OPM
[testing] to see how often we get negative values
There are fours cases got hit by the assertion,
mpi.norne_pvt mpi.compareECLFiles_flow+5_SWATINIT_MODEL2 mpi.compareECLFiles_flow+MIN_QOIL_1 mpi.compareECLFiles_flow+MAX_WATERCUT_4
and three of them are all related to invMugBg, and invMuoBo.
Maybe the invMugBg and invMuoBo are one of the tables that we should patch.
In the test_norne_pvt.cpp,
std::vector<double> mu_expected = {0.00120767750, 0.00129077352,
0.00111063039, 0.00119627038,
0.00103118116, 0.00110633521,
0.00094413471, 0.00100998373,
0.00090320931, 0.00096374536,
0.00086714481, 0.00092142974,
0.00081811098, 0.00086735227,
0.00077704364, 0.00082229010,
0.00070975205, 0.00076029164,
0.00065679329, 0.00071124175,
0.00061496175, 0.00067213642,
0.00058000381, 0.00064115346,
0.00055124739, 0.00061633274,
0.00052840888, 0.00059781928,
0.00050926184, 0.00058323394,
0.00049295739, 0.00056996321,
0.00048026810, 0.00056474486,
0.00047088998, 0.00056427878,
0.00047649659, 0.00060774836,
0.00048006188, 0.00059909192,
0.00026623648, 0.00060915386,
0.00025670489, 0.00062157315,
0.00024760210, 0.00064290735,
0.00023889979, 0.00067946283,
0.00022330662, 0.00077837223,
0.01142273040, **-0.00351292519**, **-0.00129867195**};
The test design that actually some viscosity should be negative. This is unfortunate.
testing purpose is completed.
We have issue when extrapolating the oil table for high pressure, mostly due to the invbmu tables, for both oil and gas, saturated and unsaturated.
A PR will try to pad the invBmu tables for high pressure to avoid negative values generated during extrapolation with high pressure.