To paraphrase Edwin Star ... or if you're an 80s child, Frankie Goes to Hollywood ... SCAL, what is it good for?
Last week I wrote about something that I know a lot about … this week I'm much less confident.
I recently wrote about my skepticism of absolute rock properties from core analysis and in the discussion thread a much respected ex-colleague contributed with a comment on waiting for SCAL results, so here goes.
From 1996 to 2000 I worked on Beatrice in the North Sea, a delightful field that I hold most dearly in my heart. As a mature asset Beatrice wells generally had a water cut of >85%. If a new well came on at <75% water it was deemed a raging success and champagne corks were popped. Multiple pore-volumes of water had been, and still were, injected and I and the drillers use to refer to it as "washing dirty rocks". But REs might simply call it "converging on residual oil saturation".
Residual oil saturation (Sor) and all the other elements in a relative-permeability curve are deliverables from SCAL studies, so we should be able to start this discussion with something uncontroversial, namely; that rel-perms are the single most important element in how a reservoir behaves?
When two immiscible fluids simultaneously flow through a porous medium they mutually hinder each other due to capillary forces and so the total flow is reduced. This is expressed by relative permeability curves. Rel-perm curves are usually non-linear and when substituted into Darcy's law can tell us about saturations, water and oil velocities, fractional flow and mobilities. When extended with Buckley-Leverett theory and Welge's method we can now model a water (or gas) front; its speed and the saturations at and behind the front, the time to water break-through, cumulative production, PV, STOIIP and RF. That is an awful lot of oilfield development information.
And isn't that exactly why we expend a great deal of effort and money and time to understand our rel-perms? Corey vs LET, coefficients, irreducible, critical and residual values, and end-point scaling. It's all vital information? Isn't it?
1- Collect core: millions of dollars. 2- SCAL: 100,000s of dollars. 3- Wait 2 years.
Whilst we are waiting those 2 years we will probably use Corey curves and coefficients of Nw=2, No=2 and hope it's close enough whilst we get on with our field development. We promise ourselves that we'll update with lab obtained SCAL values as soon as they're in. Uncontroversial? Surely?
But do SCAL results make any significant difference? When compared against other production variables are rel-perms important? In uncertainty studies would rel-perms make the top-5 on our tornado-chart? ... or the top-8? ... or the top-10?
Try this: take a slab of reservoir 10m thick, 100m wide and 1000m long. Make it horizontal to remove gravitational forces. Porosity 15% and permeability 10mD. Our oil viscosity is 3.0cp and our water viscosity is 0.4cp. RhoW is 1000kg/m3 and RhoO is 850kg/m3. We're going to use our preSCAL Corey rel-perms with coefficients of 2, and we'll produce at 5000stb/d.
Using your Buckley-Leverett tool of preference we can see that our time to water breakthrough (Twbt) is 64 days with a Sw at the front of 37% and produced water cut (Fwf) of 67%. We've produced 321,000stb which equates to RF of 45%.
2 years later we get SCAL results ... our coefficients are not 2.0 but now 2.5 (gasp) ... we plug the numbers in ... Twbt is now 69 days instead of 64, Sw is 43% not 37% and Fwf is 77% not 67%. We've produced 340,000stb not 321,000stb equating to 48% not 45%. That’s not a lot of difference for a 2 year wait.
The problem is that nearly everything in a fractional flow calculation is linear … except the influence of rel-perms. So if you change the Corey coefficients nearly nothing happens but if you change the production rate (Qo) then a lot happens … double the production rate and Twbt halves. If you double the porosity the Twbt doubles, and so on. Permeability has no effect on my example above because I assume it to be horizontal but once gravitational forces are included (the reservoir dips) then the absolute permeability is highly influential and yet the rel-perms are not.
If you change the viscosity of oil and water if effects the Twbt as little as the rel-perm curves but viscosity effects the mobility-ratio and that defines the very nature of the water sweep; a piston front or viscous fingering.
The Corey or LET coefficients will however affect late-time recovery, in other words, how quickly you reach Sor? Or how many pore volumes of water you have to flush through your reservoir to get to Sor? Which gets us back to Beatrice and "washing dirty rocks".
That, of course, will then be a "practical" Sor where Kro has reached a threshold value of, say, 0.001 and it will not be a "theoretical" Sor obtained from SCAL. But tell me; which subsurface team does SCAL in order to understand the late-field life?
So when it comes down to it … how useful is that 2 year SCAL programme? Personally, I find it very difficult to believe it's not useful but the evidence for is not overwhelming taking a "pragmatic view" ... of course "pragmatic" is an often used euphemism for a lack of attention to detail.
Like I, and Edwin and Frankie said at the beginning: Rel-perms? What are they good for? Say it again!