Seismic risk analysis of wind turbine support structures - Seismic Risk Analysis and Management of Civil Infrastructure Systems - page 711

26
Seismic risk analysis of wind turbine
support structures
M. MARDFEKRI, Texas A&M University,
USA and P. GARDONI, University of
Illinois at Urbana-Champaign, USA
DOI : 10.1533/9780857098986.4.716
Abstract : Probabilistic models are developed to predict deformation,
shear, and moment demands on wind turbine support structures subject
to seismic excitations, and environmental (wind, wave, and current) and
operational loadings. The probabilistic models are formulated starting
from existing deterministic models and developing correction terms that
capture the inherent bias. The correction terms and a model error are
assessed using data obtained from detailed three-dimensional nonlinear
fi nite element analyses of a set of wind turbine systems that consider
different design parameters and account for the dynamic soil-structure
interaction. The proposed probabilistic seismic demand models provide
unbiased predictions of the seismic demand on support structures and
properly account for the underlying uncertainties. The developed
demand models are used to compute fragility estimates of an example
support structure defi ned as the conditional probability of not meeting
specifi ed capacity levels.
Key words : Bayesian inference, offshore wind turbines, demand models,
experimental data, fragility, probabilistic models.
26.1 Introduction
Extensive installation of offshore wind farms for electricity production in
moderate to high seismic regions in the United States and other countries
has raised a new concern about the safety of wind turbine support struc-
tures subject to seismic loads (Prowell and Veers, 2009). The demands on
the support structure of wind turbines are typically determined using com-
putational models. Computational analyses require accurate modeling of
the structural dynamic response of the support structure as well as evalua-
tion and modeling of the environmental loading including wind, and for
offshore wind turbines, wave and current loading. Several aeroelastic simu-
lation codes are used in the industry to simulate fatigue, aerodynamics,
structural dynamic response, and turbulence. They are mostly used to
predict the extreme and fatigue loads of wind turbines. For example, FAST
(Jonkman and Buhl, 2005) and ADAMS (Laino and Hansen, 2001) are two
716