Seismic risk assessment for oil and gas pipelines - Seismic Risk Analysis and Management of Civil Infrastructure Systems - page 702

public is more accepting of severe consequences for events that are deemed
suffi ciently rare as to be considered 'acts of God'. Thus, there is more
outrage expressed if a pipeline fails under normal operating pressure than
if it fails as a result of a landslide or an extreme fl ood event, even if the
annual likelihood of failure for the three causes is the same.
The diffi culty in conveying risk-based decisions to the public at large is
often ameliorated by an offi cial governmental regulatory process in which
specifi c risk-based acceptance criteria are developed (e.g., ASCE, 2010),
often with the advice and input from experts. That is not to imply that the
regulatory requirements are always based on a rational interpretation or
assessment of risk. As with any governmental regulatory action, the process
can be subverted by individuals and organizations acting in their own best
interests. The benefi t of governmental regulations establishing risk accep-
tance criteria is primarily that the regulations can impose uniformity on
those performing the risk assessments.

25.10.5 Risk-based code provisions

At this time, there are no code requirements on the level of reliability that
is required for the design of oil and gas pipelines. Work sponsored by the
Pipeline Research Council International (PRCI) has led to a set of guide-
lines for implementing a reliability-based design framework with a recom-
mendation with respect to performance goals (Nessim and Zhou, 2005).
This work has been largely incorporated into the non-mandatory Annex O
of CSA Z662. Those referring to Annex O should be aware that the reli-
ability limits contained therein are only recommendations and not require-
ments. Considerable caution should be taken in considering adoption of the
reliability limits in Annex O for several reasons:

• The Annex O reliability targets were developed using the reliability
under normal operating conditions of the population of existing pipe-
lines as a baseline value of acceptable risk. A more appropriate approach
would have been to base non-seismic reliability targets on the minimum
performance allowed in pipeline codes (ASME, 2007).
• The Annex O reliability targets are based upon pipeline diameter,
pressure, and population density and typically require an annual likeli-
hood of pipeline failure in the range of 1/150 000 to 1/200 000 which is
unreasonable compared with other reliability targets where the poten-
tial for death and injury is far greater than what exists for oil and gas
pipelines.

• Development of the reliability targets did not address the difference in

performance expectations for extreme rare events. This is especially
relevant to seismic risk assessment. For example, design requirements