Page 291 - Design of Simple and Robust Process Plants

7.7 Site Vulnerability 277
though strictly speaking it is a raw material. The load-shedding systems need
to be included with its shortage levels. The frequency of load shedding at the
different defined shortage levels must be determined based on the reliability
of the utility system. The selected process plants to be switched off when load
shedding levels becomes activated must be known These supplies have a
major impact on site vulnerability, and therefore require an optimal balanced
but high availability and reliability.
Factors playing a role in the achievement of optimal (but high) availability and relia-
bility for common (utilities) supplies include the following:
. Availability
± Reliable design
± Inventory to cope with temporary shortage
± Back-up to cope with longer shortage times, or when inventory is not practi-
cal, as for power and gases such as hydrogen
± Load shedding to minimize the impact
. Reliability
± Reliable components
± Reliable supplies
± Redundancy
. Common mode failure
± Independent systems
± Different designs
± Independent external back-up
± Inventory
. Emergency provisions
± Independent system (emergency generator, diesel driven firewater pumps)
± Inventory (firewater pond, oil reservoir, batteries)

All the above factors require careful design as they have a major impact on site vul-
nerability. In the previous chapter reliability engineering techniques were discussed
which resulted in availability and reliability process data. For (utility) supplies, com-
mon cause failures have an additional importance in the prevention of outages.
The Center for Process Safety (CCPS) in Chemical Process Quantitative Risk
Analysis (CPQRA) book (1989) mentions that: ª¼ common cause failures events
tend to dominate system unavailability in those applications where redundancy is
used to improve system reliability performanceº. Therefore, the reliability calcula-
tion method needs to be adapted to quantify its impact. A methodology has been
developed reported in CPQRA to perform a common cause failure analysis to sup-
port designs. The basic elements of this are:

± Identification of common causes for events.
± Building of common cause event trees.
± Qualitative screening for the selection of dominant contributors to system
unavailability.
± Quantitative analysis by common cause event modeling.