Imbibition curves

[vinomarky]

Just replying to a private message received below - so that others may benefit

ECLIPSE help needed!
Dear vinomarky,

I read from the forums that you are quite knowledgable in ECLIPSE..

i need some help here i am a student doing a project in reservoir modelling using ECLIPSE..

the objective of my project is to generate the capillary pressure vs saturation curves and analyze its behavior during imbibition and drainage in IWAG.

can you guide me on how to obtain this output curve?

thank you so much for your help!

Sincerely,

Ivan.

I have never had cause to use imbibition curves, so can not offer a whole lot of help unfortunately. I can, however offer some thoughts as to why they are so rarely used.
(1) Cap pressure curves exhibit hysteresis - that is, the plot of imbibition depends upon where you started from on the drainage curve. Consequently, to map out a robust set of imbibition curves requires quite a few tests starting from different levels of initial saturation. Add into this the fact that if you simply use one core and perform drainage, imbibition, drainage, you'll end up with different drainage curves - and you start to understand that actually measuring the data to create the imbibition curves is difficult if not impossible. You'd have to use a number of cores, each with similar properties and perform successive imbibition tests from different drainage saturation points.
(2) The effects of rel perm shape often outweigh/mask imbibition capillary pressure effects

I'd be interested to hear from the rest of the community - have you had cause to use imbibition curves? did you notice a significantly different outcome over what you'd have achieved with a slightly different shaped rel perm curve anyway?

[ivancxl]

hi there.

the one who sent you the pm was me and thank you for your fast response!

i had a meeting with my supervisor just now and he mentioned something about using SCAL to generate
the capillary pressure curves to be input into the eclipse data file and from there we run the simulation to
determine the recovery rates and compare with the base case without WAG..

my supervisor said there is some keywords missing in my GRID section that i need to include (something SCAL import stuff)
so that SCAL will be able to generate the Pc curves and from there i will input it into my eclipse .DATA file and run the simulation from there.

besides that, the model that i have is only operating under normal injection and i have to convert the water injector wells into WAG injector wells..

[Shakespear]

Vinomarky comments are on target. In my experience I have never been involved in a WAG process and to be honest we simply never worried about the effect as the fields were small, late in their production cycle. There was no way in this world that the necessary data would be made available or was even possible to obtain by the contracting company. If the company is say SHELL, ARAMCO you can bet they will get the necessary data.

I was looking for a good picture of this hysteresis process and found this related to SOIL science :-) But it is clear so worth looking at

[link Point to another website Only the registered members can access]
Here is something related to WAG processes

[link Point to another website Only the registered members can access]
These guys have the ECLIPSE data files here

[link Point to another website Only the registered members can access]
And here is something that may answer vinomarky question from a place I would least expect to find it, MIT. And they did have lab data to match against.

[link Point to another website Only the registered members can access]

The results of the simulations presented here indicate
that WAG simulation is quite sensitive to the treatment
of relative permeability hysteresis. Depending on the
hysteretic model used, predictions of oil recovery may
be as much as 15% higher than without hysteresis
(Table 2). Among the different models typically available
in reservoir simulators, the largest improvements
in recovery predictions are obtained with the three-phase
WAG hysteresis model in combination with the
Stone I interpolation method. This choice is precisely
the one that agrees with experimental data best.

Again, cost , saturation history are issues that make it all very difficult to do as vinomarky has said. Issues that in the Real World you will need to get around by SIMPLIFYING the problem.

It does not mean you should ignore them, but you should keep them in mind relative to WHERE you are working. SHELL Oil, OK mention that you need it, XYZ Oil Company, forget it, they won't spend the money as they are probable just interested in knowing where to drill and move on to the next project (or they give you 2 weeks to do the work and then need the results).

How to get out of the Jail

[link Point to another website Only the registered members can access]

[ivancxl]

ouch.. >.<

ok then but i hope someone can answer my other problems..

my project here is strictly research based only my work is to analyzed the behavior of capillary pressure curve during WAG injections..

really appreciate it if anyone can teach me how to:

1. import GRID data from eclipse to SCAL (keywords to include in the GRID section so that SCAL can read the data)

2. convert water injection wells (8 in total) to 4 Water and 4 Gas which alternates its injection (using WCYCLE,WCONPROD,WCONINJE)

3. what keywords i should use in the SUMMARY part so that the Pc values are generated as an output after simulation is run

[vinomarky]

You need to spend some time learning the software yourself -
GRID section defines the shape of the model

SCAL stands for Special Core Analysis - it is the tests you perform of core to get the data required to feed ECLIPSE capillary data

Get the simulation deck that Shakespear pointed you to

[ivancxl]

@shakespear

i cant open these files can you download it and share it on another platform? thanks..

These guys have the ECLIPSE data files here

[link Point to another website Only the registered members can access]
And here is something that may answer vinomarky question from a place I would least expect to find it, MIT. And they did have lab data to match against.

[link Point to another website Only the registered members can access]

@vinomarky

i already have a working model which can be run in eclipse.. but then i am missing THINGS that prohibits me from getting the result that i needed..

so if i could get some guidelines to guide me on the right path i would really appreciate it..

i am on a tight time constraint and if i cant complete this project i cant graduate!! =(

please i am desperate in need of help.. i dont know what to do!!

[Shakespear]

Go here

[link Point to another website Only the registered members can access]
Click on the word ------> [get]
Do this for ALL the FILES

Everything is there. All The Files to run the ECLIPSE.

I would take a day and read carefully WHAT they did, THEN make adjustments to your model to do a similar thing. THIS TAKES TIME. You can NOT do this over the internet with us here, no time on this end. Sorry

[ivancxl]

thanks alot guys..

i'll try my best.. i'll post here if i encounter any prob.. =)

[ivancxl]

@shakespear

i still cant access the page..

The requested URL could not be retrieved

While trying to retrieve the URL:

[link Point to another website Only the registered members can access]
The following error was encountered:

Unable to determine IP address from host name for juanesgroup.mit.edu

The dnsserver returned:

Server Failure: The name server was unable to process this query.

This means that:

The cache was not able to resolve the hostname presented in the URL.
Check if the address is correct.

Your cache administrator is root.

[vinomarky]

replace the address with 18.181.0.46/publications/hystco2

And sorry to sound brutal, but in all student projects I have seen, there is enough time to actually learn the stuff - either you've wasted a bunch of time, or the project expectations are too high for the time you have.... the sorts of work you are looking to do(if done properly) would be significant for experienced simulators (ie years of experience) let alone someone who has problems with the basics... not quite sure what it is your supervisor expects

[ivancxl]

i see..

to confirm.. does SCAL requires an input file to run too? is it the .RCI file?

and do you have an example SCAL input file that i can refer to?

many thanks! =)

[Shakespear]

"wasted a bunch of time, or the project expectations are too high for the time you have"
Yes vinomarky, YOU ARE RIGHT about this.

So Ivan, take our advice.

1) Study this example I suggested. Once you understand what they did, modify you project to your needs.

2) Discuss this with your adviser. I know in Russia they tend to be "DO IT, come back in a week.", hence no hand holding. So, learn as much from this example and THEN go in there to discuss.

3) If you can not download the files then I will zip them up and send to your PM.



PS: I ran their CASE1 and it runs OK, so you have something THAT DOES WORK. In my time I did not have this kind of help, so you are light years AHEAD. Take advantage of it !!!!!!!!