Diverge in 2d dam breaking case
 

Hi all, I tried to use fluent to simulate the typical 2d dam breaking problem. If the viscosity of the liquid phase is set as 1cp, fluent can converge normally. However, if I increase the liquid viscosity to 10000 cp (polymer level), fluent diverge quickly. The analysis settings are PISO+geo+0.8 underrelax in pressure.Any comment or suggestion ?