Geology Reference
In-Depth Information
1. INTRODUCTION
systems (Hung et al. 2003). Among them, the
nonparametric SI approach is effective for the
complex nonlinear problems of large civil infra-
structures, in particular, one of the nonparametric
nonlinear SI methodologies that have been widely
used in the field of large civil structures is neural
network (NN) because it is more readily useful
than the parametric SI approach to identify in-
complete and incoherent measurements of large
civil structures, in general (Smith & Chase 1993;
Masri et al. 2000; Hung et al. 2003), although
conventional NN models have drawbacks of the
slow convergence rate and the potential to local
minima due to the characteristics of the black-box
model (Chassiakos & Masri 1996). On the other
hand, another popular nonparametric SI method
for modeling complex nonlinear dynamic systems
is the fuzzy logic theory because it is effective to
represent complex nonlinearities and uncertainties
of dynamic systems in a more transparent way
(Langari 1999). Since Zadeh's paper(1965), the
fuzzy logic has been widely applied to various SI
problems (Wang & Langari 1995). In particular,
there have been a number of studies on the TS
fuzzy model in recent years, which provides an
effective representation of nonlinear systems
with the aid of fuzzy sets, fuzzy rules, and a set
of local linear models (Filev 1991; Du & Zhang
2008; Abonyi et al. 2000; Wang & Langari 1996;
Johansen &Babuska 2003; Takagi & Sugeno 1985;
Chen et al. 2007). On the other hand, the fuzzy
logic theory in the field of large scale civil infra-
structures has been mainly used for nonlinear fuzzy
control system design (Tani et al. 1998; Wang &
Lee 2002; Ahlawat & Ramaswamy 2004; Dounis
et al. 2007; Loh et al. 2003; Shook et al. 2008;
Casciati 1997; Yan & Zhou 2006; Choi &Schurter
& Roschke 2001; Zhou et al. 2003; Faravelli &
Rossi 2002; Al-Dawod et al. 2004; Battaini et
al. 2004; Symans & Kelly 1999; Subramaniam
et al. 1996; Kim et al. 2004; Pourzeynali et al.
2007; Kim et al. 2009; Nomura et al. 2007; Gu &
The development of an accurate explicit math-
ematical model of dynamical systems is one
of the most important tasks in structural health
monitoring and control system design for hazard
prediction and mitigation of dynamical systems
because precise mathematical information related
to the dynamic systems is used for damage predic-
tion and/or calculation of control forces (Kerber
et al. 2007; Yen & Langari 1998; Lin et al. 2001;
Bani-Hani 1999; Kim et al. 2011). However, it is
still challenging to derive a mathematical model
of nonlinear dynamic systems (Moon & Aktan
2006). An example of such nonlinear dynamic
systems can be sought when highly nonlinear
hysteretic actuators/dampers are applied to struc-
tural systems for efficient energy dissipation: the
structure integrated with the nonlinear control
devices behaves nonlinearly although the structure
itself is usually assumed to remain linear (Yi et al.
2001; Ramallo et al. 2004). Because the structure
integrated with the nonlinear hysteretic control
device is intrinsically nonlinear, it is challenging
to develop an appropriate mathematical model
for the integrated nonlinear system including the
interaction effects between the structural system
and the nonlinear control device while it plays a
key role in both structural health monitoring and
controlling system (Dyke et al. 1998).
Such a challenging nonlinear problem has
made a number of researchers pay a great deal of
attention to system identification (SI) approaches
in recent years (Adeli & Jiang 2006). The non-
linear SI methodologies can be categorized into
two parts: parametric and nonparametric SI ap-
proaches (Bani-Hani et al. 1999). A parametric SI
method is to directly identify physical quantities
such as the stiffness and damping of the structural
systems (Lin et al. 2001; Lin &Betti 2004; Yang
& Lin 2004); while a nonparametric SI method
is to train the input-output map of the structural
Search WWH ::
Custom Search