Probing the multi-physical probabilistic dynamics of a novel functional class of hybrid composite shells

Functionally graded materials, sandwich structures and composite laminates have been widely investigated individually in last few decades for their respective set of advantages over conventional monolithic structures. However, the tremendous recent advances in manufacturing processes have opened up new frontiers of research specifically to form hybrid structures for multifunctional applications, where the advantages of each of the constituting components could potentially be exploited in a single structure. Such complex hybrid structures are often susceptible to manufacturing uncertainties and different forms of variability. To characterize the stochastic dynamic behaviour of these hybrid structural forms in a comprehensive, practically-relevant, yet efficient framework, here we present a multi-physical probabilistic vibration analysis based on Gaussian Process Regression (GPR) assisted finite element (FE) approach coupled with Monte Carlo Simulation. Integration of the GPR based machine learning model with the physics based simulation model leads to a significant level of computational efficiency in the quantification of system uncertainty. In the stochastic approach for various hybrid structural configurations, the compound effects of source-uncertainties are considered in order to assess the effect of different critical multi-physical parameters systematically involving material, geometric and physical aspects. The numerical results categorically indicate that the source-uncertainty of hybrid shells with different geometries has a significant effect on dynamic behaviour of the structure, which makes it imperative to take into account such probabilistic deviations to ensure adequate operational safety and serviceability.