[alinik]

Hi all,

I have noticed that the same transport equation for a UDS would result in totally different distributions of that scalar inside the domain if one is curved and the other case is rectangular.



Of course for the curved channel it was rewritten for curvilinear coordinate system. So that would not be the reason.
I was wondering if anyone had similar experiences and how it could be fixed?

In this picture you can see the distribution. Please note the amount of drop is no longer the same.

[AlexanderZ]

Hello.

I'm confused, did you put cylindrical material properties for UDS Diffusivity for curved case?

Best regards

[alinik]

in terms of diffusivity it would not matter since it is a constant and is coordinate invariant. I re-wrote the source term for curvilinear system and it would not produce the same distribution as rectangular case.

Thanks,

[AlexanderZ]

As a first step I'm suggesting you to try to use cylindrical properties for diffusivity and use constant values for all directions.

(so you will put radial = tangential = normal = *some_constant_value*)

In my case I had uniform source, that's why I can't make you advice on that

Best regards

[alinik]

Thanks but the problem is not the diffusivity since it is isotropic.

My problem is with source term. Do you know how we can use coordinate systems other than cartesian?

[AlexanderZ]

Hello,

Unfortunately, I do not know the way how to use other coordinate system.
If you find the way, please share

Best regards