FLOTRAN, Negative diagonal ... divergent solution

[maheshh]

Hello all,

I am modelling air flow between heatsink fins. To try a small problem, I tried it out with 2 fins and space between the fins. It works perfect.

Now I have moved on to a bigger problem where I am modeling half heatsink (symmetry) and airflow between the fins. Everthing is same as the small problem. But now I have started getting the following error 2 iterations into the flow solution:
"Coefficient matrix has a negative diagonal. Probably indicates divergent solution."

I have tried out all the help options in Ansys where they ask you to change conv, relx, etc. and none of them have worked so far.

Any hints??

Thank you in advance.

Mahesh.

 

[rikhan]

Hi Maheshh
I would suggest first to check your boundary conditions and the second is to play with Inertial relaxation factor MIR stabalization, start running your model with bigger values (1.0) for few iterations and then change it to either default values or 0.1. I hope it will help

 

[maheshh]

Thanks for the input.
Actually it was my mistake. For my Re number hand calculations I messed up the units (mm and m) and got a number of the order of 1e5 and therefore started with turbulent solver. After I got the error I re-did the Re number calculations and got the correct number which is of the order of 1e2.

Then solved the model with laminar solver and everything is fine.

Thanks,

Mahesh.