Mr. Bigdeli,

I don't know about the tracking option, but Eclipse is probably not a very useful program for simulating low sal.   And really there aren't any programs that are good at this problem.  I prefer CMG for ease of operation.  All the programs use the same math and the current math for low sal involves changing the endpoint of the rel perm curves.  The only questions are how much to change the endpoint and what triggers the change?  So far they are just fitting the curve to experiments and upscaling.  Very crude.  I believe that BP started the train in 2005 and I haven't seen anything better with time.  CMG is also pushing along this route, but IMO has stalled out on sandstones.  The Penn State group is trying to implement Brady's work, but again seem to be heading down an unrealistic pathway that ignores portions of Brady's work.

The big questions are:

1 - how do you know the initial reservoir conditions?

2 -  how do you predict the response in the reservoir?

The issue is that the low sal mechanism is chemical in nature involving several simultaneous reactions between the oil and mineral surface acting through the water film.  This is not the most popular view.  Pat Brady and I are finishing a paper that shows a 'simple' geochemical model can match low sal experiments in carbonates. And I believe that Adeel Sohal's work is showing that these chemical reactions are linked to wettability alteration.  His second paper is coming out in E&F this month.  So for me wettability is most chemical sorption of oil onto mineral surfaces.  This was once a popular view, but became confusing because engineers thought sorption meant the oil displaced the water.  This is not the general case.  Only a portion of the oil can directly contact the rock.  Most does not.  Again consider Sohal's flotation technique, you have to have wet grains before the technique works.  Dry grains coated with oil will not float.  Sorption or if you prefer adhesion occurs through the water film.

The real question is how to use a geochemical model to correctly simulate multiphase flow.  Perhaps UT-CHEM, although I am not too fond of the code.  Multiphase flow models cannot incorporate geochemistry explicitly except by split operations, which are computationally slow.

Wanted to add I enjoyed your low sal presentation, very nice work.

Geoff

Dear Dr. Thyne,

Thanks for your information. I do know five simulators that are dealing with simulation of LoSal. They are Eclipse Petrel RE(2015 version) CMG UTCHEM and PHREEQC.  As you probably know, Austin not only got too many kinds of simulators such as UTCHEM UTCOMP UTGELL UTTEST UTWELL and GPAS (general purpose adaptive simulation) code, but also they tried to couple UTCHEM with PHREEQC too. You can look at below paper (Coupling IPhreeqc with UTCHEM to model reactive flow and transport DOI: 10.1016/j.cageo.2015.06.004). On the other hand, since source code of Asutin’s simulators are available (at least to themselves), they can modified their simulations.

The problem about using Eclipse simulator is that, Eclipse use simple model to simulate LoSal WF. Because of that, for any further kinds of development we need to accesses to the source code of the Eclipse. I do follow using of CMG by researchers form the University of Calgary such as Dong and his supervisor (Dr.Chen, I guess). CMG’s flexibility is much better than Eclipse. For confirming my claim I can mention below paper (Effect of Multi-component Ions Exchange on Low Salinity EOR: Coupled Geochemical Simulation Study DOI: 10.1016/j.petlm.2016.05.004) that used CMG for their simulation and you can compare their results with Al-Shalalbi’s works (Mechanisms behind low salinity water injection in carbonate reservoirs, DOI: 10.1016/j.fuel.2013.12.045),too.

I wanted to see how eclipse responded to LoSal WF for one of my graduate course, a couple of months ago. Hence I start reading of Eclipse manuals, I did aware that Eclispe was not best option, but it was the only available document resource that I got that time. (Austin did not release UTCHEM or UTCOMP models for LoSal WF).

I like to ask a question: As we know, the only available approach for illustrating LoSal mechanisms to simulators is by changing relative permeabilities and interpolation of curves. But for field prediction approach you need two other necessary elements. Capillary pressure and residual oil saturation! So how can we do simulation of the LoSal WF, if we wanted to ignore relative perms?

By the way, thanks for kind words about my presentation.

Regards,

Alireza.

-------------------------------

Dear Prof. Mahani,

You cited that  “there is mixing between the injected low salinity brine and the insitu high salinity brine.” Here is a clue.  I have some comments regarding this reaction. We are saying that there are two kinds of brines around the oil droplet. One with high sal and the other on with loSal. Personally I do believe that one missing item that is exist there is Maragoni Effect. Because of difference salinities, there are different IFT around the oil droplet. Look at the oil droplet individually! You can consider sides of oil droplet as different boundary conditions. However to my knowledge, there is not any information in the literature regarding my statement. If my statement is true one new mechanism for loSal is Maragoni Effect and changing IFT. This mechanism can be more comprehensive compared to other proposed mechanisms, but at this moment it is just a hypothesis. I do need someone who can guide me for further explanations and experiments.

Finally, I do appreciate your explanation regarding on Simulation approach of LoSal with Eclipse Simulator.