TOUGHREACT ECO2N
Hi,
I'm quite new in using TOUGHREACT. Can anyone please clarify one question I've been having for quite a while now? The system I'm trying to model is just simple mixing between brine defined by the concentrations above, and non-condensable gases comprising of CO2 and H2S (total P of 35bars and T=25degC). TOUGHREACTv2 ECO2N module is used. No mineral phases were defined (tank reactor).
Now, I recently read in one paper (Xu et al., 2007) where CO2, H2S and SO2 were reacted to sandstone formation, that ECO2N only considers three components: sup. CO2, water, and brine, and H2S and SO2 had to be 'incorporated' into the injected brine. Was I wrong in putting in H2S as a gaseous species? Can the later ECO2N versions handle multiple gaseous species? And if I were to incorporate any gaseous species other than CO2 into the brine, how do I accomplish this?
Thanks in advance.
Alex
4 replies
-
Pre-V3 versions did not have he capability to inject trace gases with the main gas (CO2 here). So with V3 you are correct to include H2S and SO2 as gases (not as an injected brine). Enter the concentrations of injected H2S and/or SO2 (as mole fraction = volume fraction in the CO2) as an "injection gas zone" , as in the example below
#'----------------------------------------------------------------------------'
#'INITIAL gas and INJECTION zones'
1 1 !number of gas zones, number of injection zones
1 !gas zone 1
#'gas partial pressure, or zero to equilibrate with water
'co2(g)' 0.0
'*'1 !injection zone 1
#'gas mole fraction
'h2s(g)' 1000.E-6
'*'
-
Hi Nic,
Thanks for the reply. So if I understand it right, I have to represent any trace gases injected together with the CO2 as an 'injection zone' and not co-inject together with the CO2. Any reason for this?
Also since I'm still using ver2, any tips on how I can still introduce trace gases into the model (in my case H2S)?
By the way, has anybody encountered any problems when running the model at relatively low temperatures? I am modeling an actual experiment where they first had rock-brine interaction at 25degC for a week, before increasing the temperature to 200degC. The problem I'm having is that, unlike what was observed in the experiment, there seems to be too much activity going on. Precipitation of secondary minerals strip off the brine with the primary ions in the aqueous species. In particular, Mg+2, Fe+2, AlO2-, and SiO2(aq) concentrations cannot go up with the precipitation of the secondary minerals, and I end up having bad initial conditions for the 200degC run. I did try running the 200degC model alone (assuming a different set of initial conditions) and I got relatively good matches.
Again thanks.
Kind regards,
Alex
-
Hi Alex, there is no injection gas zone option with toughreact V2. However you can still approximate this by adding an infinitely large grid block to the MESH (a gas boundary block), which must be then connected to the grid block(s) into which you are injecting CO2 (i.e., the grid block(s) in your GENER) using small distances (e.g., 1E-6). Then define a separate gas zone in file chemical.inp for this large grid block, and assign it the partial pressure of CO2 plus partial pressure of other gases (the values must correspond to the concentration of these traces in the CO2, that is Pi/Ptot ~ V/Vtot using Ptot as the in-situ pressure at the well). In the flow.inp, for his grid block, set the permeability to 0, use the relative permeability and capillary pressure functions model #1 set for 0 rel perm and capillarity. Use CO2 mass fraction = 1 for this large grid block in the INCON. Depending on your problem, you can then play with the connection distance and area to make sure you have enough diffusion of trace gases from that grid block (into the injected CO2) to yield the desired concentrations at the well. It is a bit awkward to set up, but it should work (I tested this approach against the new gas injection zone option in V3 with similar results).
As far as initial conditions, if unwanted reactions in the ambient system at low temperature change the chemistry of the brine beyond what is observed, this means the choice of minerals and/or reaction rates for the modeled system is not appropriate. Obtaining steady ambient chemical conditions for any system if one of the the most difficult and often time consuming task in RT modeling and it is a critical step. The best is to first run simple reaction path simulations without transport (one grid block) to test the system. I also recommend starting with "rock titration" and/or "gas titration" type simulations at equilibrium (using geochemical codes such as CHILLER, GWB, phreeqc, etc.) to test the thermodynamic limits of the system and the types of secondary phases that could form.
- Nic
-
Thanks a lot Nic!
This is a new insight. I may try this one out on the model I'm working on now. The work-around I did for my previous models was to run an extra no-flow 'equilibrium' model which contains only one grid block. Here I react the fluid with the gas (containing the appropriate partial pressures like you defined) and assumed either (1) the gas react to equilibrium with the fluid, in which case I run the model for a sufficiently long period of time or, (2) the gas completely dissolves in the fluid, which the lab personnel doing the bench experiments claim is the one they observed.
The difficulty I encountered with the second assumption, though, is deciding to what aqueous species these gases speciates into. The aqueous species for CO2 and H2S (which are CO2(aq) and H2S(aq)) are not primary species and therefore their initial concentrations cannot be defined in the model (or is my understanding wrong?). So I have to chose HCO3- and either SO4-2 or HS-. I had to do trial and error with these to get the right secondary minerals to precipitate (for my 1D and 2D radial flow models). From here, it's just using the previous run results as INCON for the next run and so on. Was my approach to the problem correct?
Also, I am not familiar with the other geochemical modeling platforms and this is actually the first time I'm doing this (for my post-grad) so I may need to employ the help of someone good at this.
Another thing, I've been working previously with acidic environments so I just recently came across this other problem. The one I'm working on now is slightly alkaline condensate water. So initially when running the ambient (RT) model I assign ICON=3 for the concentration of H+ for me to be able to assign pH values. This gives a negative total molality in the chdump file. Now when I switch to a higher temperature (say 200degC) and use the negative molality, my run terminates or doesn't initiate at all. Should I be using the negative molality?
Again thank you very much for taking time to answer my queries.
Kind regards,
Alex