PDA

View Full Version : Kr data input in ECLIPSE and discussion about oi/wat and gas/oil kr.



DAH7542
10-29-2010, 10:51 PM
Hey fellows, I've two questions regarding 1. how should kr data (saturation tables) be input in eclipse and 2. oil/wat gas/oil kr data:

1. As you know relative permeability can be calculated using absolute permeability (e.g. klinkenberg) or oil permeability at Swir (Ko @ swir). I found in some schlumberger's forum that "SWOF (keyword for oil/wat kr data) will expect oil relative perms as per method that uses absolute permeability", however in "ECLIPSE BlackOIL Training and Exercise Guide", figure 53 (file not attached, I really hope you have it) they show that for both SWOF (oil/wat kr data) and SGOF (gas/oil kr) data, oil endpoint (Kro max in both cases) is and "should be" equal, furthermore, in that particular figure kro max is equal to 1.0 (oil/wat and gas/oil sat tables denormalized); this would only happen in denormalized kr curves if Ko @ Swir is the reference pressure (so kro max = 1.0). So my first question (finally) is: What would happen if I use saturation tables with ko @ swir as a reference? Will ECLIPSE automatically recognize that those sat tables are written that way and will provide the same results as if absolute perm was used*? *using ko @ swir as a reference requires knowledge of Ko @ swir (mD units), is it possible to input such data in eclipse?

2. Second question: Let's say a reservoir has oil/water (vs water saturation) and gas/oil (vs liquid saturation, i.e water plus oil) kr data measured at EXACTLY the same depth (same plug), however, lab results show that Kro @ Swir is different in both tests (this value is provided by the lab), what's the reason ?? furthermore, using absolute perm as a reference, should oil endpoint (i.e. kro max) in both saturation tables (oil/wat and gas/oil) be the same as ECLIPSE requieres??

Thanks in advance.....

vinomarky
10-30-2010, 02:52 PM
So my first question (finally) is: What would happen if I use saturation tables with ko @ swir as a reference? Will ECLIPSE automatically recognize that those sat tables are written that way and will provide the same results as if absolute perm was used*? *using ko @ swir as a reference requires knowledge of Ko @ swir (mD units), is it possible to input such data in eclipse?

You can put in rel perm curves relative to Kair, Kklinkenberg, Ko (@Swir) or any other reference AS LONG AS YOUR PERM IS APPROPRIATELY DEFINED. Eclipse does not require any of the endpoints to be equal to 1, so it will simply take the relative perm it gets from the lookup table given a block water saturation and multiple that number by the block perm as defined. So, if you wish to define Kro(max) as = 1 (ie normalize), then you should ensure that all your perms are appropriately reduced to a values in line with the permeability to oil at residual water saturation. Conversely, if you wish to define your perm grid in absolute (Kair) perm, then you need to ensure your Kro(max) and Krw(max) are appropriately reduced to something less than unity.



2. Second question: Let's say a reservoir has oil/water (vs water saturation) and gas/oil (vs liquid saturation, i.e water plus oil) kr data measured at EXACTLY the same depth (same plug), however, lab results show that Kro @ Swir is different in both tests (this value is provided by the lab), what's the reason ??
Many reasons. If it is the same core plug, then you are probably seeing a mixture of hysteresis and/or cleaning/handling artifacts - the more you delve into the details of SCAL testing & core cleaning/handling issues, the more you understand the breadth of uncertainty that you are dealing with....



furthermore, using absolute perm as a reference, should oil endpoint (i.e. kro max) in both saturation tables (oil/wat and gas/oil) be the same as ECLIPSE requieres??
See above - Eclipse does not require endpoints to be equal to 1

DAH7542
10-30-2010, 03:37 PM
Vinomarky, thanks for your comments, it's all clear now....

Shakespear
10-30-2010, 04:48 PM
Very very good notes on this are here



(Good that we have such good people willing to share out there :) )

One things to remember about those curves is that the End Points will be used in the initialization process of the model's saturations. When putting together the DATA file for ECLIPSE you will need to specify contact(s) location depending on the phases in your model.

The lowest value in the SWOF table will be used to initialize the SW in the Gas Zone and Oil Zone. The highest value will be used in the aquifer (very often people fail to do this), which should be krw = 1 at Sw=1. Hence the straight line one might see in reports showing the rel perms for water.

In the transition zone the Capillary Pressure will be used. So the cap data from ow/go/ is used, from OWC and up, and GOC and up. Depending what phases you have in your model.

So this has an effect on your In Place calculations and is one source of error. The end points.

Ah, one more thing, your question about the "need to be equal". This refers to the fact that "whatever" value you have in the O-W Table first value for Kro, that same numeric value must be for Kro in the G-O Table rel perms. Draw a diagram and think about it, and it will be clear why. If you don't you will be confused. I am assuming you are using SWOF and SGOF key words.

DAH7542
10-31-2010, 03:31 AM
Shakespear, thanks for your comments too, I think I undersand your point ... let me elaborate it this way: at gas saturation equals to 0 (first line in SGOF keyword) there are still irreducible amounts of water filling the pore spaces (i.e. water at Swi), so the first value of Kro as defined in SGOF would be at Sat gas = 0, which means, as explained before, irreducible water saturation, this makes that the first Kro value in SGOF data be equal to the first line (kro valued) in SWOF keyword (which starts at Swi).

Thanks again guys, it's been a very nice discussion......

Shakespear
10-31-2010, 10:26 AM
water filling the pore spaces (i.e. water at Swi),
Exactly, that lowest value of water saturation is taken by ECLIPSE into the Gas and/or Oil if present. This needs to be clear when putting these table together or else there will be confusion.
1141

dorsetlad
05-08-2011, 12:52 PM
Hi,

"The lowest value in the SWOF table will be used to initialize the SW in the Gas Zone and Oil Zone. The highest value will be used in the aquifer (very often people fail to do this), which should be krw = 1 at Sw=1". If this is the case, does this mean that relative permeability curves should be normalised to absolute brine permeability (Sw=1)?

I am not clear how this second statement reconciles with the first; "You can put in rel perm curves relative to Kair, Kklinkenberg, Ko (@Swir) or any other reference AS LONG AS YOUR PERM IS APPROPRIATELY DEFINED. Eclipse does not require any of the endpoints to be equal to 1, so it will simply take the relative perm it gets from the lookup table given a block water saturation and multiple that number by the block perm as defined. So, if you wish to define Kro(max) as = 1 (ie normalize), then you should ensure that all your perms are appropriately reduced to a values in line with the permeability to oil at residual water saturation. Conversely, if you wish to define your perm grid in absolute (Kair) perm, then you need to ensure your Kro(max) and Krw(max) are appropriately reduced to something less than unity"

So if you chose to normalise krw,kro using Kair, then you would use krw<1 at Sw=1 such that krw x Kair = Kw; have I got that right?

Any further insight here greatly appreciated.

vinomarky
05-08-2011, 01:32 PM
Ideally, perm to air (100% air saturation) and perm to water (100% water saturation) should be equal..... In actuality they can vary through (for example) things like clays dried in an oven allowing air through easily, yet re-hydrating and swelling to impair water flow..... If you have handled your core properly with humidity controlled drying etc, then after Klinkenberg corrections the air and water perms should be similar

So, assuming in your case Kair is greater than Kwater and you want to populate perms in your geomodel based on Kair, then yes your endpoint rel perms at 100% Sw should be at Kwater/Kair which will be something less than unity.

It may sound trite, but the easiest way to think about it is that you are trying to describe how the fluids flow in your system - the product of Kr at a given saturation and the populated simulation grid perm should reflect what you believe is the insitu perm to that fluid at that saturation.

dorsetlad
05-11-2011, 12:06 PM
Thanks for that Vinomarky, Cheers.