I am trying to model radial turbine in CFX.. During the run, i am getting the error message:

Floating Point Exception: Overflow

What could be the reason for this..What changes can be done in defination file, I have given mass flow inlet and refrence pressure as opening with 0 pascal

Thanks & Regards Riyaz

Very common error. Do you mean your reference pressure is set to zero? If so that has been known to occasionally cause numerical instability. The simplest thing you could do for starters would be to run double precision, as round-off errors can lead to this error. This could also be due to your boundary conditions. You might try starting with something less aggressive and ramping up over the course of a few hundred iterations iterations. The grid itself could also be the problem, if it's too fine or coarse in certain areas it could easily lead to overflow errorsn (check aspect ratios, should not be >300, definitely <500).

John

I think it is because of improper boundary condition and/or initial conditions. If you are using k turbulent models. It is likely that you have ridiculous initial values for variables such as turbulent intensity.Try changing these values. Hope this help

Pat

in cfx where to change the initial values, for turbulent intensity and other values

Regards........... Riyaz

[Pirba]

Hello, I had the same problem running k-epsilon and omega simulations in star-ccm+. When I refined the mesh to 5E6 cells I got the floating point exception error after 17-27 iterations. My initial pressure was set to 0, because I was working with low Re and pressures. When I changed my initial pressure to 8 Pa and my turbulence from length scale to k- epsilon everything went smooth again. I am simulating flow over a packed bed of spheres.
Regards