Starting a model from separate flow and transport restart files

[Orlando Silva Rojas]

Hi everyone,

I have run a steady-state flow simulation and a steady-state pure diffusion simulation, generating two separate restart files with obviously different data structures.

I was wondering if there is a straightforward way in PFLOTRAN to read these 2 files in the input deck to start a simulation of flow and transport. I have checked the manuals but did not find anything on this. I would also be happy with tricks allowing me to combine both restart files without the need of pre-processing. For instance, running very short simulations (e.g., 1 second) starting from the flow restart file to first include the flow field in the data structure for a flow and transport simulation.

I will acknowledge any comments or practical ideas.

Best wishes

Orlando

[Hammond, Glenn E]

Orlando,

Are you using the same grid?

Glenn

From: pflotra...@googlegroups.com <pflotra...@googlegroups.com> on behalf of Orlando Silva Rojas <orlando.s...@gmail.com>
Date: Friday, February 6, 2026 at 3:04 AM
To: pflotra...@googlegroups.com <pflotra...@googlegroups.com>
Subject: [pflotran-users: 8660] Starting a model from separate flow and transport restart files

Check twice before you click! This email originated from outside PNNL.

--
You received this message because you are subscribed to the Google Groups "pflotran-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pflotran-user...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/pflotran-users/CACSNO9NEmWVBBpSi3HP%2B2CPd%2BfJAh7iMs%3DgQB%2B_NJVh2LYJ5Rw%40mail.gmail.com.

[Orlando Silva Rojas]

Hi Glenn

Sorry for my delay in answering. Yes, I am using the same grid for both simulations.

Cheers
Orlando

To view this discussion visit https://groups.google.com/d/msgid/pflotran-users/SA1PR09MB10890F3D748AAA3C68040084D9A66A%40SA1PR09MB10890.namprd09.prod.outlook.com.

[Orlando Silva Rojas]

Hi again,

I realised in the past there was an option of initialising flow and transport from separate restart files, like the lines shown below

#=========================== Initialize ========================================
INITIALIZE_FLOW_FROM_FILE ./input_files/pflotran-restart_initialflow.h5
INITIALIZE_TRANSPORT_FROM_FILE ./input_files/pflotran-restart_initialtransport.h5

Do you know if the above still works or is deprecated? I've tried to use it, but got error message

ERROR: HDF5 dataset "Cell Ids" not found in file "./input_files/pflotran-restart_initialflow.h5".

Any idea?

Best regards

Orlando

[Hammond, Glenn E]

Orlando,

While the keywords INITIALIZE_FLOW_FROM_FILE and INITIALIZE_TRANSPORT_FROM_FILE are still supported, they are essentially workaround solutions. The INITIALIZE_FLOW_FROM_FILE keyword only functions with RICHARDS mode and both keywords use a different format compared to the .h5 checkpoint files. For these methods, you need to provide an integer dataset of "Cell Ids" along with companion state data: a "Pressure" dataset for flow, and datasets named after each primary species for transport.

A more reliable approach would be to restart the flow simulation using the .h5 restart file with the RESTART keyword. For detailed instructions, you can refer to this link: https://documentation.pflotran.org/user_guide/cards/simulation/restart_card.html. When configuring the transport process model, include SKIP_RESTART so that the code doesn't expect restart information for reactive transport and instead utilizes the initial conditions specified in the input file. More information can be found here: https://documentation.pflotran.org/user_guide/cards/simulation/subsurface_transport_modes/global_implicit_reactive_transport_card.html?highlight=skip_restart.

To define the initial concentrations, create datasets that capture your steady state diffusion solution within a TRANSPORT_CONDITION. For an example using uranyl (UO2++), refer to this input deck: https://bitbucket.org/pflotran/pflotran/src/7ae54275c5aa7605961da797fbfd61ef445ef5b0/regression_tests/default/543/543_hanford_srfcplx_param.in#lines-581 and the supporting .h5 dataset file: https://bitbucket.org/pflotran/pflotran/src/7ae54275c5aa7605961da797fbfd61ef445ef5b0/regression_tests/default/543/543_hanford_srfcplx_param.in#lines-286. You will need to set up separate datasets for each primary species.

Although you can theoretically use INITIALIZE_TRANSPORT_FROM_FILE combined with SKIP_RESTART for transport, placing all the free ion concentrations in a single .h5 file (with a single integer "Cell Ids" dataset and N_transport_dof floating point datasets named after each primary species), this method is not well tested, more likely to fail, and harder to support.

Regards,

Glenn

To view this discussion visit https://groups.google.com/d/msgid/pflotran-users/CACSNO9OYK3Y4%3DyqW68eXfeKADG2vQOs6xf2cV4TUp5Bj9C__ew%40mail.gmail.com.