# Thread: Question to simulation engineers!

1. Senior Member

Join Date
Mar 2009
Posts
79

## Question to simulation engineers!

I am working in field with huge lateral extent. there are no separation faults to make compartments nor geological bodies to separate and compartmentalize the field.

however,one thing I notices is the change in PVT properties mainly viscosity is drastic !

ranges from just 1 cp in the west to 20 cp on the east.and the other PVT are almost identical.

so,can I consider the field as one PVT region or should I make subdivisions based only on one PVT property?

See More: Question to simulation engineers!

2. Spons Circuit
Join Date
Always
Posts
Many

3. Member

Join Date
Dec 2008
Location
Venezuela
Posts
44
One way:
you may divide your model in 4 or 5 pvt regions, according to API variation. then you need to define a pvt table for each region. The only problem is when the fluid migates from one region to other it wil change its properties automatically.

4. -

5. Active Member

Join Date
Dec 2008
Location
Pakistan
Posts
337
Use more than 1 region and accordingly pvt tables for each region. Now it depends on you how many regions you define for a reservoir

6. Very Active Member

Join Date
Nov 2008
Location
Australia
Posts
508
Suggest that you first try to understand the physical significance of your observations.... what's causing the properties to vary so much in an ostensibly homogeneous system? Perhaps it is not as homogeneous as you think? Could it be related to change of depth/temp over that areal extent?
Last edited by vinomarky; 03-13-2010 at 03:26 PM.

7. Senior Member

Join Date
Mar 2009
Posts
79

8. Junior Member

Join Date
Feb 2009
Location
Posts
9
please how do i solve this problem?

@ LINEAR EQUATIONS NOT FULLY CONVERGED - RUN MAY GO FASTER
@ IF YOU INCREASE LITMAX (= 25 - TUNING KEYWORD)

@--PROBLEM AT TIME 1657.8 DAYS (15-JLY-2004):
@ NON-LINEAR EQUATION CONVERGENCE FAILURE
@ ITERATION LIMIT REACHED - TIME STEP CHOPPED FROM 23.41

9. Junior Member

Join Date
Jan 2009
Posts
9
ahem... try to increase Newton iterations number or something like that in the numerical parameters of your simulator (have a look at the "numerics" section in the user's guide of your reservoir simulator); and/or decrease the time step (keep it constant and very very small); and/or relax the material balance criteria
Last edited by totor; 03-16-2010 at 04:27 PM.

10. Senior Member

Join Date
Mar 2009
Posts
79
Mr perezo

I suggest you workout the problem of IMPES.vs.Fully implicit in workshop data sets in Applied reservoir simulation seminar by Schlumberger

cheers

11. Senior Member

Join Date
Mar 2009
Posts
79
Originally Posted by vergatario
One way:
you may divide your model in 4 or 5 pvt regions, according to API variation. then you need to define a pvt table for each region. The only problem is when the fluid migates from one region to other it wil change its properties automatically.
the API is identical for the whole reservoir the variation is just on viscosity !

12. Senior Member

Join Date
Mar 2009
Posts
79
Originally Posted by 06pg22
Use more than 1 region and accordingly pvt tables for each region. Now it depends on you how many regions you define for a reservoir
the problem is that : the reservoir is shared by more than one company,and the objective of the simulation is to prove the communication and calculate the oil migrated.in history match phase the field was proved to have been in communication with two PVT regions .And now i use single PVT region there is more oil migration and better history match,but the other companies still arguing of using more PVT regions! what should I do?

13. Senior Member

Join Date
Mar 2009
Posts
79
Originally Posted by vinomarky
Suggest that you first try to understand the physical significance of your observations.... what's causing the properties to vary so much in an ostensibly homogeneous system? Perhaps it is not as homogeneous as you think? Could it be related to change of depth/temp over that areal extent?
you are right to some extent!

in the measurements there are some sort of scatter variations, but as general observation (average values) there are one region with viscosity around 10 cp and another one with viscosity around 20 and the WOC difference between theses regions is about 90 ft.
One interesting point I noticed is that the region of 20 cp is the only part of the model which doesn't have aquifer support!

14. Member

Join Date
Dec 2008
Location
Venezuela
Posts
44
A good tool to prove comunication is to use geochemist analysis. You have take samples of fluids in several points (for instance one ell of your company and other sample in a well of ther company). They generate a kind of diagram for each of the samples and later it can be said if the oil is the same. you can propose that.