Abandoned X2t simulation - no convergence

[Isaac Mantelli]

I am building a simulation based on experimental work involving fluid diffusion through a fracture and into a rock matrix. This model assumes half of the fracture, where the well (injection) site is the fracture aperture. Fluid chemistry is from analyzed samples. The rock is an ultra-tight shale, hence the very large SSA.

The simulation fails after a very small reaction progress (Xi < 10^-6) with max residuals around 10^-11. I have removed all but essential ions and minerals. I would appreciate any insight!

fracture size sensitivity analysis.x2t

[Isaac Mantelli]

Rethought the design of this model. Still looking at diffusion through a fracture, but now I've edited the porosities of the middle row of nodes to be 1 to act as the fracture and removed the well sites for simplicity. I've also replaced kinetic minerals with simple minerals. However, now the model fails to converge at the same reaction progress, about 0.145. I have tried changing nearly each variable one at a time, but I still cannot find what is causing the model to stop at that point. 

fracture size sensitivity analysis simple minerals.x2t

[Jia Wang]

Hello Isaac,

Thank you for the additional files and I apologize for the delay. A 2D reactive transport modeling can be very challenging to configure correctly. I have looked through the two input files you attached and am not quite sure of the conceptual model here.

Are the minerals you are setting in the Reactants pane part of the original domain, in equilibrium with the initial fluid? In that case, you would want to swap them for components in the Basis pane and not add them in the Reactants pane. When you add minerals as simple reactants, they are added to the domain incrementally throughout the simulation to alter the composition of your system. In contrast, if you set a mass or volume % of kinetic minerals in the Reactant pane and set a rate constant, the program will treat any mass as part of the initial domain.

I think it would be better if you started with your solutions in a more simple model and then build your way up to a complex 2D reactive transport model. Maybe you would want to check that your kinetic reactions are working as expected in a one node system using React. Are the rates proceeding as expected? Should precipitation be turned off so that no new minerals are allowed to precipitate even if they are thermodynamically favorable, like how it is set now? I am not really sure what applies to your system here.

When the reactions in your single node system appear to be working as expected, I think the next step is to build a 1D model in X1t. Start with a very simple system, maybe just a 1 component system with silica and set up a diffusion system. You can take a look also at the quartz precipitation in a vein, section 3.9 in the Reactive Transport Modeling Guide. This example treats the whole domain as a single fracture. Your system is of course different in many ways, but it might serve as a starting point. You may also want to check the boundary conditions you are using to make sure they reflect the problem you're modeling. For more information on boundary conditions, please see section 2.9 in the GWB Reactive Transport Guide. Once you have your transport parameters configured, you can add in the chemical reactions from the 1 node system.

Finally, if your system requires a 2D model, then transition to X2t from what you have built in X1t.

As a side note, you can also designate many other variables with the node-by-node feature other than just porosity if necessary. For example, if you want your kinetic minerals everywhere in the domain to start but not in the fracture, you can set that using node-by-node. Other methods for setting heterogeneity are also available (e.g. equations, scripts, etc). For more information on this, please see the Appendix: Heterogeneity in the GWB RTM Guide.

Best regards,
Jia