Model Purpose - Reservoir Model Design - page 4

•Toshowthe senior manager ,orinvestment
fund holder, what the subsurface resource actu-
ally looks like. That oil and gas do not come
from a 'hole in the ground' but from a complex
pore-system requiring significant technical
skills to access and utilise those fluids.
Getting a strong shared understanding of the
subsurface concept tends to generate useful
discussions on risks and uncertainties, and
looking at models or data in 3D often facilitates
this process. The value of visualisation alone is
the improved understanding it gives.
If this is a prime purpose then the model need
not be complex - it depends on the audience. In
many cases, the model is effectively a 3D visual
data base and the steps described in Chaps. 2 , 3 ,
4 , 5 , and 6 of this topic are not (in this case)
required to achieve the desired understanding.

map-based approach, but the industry has now
largely moved to 3D software packages, which is
appropriate given that volumetrics are intrinsi-
cally a 3D property. The tradition of calculating
volumes from 2D maps was a necessary simplifi-
cation, no longer required.
3D mapping to support volumetrics should be
quick, and is ideal for quickly screening
uncertainties for their impact on volumetrics, as
in the case shown in Fig. 1.2 , where the volumet-
ric sensitivity to fluid contact uncertainties is
being tested, as part of a quick asset evaluation.
Models designed for this purpose can be rela-
tively coarse, containing only the outline fault pat-
tern required to define discrete blocks and the gross
layering in which the volumes will be reported. The
reservoir properties involved (e.g. porosity and net-
to-gross) are statistically additive (see Chap. 3 for
further discussion) which means cell sizes can be
large. There is no requirement to run permeability
models and, if this is for quick screening only, it
may be sufficient to run 3D volumes for gross rock
volume only, combining the remaining reservoir
properties on spreadsheets.
Models designed for volumetrics should be
coarse and fast.

1.3 Models for Volumes

Knowing how much oil and gas is down there is
usually one of the first goals of reservoir
modelling. This may be done using a simple

Fig. 1.2 Two models for different fluid contact scenarios built specifically for volumetrics