Fracture Modeling

[reisi...@gmail.com]

Dear Peter and Glenn,

I am trying to model a system consisting of a fracture located in the middle of a cement matrix (fracture flow with diffusion into the surrounding cement). I‌ want to track the fracture self-sealing or widening.

I would appreciate your advice on the most appropriate modeling approach for this configuration. In particular, I am wondering:

• Is there an example available for 2D fracture modeling?

• Is the multiple-continuum approach limited to 1D problems, or can it also be applied when the fracture is discretized in 2D?

Thank you very much for your guidance.

Best regards,
Fatemeh

[Hammond, Glenn E]

See my comments below in red.

Glenn

From: pflotra...@googlegroups.com <pflotra...@googlegroups.com> on behalf of reisi...@gmail.com <reisi...@gmail.com>
Date: Sunday, February 22, 2026 at 3:28 PM
To: pflotran-users <pflotra...@googlegroups.com>
Subject: [pflotran-users: 8675] Fracture Modeling

Check twice before you click! This email originated from outside PNNL.

Dear Peter and Glenn,

I am trying to model a system consisting of a fracture located in the middle of a cement matrix (fracture flow with diffusion into the surrounding cement). I‌ want to track the fracture self-sealing or widening.

I would appreciate your advice on the most appropriate modeling approach for this configuration. In particular, I am wondering:

• Is there an example available for 2D fracture modeling? I know of no simple examples for 2D fracture modeling. We have 3D fracture modeling using Voronoi meshes, but the capability is laborious and the level of complexity is beyond what we can support on a user mailing list.

• Is the multiple-continuum approach limited to 1D problems, or can it also be applied when the fracture is discretized in 2D? All PFLOTRAN simulations are inherently 3D problems (i.e., you have to describe the 3D problem in the GRID card). Multi-continuum can be considered a 4^th dimension off the 3D problem (i.e., in/out of rock matrix).

  https://documentation.pflotran.org/theory_guide/multiple_continuum.html
  With multi-continuum, a single multi-continuum grid cell contains pore space that is split between a fracture volume and a matrix volume (the latter being divided in 1D across cells in the secondary continuum [or the 4th dimension]). Your conceptual model has separate primary continuum grid cells that are assigned as fracture and matrix.

Thank you very much for your guidance.

Best regards,
Fatemeh

--
You received this message because you are subscribed to the Google Groups "pflotran-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pflotran-user...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/pflotran-users/f629d733-fefc-4817-9ab6-b0042e48123dn%40googlegroups.com.

[Fatemeh]

Dear Glenn,

Thank you very much for your comments.

Could you please advise on the most appropriate way to model the configuration shown in the attached image? My goal is to track changes in both the fracture and the surrounding matrix simultaneously. For this reason, I would need to discretize the matrix in both the x and y directions.

As I understand it, in the multi-continuum approach the matrix is represented as a 1D secondary continuum per primary cell, without lateral connectivity between matrix blocks (please correct me if I am mistaken). In that case, would you still recommend using the multi-continuum formulation?

Alternatively, would it be more appropriate to construct an explicit 2D model by assigning a thin set of cells as the fracture (with higher permeability and porosity), similar to an example I found in the PFLOTRAN user community discussions?

Thank you very much for your guidance.

Best regards,

Fatemeh

To view this discussion visit https://groups.google.com/d/msgid/pflotran-users/BY3PR09MB7555D4A98FD406B3AB7E33689A77A%40BY3PR09MB7555.namprd09.prod.outlook.com.

[Hammond, Glenn E]

See my comments in red below.

Glenn

From: pflotra...@googlegroups.com <pflotra...@googlegroups.com> on behalf of Fatemeh <reisi...@gmail.com>
Date: Monday, February 23, 2026 at 6:23 PM
To: pflotra...@googlegroups.com <pflotra...@googlegroups.com>
Subject: Re: [pflotran-users: 8682] Fracture Modeling

Dear Glenn,

Thank you very much for your comments.

Could you please advise on the most appropriate way to model the configuration shown in the attached image? My goal is to track changes in both the fracture and the surrounding matrix simultaneously. For this reason, I would need to discretize the matrix in both the x and y directions.

As I understand it, in the multi-continuum approach the matrix is represented as a 1D secondary continuum per primary cell, without lateral connectivity between matrix blocks (please correct me if I am mistaken). In that case, would you still recommend using the multi-continuum formulation? Your description of the secondary continuum is correct. It has a disconnected matrix. If you desire matrix-matrix connectivity (i.e., >1D in the matrix), you must mesh up the matrix explicitly.

Alternatively, would it be more appropriate to construct an explicit 2D model by assigning a thin set of cells as the fracture (with higher permeability and porosity), similar to an example I found in the PFLOTRAN user community discussions? I recommend narrowing the conceptual model below to a single column of grid cells for the fracture. I don’t know that using more than one column of grid cells will improve the accuracy that much as you are using Darcy flow for the fracture. If you are looking to plug the fracture through mineral precipitation, the multi-column conceptual model will likely not form mineral at the interface between the fracture and matrix any more than at the center of the fracture. So, just use one column. Once you develop some intuition with a single column, you can increase the complexity of the conceptual model. It is possible that refinement in the fracture will improve accuracy. But I doubt it will improve your ability to capture reality.

To view this discussion visit https://groups.google.com/d/msgid/pflotran-users/CACLNKk4sE%3D9FY_TcZ%3DqyGOMd255i%2B58%2BUmXdC3PZEAAGssz0sw%40mail.gmail.com.

[Fatemeh]

Dear Glenn,

Thank you very much for your prompt response. 

I have a follow-up question regarding fracture sealing.

In the PFLOTRAN, is there a way to capture preferential mineral precipitation near the fracture–matrix interface (i.e., precipitation depth across the fracture walls)?

I am trying to understand this process can be represented meaningfully in PFLOTRAN, or whether this would require a different modeling approach.

Thank you very much for your guidance.

Best regards,
Fatemeh

To view this discussion visit https://groups.google.com/d/msgid/pflotran-users/BY3PR09MB7555C9146B2DFC23721AD3EC9A74A%40BY3PR09MB7555.namprd09.prod.outlook.com.

[Hammond, Glenn E]

Mineral precipitation can only occur within a cell. Therefore, cells at the fracture-matrix interfaces (assuming you have assigned fracture/matrix to primary continuum grid cells) will experience preferential mineral precipitation if the surface area and affinity factor (a function of species concentration) favor precipitation in those regions. For example, in a calcite precipitation scenario, you could load the matrix with calcium and the fracture with high pH carbonate, resulting in calcite precipitation at the interface. However, this approach involves manipulating the system.

Glenn

To view this discussion visit https://groups.google.com/d/msgid/pflotran-users/CACLNKk7AoqxvMgQr7wrJtf9PM8AurQt5Tb_ZkkC5TmZ41vF0_A%40mail.gmail.com.

[Fatemeh]

Thank you very much Glenn.

To view this discussion visit https://groups.google.com/d/msgid/pflotran-users/BY3PR09MB7555797C5AF6A868318AACE89A75A%40BY3PR09MB7555.namprd09.prod.outlook.com.

[Fatemeh]

Dear Glenn,

I hope you are doing well.

I have developed a model for this case; however, I am currently facing convergence issues. I suspect the problem may be related to how I have defined the flow conditions, and I would greatly appreciate your guidance.

Specifically, I would like to confirm the appropriate approach for representing the injection of "brine saturated with CO₂". In this case, would it be more appropriate to use the General mode, or should I instead use the multiphase mode? Then how should I set up the boundaries? 

Thank you very much for your time and support.

Best regards,
Fatemeh

[Hammond, Glenn E]

Fatemeh,

If your simulation involves free-phase CO₂ within the brine, I recommend using SCO₂ mode. However, if you are working with single-phase brine saturated with CO₂ solely for reactive transport, RICHARDS is the more suitable choice. GENERAL mode does not offer any advantages over RICHARDS for simulating reactive transport of CO₂.

Once the above is clarified, one of us can provide further guidance.

Glenn

From: pflotra...@googlegroups.com <pflotra...@googlegroups.com> on behalf of Fatemeh <reisi...@gmail.com>
Date: Saturday, April 4, 2026 at 8:45 PM
To: pflotra...@googlegroups.com <pflotra...@googlegroups.com>
Subject: Re: [pflotran-users: 8745] Fracture Modeling

Check twice before you click! This email originated from outside PNNL.

--
You received this message because you are subscribed to the Google Groups "pflotran-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to pflotran-user...@googlegroups.com.

To view this discussion visit https://groups.google.com/d/msgid/pflotran-users/CACLNKk4o0DUeJm0ty6ZbOYn2h9mwah0vqsG%3DkD4wLKoe6ooqaw%40mail.gmail.com.

[Fatemeh]

Hi Glenn,

Thank you for your email. I was able to get the simulation running successfully, and I now have some results I’d really appreciate your input on.

As it is shown in the attached image, at first, the Calcite precipitation pattern looks reasonable, calcite starts forming at the fracture walls, which makes sense since those cells are in contact with the cement matrix and have the right conditions for nucleation.

But Calcite also starts showing up in the middle of the fracture (where porosity is 1 and there’s no solid surface), and instead of being carried away by the upward CO₂-brine flow (0.25E-5 m/s), it just stays there and keeps growing over time. Could you please guide on how can I‌ address this? 

Best Regards,

Fatemeh

To view this discussion visit https://groups.google.com/d/msgid/pflotran-users/PH0PR09MB756478052E7B6EA45E6F6A7E9A262%40PH0PR09MB7564.namprd09.prod.outlook.com.

[Hammond, Glenn E]

Fatemeh,

While I may not recall every detail of our previous discussion, your mention of nucleation suggests you are using a nucleation model. It's important to note that nucleation can take place throughout the domain, even in regions without mineral surfaces—such as the center where the porosity is 1. For example, nucleation can occur in a beaker of supersaturated water with no available mineral surface area. Therefore, observing nucleation at the center of your domain is consistent with model expectations.

Glenn

To view this discussion visit https://groups.google.com/d/msgid/pflotran-users/CACLNKk6UoAkAWZhCGHmvfmEgO3sH-HMJ7SGRBPxqXku7uPRSpA%40mail.gmail.com.

[Fatemeh]

Dear Glenn,

Thank you for your email.

I have two questions:

1. Since there is continuous injection into the fracture, wouldn’t this flow tend to wash out  calcite nuclei forming at the center of the fracture?

2. How should I set up my model so that calcite precipitates on the fracture walls and progresses toward the center, rather than initiating precipitation in the center?

Thank you for your guidance.

Best regards,

Fatemeh

To view this discussion visit https://groups.google.com/d/msgid/pflotran-users/BY3PR09MB7555B4D99228B9365C8415799A222%40BY3PR09MB7555.namprd09.prod.outlook.com.

[Fatemeh]

Dear Glenn,

There is also one more question I forgot to include in previous email. When the porosity of cells near the fracture wall decreases and approaches zero, simulation stops. Even though I have set a lower limit of 0.05 using the UPDATE_POROSITY keyword, the simulation still fails once this limit is reached. As a result, precipitation does not continue into the adjacent cells.

Could you please advise on how to handle this situation?

Best regards,

Fatemeh

[Fatemeh]

Dear Glenn,

I am sending this as a gentle reminder.

Regards

Fatemeh

[Hammond, Glenn E]

See responses in red below.

Glenn

From: pflotra...@googlegroups.com <pflotra...@googlegroups.com> on behalf of Fatemeh <reisi...@gmail.com>
Date: Wednesday, April 15, 2026 at 3:25 AM
To: pflotra...@googlegroups.com <pflotra...@googlegroups.com>
Subject: Re: [pflotran-users: 8770] Fracture Modeling

Dear Glenn,

Thank you for your email.

I have two questions:

1. Since there is continuous injection into the fracture, wouldn’t this flow tend to wash out  calcite nuclei forming at the center of the fracture?

1. Mineral species cannot be transported. Therefore, nuclei are considered to be immobile.

1. How should I set up my model so that calcite precipitates on the fracture walls and progresses toward the center, rather than initiating precipitation in the center?

1. I don't think this approach is feasible. If I were to tackle the problem directly, I would design a new process model that updates a flag at each timestep, depending on whether any neighboring cell has exceeded a specific threshold for calcite volume fraction. Essentially, a cell would only be allowed to precipitate minerals once a neighboring cell has met this requirement. Afterward, I would implement a new reaction sandbox that relies on this flag to control mineral precipitation. However, it’s important to note that implementing this strategy would demand extensive code modifications, particularly for managing the flag updates.

To view this discussion visit https://groups.google.com/d/msgid/pflotran-users/CACLNKk7U%2BOz4_bHVMEM6EUtYB3KRsFmiY06wM4KNLRrdmQ-6Hg%40mail.gmail.com.

[Hammond, Glenn E]

This appears to be a bug. Please send me an input deck that reproduces the issue and executes quickly. I should mention that this capability has not been used extensively.

Glenn

To view this discussion visit https://groups.google.com/d/msgid/pflotran-users/CACLNKk42nN1qoPZtmaFxwHXh-qrY2BgGj4PUjijRMVbDL0BYKg%40mail.gmail.com.

[Fatemeh]

Hi Glenn,

Thank you very much. I applied changes and managed to resolve the issue, and minimum_porosity is now working perfectly as well.

Regards,
Fatemeh

To view this discussion visit https://groups.google.com/d/msgid/pflotran-users/CO6PR09MB7558ABE5836B8CF763C7C0CF9A2B2%40CO6PR09MB7558.namprd09.prod.outlook.com.