[botman]

Hi,


I am trying to run the "Irregular wave sea state in 2D" example provided with REEF3D::FNPF with a harsher seastate (JONSWAP: Hm0=16.70 m, Tp=18.60 s, gamma=1.67). I ran 30 3-hour simulations with different wave seeds. Of the 30 simulations, only 6 ran the full 3 hours. Most crashed early on in the simulation due to issues at the inlet boundary (N 61 problem).



I have blindly tried a few options but to no avail, such as

• adaptive time stepping (default setup)
• reduced fixed time step (0.03 s -> 0.005 s).
  • It may be that this helps stabilize the simulation but I still see nonphysical vertical velocities at the inlet after a short period of time, wmax > 100 m/s, so I have never tried to run a full 3 hours with dt=0.005 s. The high velocity occurs in the top left of the mesh at the intersection of the inlet and free surface boundaries (see figure). For info, the wave seed used in this simulation is 5193 (i.e. B 139 5193)
• initializing the wave field (i 30 1)
• increasing number of elements in the sigma-direction (20 -> 60)
• ...

I was hoping some kind soul might point me in the right direction.



Thanks in advance

[valgrinda]

Hi Tim,

can you send me the input for one of the cases that crashes?

[botman]

Hi Hans,

Attached is one of the cases that fails (ctrl.txt, control.txt).

For reference, I have provided a case that ran through (control_s.txt, ctrl_s.txt). With the exception of the wave seed, the two attached cases are essentially identical. The only other differences are that, in the failed ctrl.txt/control.txt files, I have tried to refine the discretization (time-step and sigma direction)

The list of wave seeds that succeeded [7814,5375,12646,3563,10821]
The list of wave seeds that failed were [13350, 236, 12173, 5193, 906, 10956, 2896,5057,145,4226,7752,10990,14845,3463,15642,939 5,5397, 2963,2517,15244,8445,4765,8094,6543,563]

Thanks,
Tim

[botman]

Originally Posted by valgrinda

Hi Tim,

can you send me the input for one of the cases that crashes?

Hi Hans,


I don't suppose you've had a chance to look at this case?


Br,
Tim