Dear all: I want to give a velocity inlet boundary condition by using UDF. As I have read, the velocity profile is given by F_Profile macro. So it is based on the face centroid corridnate. When I export the inlet velocity profile, I found the velocity is not as I want at the edge. I think it is because the Fluent software do interpolation process there, Is there any method which can give node-based velocity profile, So we can avoid the shift on edge.

The problem is u=1.5(y^2-2y+1) the velocity u at y=0 and y=1 should be 0. but in fact, when I use the follwing UDF.Afterig 1 step iteration, I export the inlet velocity profile and found the velcoity u is not 0 at y=0 and y=1. {... F_CENTROID(x,f,thread); y=x[1]; F_PROFILE(f,thread,position)=1.5*(y*y-2*y+1); ...}

Hi,

Have a look at the UDF manual about how to define a velocity profile (I think in the section "sample problems").

If you really want this mathematical perfection, refine the mesh at the edges. But is it really indispensable to get this perfection to obtain reliable results ??????

I think you simply lose time in details.....

Fab.

Dear Fab: Thanks for your suggestion! I have looked the section you mentioned. In fact, I think I use the ame method to define the velocity profile as the example, just use different maths function. When I plot the velocity profile at the inlet boundary, I find the velocity magnitude is what I wanted, but the X-velocity is differnet at the edge (seems to be explotated from inside). Whereas, in the boundary definition interface, I choose "component" method and just specify x-velocity using UDF, Y-velocity keeps to be 0.

What's wrong with it?

Thanks again! Regards!