Geoscience Reference
In-Depth Information
0.035
displacement: m
0.03
Chalfont Valley
0.025
0.02
Mammoth Lakes
Kocaeli
0.015
0.01
Trinidad
0.005
0
0
0.5
1
1.5
2
2.5
k
h
/
A
max
Fig. 6.17. Influence of relative earthquake acceleration on wall displacement (after
Kalasin, 2004)
ordertounderstandtheirinteraction.Theresponseischaoticinthesensethattheoutcome
is sensitive to the detail of the input motion and the time history.
6. Conclusions
Whatever the category of modelling that is undertaken, some detailed constitutive model
is required: constitutive modelling plays a key role in defining characteristics of soil
response to be simulated and in establishing the route to extrapolation from model or
fieldobservations.Keycharacteristicsofthemechanicalbehaviourofsoil
elements
under
non-monotonicloadingprovideastronghintastothewayinwhichmacroelementmodels
shouldbeconstructedtodescribegeotechnical
systems
.Evenwiththeserealisticfeatures,
macroelement models are considerably simpler numerically and more rapid to use than
full finite element analyses from which there seems little prospect of escape for study of
complex systems. Laboratory geophysics can provide a means of probing the evolving
elastic anisotropy of the soil, but even in this controlled laboratory situation (or perhaps
becauseofthis)thedynamicsignalsarenoteasytodecode—theycontainmuchrichness
ofinformationaboutthedetailofthepassageofwavesthroughtheparticulatesoilwhich
has not yet been greatly exploited. Nevertheless, the benefit of using such techniques for
anchoring the elastic properties and for revealing information about evolving fabric is
clear.
In the end any form of numerical modelling of static or dynamic geotechnical problems
will need to be validated against observations from physical experiments at some scale.
There is a role for all forms of modelling. Macroelement modelling—provided it can be
calibrated for correct values of controlling parameters—provides a useful and effective
