Smaller time step leads to divergence?

Feidao Li

I am modeling the multiphase flow using Eulerian method. The thresholds for my convergence are 1.e-3. I found the continuity index fluctuates around 1.e-2 without further decreasing when I use deltaT=1.e-4, but the simulation can be convergent when I use deltaT=5.e-4. It is so weird. Can anybody explain this for me? I am really appreciated for that.

By the way, sometimes when I use deltaT=5.e-4, the continuity index can only decrease to 1.05e-3 and fluctuates around this value without going down below 1.e-4. But sometimes it can pass threshold very smoothly. What's the reason for that?

Thanks for reading my long questions