Page 391 - Hydrocarbon Exploration and Production Second Edition

378 Risk Analysis for Major Capital Investments in Projects

Best Guess anchored
on preferred
‘base case’

Multiple Multiple
Stochastic Deterministic
Models selected statistically based Models designed manually,
on boundary conditions and based on discrete
parameter inter-relationships. alternative concepts.
May or may not include a base case. Ideally no base case.

Figure 15.14 Alternative approaches to subsurface uncertainty handling.

the result by

Listing the perceived uncertainties in the field. This is done for both the static and
dynamic realms.
Rank the uncertainties. This can be done by assigning a low, medium or high
impact label to each of the uncertainties after a round table discussion involving
the entire subsurface team, or by building some very simple models to determine
the relative impact of each input listed.

If suitable simple models can be built for this initial investigation, the spider
diagram is a useful way of representing the impact of the ranges of uncertainty of the
input parameters. The ones which need to be captured in the risk analysis are those
which move the result furthest from the base case (Figure 15.15).
Alternatively the result of the ranking exercise could be as simple a table listing
the input parameters and flagging them as having high, medium or low impact.
In the base case-dominated approach (Figure 15.16), the models are ‘anchored’
around a favoured interpretation of the raw data, and variations on this model are
created by adjusting the values of those input variables which rank highest in the
exercise above. This method will typically yield a low, base and high case model. Its
advantage is that it is relatively quick to generate, is broader than a single case and is
easily understood. The disadvantage is that it may not be broad enough to capture
the true range of uncertainty, and may yield a false sense of accuracy in the models
derived.
A multiple stochastic approach (Figure 15.17) uses statistical techniques to infill
data between fixed points, or boundary conditions, such as well depths, top reservoir