[thapostol]

Hello guys. First of all thumbs up for a great forum. Superb job!

I am trying to simulate a laminar, steady state, incompressible liquid (water) flow through a micro-channel (trapezoid cross section), with a uniform inlet-velocity (couple of mm/sec).

The simulation's velocity results are then compared to actual microPIV measurements on a super-hydrophilic and a super-hydrophobic microchannel.

The fluent solution, with no-slip wall boundary conditions, provides velocity profiles in accordance with the measurements on the super-hydrophilic microchannel.

For the actual super-hydrophobic channel, experiments provide the slip velocity (and slip length) of the flow for the small base and the sidewall of the microchannel.

Now I need a way to fix these channel-height-related slip values on the fluid domain and calculate a solution.

I tried using DEFINE_PROFILE UDFs to impose wall motion velocities (but the results are nowhere near the measurements). Also I have been trying to use fixed velocities in the fluid domain with no luck so far.


Some questions:
1) Is FLUENT able to handle non-uniform wall motion for wall boundaries (in reality no wall can move that way...)?
e.g.

Vel(x)=a*y+b // a=a1, b=b1 for 0 <= y <Y1
                            a=a2, b=b2 for Y1 <= y <y2   ...

                            a=an, b=bn  for Yn-1 <=y <=Yn.

2) Having a universal XYZ coordinate system (X along the length of microchannel, Y height between parallel bases and Z along small base), how can I input the Y coordinates of the sidewall (54 degrees angle against the parallel bases) in a UDF? Will the use of universal x[1] coordinates for the sidewall ID thread be enough for the above mentioned velocity functions??

3) Fixed velocities values in the fluid domain.
I am trying to find out whether can I fix velocities in the cells adjacent to the thread zones of wall boundaries, or not.

I think maybe something like this udf not used yet (in part used for the aforementioned velocity profiles).

#include "udf.h"
DEFINE_PROFILE(fixed_velocities,t,nv)
{
	real x[ND_ND];
	double a;
	double b;
	double y;
	cell_t c;
	Thread* t;
	int idsidewall = 6;
	int idbottom = 0;
	
	if (THREAD_ID(t)==idsidewall)
	begin_c_loop(c,t)
	{
		C_CENTROID(x,c,t);
		y=x[1];
		if ((y >= 0.) && (y < 0.000041027))
			{
			a = 0.731231;
			b = 0.00125;
			}
		else if (y < 0.000049232)
			{
			a = 12.181716;
			b = 0.00078;
			}
		else if (y < 0.000065643)
			{
			a = 100.544204;
			b = -0.00357;
			}
		else if (y < 0.000082053)
			{
			a = 12.796535;
			b = 0.00219;
			}
		else if (y < 0.000098464)
			{
			a = -38.389605;
			b = 0.00639;
			}
		else if (y < 0.000114875)
			{
			a = -91.403822;
			b = 0.01161;
			}
		else if (y < 0.000147696)
			{
			a = -22.850956;
			b = 0.003735;
			}
		else if (y < 0.00015196)
			{
			a = -30.435494;
			b = 0.004855;
			}
		else
			{
			a=0.;
			b=0.;
			}
		F_PROFILE(c,t,nv) = a*y + b;
	end_c_loop(c,t)
	}	
	
	
	if (THREAD_ID(t)==idbottom)
	begin_c_loop(c,t)
	{
		C_CENTROID(x,c,t);
		F_PROFILE(c,t,nv) = 0.00125;
	
	end_c_loop(c,t)
	}
}

If I haven't explained appropriately my case please ask for clarifications.
Any ideas, pointers and/or corrections will be much appreciated!