Optimization Under Uncertainty of Nonlinear Energy Sinks

Abstract

Nonlinear Energy Sinks (NES) are used to passively reduce the amplitude of vibrations. This reduction is made possible by introducing a nonlinearly stiffening behavior in the NES, which might lead to an irreversible transfer of energy between the main system (e.g., a building) and the NES. However, this irreversible transfer, and therefore the efficiency of the NES, is strongly dependent on the design parameters of the NES. In fact, the efficiency of the NES might be so sensitive to changes in design parameters and other factors (e.g., initial conditions) that it is discontinuous, switching from efficiency to inefficiency for a small perturbation of parameters. For this reason, this work introduces a novel technique for the optimization under uncertainty of NES. The approach is based on a support vector machine classifier, which is insensitive to discontinuities and allows one to efficiently propagate uncertainties. This enables one to efficiently solve an optimization under uncertainty problem. The various techniques presented in this paper are applied to an analytical NES example.