Possible inconsistency in region numbering

[olwijn]

The number of regions, for example equilibration regions (EQLNUM), recognized by FlowNet is based on the region ID's that are mapped to the flownet cells from the corresponding ECLIPSE grid. In that process it is possible that some regions are lost (i.e. they are present in the ECLIPSE model, but there are no cells in FlowNet with that region number). The region table dimensions are based on unique elements that are defined in FlowNet so this number could be smaller than region ID's used in the parameterization when using the regions_from_sim option. For example, parameters may be defined for up to 20 regions, but if for example no FlowNet cells have region ID 5, only 19 regions will defined in the Flow deck, and Flow will throw an error when it encounters cells with region ID 20.
If and which regions ID will dissappear during the mapping process cannot be predicted on forehand since it depends on the distribution of nodes, which in turn depends in a complex way on various configuration settings.

The dimensioning of the region tables in FlowNet should therefore ideally be based on the total number of regions appearing in the orginal model, and not only on the regions that are mapped to the FlowNet cells. This number can be extracted automatically, or it could be provided explicitly as input in the config file.