Poor convergence with high streamwise loss multipler
 

Hi,
I am doing flow distribution analysis where I am using a porous media to simulate many small channels within a heat exchanger. I want to force the flow to be parallel within the porous media, which necessitates a high streamwise coefficient multiplier. The difficulty I am having is when I make the multiplier much bigger than 10, the solution convergence becomes highly oscillatory and I can't get good convergence.
Any ideas?

I suspect the tips here will be relevant:

http://www.cfd-online.com/Wiki/Ansys...gence_criteria

Apologies if I am not remembering this correctly, it's been a couple years since I did this kind of simulation, but:

you are saying you need a large stream-wise loss coefficient. I would think that in fact the directions _normal_ to the stream should have the very high loss coefficients to suppress crossflow. Your stream-wise coefficient should be whatever you need to simulate the expected pressure drop through the HEX for a given flow rate.