Porepressure in PorousFlow Modulue

[Adam J.]

Hi all, 

First, I would like to thank Andy and the other developers for the great work that they did in the porous flow module and the well documentation of it.

I just have a few inquiries that require some clarification as I am new to this subject.

1) What does the negative pore pressure actually mean in the tests? For example, in the first test for single fluid, why there is negative pore pressure? In general, is it normal to have negative pore pressure in these kind of tests?

2) Can the porous flow module be used to model the injection and production process at the same time? Let's say we have an injection source at one boundary and a sink at another boundary. Do we assign a negative pore pressure at the sink boundary or does the code already takes care of that?

3) Let's say we have natural fractures in the system that are simply created by modifying the geometry and mesh properties, can we use some other physics to simulate the fluid flow inside the fracture rather than using Darcy's law? for example the NS equations. Just curious about this. 

Thanks again.

[Robert Podgorney]

Adam,

I'd be happy to help to the degree that I can while Andy is still asleep in Australia, but I'll need a few more details

1. Can you please give the exact name of the test you are referring to?  I'll check it to be sure but if it is an unsaturated flow problem then there may be some capillary pressures.  Else you might be looking at an extraction/production test with a negative flowrate?

2. This is easy and done regularly.  There are a plethora of ways to approach it, depending on your specific problem. You can use a combination of Neumann or Dirichlet BCs or choose between a number of source/sink formulations.  I'd recommend you look at dirac kernels in Porous Flow or in the MOOSE framework proper.

3. Short answer is yes, long answer is that from my experience it isn't easy to do a mix of NS and Darcy equations in different blocks of the same simulation.  I'll let others weigh in here, but fluxes/exchange between the two blocks/domains (fracture/matrix) is difficult.  Using stabilized Darcy equations is usually adequate.

What are you wanting to simulate?

Best,

Rob

[Adam J.]

Thank you, Rob. This is helpful. 

The examples that I was referring to are the first example in the tutorial of the porous flow module and example 8. You can see the negative porepressure in the figures. Link below: 

https://www.mooseframework.org/moose/modules/porous_flow/tutorial_01.html

https://www.mooseframework.org/moose/modules/porous_flow/tutorial_08.html

For now, I want to simulate a single phase fluid and look into the fluid flow behavior in complex fractured reservoirs where injection and production occur simultaneously. I want to study how the fractures' geometry and directions affect the overall flow behavior considering the effects of rock mechanics as well. Later on, I want to study the injection of CO2 in such systems. 

Based on your experience, could you please comment on how accurate and representative Darcy equation is if used to simulate the flow in fractures compared to NS equations? Is assigning high permeability values in the fracture's domain sufficient enough? 

Thanks again Rob. 

[Robert Podgorney]

I wouldn’t put too much thought into the absolute values from the example sims in the tutorial, they are for illustration of the setup process.  For a case specific simulation, you can set the values on the boundaries to your problem.

 

“Complex” fractured reservoir…….can be done but some tradeoffs will be necessary.  How many fractures?  What is their size distribution?  Meshing a complex DFN will be your biggest challenge.  While meshing the DFN itself is doable, meshing the rock matrix that the fractures are embedded in is all but impossible except for simple cases with few fractures.  The issue is mesh quality and holes in the mesh.

 

I’d recommend upscaling the fractures onto a structured mesh, this can still be a challenge, but it is manageable.  Here’s an example of a simulation of injection into an EGS reservoir.  Stream traces are colored by pressure, the slice is colored by temperature.  Injection is the origin of the traces, from a short segment of a well.  Outflow exists the reservoir from three main stimulated fractures.  All run using a Darcy formulation.

 

--
You received this message because you are subscribed to a topic in the Google Groups "moose-users" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/moose-users/hiu2wHrhUEs/unsubscribe.
To unsubscribe from this group and all its topics, send an email to moose-users...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/moose-users/be4d9764-92f1-49a2-9ddf-51812691eb44%40googlegroups.com.

[Wilkins, Andy (Mineral Resources, Pullenvale)]

Thanks for providing such comprehensive answers, Rob!

Regarding the negative pressure: in the case of page 2, it is due to no numerical stabilization.  There is a mention of this at the bottom of tutorial page 2 (https://mooseframework.inl.gov/modules/porous_flow/tutorial_02.html) and page 6, while >=page 8 will show you how to remove these unphysical negative pressures by using stabilization.

The negative pressure on page 8 is due to removing fluid "ad infinitum", and is actually perfectly physical as it simply means saturation<1  (page 8 is using saturated + unsaturated physics employing capillary pressure and relative permeability).  But from the sounds of your problem, it is unlikely you'll be working with saturation<1.

Having said all this, in many settings the negative pressure resulting from no numerical stabilization is rather irrelevant (in contrast to the negative pressures when working in the unsaturated zone) and often may be simply accepted and glossed over.

a

---

From: moose...@googlegroups.com <moose...@googlegroups.com> on behalf of Robert Podgorney <rpodg...@gmail.com>
Sent: Wednesday, 30 October 2019 7:29 AM
To: moose...@googlegroups.com <moose...@googlegroups.com>

You received this message because you are subscribed to the Google Groups "moose-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to moose-users...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/moose-users/BN6PR01MB3267587883D95E3771739006AC610%40BN6PR01MB3267.prod.exchangelabs.com.