I received some messages asking for clarification on the relative permeability end-point scaling method (use of ENDSCALE keyword) in Eclipse. I decided to share my comments here:
A previous thread discussed the topic:[link Point to another website Only the registered members can access]
Visit it again and read it. I will briefly give you and introduction anyway but you'll need to search for more and built your own spreadsheets for some of these calculations.
Endpoint scaling is an option that allows the user to use the same normalized relative perm curve to a set of different regions in the reservoir (or blocks in a simulation model) while still honoring variations in rock properties such as connate water saturation (which is strongly and inversely correlated to absolute rock perm), maximum values of relative perms to flowing phases (i.e. end-point relative perms), trapped phase saturations (such as residual oil saturation). I will give you an example that will make this clearer for you. In first place you should know that, for instance, absolute rock perm usually correlates very well with connate (irreducible) water saturation for clastic reservoirs. See for example the book by Amyx et al "Petroleum Reservoir Engineering", page 151. This is true in absence of capillary pressure effects. Now, let's imagine that you have two clastic rocks composed of exactly the same minerals and filled with oil and water of the same compositions; in addition, wettability is the same for both rocks. However, the only thing that makes them different is permeability (even porosity is the same; since porosity is unaffected by grain size you might have two rocks of the same porosity but different absolute perm). Then you take rock samples (plugs) from both rocks and measure, by any method, the water/oil relative perm curve. Would you get the same water/oil kr curve? The answer is NO, because if perms are different then the connate (irreducible) water saturation must be different. Read SPE 2588 on the “Influence of pore geometry on water-oil relative permeability” for a more comprehensive discussion on this.
Now move on to another aspect. Let’s suppose you are asked to build a geologic model for a reservoir you work with. Think of a simple case in which you only have one geologic layer of interest. However, across this one layer there are still spatial variations of rock permeability, which is the common thing in real life. Thus you have hundreds or even thousands of grid blocks in this one layer, each grid block with a different absolute rock perm value. Since absolute rock perms are different in each block then the connate (irreducible water saturations) are different then you would need hundreds or thousands of relative permeability curves to honor each single rock perm in each block. Would you? Is it practical to input hundreds or thousands of relperms in your model? Of course not, that’s why we use End-Point Scaling.
The basic process of the Endpoint scaling option in Eclipse is:
1. Get a normalized/average relative perm curve that best honors your measurements. This is easily done in Excel, so built your own spreadsheet. Read page 313 in Ahmed’s book “Reservoir Engineering Handbook, 3rd edition”. There you’ll learn how to do it.
2. In Eclipse, insert the ENDSCALE keyword in the RUNSPEC section.
3. In the PROPS section input the following properties depending on the fluid system you are dealing with (i.e. water/oil only, gas/water, gas/oil/water):
• SWL - Connate water saturation
• SWCR - Critical water saturation (largest Sw with zero Krw)
• SOWCR - Largest oil saturation for which oil is immobile
• SGCR - Largest gas saturation for which gas is immobile
• KRWR - Krw at residual oil saturation
• KRO - Kro at connate water saturation
• KRG - Gas rel perm at maximum gas saturation as defined in rel perm table (usually connate water)
• KRO - Kr to oil at connate water saturation
4. Check the Eclipse manual all the time while doing points #2 to #3 to avoid mistakes.
This is it in simple words.
A
n additional advice for you, for anyone actually: I learnt a while ago that before asking for any help, you must do everything in your hands to come up with the solution by yourself. You must be able to prove yourself, and the person you’re asking, that you searched and looked for the solution with keen curiosity. This sometimes involves easy things like searching in Google for example. However, most of the times in Reservoir Engineering this involves going to the books and papers to read them. Do this, it will make you a better professional in your field.
PS: attached you'll find a excerpt from an Eclipse manual with some details on the topic.
Best regards,
Reply With Quote
Bookmarks