- Ava Saturation
- Article 3
Date of publication: June 22nd 2018, by Matt Bowyer
Thank you for taking the time to read my first article, in which a geologist was let loose deriving a saturation height function… hopefully you're not too traumatised and thanks for coming back!
An interesting follow-up to that article on water saturation modelling is to consider the task from a multidisciplinary approach, as opposed to an 'individualistic' approach. Rather than one person doing the work, and someone else applying a function, I think there are a lot of benefits to getting geologists, reservoir engineers and petrophysicists to work together on the saturation model. Is this too idealistic, I wonder?
What are your thoughts… is there a benefit to a multi-disciplinary approach when modelling water saturation?
I think it's worth considering a few things when thinking about this particular task, but keep in mind this is not an exhaustive list. Do tell me what I'm missing, or if you disagree with my thought process. My aim is to start a discussion.
So what are the benefits of collaboration?
From my (a geologists) perspective collaboration helps get my m's and Rw's right - working with my engineering and petrophysical colleagues to understand wettability (thank you to those who mentioned this in my previous article!), building cap curve functions and comparing them back to log based functions. There's also the requirement for full QC of capillary pressure measurements prior to use in saturation height modelling - I certainly require, well, let's just say "a bit" of support here!
From a reservoir engineers point of view, getting a good match between static and dynamic volumes, I think, is important. Building a shared saturation model so that I'm not doing one thing and my RE doing another. We probably need to have a chat about grid size too, to make sure that we get the right compromise… not too fine, not too coarse. Talking about grid size before modelling might alleviate the need for upscaling from static to dynamic. I'm not a big fan of upscaling the static model, but I'm happy to be persuaded otherwise. I actually prefer to build a grid at a scale I'm happy with and my RE is happy with. I remember an instance when I didn't talk to my RE and I gave him a model with about a million cells. I wonder if he's forgiven me yet. I digress… what really matters to the RE? Matching the STOIIP between static and dynamic models? Initialisation times? Predicting water cut? Modelling multiple scenarios and testing uncertainites? How do you want to tackle Sw uncertainty? I have a few thoughts… I'm sure we all do, but what is priority number 1?
My petrophysical colleagues have the knowledge to be able to calculate our reservoir parameters - applying the important and necessary corrections and be able to deal with complex environments where we have low contrast, low resistivity pay, variable Rw and so on… QC of core data as I mentioned before, understanding of the saturation history of a reservoir, imbibition, drainage… you know what I mean. However, I'd like to think that, as a geologist, I have something to offer here and lend a hand with looking into reservoir quality, tying in my knowledge of the structure and depositional environment.
Between the three disciplines, and integrating all forms of data I think this makes for a pretty good saturation model. Heck, I think it makes for a good understanding of the reservoir! And an important step toward understanding where the uncertainties are and challenges we face.
Bear with me, digression required (again!).
My apologies to the geophysicists for not giving you a look in here - I know I have missed your contribution, so please shout at me and set me straight; I'm being naughty and focusing only on the logs and reservoir properties.
Getting a model that works from static to dynamic (think audit trail and repeatability), where we have started to take a look at where the uncertainties are, should enhance our confidence in a number of things: STOIIP/GIIP, production profile, water cut, planning for secondary recovery (insert other key understandings here), and importantly, gives the Asset Manager that warm fuzzy feeling they're looking for - the feeling of confidence that the problem has been addressed from differing angles and different disciplines.
Is what I've outlined above radically different to how people work today? I'd be very interested to know! How would we go about getting these three disciplines working together to handle this task? Is a cross discipline saturation model as important as I think it is or am I wrong?
Without biasing your answers, for me it would be different to how we currently work and very beneficial.
Thank you again for reading and for you interest!
All comments welcome!