Seismic risk assessment for oil and gas pipelines - Seismic Risk Analysis and Management of Civil Infrastructure Systems - page 687

estimates presented in Lee et al. (2000), for a predicted peak ground accel-
eration value of 0.5 g , the actual peak ground acceleration would lie within
0.4 g to 1.0 g , 50% of the time. The actual variation would be larger if vari-
ability other than that associated with attenuation were considered. From
this simple example, it should be clear that defi ning levels of ground shaking
to the nearest 0.05 g is adequate for most risk assessment applications.
Similarly, ground displacement hazards should not be defi ned with artifi cial
precision.
25.6 Determiningseverityofhazard
The assessment of site-specifi c geotechnical hazards forms an important
part in overall vulnerability assessment of a pipeline system and facilitates
the development of potential mitigation measures. Permanent ground
deformations triggered by earthquakes have been recognized as one of the
major causes of system damage and associated service disruption to life-
line facilities. For example, signifi cantly large permanent lateral deforma-
tions are expected to occur even under gently sloping ground conditions in
areas of liquefi able soil. Therefore, an estimation of the extent of such
ground displacements is important in the assessment of pipeline system
vulnerability. In particular, adequate knowledge of site-specifi c soil and
groundwater conditions is critical to the success of the design and installa-
tion of pipelines, as well as in predicting its anticipated performance under
fi eld conditions.
The methods available for the computation of earthquake-induced per-
manent lateral ground displacements can be broadly classifi ed into: (a)
empirical approaches developed based on measured displacements (e.g.,
Youd et al. , 2002); and (b) mechanistic approaches which rely more on the
principles of engineering mechanics (e.g., Byrne et al. , 2004). Mechanistic
methods mostly involve fi nite element or fi nite difference analyses, which
are more appropriate for detailed site-specifi c analyses requiring a greater
level of confi dence. The estimation of earthquake-induced ground deforma-
tions, particularly from the viewpoint of regional assessments, still relies
heavily on empirical correlations. It is important to note that both empirical
and mechanistic approaches (or any combined approaches thereof) would
require the selection of earthquake magnitude and epicentral distance that
is consistent with the probabilistic seismic hazard considered in the design.
Except under high ground shaking intensity levels (i.e. in excess of peak
ground acceleration of about 0.4 g ), propagation of seismic waves through
soft soils generally causes the ground motions to amplify as they reach the
ground surface. This aspect is important and should be accounted in the
liquefaction assessment and estimation of ground deformations in soft/
loose soil zones.