[mnabi]

Dear all;

I am solving full NS equations by implicit-explicit method. I am applying Crank-Nicolson for diffusion, and second order Adams-Bashforth for convection term. But I am still annoying from a small CFL number. The time step becomes one order smaller than the smallest time scale of turbulence.

and my question: Is there any method to increase the CFL number, and at the same time conserve the second order of accuracy in time?

I tried Crank-Nicolson for both, diffusion and convection. But I am still annoying from a small CFL. Is CN method is not unconditionally stable for convection term?

Thank you in advance