Gaussian process assisted stochastic dynamic analysis with applications to near-periodic structures

This paper characterizes the stochastic dynamic response of periodic structures by accounting for manufacturing variabilities. Manufacturing variabilities are simulated through a probabilistic description structural material and geometric properties. The underlying uncertainty propagation problem has been efficiently carried out functional decomposition in space with help Gaussian Process (GP) meta-modelling. is performed projected onto eigenspace involves nominal number actual physics-based function evaluations (the eigenvalue analysis). allows evaluation to be solved low computational cost. Two numerical examples, namely an analytical model damped mechanical chain finite-element multiple beam-mass systems, undertaken. key findings from results that proposed GP based approximation scheme capable (i) capturing systems well-separated modes presence high levels uncertainties (up 20%), (ii) adequately sets identical 5–10% uncertainty. validated Monte Carlo simulations.