Page 105 - Numerical Analysis and Modelling in Geomechanics

86 D.S.JENG
strange variance.

Wave-seabed-caisson interaction
Considerable efforts are devoted to the protection of marine structures (such as
caissons, seawalls, etc.). The major reason is that marine structures such as
caissons and seawalls are commonly adopted for coastal defence. Recently,
caissons have also been used more widely as foundation elements for offshore
structures. Although the protection of caisson-type breakwaters has been
extensively studied in recent years, the understanding of their interaction with
waves and the seabed is far from complete.
Some investigations into wave-induced pore pressure in the vicinity of a
caisson have been carried out through analytical and numerical approaches
(Mynett and Mei, 1982; Tsai et al., 1990; Mase et al., 1994; Mizutani
andMostafa, 1998; Jeng et al., 2000, 2001a). Among these, Mynett and Mei
(1982) proposed a boundary-layer approximation for a rectangular caisson
located on an isotropic seabed without a rubble mound base. Later, Tsai et al.
(1990) extended the model to a seabed of finite thickness, and also including
three different mechanisms of interaction between waves, caisson and seabed.
However, it has been demonstrated that boundary-layer approximation is only
suitable for finer sandy beds, not for coarser materials (Hsu and Jeng, 1994).
Numerical modelling has been widely used for the wave-seabed-caisson
interaction problem, due to the complicated configuration of such a problem,
which is difficult to handle by analytical approximation. Mase et al. (1994)
proposed a finite element model to investigate the wave-induced seabed response
around a composite breakwater, including a rectangular caisson and rubble
mound base. In their model, the lateral boundary conditions were directly given
by the analytical solution proposed by Yamamoto (1977), and they treated the
derivative terms with respect to time by finite difference methods. The initial
values of the pore pressure and soil displacements were assumed to be zero in
their model. This may not provide a solution that accurately performs the
oscillatory fluctuation in the initial stage of the time series. Furthermore, they
considered the rubble mound base as gravels, but they took the values of
−1
−2
permeability (K) as 10 m/sec and 10 m/sec. In fact, the common value of the
−1
−2
permeability of gravel is 10 m/sec, while 10 m/sec is the common value of
coarse sand. Mase et al. (1994) did not discuss in detail the effects of wave and
soil characteristics on the wave-induced soil response near to the composite
breakwater. Furthermore, the lateral boundary conditions they used were only a
simple isotropic solution, not a general solution for more complicated soil
behaviour.
Later, Mizutani and Mostafa (1998) developed a combined boundary element
model and finite element model to investigate the wave-seabed-caisson
interaction in an isotropic homogeneous seabed. In their model, the wave field
and porous seabed are coupled in a combined numerical model. However, their