code_saturne User's Forum

Hello,
In case of diverging calculations, using a relaxation of pressure increase is often what helps the most. A relaxation coefficient of 0,7 to 0,9 usually works (you seem to have used 0,05 which is very small).
Be careful, the relaxation coefficient for steady or unsteady calculations is not the same. Are you sure you used the correct one ?
Also, if your calculation starts diverging before it crashes, you may postprocess it at time step 18 or 19 to see where the divergence initially appears. Chances are this occurs in an area where your mesh is warped or otherwise non-orthogonal.
Best regards,
  Yvan

thanks
     The divergence appears at the inlet where the pressure is larger. I met the same problem in a simple case. It's a 2D symmetry gas combustion,  The pressure and the velocity Z at the fuel inlet is much greater than I set in the boundary conditions.  The details see attachments.