Convergence problem (small Xi)

[Neda]

Hello,

I have been trying to set up an equilibrium model in React to calculate initial chemical composition status of a system composing of brine and host rock minerals. I added water components in the Basis pane and minerals quantities in the Reactant pane. During the course of modelling I suppressed couple of minerals which I think should not present at initial equilibrium state. I'd like to use "Pick up" option later and add some gases to study evolution of the system in presence of them. When I suppressed the minerals I faced the convergence problem and I get the following error.

-- Can't converge, abandoning path.

-- Xi step is too small

If I don't use the suppress option there will be no problem and I can use " Pickup->system->entire". Could you please advise on how this suppress option may be causing this problem and how I can fix it without neglecting this option? I have attached the the React file for more details. Thank you.

Minerals-initial status-supress minerals.rea

Regards,

Neda

[Brian Farrell]

Hi Neda,

 

I noticed that the last few basis entries (Al+++, O2(aq), SiO2(aq)) were set to non-zero but small concentrations, presumably because they were not analyzed or detected in the fluid. As you know, it’s necessary to include these entries in the initial system if you’re going to introduce any minerals whose reactions include any of those species (Muscovite, Pyrite, etc.). One potential issue that I notice is the way you’ve constrained the oxidation state of the initial fluid with a small concentration for the O2(aq) component. Simply specifying a small value for the Al+++ and SiO2(aq) components is probably okay, but O2(aq) can sometimes be problematic. Do you have any idea about the oxidation state in the real system? Is it generally oxidizing or reducing? If you don’t know, you could make an assumption. For example, equilibrium with a mineral might control the oxidation state, or equilibrium with a gas of known fugacity, or you could even assume that the ratio of NO3- to NH4+ (since you have analyses for both) controls the system’s oxidation state. You should also keep in mind that it’s generally common to use a “free” constraint for O2(aq) instead of setting the concentration of the “bulk” component.

 

In playing with your script, I enabled the coupling reaction between NO3- and NH4+, thus deleting the NH4+ basis entry, then swapped out the O2(aq) for NH4+. By doing this, I assume that the ratio of NO3- to NH4+ sets the oxidation state of the system. This may or may not be appropriate, though. After doing this and decreasing the step size of your model (change delxi from 0.01 to 0.0001 in the Config – Stepping dialog) I’m able to complete the reaction path.

 

For more information on mass transfer, simple reactants, and reaction progress, take a look at Chapter 13, Mass transfer, in the Geochemical and Biogeochemical Reaction Modeling text by Craig Bethke. For controlling the step size (reaction progress) in the GWB programs, see 6.19, delxi, in the GWB Reference Manual. As for constraining oxidation state using ratios of oxidized to reduced species, there are various examples in the GBRM text. Section 23.3, Geothermal fields in Iceland, is one such example.

 

Hope this helps,

Brian Farrell

Aqueous Solutions

[Neda]

Hi Brian,

 

Thanks for the explanations, really appreciate it.

I don'e have any information about the oxidation state of the real system. I had to add O2(aq) since React needed it for smectite to get involved in the reactions so I just added small concentration of it. In fact I don't have enough knowledge of making a proper assumption and I am not willing to disturb the Basis which is made based on the water sample analysis I have. I added some CO2 to my system to account for degassing that most probably occurred during water sampling and to get a pH of 6.8 in the end which I am expecting for the studied reservoir and now reaction path can be completed. I am going to read the materials you mentioned to get better insight on what React is doing. Thanks again.

 

Best,

Neda