What do you use in simulation Pc or Jfunction?

[catapam]

I know you can use one or the other.
Still what will be the difference?
I'm planning to test it myself on a model but I would like some opinions.

If u are using core data split on RQI =0.0314SQR(perm/poro) then Pc points will be more scattered requesting more saturation regions compared with Jfunction.

So what are u using Pc or Jf and why?

[vinomarky]

Which ever one works for you. If you have a wide range in perms, then you may need a bunch of capillary pressure curves. If I can fit one or two j-functions, then this is my preferred approach IF I need to use capillary pressure. Often (high relief and/or high perm) you can get away with lognormal relationships of SWL vs PERMX (calculated at run-time with OPERATE) and zero Pc, which makes life even easier (essentially assuming transition zone has negligible effect on the outcome of your simulation)

In terms of when you'd use Pc over J-function, would be if you simply didn't trust the J-Function (not enough tests to fit etc) in which case you may get silly initialized saturations

Important to note that unless you set the porosity exponent to zero, you NEED TO ENSURE THAT PERM AND POROSITY ARE POPULATED APPROPRIATELY THROUGH THE MODEL. That means developing a facies model (driven by your RQI classes) and ensuring that for the given poro or perm the corresponding poro or perm is used.

[catapam]

developing a facies model (driven by your RQI classes) and ensuring that for the given poro or perm the corresponding poro or perm is used.

regarding facies model and RQI classes
we could have , as example, 3 rock types limestone, dolomite and silt, as were identified by facies modeler guy
on the other hand I could identify inside of limestone more rock types based on poro-perm data and that mean I will have a higher SATNUM regions than lithological defined classes.
Is this OK or if I'm using jfunction I should have only 1 jfunction for each lithological defined rocktype?

[vinomarky]

You can have as many or as few J-function regions as you need to adequately describe the various rock classes from a saturation & flow characteristics perspective. There are many types of 'facies', used to aid different disciplines to appropriate lump rocks into families - ie electro facies for log analysts, depo facies for geos, hydraulic units for res eng's etc..... They don't always correspond one to one.

Most often though, you normally find that one or more of (for example) depositional facies will fall into a single flow facies (or RQI group). So you simply need to figure out which of the flow facies as described by the geos fits into your RQI classification(s). If (for example) you find facies 1 and 2 both fit nicely in RQI index 1, then you can either calculate a new SATNUM property in Petrel to reflect this, or simply specify identical SCAL/PC tables for SATNUM values of 1 and 2

In your case, if you can identify subgroups of facies that need more granular RQI classes then it may well be appropriate to honour those, although I would question how you intend to identify the extra groups through the entire model - if it is a simple perm/poro cutoff then a simple Petrel property calculation operation would be all you need for a more granular SATNUM array

[catapam]

what I have is poro / perm / Swi data from cores.
the table was already split in 2 by geologist dolomite and limestone
I could ignore geologist split and put everything in 1 table and then to try to split on classes based on RQI
I could also keep the split and to create separate classes for dolomite and separate for limestone
At the end I could have , as example, 3 Pc curves for limestone and 2 Pc curves for dolomite, that will mean 5 SATNUM regions in total.
On the other hand Jfunctions narrow the range and maybe I could have 1 Jfunction for dolomite and 1 for limestone and then 2 SATNUM regions

So my understanding was that using Jfunctions you can have less SATNUM regions compared with using Pc functions.

I have not much experience in simulation so pls tell me if this make sense.

[vinomarky]

I get the distinct impression we are talking in circles here. Here's a real easy way to look at it... To paraphrase Einstein, you need to make things as simple as possible, but not one bit simpler.... If you need to add more SATNUMs then do so, if you don't then don't. Only you can determine exact number needed

[catapam]

OK,

thank you

when you starting something new is always better to ask twice

[dante89]

Great