Assessing and managing the risk of earthquake-induced liquefaction to civil infrastructure - Seismic Risk Analysis and Management of Civil Infrastructure Systems - page 120

Once the potential for liquefaction hazard to occur has been assessed, a
key stage of the risk assessment framework shown in Fig. 5.2 is to quantify
the hazard, e.g. the amount of potential permanent ground deformation
(PGD) that could occur in an event. The principal modes of PGD due to
liquefaction are lateral, either fl ow failure or lateral spreading, or vertical,
due to either bearing capacity failure or volumetric strain-induced settle-
ments. The extent of PGD and its mode, will dictate the response of infra-
structure and hence the impact side of the risk equation. If ground
deformations are not expected to cause damage, then the risk is negligible
and there is no need for further assessment.
The evaluation of liquefaction-induced ground deformations, both verti-
cal and horizontal, is complex. The amount of movements can be estimated
using empirical relationships or numerical modelling. The empirical rela-
tionships, although widely used, have a signifi cant degree of uncertainty,
which is not always rigorously considered.
Uncertainty associated with the assessment of PGD arises from the fol-
lowing sources:

• the level of earthquake hazard (which itself is correctly described
probabilistically);
• the likelihood of liquefaction triggering as discussed in Section 5.2.1;
• the calculation methodology used (see following sub-sections);
• the in-situ soil parameters (e.g. penetration resistance), which are often
based on limited data and variable ground conditions.

5.3.2 Lateral spreading

Lateral spreading results from a combination of static disturbing forces
on a slope, additional inertial forces due to an earthquake, and reduced
strength of the liquefi ed material. According to Seed et al. (2003) there
are three main approaches to estimate liquefaction induced lateral
deformations:

• Numerical methods (fi nite difference/fi nite element, see Section 5.2.4)

employing advanced constitutive models (e.g. Valsamis et al. , 2010). The
majority of the existing publications though focus on simulation of cen-
trifuge test results.

• Statistically derived empirical methods based on back-analyses of earth-

quake case histories. Rauch & Martin (2000), Youd et al. (2002) and
Zhang & Zhao (2005) rely on magnitude and epicentral distance for the
prediction of ground surface displacements, while Shamoto et al. (1998),
Hamada (1999), Zhang et al. (2004) and Aydan et al. (2005) employ
either the peak ground acceleration or velocity (PGA/PGV) as an inten-
sity measure for the lateral spreading prediction.