Multi-scale uncertainty quantification of composite laminated plate considering random and interval variables with data driven PCE method

Abstract

The paper proposes a multi-scale uncertainty quantification framework of composite laminated plates. The micro and macro uncertainty parameters are represented by random and interval variables according to available uncertainty information. The data-driven Polynomial Chaos Expansion model is constructed, where polynomial coefficients of random variables are determined based on their high-dimensional statistical information, and coefficients of interval variables are calculated using sampling analysis method. The proposed framework is applied for multi-scale buckling analysis, natural frequency calculation and reliability estimation. The comparison results with Monte Carlo Sampling method demonstrate the computational superiority of the proposed algorithm to achieve high accuracy in multiple conditions.