I use FIDAP to simulate a flow of molten plastic through a nozzle. Typically the viscosities of the plastic is very high (approx. 1000 to 10000 Pa s) and the velocities are very low (approx. 0.010 m/s), high pressure losses have to be expected (approx. 100 bar to 500 bar). I have severe problems to get convergence of the solution and if convergence is achieved the solution does not appear physically meaningful. I used the segredated solver as well as a coupled solver. For both solvers the solution diverged after more or less iterations. I have already tried several methods of stabilization and convergence acceleration etc. (e.g. INCREMENTAL, RELAXATION, UPWINDING). Also I have increased the viscosity incrementally via RESTART-Runs to get an initial solution for the subsequent iterations. Nothing of all helped so solve the problem. Does anybody has an idea.

I'm not sure if you should set up a nonNewtonian fluid in the PROPLEM difinition. Also, try to run the code for one iteration and see what's going on. As I always suggest on FIDAP: try to run the first 5 iterations using (S.S.) solver and then switch to N.R. solver. Good luck

With very low Reynolds number flows due to fluids with very high viscosity like molten plastics, I have had success by non-dimensionalizing the problem. Chapter 10 of the FIDAP 8 Tutorial Manual explains how to non-dimensionalize your problem and how this affects the parameters in your simulation. For your case the low Reynolds number formulation would be appropriate.

Let me know if you would like more information.

Viele Gruesse,

Elliot