Abstract: Due to the large number of design variables that can be present in
complex systems incorporating visco-elastic damping, this work examines the
application of genetic algorithms in optimizing the response of these
structures. To demonstrate the applicability of genetic algorithms (GAs), the
approach is applied to a simple viscoelastically damped constrained-layer beam.
To that end, a finite element model (FEM) derived by Zapfe, which was based on
Rao's formulation, was used for a beam with constrained-layer damping. Then, a
genetic algorithm is applied to simultaneously determine the thicknesses of the
viscoelastic damping layer and the constraining layer that provide the best
response. While the targeted response is ultimately at the discretion of the
designer, a few different choices for the fitness function are shown along with
their corresponding impact on the vibratory response. By integrating the FEM
code within the GA routine, it is easier to include the frequency-dependence of
both the shear modulus and the loss factors for the viscoelastic layer.
Examples are provided to demonstrate the capabilities of the method. It is
shown that while a multi-mode optimization target provides significant
reductions, the response for that configuration is inferior to the response
when only single-mode reduction is considered. The results also reveal that the
optimum configuration has a lower response level than when a thick layer of
damping material is used. By demonstrating the applicability of GA for a simple
beam structure, the approach can be extended to more complex damped
structures.
