[Opm] Error with upscale_relpermvisc

[Bård Skaflestad]

Hi Alun,

Sorry about the late reply.  You're essentially correct regarding 'MULTZ'.  The parser layer knows about this keyword and is able to read the data, but we do not use the information contained therein as part of the upscaling process in 'upscale_relpermvisc'.

I'm afraid I don't have suggestion for how to incorporate these multipliers into the calculation right now.


Best Regards,

Bård

-----Original Message-----
From: Opm <opm-bounces at opm-project.org> On Behalf Of Alun Griffiths
Sent: Wednesday, May 4, 2022 3:16 PM
To: opm at opm-project.org
Subject: Re: [Opm] Error with upscale_relpermvisc

Hi

I have tried converting the MULTZ keywords in my test ECLIPSE input file to PERMZ so that upscale can read the fine-scale grid.  Unfortunately, it does not seem possible to do this.  Consider a 1*1*N model with MULTZ(I) = 0.0.  In order to reproduce TRANZ(I) = 0 using PERMZ  only, then either PERMZ(I) = 0.0 or PERMZ(I+1) = 0.0.  If the former, then TRANZ(I-1) = 0.0 also and if the latter then TRANZ(I+1) = 0.0.  If either TRANZ(I-1) > 0 or TRANZ(I+1) > 0 in reality (which they are in my case), then conversion is not possible.  Am I correct in assuming that upscale does not read MULTZ* keywords?

Best regards

Alun

-----Original Message-----
From: Alun Griffiths
Sent: 28 April 2022 14:29
To: opm at opm-project.org
Cc: Bård Skaflestad <Bard.Skaflestad at sintef.no>
Subject: RE: Error with upscale_relpermvisc

Hi Bard

Looking at the source code was really helpful - thanks!   The .grdecl file was missing SATNUM and PORO and once I added them, upscale_relpermvisc ran OK.  I now need to check how the fine grid model + rock curves compares to the coarse grid model + pseudos.  I have MULTZ keywords in the .grdecl file but I don't think upscale_relpermvisc will read these so will probably have to adjust the PERMZ array to incorporate them.  Will let you know how I get on

Best regards

Alun

-----Original Message-----
From: Bård Skaflestad <Bard.Skaflestad at sintef.no>
Sent: 28 April 2022 13:28
To: Alun Griffiths <Alun.Griffiths at rockhopperexploration.co.uk>; opm at opm-project.org
Subject: RE: Error with upscale_relpermvisc

Hi,

Thanks for bringing this issue to our attention.  The first problem is *probably* that the diagnostic message is incomplete.  For what it's worth, the upscale_relpermvisc utility looks for the following input data:

* SPECGRID
* COORD
* ZCORN
* PORO
* PERMX
* SATNUM

If you're comfortable looking at C++ source code, you can review the keyword existence check on lines 351--356 of upscale_relpermvisc.cpp (GitHub URL: https://github.com/OPM/opm-upscaling/blob/78bf78a8373f59d5d3d0394a0d8de12e822c9c5a/examples/upscale_relpermvisc.cpp#L351-L356)

At minimum the diagnostic message should mention PORO, PERMX, and SATNUM too, but ideally we should only identify the requisite input data that's actually missing.  Maybe your case is missing SATNUM?  In that case you can insert

SATNUM
N*1 /

with 'N' being the product of the grid dimensions instead.  That is still a little superfluous however, because the upscale_relpermvisc utility will use that definition as a fallback if it's not actually present in the input file.

Please let us know if you get any further if your input file provides porosity, permeability and saturation region information.


Best Regards,

Bård Skaflestad
OPM Developer
SINTEF Digital, Mathematics & Cybernetics Computational Geosciences group

-----Original Message-----
From: Opm <opm-bounces at opm-project.org> On Behalf Of Alun Griffiths
Sent: Thursday, April 28, 2022 11:47 AM
To: opm at opm-project.org
Subject: [Opm] Error with upscale_relpermvisc

Hi

I have created a .grdecl file which was successfully upscaled using upscale_perm.  However, when I use the same file in upscale_relpermvisc, I get the following error message:

Error: Did not find SPECGRID, COORD and ZCORN in ECLIPSE file facies_2a.grdecl

I checked the .grdecl file and it contains these keywords so not sure what the issue is.

Best regards

Alun Griffiths