CFD Online Discussion Forums

CFD Online Discussion Forums (http://www.cfd-online.com/Forums/)
-   CFX (http://www.cfd-online.com/Forums/cfx/)
-   -   CFX Solver error: floation point exception (http://www.cfd-online.com/Forums/cfx/24374-cfx-solver-error-floation-point-exception.html)

Kushagra Mittal August 6, 2007 13:47

CFX Solver error: floation point exception
 
Hi, After 4 iterations in my simulation, the CFX11-Solver throws the following error. I am unable to understand it. I will highly appreciate if some one can help me out.

Thanks, Kushagra.

ERROR #001100279 has occurred in subroutine ErrAction. | | Message: | | c_fpx_handler: Floating point exception: Overflow

Glenn Horrocks August 6, 2007 18:22

Re: CFX Solver error: floation point exception
 
Hi,

Read the documentation - "obtaining convergence"

Glenn Horrocks

Kushagra Mittal August 6, 2007 22:29

Re: CFX Solver error: floation point exception
 
Hi Glenn,

Thank you very much for your help. The document says that the solver failure could mean that non-physical boundary condition or initial values were applied.

I was trying to see the effect of restitution coefficient in 2-phase (liquid-solid) Eulerian-Eulerian simulation. And I see that the floating point exception error is thrown until the restitution coefficient is lowered than 0.9

Is the restitution coefficient greater than 0.9 not feasible? Could you please help with your valuable suggestions. I am very new in the CFD world.

Many Thanks, Kushagra Mittal.


Glenn Horrocks August 8, 2007 06:57

Re: CFX Solver error: floation point exception
 
Hi,

You would imagine particles with such high restitution factors would bounce around like billard balls, this sounds difficult to converge to me. A solver overflow is one type of convergence failure.

Glenn Horrocks

monkey1 January 10, 2013 07:46

As I just noticed this error is not solely due to a too big time step. It also occurs when using only "single precision" instead of "double precision override" option. With single precision the possible time steps are a lot smaller than those you can use with double precision while still getting convergence.
At least my overflow problem has just been solved by switching from single to double precision.

ghorrocks January 12, 2013 16:27

Yes, that's why double precision is listed as one thing to try on the FAQ: http://www.cfd-online.com/Wiki/Ansys...gence_criteria


All times are GMT -4. The time now is 12:19.