Hello,
In resolp, we compute a pressure variation to update the velocity.
To my knowledge, the Poisson equation is solved by a jacobi.
The REX about it is to use a precision around 10^(-5). Is the stop criterion applied on ||\delta p^{n+1}-\delta p^{n}||/||\delta p^{n}|| for the Jacobi method?
Best regards,
Laurent.
Hello,
The solver for pressure correction is a PCG or multigrid solver, not Jacobi.
The default tolerance is 10^(-5) for LES, 10^(-8) otherwise (but 10^(-5) should be OK in most cases).
I’m not sure about the details of the stop criterion.
Regards,
Yvan
Hello Yvan,
It’s a PCG because it’s a positive-definite matrix…
Thank you for answer.