Source Term magnitude problem (Aquifer Thermal Energy Storage)

Usability
1

Hi,

I am trying to simulate Aquifer Thermal Energy Storage, by using th-decovalex Benchmark (HT process) and introducing a source term under “p process variable”. my model is axially_symmetric, so my source term is Line type. the thickness of aquifer is 50m, the injection rate is 35kg/s. As it is suggested in the scaling source term section, I should divide the magnitude of source term by the injection area which in my case is a borehole. the diameter can be around 0.2m. therefore the final magnitude of source term would be around 1.1kg/s/m^2. but the result with this source term does not make sense and the whole 500m (radius) aquifer become full of water after couple of days. the magnitude that make sense could be around 10e-5 to 10e-6. (I used SI units in the whole project file).
I would appreciate it if someone can help me.

2

Hi,

you are using axial symmetry, i.e. you have a 2D simulation domain in the x-y plane. The symmetry axis is the y axis. I.e., the x coordinate is the radius. Just to repeat OGS’s convention.

When you apply a line source, OGS will integrate the source term along that line.
So the correct scaling should be injection_rate / length = 35 kg / s / 50 m = 0.7 kg / m / s.
And when your source is at the symmetry axis, this should be the only scaling to apply.

You might use this example as a reference:

grafik
grafik1061×802 51.9 KB

note that the mesh at the symmetry axis is not axially_symmetric="true":

        <mesh axially_symmetric="true">square_1x1_quad_100x100.vtu</mesh>
        <mesh>line_inner.vtu</mesh>
        <mesh axially_symmetric="true">line_outer.vtu</mesh>

otherwise OGS would multiply with a factor of radius r, which is zero at the symmetry axis.

I hope that helps.

Best regards,
Christoph

1 Like
3

Dear Christoph,

Thank you for the information.
If I insert the source = 0.7 kg/m/s (with all boundaries axially symmetric except the source line) the injection pressure is too high (the aquifer is 270 bar and the injection pressure becomes 280000 bar, about 1000 times higher than expected).
If I insert the source in volumetric unit = 0.0007 m^3/m/s it perfectly works and matches the real data.
I highly doubt that in HT process also we should insert volumetric.
Thank you very much.

Best
Mohammad

4

Hi Mohammad,

I just had a look at the source code of the HT process.
The pp block of the assembled stiffness matrix is
\mathbf K_{pp} = \int_\Omega \nabla N^T K/\mu \nabla N \, \mathrm{d}x .
So the unit of \mathbf K_{pp}p is [\mathbf K_{pp}p] = \mathrm{m}^3 / \mathrm{s}.
I.e., HT has a volumetric formulation and the source term has to be given as a volumetric quantity, which is what you already found out:

If I insert the source in volumetric unit = 0.0007 m^3/m/s it perfectly works and matches the real data.

Best regards,
Christoph

1 Like
5

Hello,

Thank you very much for your help.
It would be nice if it is also added to this section of the website as it makes a big difference.
https://www.opengeosys.org/docs/userguide/basics/conventions/

Best,
Mohammad