[Ritah]

Hi,

I do simulations of a hot-dip galvanizing process ( a steel strip is immersed in a molten Zn alloy bath and is guided through the bath by stabilization rolls; Aluminum is added to the bath to obtain a good adherence of the liquid Zn to the steel surface). The bath temperature is done by inductive heating with a PID Control scheme and Zn alloy Ingots containing Al compensate the zinc and aluminum consumption.

The model takes into account the flow (N-S eq), turbulence (due to steel strip movement), energy and species Transport (fe and al Diffusion, Fe2Al5Znx and FeZn7 Formation) . Previous calculations proved that the K-Omega-SST model gave best results for the Zn bath Simulation , so we use K-Omega-SSt model with low Re corrections, Curvature corrections and Production limiters.

Since the transient calculation is too long (1-2 weeks) we are trying to obtain a Steady state solution. We have used a SIMPLEC scheme ( in the pressure-velocity coupling solver) with reduced under-relaxation factors and the results were promising regarding convergence. So we didn't have divergence problems, and the average temperature values of temp and al conc were stable. But there were still changes in the Velocity field (contour plots) -after many hundred thousend iterations-, so the Steady state was not reached.

This problem persists when we consider only flow and turbulence (without Energy and species Transport). In this case we have also tried the Coupled scheme ( in the press-vel coupling solver), but the results were even worse (fluctuations and Velocity field stability ) than in the SIMPLEC scheme.

Does anyone has experience or any ideas with Steady state stability and different computational schemes in Fluent?

Any answer would be appreciated. Thank you for your help.