Uncertainty quantification and global sensitivity analysis for composite cylinder shell via data-driven polynomial chaos expansion

The mechanical properties of composite material exhibit inherent variation with uncertainty. Uncertainties in material properties propagate and result in uncertainties of mechanical performance of structure made of composite material. Polynomial chaos expansion (PCE) is implemented to carry out uncertainty quantification (UQ) and global sensitivity analysis (GSA) of cylinder shell made of composite material for this paper. A case study concerning eigenvalue buckling load of composite cylinder shell is investigated. Design of experiment (DOE) is conducted by utilizing Latin hypercubic sampling. Then data-driven PCE is established and later validated. Statistical moments (mean and standard deviation) and Sobol sensitivity indices of eigenvalue buckling load are obtained respectively. It is found that the PCE can serve as an efficient approach to handle UQ and GSA in engineering applications.