companies are utilising stochastic techniques for
facies and petrophysical property modelling, as these
produce more realistic variability. This means that the
modelled distribution of properties and resulting
volumetrics represent one possible outcome within a
larger outcome space. Hence it will be necessary to run
multiple realisations of the geomodel, to be able to
visualise the outcome space and to run volumetric
calculations on all of these model realisations.
Often tens to several hundreds of stochastic model
realisations are generated. For computer capacity
reasons, it is rare that all realisations are taken further
to flow simulation models, hence a ranking of
realisations is required. Picking representative model
Fig. 22.14 Example from an offshore Norwegian field model, with facies realisations (left) which are input to porosity realisations
(right). Both the facies and the petrophysical modelling within each facies are based on stochastic methods
574
J.C. Rivenæs et al.