log(K) vs. phi, looking for K-Predictor for geological model

[petengr]

hello,

if you have plotted log k vs. phi, or log k vs. log phi, and there was no trend, what other option can you do?

[vinomarky]

Split it out by facie and re-plot log k vs phi for each facie

[reservoir_engineer]

i agree with Vinomarky, you should split it by facies, flow unit, porosity ranges,....

[paolomaldini]

PETROPHYSICAL MODELING IN PETREL - PERMEABILITY PROPERTY

PLEASE I NEED TO CREATE PROPERTY OF PERMEABILITY DATA FROM CORE DATA.
I THINK THAT IS WITH A FUNCTION XY LOG NORMAL X=K LOG ESC Y = POROSITY NORMAL ESC.

IF ANYONE CAN HELP WITH WORKFLOW.
THANKS IN ADVANCE.
I HOPE YOUR HELP.

pmaldini85@hotmail.com

[vinomarky]

Bit confused, you talk about petrophysically modeling, yet you explain a more stochastic modeling process... which is it you are after? Modelling at the wells, or populating the rest of the model? Co Kriging?

[paolomaldini]

YES, I´M TALKING ABOUT PETROPHYSICAL MODELING, THE OBJECTIVE IS TO CREATE PERMEABILITY FOR STATIC MODEL (K ABSOLUTE) AND THEN POPULATE THE DATA WITH SEQUENTIAL GAUSSIAN SIMULATION.BUT I have seen in some papers that for create K absolute the autors use core data, Log K vs porosity then fit the data and the equation is used in petrel calculator for create the property.
another that method that I have seen is create the K absolute property with empirical equation from well log data for example Eq of Timur.

which would be the correct flow?
if I'm wrong please correct me.
sorry and thanks for help. pmaldini85@hotmail.com

[Godson]

Using a log derived K/Phi transform is quite tricky if not validated in part by core data. Several transforms have been aduced by several authors but recall that unlike Phi, K is much more unpredictable and relys heavily on stratigraphic and structural architecture (at the static level) Reservoir fluids rel perm also contribute in the dynamic sense.

So to answer your question Paolomaldini, I would start by considering core data from analogue reservoirs (this would reduce my uncertainty albeit not eliminating it entirely) Plot log derived K from core derived K and establish a relationship (Linear, log, exponential etc). This should be done if you can correlate the cored interval in the analogue field to any level in your reservoir/field. If there is a relationship you can extend this to your uncored interval using the appropriate equation.

[ahmedm]

usual method for calculating permeability from porosity and vcl
by plotting logk vs. porosity with function of vclay to represent the effect of clay on the permeability
you can plot logk vs vclay then combine it in the porosity plot so you will have lines on the data presenting vcl
lets say fro 0 to cutoff lets say 40%

usually this is done by the res eng then given to the geologist to be anticipated in the modeling of permeability
i think the book of Amyx- Petroleum Reservoir Engineering-Physical Properties gives a good reference in this case


all the best

[catapam]

This is a simple example of a plot Log(K) vs poro on real data and how can be fitted based on mathematical equation

[dipak_m]

Dear Catapam,
What is the basis to chnage C1, C2 & baza log? Is it only for curve fitting?
thanks

[catapam]

Dear Catapam,
What is the basis to chnage C1, C2 & baza log? Is it only for curve fitting?
thanks

yes, this is the issue, to find a curve describe by a mathematical equation that fitting with your points.
then this equation could be used in Petrel

[ahmedm]

as catapam showed it is ok approach as a stight line to fit the data !!!!

i think the best way to the set attached and also to any data is curvey shape which build up with low porosity values
and bend to somehow flatten or still building up based on core data to high porosity values

to remark, i think applying cutoff of the porosity, permeability and Vcl data might refine the data furthermore

good luck