The extent of CO2 injection cooling using ECO2N
Aiming at thermal QC of another code by using Tgh2-ECO2N, I made a simple coarse model with dimension 25x5x100 where dx=dy=200m and dz=4m, por & perm were 0.3 and 500md, fluid in place were only pure water at hydrostatic pressure (Pavg. 288bar) and (avg. Tres=91’C) where top reservoir was at 2680m. A CO2 injector put in the center of the model, with 160m open interval in the middle of grid cell column. 2 mill ton annual rate with a water enthalpy corresponding to T=35’C and P=300 bar (about 250000 j/kg) was injected. Instead of physical boundary, a horizontal producer put in the bottom of the model with enough withdrawal to avoid pressure build up.
Running under this injection enthalpy, injection- rate and interval conditions, some of the grid cells hit the lower limit of 3’C after 4.35 years. How realistic this is; given relatively low expected J.T. cooling (large inj. depth Interval), would it be too much for CO2 dissolution/water vaporization part of the heat exchange?
Running the model with only 12m injection interval and 0.5 mill ton per year injection, the 3 degree limit reached by some cells after 2.1 years. How realistic this is; given relatively high expected J.T. cooling, but less rate?
I searched literature for experimental results hoping to find comfort regarding these results, but I found few works and they were done on limited scale, not enough to persuade me. Going back to the fundamental of EOS and knowing that Battistelli et al. formulation deviates when temperature goes below 20°C, would it be an additional element in overestimation of cooling?
Hoping that you may have some advise for me, I attach the 2 input files here.
Thank you reading this.
I checked your input. The enthalpy you entered does not correspond to the T, and P (30C, and the injection pressure in the output file) you mentioned. When I changed to the correct enthalpy (512760 based on NIST webbook), I got 40+C for the injection element after 5 yr.
An easier way to do this is to use constant temperature injection (if you do not want to figure out the enthalpy), i.e., assign the element with a different rock type with a very large rock density so T does not change.
What a relief, thank you.
I used this:
In addition, QCed by this:
Both gave me about 250kj/kg for 35'C and 288 bar. Could these site be so off!??
I haven't run it with the right DH yet, I will do that.
I tried the constant temperature approach and the results look much more realistic, it is also in line with what you obtained by using enthalpy.
But I still have a problem to get as high enthalpy as you have got from NIST, please see the attached image from NIST calculator, where enthalpy shows almost no sensitivity with pressure and it is about 260 kj/kg. Is there something that I miss here.
Appreciate your further comment.
I am glad the constant temperature approach works. I checked the few sources again and yes, the enthalpy they provided are all about 260kj/kg. I run a one-element problem (saturated with CO2) for 1 timestep and output the enthalpy, it is 563 kj/kg. Apparently TOUGH-ECO2N uses a different standard state (reference enthalpy) than the sources you and I used. However, the enthalpy change due to T change, calculated from those sources or TOUGH should be similar. Therefore, using a different reference state is OK. But of course, it brings inconvenience for the user who provides enthalpy in the GENER block.
So to summarize, a safe way to do this would be either using the constant temperature injection (preferred), or do a one-element problem for one timestep and output the secondary parameters (set MOP(5)=9) to figure out enthalpy at the desired T.
In addition, I noticed you had very large element. The pressure build-up is not much. The CO2 is mainly in the injection element after a couple of years. I hope this is only for test purpose.