[pi120]

Hello,

I can't understand conceptually how Star-CCM+ allows to set any arbitrary CFL number (Courant number under solvers) for the coupled implicit solver.

I already have a grid (so a Dx) and I already set a time step (Dt). I also have a flow velocity over the domain (u). So the CFL number should already be defined as

CFL = u*Dx/Dt

What happens when I set an arbitrary CFL (e.g. 50)? It doesn't seem that my time step gets overwritten, and for sure the grid and flow velocity do not either...

Thank you
Pietro

[pi120]

I meant

CFL = u*Dt/Dx

[LuckyTran]

It is a dual-time marching scheme. The dt you specify is the global physical time specified across all cells but under-the-hood you are taking multiple pseudo time-steps at each individual cell all converging at different rates to eventually arrive at the solution at the physical time across all cells that you specify. The relative size of the underlying pseudo/virtual steps is controlled by the Courant number. This is property of coupled/density-based solvers. If you notice, you don't specify how many iterations are performed at each time-step like you would in a segregated flow solver.

[pi120]

Ok, makes sense, even though you still specify the number of inner iterations for each time step. I have found now some more info on the user guide under 'implicit unsteady'.
Thank you!