primary species that do not transport

[Dipankar Dwivedi]

Hello,

 How do I specify primary species in PFLOTRAN input deck that do not transport (e.g., lignin)? 

Thanks,

Dipankar

[Peter Lichtner]

Primary species must be aqueous species. Non-transported species are solids and sorbed species, for example. You can add a new solid to the database and use it under MINERALS. You also need to specify a kinetic rate constant, volume fraction and surface area. Hope this helps.

…Peter

[Dipankar Dwivedi]

Thanks Peter for the response. I want to use biomass as non-transported species (it will have a decay rate and Monod Kinetics). If I use biomass under MINERALS card that means : I need to define biomass in database as minerals and its decay rate same as dissolution of minerals. Is that right?

Thanks,

Dipankar

[Hammond, Glenn E]

Dipankar,

 

Use the IMMOBILE_SPECIES card as shown in the attached input file (that may not run).  Note that you must have at least one mobile (transported) species, and that species can be a dummy species that has nothing to do with the problem.

 

Glenn

[Peter Lichtner]

Here is a updated file that runs.

…Peter

[Dipankar Dwivedi]

Hi Glenn,

I was able to modify my input file and run the reaction network. I also noticed if we have  some species under IMMOBILE_SPECIES  card, then we need to remove them from the thermodynamic database. Otherwise, the model does not run.

Secondly, I should not have biomass in stoichiometric reactions; instead, I need to put them as BIOMASS with Rate_reaction as positive for growth and negative for decay. 

 I also found that my model takes very small time steps when I defined BIOMASS as compared to when I had all aqueous species. I am not why this should happen.

Thanks,

Dipankar

[Hammond, Glenn E]

The small time steps are probably due to BIOMASS reaction not being fully implemented/debugged.  The IMMOBILE_DECAY_REACTION should only be used for decay, not growth.  All growth should come through a yield prescribed in the MICROBIAL_REACTION, but I have not completed implemented of all this functionality.

 

Coincidently, I plan to work on this later today in support of the PNNL SFA work.

[Peter Alt-Epping]

Hi all,
I am running a simulation in which I inject water into a container at a
constant rate. For a short period of time I add a tracer to the water.
When I look at the mass balance in pflotran-mas.tec, the amount of
tracer injected at the inlet ("inlet Tracer") shows the correct amount
of tracer and the correct timing of the tracer pulse (the injection of
the tracer stops after 4.76e-5 y). However, the global tracer
concentration continues to increase even though the injection of the
tracer has stopped. I am not sure about the reason for this - there is
no other source or any other mass influx through the boundaries ...

Thanks,
Peter

--
_____________________
Dr. Peter Alt-Epping
University of Bern
Institute of Geological Sciences
Baltzerstrasse 3
3012 Bern
Switzerland
____________
T: +41 31 631 4531
F: +41 31 631 4843

[Peter Alt-Epping]

Hi again,

I am trying to convert a NASTRAN grid to a PFLOTRAN grid via
Cubit/Trelis and exodus2pflotran.py (I modified the script slightly by
including TETRA4 element types) . The conversion seems to work ok (i.e.
no error message or crashing up to this point) but when I read the grid
into PFLOTRAN I get the following message:

ERROR: face not found in connection loop 0 57 6934

There is obviously a face missing but I was wondering if someone could
give me a hint where in the conversion process the error may have occurred.

[Gautam Bisht]

Peter,

If you send me the original grid data and your generated PFLOTRAN hdf5
file, I can take a look at what is going wrong. Unfortunately, our
error messages regarding why an unstructured grid fails are not very
helpful to chase down the bug, unless one digs deep inside the code.

-Gautam.

[Hammond, Glenn E]

Peter,

Your simulation is running correctly according to what you have prescribed in the input deck. You have prescribed that at 4.7564688d-5 y, the constraint change from tracer_injection to no_tracer, while the water flux remains the same. The assigned tracer concentrations are:

tracer_injection: Tracer 3.0
no_tracer: Tracer 1.e-10

Therefore, after 4.7564688d-5 years, the rate of tracer injection will be ~1.e-10 mol/L * 0.010512 m^3/y * 1000 L /m^3 = 1.0512e-9, which is consistent with what I see in the mass balance file.

At 2.739726e-3 years, the water mass influx goes to zero and the tracer mass flux goes to zero along with it.

Since you cannot assign a zero concentration in a constraint, I recommend that you use a concentration of 1.e-40 in 'no_tracer', which is essentially zero. That should solve the problem.

Glenn

[Peter Alt-Epping]

Hi Glenn,

I agree, the injection rate and the tracer concentration at the inlet
are correct. However, the global tracer concentration seems too high
after 4.7564688e-5 years. Shouldn't the global tracer concentration in
mol be 'inlet Tracer (mol)' - abs('right Tracer(mol)') ?

Cheers,
Peter

[Hammond, Glenn E]

The global mass balance has nothing to do with the boundary fluxes and source/sinks; it is solely a summation of the water/solute mass in all grid cells.

Glenn

> -----Original Message-----
> From: pflotra...@googlegroups.com [mailto:pflotran-
> us...@googlegroups.com] On Behalf Of Peter Alt-Epping

> Sent: Tuesday, April 14, 2015 8:55 AM
> To: pflotra...@googlegroups.com

[Peter Alt-Epping]

Hi Glenn,
What I meant was that after 4.7564688d-5 years the global Tracer
concentration increases at a higher rate than the 1.0512e-9 mol/yr.

Cheers,
Peter

[Peter Lichtner]

Here is what I see using 1.e-60 for the tracer.
…Peter

[Peter Lichtner]

Peter: here is the latest with tighter tolerances for transport: ITOL 1.e-12.
…Peter

[Peter Lichtner]

Also turning off the log formulation helps.

…Peter

Begin forwarded message:

[Hammond, Glenn E]

Peter,

Peter and I have been iterating on this. The issue is with your solver tolerances. They are too loose. Remove the ITOL from transport. Also, if you use a non-log formulation for transport, the solution gets even better, but that may not be possible with full geochemistry.

Regardless, your issue revealed a glitch in my update to time steps for transport when run with flow. The glitch does not affect the solution, but introduced unnecessary sub-steps.

Thanks,

[Peter Alt-Epping]

Great! Thanks Glenn and Peter.

Cheers,
Peter

[Gautam Bisht]

Peter,

There appears to be an issue with your grid. The cell-id 19510 has 5 duals (see below); but since all control volumes are tetrahedron, each cell can only have 4 duals.

Cell ID         #vertices        VertID-1    VertID-2  VertID-3  VertID-4

       19510           4        4573        4574        4575        4576

       19511           4        4576        4573        4577        4575

       19514           4        4576        4578        4574        4575

       19516           4        4575        4574        4573        4581

      125588           4       24410        4574        4573        4576

      227352           4       40152        4574        4573        4576

(It maybe the case that vertex 24410 and 40152 are same).

On a different note, at a cursory glance it appears to me the control volumes in your unstructured grid are non-orthogonal (i.e. line joining the centroid of two neighboring control volumes is not perpendicular to the plane of shared face). If you have non-orthogonal control volumes, you will have numerical error because the discretization in PFLOTRAN is only first order accurate.

-Gautam.

On Tue, Apr 14, 2015 at 9:06 AM, Peter Alt-Epping <alt-e...@geo.unibe.ch> wrote:

Hi Gautam,

I have attached the original Nastran file which contains the grid, the converted exodus grid (mesh1.e) and the converted pflotran grid (mesh1.h5).  I have also attached the conversion script ex2pfl.py with the TETRA4 elements types and the pflotran input file.

Thanks very much for your help!
Peter

[Peter Alt-Epping]

Hi Gautam,

Thanks for taking a look at the files! I am fairly new to using unstructured grids and I was wondering how can you see that cell-id 19510 has 5 rather than 4 duals? 

Thanks,
Peter

[Gautam Bisht]

Hi Peter,

On Wed, Apr 15, 2015 at 4:48 AM, Peter Alt-Epping <alt-e...@geo.unibe.ch> wrote:

Hi Gautam,

Thanks for taking a look at the files! I am fairly new to using unstructured grids and I was wondering how can you see that cell-id 19510 has 5 rather than 4 duals? 

Since a pair of 3 vertices (out of 4) that form cell-id 19510 are present in 5 other cells.

-Gautam