CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   Overflow Error Because of defining a New Coordinate Frame (http://www.cfd-online.com/Forums/cfx/100274-overflow-error-because-defining-new-coordinate-frame.html)

ashtonJ April 22, 2012 23:26

Overflow Error Because of defining a New Coordinate Frame
 
Dear All,

I got "Floating point exception: Overflow" error, I have read all the post about this problem. My problem is because of defining a new coordinate frame.

I defined a Poiseuille velocity profile as inlet boundary condition of my model but the original coordinate frame is not in middle of the inlet surface, so my velocity is not zero at the centre, it is zero somewhere close to the edge of the inlet. To resolve this problem, I defined a new coordinate frame on the inlet surface and used the new coordinate frame for defining the inlet boundary condition. The problem is solved and I got the zero in the centre, however, in the first iteration of run, I got the "Floating point exception: Overflow" error.

Does anybody know where the problem originates from.

Kind regards,
AshtonJ

ghorrocks April 23, 2012 01:26

I have just written an FAQ on floating point error because it comes up so much:
http://www.cfd-online.com/Wiki/Ansys...do_about_it.3F

But you comment that in this case it appears the problem is due to local coordinate system. Have you tried to define the velocity profile without the local coord? Sure the maths is a bit more tricky but I think you are right when you say it is probably where the problem lies.

ashtonJ April 23, 2012 01:44

Thanks so much. Without a local coordinate frame, I don't get a symmetric poiseulle flow profile, the zero velocity is not in the middle of the inlet, it is near the edge.

ghorrocks April 23, 2012 08:08

The coordinate frame is a simple transformation. You just include the transformation in your CEL expression then you do not need the local coordinate system. The end result is identical.

And if your problem is due to some problem with local coordinate frames this will work around it.


All times are GMT -4. The time now is 13:44.