[heather]

Hi,

I attempted to create a very simple steady state solver for scalar transport to test the various div schemes. However, I've been getting some strange results...

I've attached the solver and a little test case - is just a unit square, pure convection, with a uniform flow from south-west to north-east, and the scalar is set to 1 on the west and north, and 0 on the south and east (all fixed value). The solution appears to converge, but instead of the one-way transition of the scalar from 0 to 1 across the north-west to south-east diagonal, I'm seeing a 'pocket' in the north-east corner...

steadyScalarTransportFoam.tar

test1010_upwind.tar

Any comments are appreciated

Andy

[niklas]

Just a question.

Is the problem well-posed if you specify properties on outlet boundaries?

N

[heather]

Good point - no...

changed outlets to zeroGradient for U/T and behaving more as expected.

Thanks,

Andy