sources but I am struggling to find some sort of settings which result in some kind of flow behaviour which looks ok.

The attached example mesh (.bdf) is a simple 2D flow channel with a circle in the centre. The elements used are triangular and locally refined around the cylinder. The fluid flow is entering from the pressure inlet boundary on the left and the air is removed through the pressure outlet on the right hand side. On the top and bottom boundary, symmetry is assumed.

For this simple impregnating flow example I use the explicit VOF model in order to visualize an accurate time dependent transient flow front. If I run this case with the settings described in the attached journal file (.jou), I get a flow profile of the entering fluid which does not look like something which would happen in reality. The fluid flow front is already deforming quite significantly before it touches the cylinder (.jpg). I would expect, however, that the fluid would ‘hit’ the fibre in quite a straight line and then forming a partially parabolic flow front…

I tried to play with the material properties (increased the fluid viscosity by two orders of magnitude) or played with the inlet pressure (=velocity). None of these settings seem to change the flow profile. I tried to change the solver to Coupled as well and the flow front becomes a bit more realistic but still not the way I would expect it should look like. Of course, I tried to increase / decrease the time step as well…

I’m kind of wondering what is going wrong at the moment. Maybe someone got a good idea of what I could try next? The version of Fluent I’m using is 12.1.4 in parallel or serial.

Thanks a lot for your help!
Frank

PS: Unfortunately, I just had to learn that a file size of 1.4MB for the mesh (.bdf) is too large for this forum. If you like to give the mesh a go, send me a mail and I will email you a copy...