Extraction well sanity check

[Kenton Wu]

Dear Glenn et. al.,  
I am developing a simple model for salinity contamination in an extraction well in a coastal region and would like a sanity check for our methodology of simulating freshwater extraction.

 
For context, the solver is running in RICHARDS mode, while the domain consists of a 2D transect with a dirichlet pressure condition on the left boundary, a 'far field' diriclet pressure condition on the right, and a 'surf' condition on part of the top boundary representing the shelf that sits below the water surface. An example frame of the salinity is shown below with the pressure condition highlighted in green and the neumann condition in light blue.  

We are relatively confident that the subsurface model is correct, and now we would like to add extraction well representation via a point sink with volumetric extraction rate. The extraction rate will be set to be a percentage of the water content of the whole 2D slice. I was wondering if this is a reasonable approach, or if you see something that could be improved.

 

Here is the part of the input file defining the well: 

REGION testwell
        COORDINATE 0.000000 0.d0 0.000000
/

FLOW_CONDITION Injection_test
        TYPE
                RATE VOLUMETRIC_RATE
        /
        RATE -10.000000 m^3/day
/

TRANSPORT_CONDITION sink
        TYPE EQUILIBRIUM
                CONSTRAINT sink
                        CONCENTRATIONS
                                1e-10 T
                /
        /
END

SOURCE_SINK test
        FLOW_CONDITION Injection_test
        TRANSPORT_CONDITION sink
        REGION testwell
/                                

I am also wondering if the concentration is relevant (since it's extracting) and if I could be enlightened on what the equilibrium transport condition does. 

Thanks!

Kenton 

[Kenton Wu]

Hi Glen et. al, 

After some thinking I actually changed the transport condition to dirichlet_zero_gradient. Any thoughts? 

Best, 

Kenton 

[Hammond, Glenn E]

Kenton,

Please send a simple schematic of the conceptual model. Do you want saline water to enter only from the top? If so, why use Neumann instead of Dirichlet?

For the outlet well, the concentrations listed in the constraint specified in the transport condition are irrelevant because the concentration in the cell (through which the well intersects) is used. But I would not assign EQUILIBRIUM to anything. That is used for scenarios where one wants to hold the concentrations in a grid cell equal to that of the constraint, not what you want in this case.

As for how much water to extract, I would need to better understand the conceptual model or what you are attempting to achieve. The value for a pumping well can be arbitrary, though the higher the rate, the greater the intrusion of saline water (obviously).

Glenn

From: pflotra...@googlegroups.com <pflotra...@googlegroups.com> on behalf of Kenton Wu <wuke...@gmail.com>
Date: Tuesday, September 16, 2025 at 4:33 PM
To: pflotran-users <pflotra...@googlegroups.com>
Subject: [pflotran-users: 8557] Re: Extraction well sanity check

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/0002c989-ea60-4f59-978c-9386c5a909f1n%40googlegroups.com.

[Kenton Wu]

HI Glenn, 

Here is a mockup of what we're trying to model. The freshwater recharge is modeled using a neumann condition, while the saline and freshwater flux are modeled using a hydrostatic pressure dirichlet condition. Three extraction wells are modeled using a point sink with the dirichlet_zero_gradient condition. The computational domain is 1100x20x1 m. Currently I'm unsure how to set a realistic range for the extraction rate of the wells. Just inputting estimates of water usage rates in the location we are modeling (>10,000 cubic meters per day), it seems to be too intense and the solver has trouble converging. I figured this is because we are modeling this as a 2D transect, and there is some conversion that needs to happen, but I'm not sure how. So, our second idea was to just read off the total water content off the mass balance file and then set the rate to be a percentage of that. 

 

Thanks for the input, 

Kenton 

[Kenton Wu]

Here's the image as an attachment since it seems like the inline photo has trouble rendering.