The reactive transport benchmarking including precipitation/dissolution equilibrium reactions in a saturated calcite column using X1t and TEdit 12.0 cannot be read in GWB 2021

[Polly Tsai]

Hi!

 I am carrying out a reactive transport benchmark including precipitation and dissolution reactions along a saturated column of calcite (namely, a typical flow-through column) which is fluxed with a diluted solution of MgCl2 using X1t simulator. The model setup and parameters were described in the ppt file and the script was also attached in the X1t input file. However, the modeling results cannot be converged. How can I get the same results simulated by PHREEQC show in the ppt file.

 In addition, I have edited and added some radionuclides in the thermo dataset last year using GWB 12.0. However, it cannot be read in GWB 2021. I don’t know why. Enclosed please find my t.dat extension file of thermo_phreeqc_Polly .

 Your further response will be highly appreciated.

 

Cheers,

 Polly Tsai

Input parameters, IC, BC and modeling results_20210711.pptx SA case_20210711.x1t thermo_phreeqc_Polly.tdat

[Polly Tsai]

Hi, GWB technical consultant,

After discussion with my colleague, some fluid and interval parameters were revised and the modeling results could not be converged. Please refer to the updated X1t input file.

 

Cheers,

 

Polly

SA case_20210714.x1t

[Jia Wang]

Hello Polly,

Thank you for the new input file. Here are a few suggestions to help you get started. 

I think the main issue here is that adding calcite as a simple mineral doesn't really represent your system. Simple reactants are typically used in titration pathways. If you are designating a portion of your initial domain with calcite, you would need to swap it into the basis instead of reacting a quantity in as a simple mineral. You can see an example of this in section 3.6 Rainwater infiltering a quartz aquifer of the Reactive Transport Modeling User Guide. 

Also note that the component chosen for charge balancing ion is generally the species with the most abundant concentration for which the greatest analytic uncertainty exists. Cl- is generally chosen because it typically exists in fairly high concentration in natural waters and unless specified, commercial labs report a Cl- concentration through charge balancing. In your case, your fluid in the basis pane has a really low Cl- concentration, so it might make more sense to pick one of the components with a much higher concentration as your charge balancing species. 

You should also check that the log Ks in your database match the one that are used in thermo.tdat. 

With regards to the dataset, there are several entries that are missing some important information, resulting in TEdit in GWB 2021 to crash when opening the file. Cs2U2O7 is missing Log K. H4SiO4(aq) does not have any species in reaction. I also noted that you have several silica basis species which is very unusual. Typically, a database is put together such that there is only one basis species for that element and other aqueous species of the same redox state should be written with a reaction using the basis and/or other species. To fix your dataset so that Tedit can open it, I suggest locating the entries ( Cs2U2O7, H4SiO4(aq) ) mentioned above and correct them in a text editor (e.g. Notepad).

Hope this helps,
Jia Wang
Aqueous Solutions LLC

[Polly Tsai]

Dear Jia Wang,

 Thanks so much for your response. According to your suggestion, it could work. However, some questions still raised.

 1. The results were shown as pages 4-6 of ppt file. But , as you can see, some differences exist, when compared to page 2, modeling results by PHREEQC. I think the log Ks in thermos database do not completely match PSI/Nagra database that are used in the PHREEQC modeling.

2. In the Xtplot, the captions of x and y axis could not be edited and primary coordinates sub-coordinates could not also be merged into a figure. Only I could do is to export the raw data to Excel and replot, shown as page 7 of ppt file.

3. The default TDB seems to be thermo.tdat. But, when I changed the other TDB, like thermo_phreeqc, the results could not be converged.

 

As to the TEdit 12.0 not read in GWB 2021, I have corrected it by adding Log K of Cs2U2O7 and adding species of H4SiO4(aq) in reaction. It can finally be read in GWB 2021.

 Enclosed please find the updated input and ppt files.

 

 

Warm regards,

 Polly

SA case_1100729.x1t Input parameters, IC, BC and modeling results_20210729.pptx

[Brian Farrell]

Hi Polly,

If I follow the benchmark while using thermo_phreeqc.tdat, I get almost the same results as PHREEQC-OGS. The only difference is a slightly wider zone of dolomite stability, and consistent with that, a few shifted characteristic points in the Mg++, Ca++, C, and pH curves. If you can find the PSI-NAGRA database used in the benchmark, your results might be even closer. You might only need to alter the log Ks of calcite and dolomite, or you could convert the entire PSI-NAGRA dataset from phreeqc to GWB format in TEdit if desired. 

Here are a few comments on your current input file. Calcite should be swapped into the basis on the Initial pane, as you’ve done, since it is in equilibrium with the pore water at the start of the simulation, but its abundance should be set to 2e-4 free mol/l to match the original simulation. Calcite should not additionally be set as a simple reactant. I don’t believe a non-default value for the diffusion coefficient is warranted, although its effect is negligible in this simulation. There is no reason to set an inert volume here. There is no reason to change the convergence criterion “epsilon” variable, as far as I can tell. You can set Nx to 100 to match the original simulation and to speed up program execution. I took care of these issues and didn’t have any convergence issues using thermo_phreeqc.tdat, but if you continue to have difficulty you can attach a revised version.

If you want to overlay different results (mineral abundances, fluid composition, pH) in a single plot, you’ll have to do so in Excel, PowerPoint, Adobe Illustrator, or some other graphics program, as you’ve done.

Finally, regarding your previously posted custom thermo dataset, you had 3 basis species SiO2(aq), Si(OH)4, and H4SiO4, along with an aqueous species H4SiO4(aq). Those species all represent the same thing, so there should only be one of them. I’d be very careful to clean up those entries and any dependent reactions so that you have a complete, consistent, and non-redundant set of reactions.

Hope this helps,

Brian Farrell
Aqueous Solutions