Geoscience Reference
In-Depth Information
5. Conclusions
Thestrategicroleofundergroundstructuresandburiedpipelinenetworks,andtheircriti-
calresponseduringstrongearthquakes,requiremoreadequatetoolsforthedetermination
of design loads and to support seismic riskassessment analyses.
In the first part of this paper, we have addressed the recent developments made to derive
soundandwellestablishedformulasfortransientgroundstrainevaluationasafunctionof
peak parameters of earthquake ground motion. For this purpose, recent transient ground
strain evaluations have been introduced, obtained by a suitable spatial interpolation of
thedisplacement wavefieldrecordedbydenseseismicnetworksduringearthquakeswith
different magnitude and epicentral distance. These data support recent research works
on this subject, and highlight that transient ground strains are not only affected by wave
passage effects, as generally assumed in formulas for design, but also site effects, espe-
cially in terms of lateral variations of soil properties, and spatial incoherency of ground
motion.Thelattereffectsplayanimportantrole,atleastcomparabletothewavepassage.
Further experimental studies on dense seismic arrays and numerical parametric investi-
gations are recommended, especially with reference to: (i) the dependence on azimuth
of the peak ground strains, that was found to affect results by at least a factor of two
(see Figure 18.1); (ii) the quantification of the ratio between peak values of shear strain
vs. axial strain, as a function of magnitude, depth, and ground conditions; (iii) the role
of non-linearity in the
PGS-PGV
relationships, and that of the spatial incoherency of
earthquake ground motion as well.
The previous studies will also provide material for production of more reliable hazard
maps in terms of peak ground strains and their application to seismic risk assessment of
underground structures: at present, these are affected by a significant degree of uncer-
tainty and limitations, extensively discussed in this paper. In spite of such uncertainties,
the seismic risk assessment studies presented in this paper, mainly based on 1D linear
equivalentwavepropagationanalyses,seemtoberobustandingoodagreementwiththe
results shown in the first part of this paper, at least for peak ground strain evaluations.
PGV
and/or
PGS
basedvulnerabilitycurvestoestimatetheexpecteddamagesforburied
pipelines,besidestheirempiricalnature,arereasonablyaccuratewhenextensiveandwell
focused site effect studies are available and ground conditions and surface geology are
also well known. Hazard and risk map resulting from the application of
PGV
and
PGS
values may be used for the risk management of lifeline systems. Vulnerability functions
using
PGS
are very promising but they need advanced studies to reduce uncertainties,
whiletheempiricaldatacorrelatingstrainswith
RR
/kmarestillverylimitedandthusthe
reliability of the proposed function is rather lowfor the moment.
Acknowledgments
ThisworkwaspartiallyfundedbytheEuropeanCommissionwithintheLesslossProject
(GOCE-CT-2003-505488). The first author (RP) acknowledges funding from the DPC-
RELUIS National Research Project No.6 Sub-Project “Costruzioni in sotterraneo—
