This book combines a model reduction technique with an efficient parametrization scheme for the
purpose of solving a class of complex and computationally expensive simulation-based problems
involving finite element models. These problems which have a wide range of important
applications in several engineering fields include reliability analysis structural dynamic
simulation sensitivity analysis reliability-based design optimization Bayesian model
validation uncertainty quantification and propagation etc. The solution of this type of
problems requires a large number of dynamic re-analyses. To cope with this difficulty a model
reduction technique known as substructure coupling for dynamic analysis is considered. While
the use of reduced order models alleviates part of the computational effort their repetitive
generation during the simulation processes can be computational expensive due to the
substantial computational overhead that arises at the substructure level. In this regard an
efficient finite element model parametrization scheme is considered. When the division of the
structural model is guided by such a parametrization scheme the generation of a small number
of reduced order models is sufficient to run the large number of dynamic re-analyses. Thus a
drastic reduction in computational effort is achieved without compromising the accuracy of the
results. The capabilities of the developed procedures are demonstrated in a number of
simulation-based problems involving uncertainty.
