(71ay) An Optimization Formulation for Risk Reduction of the Layout of Offshore Platforms

An Optimization Formulation for Risk
Reduction of the Layout of Offshore Platforms

Joshua Richardson and M. Sam Mannan

Mary Kay O'Connor Process Safety Center, Artie McFerrin Department of
Chemical Engineering, Texas A&M University, 3122 TAMU, College Station, TX
77843

Facility
siting and layout, whether onshore or offshore, is one of the key factors in
the safety of a process operation. Fire, explosion, toxic release, as well as
cost considerations drive the layout of onshore facilities, while offshore
facilities present their own challenges when faced with the same hazards.
Unfortunately in the offshore industry, spacing is many times not an option.
Platforms are relatively small, compact, confined, and congested. Layout is
more difficult and a fire or explosion event on a poorly laid out platform can
cause the loss of utilities, destruction of mitigation systems, and a blockage
of escape routes. This has contributed to two of the worst offshore incidents
in history: Deepwater Horizon and Piper Alpha, where questionable layout
decisions led to escalation of the disasters. However, the layout flaws that
led to the escalation were not necessarily apparent at the time of design. The
need for a more efficient and automated approach to facility siting of offshore
operations is apparent.

A
MINLP optimization model has been created that optimizes the layout of a set of
sections with fixed footprint areas bound to a platform of a given size based
on safety considerations due to fire and explosion. Process parameters are used
to estimate the probability of an event as well as the magnitude of the
possible impact on other sections, which can be weighted in importance in the
objective. The magnitude of the impact is directly dependent on the spacing
between the sections ? a larger spacing will lower the impact compared to a
smaller one. The explosion overpressure and fire radiative flux calculations
are based on impact correlations widely used in industry for layout evaluation.

Explosion
modeling is done using an approximation to the TNO multi-energy method. This
allows for the amount of energy in a section to be specified, as well as a
blast strength level that is determined using TNO guidelines based on the
amount of congestion and confinement in the area. In this way, different
sections may have different magnitudes explosion hazards or no hazard at all.
The explosion overpressure from one section on another is calculated and then
converted into a probit value, which can be converted into a probability of
destruction of a section.

Fire
modeling is used to ascertain the adequacy of layout of both sections and
muster points that the sections are assigned to. Modeling is done using three
different correlations for different fire scenarios: pool fire, fireball, and
jet fire. The process parameters that are specified can be used to calculate
the emission of heat from each scenario and the effect of the heat flux on
points of interest between the section and its muster point which can be
converted to a probability of failure to escape in the case of a fire event.

Two case studies
are presented to demonstrate the applicability of the model: the first a total
layout optimization for the Piper Alpha platform, and the second an assignment
of muster points to sections for an existing layout. It is shown that the model
is able to improve layouts dramatically in the case of a poor layout,
noticeably in the case of an adequate layout, and can be practically used to
assign muster points to sections and determine which route should be taken by
personnel trying to escape. It also implicitly corrects common problems
associated with tradeoffs that are often found in offshore platforms.