In all shelf marine depositional environments, and it is specially true in Carbonates, preserving 3D geological consistency while building reservoir models is the key for a realistic extrapolation of Petrophysics.
Stochastic simulation techniques are usually required for preserving heterogeneity in models, through the simulation of depositional facies prior to the simulation of petrophysics. A cube of facies proportions is usually needed to input in the simulation the 3D drift that describes the depositional model. The assessment of this cube, together with the building of a real stratigraphic grid, is a critical stage in the modeling process, as it drives the quality and the realistic aspect of the model. However, whatever the input data used for its computation, the cube of proportions constitutes up to now one of the most uncertain parameter. As well, the available methologies for gridding produce stratigraphic grids far from reality.
Neptune is a methodology and a tool (gOcad™ plug-in developed in house by TotalFinaElf) which enables the building of a realistic and consistent stratigraphic layering, and the computation of the facies proportions from hard data and geological knowledge.
Neptune is based on the principle that, in shelf marine sedimentary environments, the depositional facies depends on the water depth, the energy of deposit, and the stratigraphic context. Therefore, Neptune provides models of depositional facies constructed on real stratigraphic grids, which reproduce in the whole field the prograding or transgressive features recorded at wells. This is possible thanks to a prior modeling of palaeobathymetry and also of accommodation potential, the easiest parameter to extrapolate.
When this approach is associated with the modeling of diagenetic features that usually overprint depositional facies, it can drive the 3D distribution of petrophysics, and renders very realistic images of reservoir heterogeneity.
Populating 3D reservoir models with Petrophysics in shelf marine sedimentary environments constitutes one of the main challenges in the Geomodeling world.
In such depositional environments, and mainly in carbonate reservoirs, the direct extrapolation of petrophysical properties with respect to geology appears as an unreachable quest.
Seismics usually cannot provide the same help as in clastics for filling 3D grids with net/gross or porosity values. Indeed, in carbonates, this soft information is much more difficult to assess with the same resolution than requested by detailed reservoir models.
As well, the use of stochastic simulations of depositional facies for building reservoir models is much less frequent than in the case of alluvial or turbiditic systems, for -at the least- two main reasons :
the assessment of the proportions cube required for stochastic modeling of depositional facies is far from easy, and the attempt to obtain 3D drifts from hard data only offers the illusion of objectivity. Proportion of facies is a 3D-non stationary random function at any scale, for which the 3D drift is difficult to assess from well sampling.
in carbonates, depositional facies does not constitute the main parameter that drives petrophysics, since diagenesis usually significantly overprints depositional features. Original properties, and mainly permeability, can be strongly modified by early or late diagenesis. Therefore, actual permeability fields result from the combination of several successive geological processes, each of the process having its own spatial organization.
