[ashish.svm]

I am solving scalar transport in a thin Y shapes channel. Mesh has a very small cell at the at the intersection point.

This is my problem set-up
problemDefinition1.jpg

This is the mesh zoomed at the intersection point

intersectionCell.PNG

I think calculation of pressure gradient (velocity) is wrong at the cells near the intersection, because of very large contrast in cell size.

velocityMagnitude.jpg
Velocity in the outlet section should be equal to the sum of the velocity in the inlet section. But that is not the case with my simulation.

I have tried different gradSchemes such as
-extendedLeastSquares 0
-cellLimited leastSquares 0.0

What type gradScheme should I use for this type of problem? Or is there any other method to calculate velocity correctly?

[piu58]

If I understand your sketch right, you have very long cells in direct contact with the tiny cell in the intersection point. This is not recommended in fluid dynamics. I recommend to reduce the cell size stepwise toward the intersection point. BlockMesh supports this, it is easy.

[ashish.svm]

Originally Posted by piu58

If I understand your sketch right, you have very long cells in direct contact with the tiny cell in the intersection point. This is not recommended in fluid dynamics. I recommend to reduce the cell size stepwise toward the intersection point. BlockMesh supports this, it is easy.

Can you please explain the reason behind not using the tiny cell at the intersection point? Or if you can give some reference also that will be useful.

Thanks.