Porous_flow simulation with Segmentation fault: 11

[wjin33]

Hi,

I am using Porous Flow/Falcon to run brine flow and heat transfer problems. Prepared input file with water as the fluid runs ok, yet, switched to brine as the injection/extraction files have this error:
Segmentation fault: 11

Please help.

Many thanks!

Attached is the input file, I have to change the file type to be able to upload.

Cranfield_HT_season_caps_brine_newboundary_coarse.log
input.

[WilkAndy]

Your input file uses an exodus mesh, so i can't be exactly sure what your problem is, but i believe your problem is in your PorousFlowUnsaturated block. Instead of fp = brine, you need to specify

  fluid_properties_type = PorousFlowBrine
  nacl_name = f_nacl

See the doco at https://mooseframework.inl.gov/source/actions/PorousFlowFullySaturated.html .

There are other things you might like to change about your input file. Eg, the f_nacl is mass fraction, so setting it as 0.62 means you're simulating with 62% of salt. You'll probably also want to use PorousFlowFullySaturated, as that is more efficient and suited to your case (above doco).

When you use these Actions, you do not need to supply any FluidProperties, since PorousFlow does it all for you (see the "Materials Added" section of the above documentation). If you like, you can supply a custom fluid_properties_type if the ones automatically built by PorousFlow are too slow or don't suit you for some reason.

a

[wjin33]

No, it does not work, @WilkAndy. I tried both the fullysaturated action and the fluid-type lines. Before posting the question here, I have already tried different actions and input lines for brine. All failed.

I combined both the exodus mesh file and the input file here. Please kindly take a look.

Archive.zip

[WilkAndy]

Cranfield_HT_season_caps_brine_newboundary_coarse.txt

Hi @wjin33 , i wonder whether you need to update MOOSE. Your input file works fine for me, after i've removed the EnthalpySink stuff, which i don't have in standard porous_flow. So there could be something wrong with the EnthalpySink. The input file that works is found attached (it doesn't converge well, but that's probably due to the missing EnthalpySink).

a

[WilkAndy]

Did updating MOOSE work, @wjin33 ?

[wjin33]

Dear Andy @WilkAndy, I have already updated the MOOSE, and it works for water with my kernel "EnthalpSink". Upon your suggestions, I removed the EnthalpSink with PorousFlowPolyLineSink, and I can simulate the brine without error. Many thanks.

I double-checked my kernel and attached them for your reference. I think there are two aspects that may contribute to the error:

• I am using the SinglePhaseFluidProperties to calculate the enthalpy injection, which works fine for water97 equation of state. Yet, for brine, this is not properly implemented.
• I did not calculate the Jacobian for this kernel.

Any comments are appreciated.

Thanks,

Wencheng

EnthalpySink.zip

[WilkAndy]

Is EnthalpySink trying to accomplish the same thing as @andrsd #17235 ? If so, just take something like https://www.ratebeer.com/Ratings/Beer/Beer-Ratings.asp?BeerID=84035 and sit with him until it's done.