[Sun qi]

Hi,
Hello, my program was running normally when I adopted the adaptive time step, but when I changed it to a fixed time step, the program stopped running one step and the display speed exceeded N 61.
My changes were to remove N 47 0.3 and add N 48 0 and N 49 0.5.
The time interval is about 0.01-0.05 when the adaptive time step is adopted. The total running time is 80 seconds.I made the change in the hope of reducing the output data. Because my force output diagram is not very accurate, the attached diagram is as follows.
Looking forward to hearing from you.

[kamath]

Hi Sun qi,
I think you are applying the wrong medicine to your problem.

Using N 49 0.5 means a time step that is 10 times higher than the required time step when the courant number is satisfied with a factor of 0.3.
Therefore, your simulation will crash with too high velocities (N 61 exceeded).
This is how it is supposed to be.

Your problem, if I understand correctly is that you want to reduce the output produced / printed out from the simulations.

For the vtu output, use P 30 or P 35 to define how often you want the output.
For .dat (text files) which print out spatial data (eg. wsfline) use P 55.

For time series output (force, wave gauge), there is no point in restricting the output (you will get data that can't be plotted nicely) and there is no option to restrict this. The output is every time step of the simulation.

I hope this solves your problem, which I think is to save disk space in runtime.

[Sun qi]

Originally Posted by kamath

Hi Sun qi,
I think you are applying the wrong medicine to your problem.

Using N 49 0.5 means a time step that is 10 times higher than the required time step when the courant number is satisfied with a factor of 0.3.
Therefore, your simulation will crash with too high velocities (N 61 exceeded).
This is how it is supposed to be.

Your problem, if I understand correctly is that you want to reduce the output produced / printed out from the simulations.

For the vtu output, use P 30 or P 35 to define how often you want the output.
For .dat (text files) which print out spatial data (eg. wsfline) use P 55.

For time series output (force, wave gauge), there is no point in restricting the output (you will get data that can't be plotted nicely) and there is no option to restrict this. The output is every time step of the simulation.

I hope this solves your problem, which I think is to save disk space in runtime.

I am very glad to receive your suggestion. My understanding is as follows.
My data was exported from reef3D_6DOF_forces_0.dat in the REEF3D_CFD_6DOF file. You mean the amount of this data cannot be changed, right? Can't even change the time step to get better results?

[kamath]

For the file that you mention:
Yes, that is a time series, so cannot manually change the output.
You can printout more data by changing the time step, but you have to decrease the time step size and not increase like you have done.

The simulation will be slower, without much improvement in the results, in my opinion

[Sun qi]

Originally Posted by kamath

For the file that you mention:
Yes, that is a time series, so cannot manually change the output.
You can printout more data by changing the time step, but you have to decrease the time step size and not increase like you have done.

The simulation will be slower, without much improvement in the results, in my opinion

I see. Thank you for your help

[kamath]

Something I forgot to mention, since you started with N 47 0.3, you still have room for improvement up to N 47 0.1.

I mostly use N 47 0.1.
I have seen decent results with N 47 0.3 for wave propagation but have not thoroughly tested it. You can carry out a time step convergence study if you wish by varying the N 47 parameter- 0.3-0.2-0.1-0.05 etc.

[Sun qi]

Originally Posted by kamath

Something I forgot to mention, since you started with N 47 0.3, you still have room for improvement up to N 47 0.1.

I mostly use N 47 0.1.
I have seen decent results with N 47 0.3 for wave propagation but have not thoroughly tested it. You can carry out a time step convergence study if you wish by varying the N 47 parameter- 0.3-0.2-0.1-0.05 etc.

Ok, I understand what you mean, and I will try your proposal.