[prikeyma]

Hi,

I have a curved pipe with a half circular cross-section with a symmetry plane defined in the plane of the curvature. The y-direction is normal to the symmetry boundary and as expected the nodal values of the v-component of velocity are zero at the symmetry boundary. Strangely, the velocity gradient dv/dy does not equal to zero at the cell-center or nodal locations on the symmetry boundary. In addition, the pressure gradients are zero in the symmetry plane. Based on the description in the Fluent manual I expected the gradients d/dy to also be zero.

Can someone please explain why the velocity gradients wouldn't be zero?

Thanks

[Amir]

Originally Posted by prikeyma

Hi,

I have a curved pipe with a half circular cross-section with a symmetry plane defined in the plane of the curvature. The y-direction is normal to the symmetry boundary and as expected the nodal values of the v-component of velocity are zero at the symmetry boundary. Strangely, the velocity gradient dv/dy does not equal to zero at the cell-center or nodal locations on the symmetry boundary. In addition, the pressure gradients are zero in the symmetry plane. Based on the description in the Fluent manual I expected the gradients d/dy to also be zero.

Can someone please explain why the velocity gradients wouldn't be zero?

Thanks

Hi,
I think that there is a problem in setting B.Cs; in curved pipes you have secondary flow as a result of centripetal acceleration and in such cases you cannot use symmetry B.Cs. Am I right?

Bests,

[prikeyma]

The secondary flow structure has been shown in the literature to be symmetric across the horizontal centerline (from the inner wall to the outer wall) of the pipe cross-section. This is why I used the symmetry plane along the horizontal centerline. The results show the expected secondary flow structure within the cross-section when I use the symmetry plane but the velocity gradients at the symmetry plane are not zero for some reason.