Databases Reference
In-Depth Information
FIGURE 9.12: An iteratively refined adaptive finite element discretization
showing the distribution of the horizontal loading. (This figure is a copyright
of and reproduced with permission of Civil-Comp Ltd. Previously published
in [98].)
9.13. Note that each sub-surface is assigned to a specific processing node in
the grid.
Once the mesh is generated from the initial structural pattern obtained
from the GN pattern matching scheme, the next procedure in our integrated
grid-sensor network framework is the pattern recognition process involving
the meshes (both normal and damaged structural patterns).
9.2.2.3
DPR Scheme for Damage Detection Using Mesh Repre-
sentation
To provide a continuous measure in structural analysis, design, and moni-
toring, we propose a second-level recognition for the structural pattern. At this
level, the recognition process involves the recognition of a damaged structural
mesh representation using the stored meshes of normal structural conditions.
The meshes generated from the previous stage of the framework are treated
as images, and a distributed image recognition scheme, such as the DHGN, is
deployed.
Using the sub-domain meshes generated from the parallel adaptive finite
element mesh generation, we compare the meshes obtained from the WSN
network with the pre-stored meshes in the computational grid network. Fig-
ure 9.14 shows a simple mesh recognition process using the DHGN approach.
