Stochastic design of double-skin façades as seismic vibration absorbers

Double-skin façades (DSFs), usually designed for aesthetic reasons or energy saving purposes, have been recently proposed as an alternative mean of passive control, able to reduce the effects of dynamic actions on building structures. This paper presents a novel approach to optimise their design as distributed vibration absorbers (VAs). Four key design parameters have been chosen to representing the façade, namely its flexural stiffness and viscous damping ratio, along with the stiffness of the elements connecting the façade to the primary structure. The optimisation is achieved by minimising the variance of the building’s dynamic response, conveniently computed in a stochastic framework. Solutions are obtained using genetic algorithms (GAs), including nonlinear constraints limiting the relative displacements between primary and secondary structures. Computational efficiency of the optimisation procedure is largely improved, compared to previous works, by characterising the seismic action as a stationary random process, fully defined by some closed-form analytical expressions for the power spectral density (PSD) function consistent with target response spectra. Four configurations of double-skin façades have been analysed, including single and multi-panel layouts, spanning one to six storeys, and their efficiency has been quantified. Results are compared with those obtained directly in the time domain by numerical integration of the equations of seismic motion for a suite of recorded accelerograms, showing a good level of consistency.