some polymer test cases seem to be broken

[andlaus]

This might also be a problem that is located between my ears, but if I try to run the polymer cases with shear thinning or thickening with flow, I get the following:

IOConfig: Reading GRIDFILE keyword from GRID section: Output of GRID file is not supported. Supported format: EGRID
**********************************************************************
*                                                                    *
*                 This is flow (version 2017.10-pre)                 *
*                                                                    *
* Flow is a simulator for fully implicit three-phase black-oil flow, *
*            and is part of OPM. For more information see:           *
*                       http://opm-project.org                       *
*                                                                    *
**********************************************************************


***************Saturation Functions Diagnostics***************
System:  Black-oil system.
Relative permeability input format: Saturation Family I.
Number of saturation regions: 1



================ Starting main simulation loop ===============


Error: [/home/and/src/opm-simulators/opm/autodiff/WellInterface_impl.hpp:717] Cell with i,j,k indices 12 9 3 not found in grid (well = INJE01)

Error: Program threw an exception: [/home/and/src/opm-simulators/opm/autodiff/WellInterface_impl.hpp:717] Cell with i,j,k indices 12 9 3 not found in grid (well = INJE01)

I assume that this means there's something wrong with the well specification of case-shear-thickening/3D_POLYMER.DATA!? (note: I compiled this in debug mode.)

[pgdr]

For what it's worth, ecl reports that the EGRID file contains said coordinates:

from ecl.ecl import EclGrid
basepath = 'polymer_test_suite/3D_plyshlog/case-shear-thickening/'
path = basepath + 'opm-simulation-reference/' # or eclipse
g = EclGrid(path + '3D_POLYMER.EGRID')
i,j,k=12,9,3  # I don't know if i,j,k are 0 or 1 or -1 indexed
g.get_node_pos(i+1,j+1,k+1)
#>>> (433.3333435058594, 375.0, 2096.109619140625)
g.get_node_pos(i,j,k)
#>>> (400.0, 337.5, 2051.396240234375)
g.get_node_pos(i-1,j-1,k-1)
#>>> (366.6666564941406, 300.0, 2026.7259521484375)

[GitPaean]

The legacy flow_polymer is still working okay, that part of the code should be same unless changed when migrating from the old one to ebos.
It might be related to something related to grid. I am a little too busy to look at this. Hopefully someone can take a look at it.

[totto82]

@andlaus Could you try it again? It work both in debug and in release mode for me.

[andlaus]

now it works on my machine as well. it's quite possible that this was a false alarm anyway. closing

[GitPaean]

I reproduced the issue, while it looks like fixed from somewhere.