Need help with a react model convergence- GWB 2021!

[Mojo]

Hi All, 

I am running a a React model for studying CO2-brine-rock reactions. I increased the CO2 pressure linearly to my desired number in 2 days and I used the model outcomes for running a new model under constant CO2 pressure from 2 days to 20 yrs. The 20 yrs model has not converged even after 3 days running! I have run parallel models for which I tuned the time step and even put Siderite and Anorthite (the most reactive minerals in the model) at equilibrium condition to help the model converge faster, but it was un-effective. Non of the 20 yrs models converge yet and they are still going on....

Any idea why is that? I am using the most recent version of GWB (2021). Not sure what causes this issue. I faced this issue once in the past when I had Calcite (highly reactive mineral) in my model. And I found that I need to set Calcite in equilibrium so the model that can converge in a few mins. But as I said this trick did not work this time. Attached are both 2 days and 20 yrs models. 

 

Also, I wonder why the calculated porosity change in 2 days model is wrong. When I calculate the porosity using the minerals volumes data, and a fix bulk volume (initial bulk volume value), i end up to a different number. It seems bulk volume is changing in the model in each time step. 

 

Thanks

Seyed

20 yrs injection.rea 2 days injection.rea

[Jia Wang]

Hello Seyed, 

I couldn't replicate your exact output since I don't have your modified thermo database, but I was able to run your input file without the Mg++ component and Montmor-K. Setting fast reacting minerals as equilibrium reactions is a good strategy in geochemical modeling because the reaction proceeds to equilibrium so quickly that a kinetic description is not necessary. I suggest checking your longer simulation by shortening it so that it runs to completion and you can plot dissolution rates under Reactant Properties using Gtplot to see which mineral changes the most abruptly. I shortened your 20 year simulation to end on day 5 and saw that your have a relatively high rate of precipitation for pyrite initially, then goes to 0 very quickly as pyrite reaches saturation early on with respect to the system. I suspect that this was causing your simulation to take very small time steps. I suggest setting pyrite as a simple reactant to titrate into your system to see if that affects the result of your simulation relative to the kinetic approach.

I would also recommend returning to your 2 day simulation and examine the rates for your mineral reactions as well to see if you can just use an equilibrium approach for any of these minerals. Please note that swapping a mineral into the basis pane sets the concentration of that component in equilibrium to that mineral, however, if you set the swapped mineral as a kinetic reactant, React would treat that mineral kinetically as your reaction path begins (i.e. sliding CO2 fugacity to 322.7). 

The porosity is calculated as the fluid volume divide by the bulk volume of the system. The bulk volume is composed of the mineral volume, fluid volume, and any inert volume that may be prescribed. So if your minerals dissolve, the total bulk volume will decrease and the porosity will increase. Running your model (without the input of Mg++ and Montmor-K), it seems like your total mineral volume decreases slightly and your fluid volume increased.

Hope this helps,
Jia Wang
Aqueous Solutions LLC