Re: [PFLOTRAN] Stochastic --multi_realization option
Hammond, Glenn E
Marino,
Once you have created all the combinations of porosity and permeability (assigning a unique integer ID to each combination), you can run the scenarios using two options:
(1) run all scenarios with --multi_realization
(2) run each scenario separately using --realization_id #
I prefer the latter as there is more control if a given realization fails to run to completion.
Examples of each located in the regression tests. The links below are to the command lines used to execute each
--multi_realization
--realization_id
These scenario reference the datasets in https://bitbucket.org/pflotran/pflotran/src/master/regression_tests/shortcourse/regional_doublet/perm_fields_10x5x4.h5
You must add the integer ID to the dataset name (e.g., Permeability2, Porosity5).
Please take a look at these examples and ask further questions.
Regarding the 2015 PFAS scenario, I need more information for context. I do not recall that simulation.
Glenn
From:
Marino V. Zuccolini <marino.z...@unige.it>
Date: Friday, June 13, 2025 at 2:41 AM
To: Hammond, Glenn E <glenn....@pnnl.gov>
Cc: Micaela Raviola <micaela...@edu.unige.it>, marianna.miola <mariann...@ge.imati.cnr.it>, DANIELA CABIDDU <daniela...@cnr.it>, SIMONE PITTALUGA <simone.p...@cnr.it>, TOMMASO SORGENTE <tommaso....@cnr.it>
Subject: [PFLOTRAN] Stochastic --multi_realization option
Check twice before you click! This email originated from outside PNNL.
Dear Glenn,
We have developed open-source geostatistical code for multidimensional unstructured meshes called MUSE (Modelling Uncertainty as a Support to the Environment), which can be used in real time for continuous and indicator variables with (un)supervised variography
automation. The code is not yet available, but it will be soon. The next step is to integrate some tools to prepare meshes for use with PFLOTRAN.
The plan is to make the computing chain transparent, starting from boreholes and ending with Paraview visualisation, so that uncertainty can be evaluated. To do this, the stochastic --multi-realisation option should be used.
Specifically, we have in mind two meshes: one for average porosity and one for class permeability, both of which are embedded in the pflotran.in file. If we need to compute uncertainty, we use one equiprobable realisation for porosity and one unique (best guess)
permeability.
As I understand it, we need to embed all the porosity simulations in an HDF5 file using the --multi_realization option.
Could you please send me a template with the correct format for identifying the ID for each scenario? The permeability is a best guess and should always be the same for each PFLOTRUN run.
I have another question about a file that I found on Google Groups, and as I can understand it was written by you in 2015, about a sandbox used for PFAS. Could you please send me the reference paper about the meaning of the constant you defined for computing
adsorption?
Thank you for your attention.
m.
*******************************************************
Marino Vetuschi Zuccolini
marino.z...@unige.it
Professor / Geochemist
Laboratory of Geochemistry [LGCH]
ORCID ID: 0000-0002-2874-631X
WOS ResearcID ABC-1336-2020
Associate to the CNR-IMATI Department
https://gcc02.safelinks.protection.outlook.com/?url=http%3A%2F%2Fwww.ge.imati.cnr.it%2F&data=05%7C02%7Cglenn.hammond%40pnnl.gov%7C9e50b362e9644c8ec81308ddaa5e2bf7%7Cd6faa5f90ae240338c0130048a38deeb%7C0%7C0%7C638854044745025671%7CUnknown%7CTWFpbGZsb3d8eyJFbXB0eU1hcGkiOnRydWUsIlYiOiIwLjAuMDAwMCIsIlAiOiJXaW4zMiIsIkFOIjoiTWFpbCIsIldUIjoyfQ%3D%3D%7C60000%7C%7C%7C&sdata=9z7nQBEcJpcqyzonMkjLXxuJeTrJRkLYqx%2BJRIe2Mjo%3D&reserved=0
DISTAV - Università di Genova
Tel. ++39 010 3538308 Fax. ++39 010 352169
Corso Europa 26, 16132 - Genova - Italy
Marino V. Zuccolini
On 18 Jun 2025, at 17:05, Hammond, Glenn E <glenn....@pnnl.gov> wrote:
Marino,Once you have created all the combinations of porosity and permeability (assigning a unique integer ID to each combination), you can run the scenarios using two options:(1) run all scenarios with --multi_realization(2) run each scenario separately using --realization_id #I prefer the latter as there is more control if a given realization fails to run to completion.Examples of each located in the regression tests. The links below are to the command lines used to execute each--multi_realization--realization_id
These scenario reference the datasets inhttps://bitbucket.org/pflotran/pflotran/src/master/regression_tests/shortcourse/regional_doublet/perm_fields_10x5x4.h5
Prof. Marino Vetuschi Zuccolini
*******************************************************
Marino Vetuschi Zuccolini
marino.zuccolini@unige.it
Professor / Geochemist
Laboratory of Geochemistry [LGCH]
ORCID ID: 0000-0002-2874-631X
Associate to the CNR-IMATI Department