[alextau]

Hello,

I ran a simulation till 30 seconds run time, reconstucted the case removed all processors folders. Now I'd like to change velocity inlet boundary and continue the run from the same run time (30 seconds).

What are my steps for changing the velocity? Is it right to change the U boundary in 0 folder, then apply the command decompose -latestTime and then run my application at startTime latestTime in controlDict? Or the change must be made in the 30 folder? Please help, I am confused.

Thank you,
Alex.

[Tobermory]

Typically I would start by making a copy of the old 30 folder somewhere, so you have a copy of it, in case you need to go back to it. Now just edit the U file in the 30 time folder ... the search for the word "boundary", since you'll have all the internal field data first (this may be in binary, but the boundary names etc will still be in ASCII). Then just restart from this 30 time folder, with the latestTime flag as you noted (you'll need ro decompose again, with latestTime flag).

Good luck.

[alextau]

Great thank you!!! How do I search for text in terminal? For now I opened U file usng gedit command, however I could not find boundary there...

[Tobermory]

For large files, I find that gedit struggles to open them. From a terminal, I use vim (I am old school), and you can search for the word boundaryField very simply by just typingh

/boundaryField

Another terminal tool that I see being used often is nano ... I haven't tried that myself though.

[Yann]

Hello,

Another way around would be to use the "changeDictionary" utility to update your boundary conditions. You can directly use it in parallel, on the latest time, in order to avoid reconstructing your case.
Have a look to the tutorials to get more details, but your changeDictionaryDict should look like this, or something similar, depending on the OF version you are using :

U
{
    boundaryField
    {
        my_inlet_patch
        {
            type            fixedValue;
            value           uniform (0 0 0); //just set your desired velocity vector
        }
    }
 }

Cheers,
Yann

[alextau]

Thank you. This is also very helpfull!

[PSander]

Hello,
I am facing a similar problem. I run a case in parallel until a steady state equilibrium is reached. Then I would like to change my inlet velocity and as well calculate transient.
Therefore I did what you explained, additionatly I change in system folder from "steady state" to "Euler". After a few iterations the simulation crashes with a negative initial temperature. Do you know, what the problem could be?

[alextau]

Well, I faced the same problem and there are many suggestions how to solve it here. Decreasing the maximum Courant number is what was helpful for me, but obviously it will increase your computation time…

[PSander]

Thank you for your fast reply. My courant number should be low enough, since my maxCo =1 and adjustTimeStep = true, or not?

I managed to change my inlet, but it is only working while I stay on steady state calculation. The negative initial pressure is occuring, if I continue the run with euler. And it happens running it parallel and non parallel, this should not be problem.