Obstructed internal flow - convergence stall
 

Hi everyone,

I'm modelling an unsteady 2D channel flow with a semi-circular obstruction located midway through the channel. The inlet velocity is uniform but is gradually increased (from zero) with time.

I noticed that after a certain inlet velocity, the solver residual is not monotically reducing as before - instead the convergence is somewhat stalled !?

Why is that ?
And, how could this be overcome ?

Appreciate all the response.