Convergence Problems

[kingfisher]

Hey all,

I am facing a lot of convergence problems and I have no idea where can be the origin of these problems...
I have about 135k cells, a couple of faults, i use corner point cells, 3 OWCs, 3phases (but with gas still dissolved because P > Pb).
I left perm, poro and NTG homogenous for now to keep it simple...

Then i introduced one well and let it produce for 1 month.
When i start now the simulation, it takes a loooot of time, because eclipse is reducing the timestep due to the convergence problems...
Even if i just simulate three days it takes a hell of a time... I played a little bit with the tuning keyword, but i guess the convergence probs are too big and they originate from somewhere else...
I am totally new to eclipse, so i really dont know where to look ? What might be the problem for it ?
If it is helpful i could post the "problem" "warning" and the convergence related messages...


I appreciate any advice !

[Bang Gaol]

Maybe it's related with the corner points. And how about the PVTs? Do you mention for each phase? 3 OWCs? To make more simple, put only one OWC. And look what happens next.
And you can discuss with us afterwards.

[kingfisher]

Yes i mentioned all of them PVDG, PVTO, PVTW... when i just define one OWC its way faster and no convergence problems so far... :/

[ivan_aljeburi]

This problem can be related to PVT... is it black or comp?

[kingfisher]

black

edit: sorry there are still convergence problems also with just one OWC,but not that much like before...

[dhavaldesai4u]

----

[ivan_aljeburi]

Check the PVO file .... it can often be a problem...how did you created it ? in PVTi or by correlations? The most often problem is that bubble point pressure in PVO must be equal to your maximum pressure.

[kingfisher]

I did it by correlation...
"The most often problem is that bubble point pressure in PVO must be equal to your maximum pressure. "
You mean, that i should give eclipse just the correlated values up to Pb ??? I assume, that i got you wrong, because that doesn't make sense to me.


OK...I went one step back and left my schedule section out and just initialized the model: I got several warning of this type:


@ NEGATIVE COMPRESSIBILITY FOUND IN GAS
@ PRESSURE TABLE 1 AND OIL PRESSURE
@ TABLE 1 AT A SAMPLE PRESSURE VALUE
@ 182.31034 . ADJUST SATURATED FLUID
@ PROPERTY VALUES AT THIS PRESSURE.
@ NEGATIVE COMPRESSIBILITIES OCCUR
@ FOR GAS SATURATIONS LESS THAN 0.03998


Obviously there is sth. wrong in my PVT data... any idea what is the reason for the "negative compressibility" ?

[ivan_aljeburi]

First of all your PMAX in the .data file should be the bubble point pressure in the PVO file...that's mean that you should increase Rs until you reach a bubble point pressure that equals your PMAX...
You can use a small reservoir simulation trick, copy the compressibility data table (dBo/dP ,..) from the debug file into excel and try to change the Bo that cause a negative compressibility then change that value in .PVO..i will upload an excel file, i believe it will help you to solve the problem...

[ivan_aljeburi]

You can use this .xls file to correct your Bo

[link Point to another website Only the registered members can access]

[mkhurram79]

You can use this .xls file to correct your Bo

[link Point to another website Only the registered members can access]

thnx

[kingfisher]

thx a lot, but already got rid of the problem !!

When i change PMAX to my Pb, then during my simulation the control mode of the wells is changed to BHP... Any idea how to solve that ?

right now i am again messing around with convergence probs again... any idea on that ?

@--PROBLEM AT TIME 4.0 DAYS ( 5-JAN-1988):
@ SOLUTION IN WELL H19 NOT CONVERGED AFTER 14 ITERATIONS
@ RESIDUAL ERROR IN OIL, WATER AND GAS FLOWS =
@ 5.40361E+04 -1.04940E-09 9.30135E-04
@ TRY INCREASING MXWSIT IN TUNING RECORD 3