Hi all,

I have switched from STAR to FLUENT some time ago. I see some strange behavior of the solver which I have never seen in STAR/ccm+. Residuals "peck" up for few interations and afterwards fall back within some iterations to the former lever. However, sometimes it results in an AMG error. I have external flow, T-Grid Hexcore mesh with qualtity much better than in STAR, coupled solver, steady state, realizable k-eps Turbulence with enhanced wall threatment. It is also very strange that this happens only on few cases, most of the cases (all similar) run fine. Anybody seen similar things? Thanks.

To add some more stuff: I use the coupled solver of FLUENT 6.3 (not the "density based" former coupled solver) Boundary conditions are velocity-inlet, pressure-outlet and symmetry on far walls (bounding box). I do incompressible calculations.