Cooling cell
Hello and thanks for maintaining this forum, very useful indeed.
I am running a variation of the 5 spot problem in Toughreact. Essentially this is a polygonal mesh and there is a 10 meter high fractured zone in the centre where most of fluid flows between an injection well and production well. The attached figure shows temperature isosurfaces after 10 year injection of water in the left lower corner and production on the far right corner. I have no complaints about what happens during these 10 years and the chemistry and evolution of pore space seems reasonable.
However, after 10 years, I restart the simulation, now with new injection fluid and then one cell in the right lower corner starts to cool down. I decided to run few restart simulations but after approx. 12 years total simulation the temperature in the cell went below zero and the simulation halts. The second figure shows the temperature decrease just below the fracture zone after 12 years and 1 day simulation (just after restart).
Can you advise what could be causing this? There are no significant porosity changes observed in this region (only minor changes in minerals due to the temperature and pH changes). I do not treat the cells in this region any differently than other cells in the model.
Many thanks for taking the time of looking into this.
Bergur
Details:
I am running a Petrasim 2015 with Toughreact 1.2
Initial T and P, 200 °C, 3e7 Pa
Matrix initial porosity: 0.01
Fracture initial porosity: 0.1
Most minerals (The Soultz reservoir) are kinetic.
I allow reactions to affect porosity and flows and apply Verma-Pruess Perm-Por relation.
Pumping procedure:
Initial 10 year pumping
Reservoir water cooled down to 60 °C in PHREEQC and then injected into injection well.
Water produced: Well on delivery with PI 2e-12 and pressure 3e7 Pa
Restart:
Solution from 10 year pumping mixed with HCl in PHREEQC and injected for 1 day.
2nd restart:
Water injection followed by acid solution and so on.
4 replies
-
Hi Bergur,
Yes, this is a little funny. Which EOS are you using? I don't think this has anything to do with chemistry, but you may want to check that on restart the porosity didn't increase too much.
I looks like a very large grid block adjacent to a much smaller grid block? See what the connection lengths are. If PetraSim used AMESH to do the Voronoi tessellation, then it puts the grid block centers equidistant to the interfaces. That could lead to problems in the pressure gradients. Now if I use AMESH I move the coordinates back to the center and revise the connection lengths.
Let me know what you find!
regards,
Eric
-
Bergur,
One more thing -- Did you try the same problem with the restart turning the reactive transport off (just flow)?
Eric
-
Hello and thanks for the reply.
I am using EOS1. This does not happen when I turn off the reactive transport.
I changed the grid so there is more gradual shift of the grid block sizes.
There are no sudden porosity changes observed during the restart.
It is as you suggest, the temperature drops in the first grid block that has considerably larger volume than the fracture (the fracture blocks are yellow). The upper picture shows the original model. In the lower picture, I have divided the matrix layers into finer blocks and then the temperature drop is not observed immediately adjacent to the fracture. These pictures are both from 1 day after first restart. What I don't understand is why the temperature drop should only be in one corner of the grid. I will have a better look at the input files tomorrow.
Thanks again, Bergur
-
Hello Eric,
Problem is solved. It was as you suggested, problem with the mesh. I changed the fracture configuration (and the adjacent matrix as well) as the figure shows. Now the temperature stays constant during restart. Thanks for pointing me in the right direction and have a nice day.
Bergur
