Issues with time series data file

[Lucien Stolze]

Hello,

After updating Pflotran to its latest version, I observe that when reading time series data from file (e.g., LIQUID_FLUX FILE xx-yy), the simulation time step strictly follows the time steps provided in the data file. And this occurs despite having specified a larger MAXIMUM_TIMESTEP_SIZE.

The CYCLIC option leads to another issue as the stored time series becomes too large and raise an exception in line 1736 of realization_subsurface.F90. I have increased the criteria to override this exception (i.e., if (size(times) > 1e9) then). However, I do not think this should occur as the simulation takes ages to initialize when CYCLIC is in use.

I did not observe these issues with previous Pflotran versions. Could that be related to a bug or should I change the way to write the input file according to latest commits?

I attach some basic input files that should work with the main branch of Pflotran so that you can have a look.

Thanks,

Lucien

[Hammond, Glenn E]

Lucien,

 

Please see my comments in red below.

 

Glenn

 

From: pflotr...@googlegroups.com <pflotr...@googlegroups.com> On Behalf Of Lucien Stolze
Sent: Friday, July 12, 2024 9:05 PM
To: pflotran-dev <pflotr...@googlegroups.com>
Subject: [pflotran-dev: 6269] Issues with time series data file

 

Hello,

After updating Pflotran to its latest version, I observe that when reading time series data from file (e.g., LIQUID_FLUX FILE xx-yy), the simulation time step strictly follows the time steps provided in the data file. And this occurs despite having specified a larger MAXIMUM_TIMESTEP_SIZE.

This is by design. We must align time stepping with the changes in boundary flux for mass conservation, even if the flux is the same for successive times in your list. It would be too cumbersome to check for change as opposed to assuming that it changes.

The CYCLIC option leads to another issue as the stored time series becomes too large and raise an exception in line 1736 of realization_subsurface.F90. I have increased the criteria to override this exception (i.e., if (size(times) > 1e9) then). However, I do not think this should occur as the simulation takes ages to initialize when CYCLIC is in use.

I did not observe these issues with previous Pflotran versions. Could that be related to a bug or should I change the way to write the input file according to latest commits?

You are running for 11 years with hourly data. (11*365*24 = 96360). This should work fine. When I create the threshold number in the error message to 100,000, it is a bit slow as the waypoints are created, but it runs.

I recommend removing the redundant entries in the flux list when the value does not change. This will decrease its size.

I attach some basic input files that should work with the main branch of Pflotran so that you can have a look.

Thanks,

Lucien

--
You received this message because you are subscribed to the Google Groups "pflotran-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pflotran-dev...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/pflotran-dev/fc7efddf-c2d6-43df-a33d-e5ad38036138n%40googlegroups.com.