Assessing and managing the risk of earthquake-induced liquefaction to civil infrastructure - Seismic Risk Analysis and Management of Civil Infrastructure Systems - page 132

Design criteria for any project should have defi ned levels of risk toler-
ance for different types of risk. Risks associated with liquefaction are no
different from any other sources of risk, whether they relate to other natural
hazards, or man-made events. All project risks are assessed relative to the
tolerable cost, delay or safety risks, and where the risks are beyond the
project tolerance, risk mitigation or reduction will be required. In assessing
the tolerability of liquefaction risks, the overall performance requirements
for the project in the event of an earthquake should be considered. Seismic
design codes clearly defi ne acceptable performance under different levels
of ground shaking. Where life safety (i.e. a structure may experience signifi -
cant damage, and may not be functional after an earthquake, but it will not
collapse and will not present a risk to life) is the performance requirement
for ground shaking with a 10% probability of exceedance in 50 years, sig-
nifi cant damage due to liquefaction should also be acceptable, provided that
life safety can be demonstrated. Uniform risk due to various potential
hazards should be the aim of a risk-based approach to design.
5.6 Conclusions
Liquefaction is widely understood to have the potential to cause signifi cant
damage, delay and disruption to civil infrastructure. Methods for the assess-
ment of liquefaction potential and of the resulting PGD have been pre-
sented in this chapter. As with any natural hazard, earthquake-induced
liquefaction must be treated as a risk item in project planning and design.
If the occurrence of liquefaction was certain, then the design approach
would be more straightforward. However, as has been discussed in this
chapter, the occurrence and the impact of liquefaction is an extremely
uncertain area of geotechnical earthquake engineering for the following
reasons:
• The occurrence of the design earthquake in the fi rst place is uncertain,
and must be considered probabilistically.
• The ground shaking induced by the design earthquake, although typi-
cally defi ned in terms of a single acceleration or response spectrum
should be defi ned in terms of a probability distribution.
• The potential for liquefaction to occur is uncertain in terms of hazard,
in-situ properties of the soil and methodology itself. Each of these areas
can be treated as certain but clearly the results are then a single point
estimate where there is actually signifi cant variance around this value.
• The permanent ground deformation, either lateral or vertical, which
results from liquefaction, is the damaging hazard, not merely the occur-
rence of liquefaction. A similar set of uncertainties exist for the assess-
ment of PGD, in terms of input data and methodologies.