Hi every body
I saw in cfd book and cfd references " the cfl (Courant) number should below than 1' but in official star-ccm+ suggested bigger cfl for outer current to active HRIC scheme. whats your suggestions.
Thanks in advance.
To understand the role (numerical but also physical) of the CFL number, you can consider the 1D case CFL=u*Delta_t/h. Just rewrite u=dx/dt so that you can write CFL=(dx/dt)/(h/Delta_t). This way you see that the number is the rate between the exact slope of the characteristic line and the slope of the discrete approximation. When CFL
Dear Mohammad,
CFL condition is only meaningful when you are dealing with an "explicit" scheme, or part of the scheme is explicit (e.g. imex). The internal switch of Star-CCM+ automatically finds the largest possible time step and refine your time step based on that, so in practice you do not need to worry about that CFL# larger than 1.
Hope it helps,
Kaveh
To understand the role (numerical but also physical) of the CFL number, you can consider the 1D case CFL=u*Delta_t/h. Just rewrite u=dx/dt so that you can write CFL=(dx/dt)/(h/Delta_t). This way you see that the number is the rate between the exact slope of the characteristic line and the slope of the discrete approximation. When CFL
Hi Kaveh Zamani
Thank you for attention. what switch star use for refining time step? and where I can find this switch? how can I find results for maximum possible time step from a star?
thank in advance
It is an internal switch, you usually cannot adjust it, you may put a cap on your acceptable Delta t (or sometimes CFL).
In many numerical schemes error is proportional to a constant time (1-CFL^2) so you would rather use CFL values close to one and not reduce it, unless there is solid reason for that
You might also want to change the CFL number depending on the actual simulation you are dealing with. For example 2D steady simulations of airfoils might run with CFL numbers as high as 500 leading to very fast convergence for steady cases. However such a high CFL number is completely inappropriate if you deal with unsteady aerodynamics. So it really depends. What do you actually want to simulate?
- Badges
- Science topic
- Similar topics
- Books