equinor/TimeSeriesAnalysis

Normalization (Unorm) different from one causes incorrect plant-simulations

steinelg opened this issue · 2 comments

it appears that simulations work best if Unorm is forced to one, otherwise if UnitIdentifier itself pick Unorm, it may cause PlantSimulations to miss the data, even though identification indicates a good match.

Unorm different from one is only used in the unit test "I2_Linear_OnlyOneInputIsExcited_EstGainShouldBeZero"
where Unorm is positive infinity for u2, causeing the second gain in essence to be forced to zero, otherwise it is not used.

workaround currently is to just for UnitIdentifier to always use Unorm =1

fixed