unsteady flow problems

Daniel

Dear all, I am modeling flow around a cylinder at varying reynolds numbers to try to demonstrate the varying flow phenomena. When modeling laminar flow with steady wake I had no problems, but I am now trying to model unsteady flow (karmen vortex street, subcritical regime, critical regime and supercritical regime) and am finding that after the first time step I get the error:

'Error: Floating point error: divide by zero

Error Object: ()

Then after changing any parameters, if I try to re-initialize and start again, after the first time step, the following message appears:

Reading existing force monitor file cd-cylinder reynolds 1000B... Error: read-data: invalid plot file format Error Object: *eof*

An error message alert reading 'read-data: invalid file format' appears also.

I can avoid the second error by creating a new cd data plot file, but I shouldn't have to do that. Even when I do that and try again the first error appears again. Any ideas?

I have read advice on this site on the floating point error and tried many things.

I dont think its to do with boundary conditions, grid problems or invalid numerical computation because the same model worked for steady flow, all that is different is the viscosity and the fact that it is unsteady.

I have tried laminar, k-epsilon , k-omega, default time steps of 0.001, 0.02, 0.5, varying iterations per time step, simple p-v coupling, coupled p-v coupling, courant number = 1, 2, 200, increasing/decreasing under-relaxation factors, defining the turbulence in different ways...

not really sure where to go from here. Please help! any help would be really appreciated!