Summary
This paper presents an integrated method for identifying and inserting
valuable flexibility into major projects. It builds upon recent work that (1)
documents how errors in estimates can bias the selection of design concepts,
(2) shows how concept flexibility can improve the project performance, and (3)
usefully illustrates the probability distribution of outcomes. It involves: (1)
developing and evaluating a base case design, (2) exploring the outcomes this
design might generate, (3) identifying opportunities for flexible design, and
(4) evaluating and selecting the most valuable flexibility to incorporate into
the design. It embodies a paradigmatic change in the way designers deal with
uncertainty: instead of basing a design on fixed assumptions and then testing
its sensitivity to risks, the approach recognizes risks in the design process
and thereby develops valuable flexibility that increases the expected value of
projects. A case study of an oil platform development in the Gulf of Mexico
demonstrates the method.
Introduction
Standard Design Process Leads to Losses. As observed in practice, the
standard approach to the design of major development projects is a deeply
technical process that focuses on the highly complex physical arrangement of
the system. Indeed, the major elements of the platform, the risers, the
tiebacks, and so on can be combined in literally millions of ways, sizes, and
locations at different times. Both to simplify this process, and because
physical design focuses on engineering, the usual approach assumes that major
parameters are known. Most obviously, the design process normally takes the
price of oil as given; in fact, senior management typically requires that
designers across the company all use the same value for oil in their
evaluations. Equally important, the design process typically works with a best
estimate of the original oil in place (OOIP). In short, the standard approach
to design is a technical process that generally does not explore in detail the
consequences of the major variations in two of the prime drivers of value of
any project.
The actual conditions that prevail, however, almost never equal point
forecasts that have been assumed for design, whether one refers to the price of
oil, the amount in the reservoir, the time and cost of construction, etc. The
general rule, as demonstrated by experience in many fields of design, is that
the "forecast is 'always' wrong" as demonstrated by observation of the
discrepancies between prior point estimates and subsequent reality (de
Neufville and Odoni 2003). Note that this fact is not inconsistent with the
notion of a most likely estimate, which represents a balance between the
possible upside and downside values of a parameter. A ‘most likely’ set of
parameters may correctly represent some middle or average value and yet be
unlikely to occur in practice.
© 2009. Society of Petroleum Engineers
View full textPDF
(
941 KB
)
History
- Original manuscript received:
10 November 2008
- Revised manuscript received:
24 December 2008
- Manuscript approved:
3 January 2009
- Published online:
19 February 2009
- Version of record:
1 June 2009
View Cart
