X1t Not Converging

[shensel]

Hello,

I'm trying to run a flow path model using X1t and can't seem to get it to converge. I've attached my input files here. Any thoughts on how to fix this? I did get my Initial fluid to converge in React by reducing the alkalinity, but that doesn't seem to work for my Fluid 1 or for the Initial fluid in X1t. Any suggestions would be a great help!

 

Thanks,

Scott

X1t for Quesiton.x1t

[Jia Wang]

Hello Scott,

Thanks for attaching your input file. I wasn't able to run your file exactly because I didn't have your custom dataset. I switched to thermo.V8.R6+.tdat, which contains Molybate.

When I run "go initial" with your input file, X1t tells me that "fluid_1" is the issue. Looking at this fluid more closely, I noticed that you have more Na+ than that of Cl- in your simulation. I switched the balancing ion to SO4-- and then did another "go initial" run and X1t converged. Typically, you would choose the most abundant anion or cation as the balancing species. By default, the program uses Cl-. You can change the balancing ion to another species by either clicking on the unit for the species in the GUI or use the command line "balance on" the species that you wish (e.g. "balance on SO4--"). I would suggest trying to troubleshoot by switching your balancing ion and see if that helps.

If you are troubleshooting an issue that arises in the Fluids pane, you might find it easier and more direct to use React. If you have only one fluid, you can conveniently right-click on the Fluids tab and drag and drop into the Basis tab of a new React window.

If this didn't help resolve your problem, please attach your custom thermo dataset so I can run your script directly.

Hope this helps,
Jia Wang
Aqueous Solutions LLC

[shensel]

Hi Jia,

Thanks you for the reply and the help. I'm sorry I forgot to attach the thermo database, I've attached it here - it is a converted .geo EQ3/6 database that I believe is fairly similar to the thermo.V8.R6+ database.

I've experimented with balancing on different ions and still can't seem to get my Fluid 1 to converge. Perhaps you'll have better luck? I did make a couple other changes to the fluids based on my data, so I've re-attached an updated X1t file here.

 

Thanks,

Scott

X1t for Quesiton2.x1t data0.geo.tdat

[Jia Wang]

Hello Scott,

Thanks for attaching your custom thermo file. I opened your input file again and noticed that most of your components are in moles instead of relative units (e.g. mol/kg). Were the constraints created by picking up a fluid at the end of a React simulation? When you have relative units like mol/kg and you scale your system's volume by changing the node volume or change it's porosity, the solute concentration is scaled so that your mole of solute relative to kg of water is fixed. When you specify a mass unit like mg or mol, then scaling the volume of your bulk system in the Reactive Transport apps won't scale your solute concentration. I would suggest double checking that your units are what you intend them to be.

A side note, if you are using results from a React simulation to set your initial fluid composition, you can right-click and drag the Results pane (from React) and drop it in GSS to convert from mass units like mg and mol to relative units like mol/kg and mg/kg. To transfer from GSS to the Basis or Fluids tab, right-click on the sample header in GSS and drag and drop into the appropriate pane in X1t.

Hope this helps,
Jia

[shensel]

Hi Jia,

Thank you for the suggestions. I have now converted both fluids to a relative unit (mol/kg) and have attached an updated X1t file here. I still could not converge a reaction between my two fluids, so I tried them individually in React. I did get my Fluid 1 to converge when balancing on sulfate (among others), but could not get my Initial Fluid to converge. I'm not sure why it isn't working, maybe I'm not representing the system properly somehow? I've attached the React file with the fluid I can't converge.

 

Thanks,

Scott

Initial Fluid_molkg.rea X1t for Quesiton3.x1t

[Jia Wang]

Hello Scott,

Thanks for attaching the additional files. I am glad to hear that you managed to get one fluid to converge. Going back to your previous attachment (X1t for Quesiton2.x1t), changing the absolute units to relative units seems to converge just fine when I also reset the epsilon value to the default value. The epsilon value is used as a convergence criterion for the Newton-Raphson used to solve a system of equations. Setting the epsilon to a value smaller than the default would produce a slightly more accurate result at the expense of smaller step sizes. In this case here, I think the small epsilon value is affecting the program's ability to converge since it's extremely small. I would not recommend changing it unless you have specific reason to do so.

I also noticed that the fluid composition in the new attachments are not the same fluid from before. Was that intentional? In your Initial Fluid_molkg.rea file, I noticed that you had included a small amount of total iron concentration and you have set an alkalinity value instead of a total concentration for your bicarbonate component. I suspect that the oxidation state set by your aqueous oxygen concentration probably results in a very small amount of Fe++ in solution. To help the program's initial calculation of the equilibrium state, it is sometimes helpful to swap in a higher abundance species in solution. In this case, I tried swapping in Fe+++ and React had no issues with convergence. The program is still calculating the mass distribution between the different redox state of iron.

For your input file "X1t for Quesiton3.x1t", if I reset the epsilon value to default, swap Fe+++ for Fe++ and set SO4--- as the balance ions in both the Basis and Reactants pane, your input file runs without issues.

As a side note, you can disable redox equilibrium between oxidation states if you suspect that redox species are not in equilibrium. To do so, you can go to Config -> Redox couples... and then decouple Fe++/Fe+++. After doing so, you can directly add Fe+++ as a separate entry in you Basis, allow you to constrain the total concentration of Fe+++ species. For more information on redox disequilibrium, please see section 7.3 of the GWB Essentials User Guide.

Hope this helps,
Jia

[shensel]

Thanks so much! This really helped. I believe it was my misuse of iron that had me stuck.

[Jia Wang]

Hello Scott,

Glad to hear that was helpful. Hope the rest of your project went smoothly.

Best,
Jia