||March 28, 2011 09:35
Blowing up in the DNS of boundary layer transition
AS the disturbance wave propagates to the high Reynolds number region where breaking down to turbulence occurs, my algorithm starts to blow up. I am using finite difference methods in downstream and wall-normal direction and spectral method in spanwise direction, and explicit RK method for temporal integration. I notice before the blowing up, there are strong oscillations in the near wall region. So I double it may be due to the poor resolution in the wall-normal direction or the violation of CFL condition in that region. Anyone has a clue? Thank you.