Stochastic buckling quantification of laminated composite plates using cell-based smoothed finite elements

In this work, uncertainty in fiber angles, laminate dimensions, and constituent material properties of composite laminates is investigated to quantify their effects on the buckling load of different laminated composite plates, consisting of four composite plies. For stochastic buckling quantification, typical fourteen scenarios of randomness are investigated. The convergence study helps to determine the effective number of pseudorandom Monte Carlo samples, that are small compared to conventional ten-thousand samples showing the effectiveness of the Monte Carlo simulation to the present problem. The triangular cell-based smoothed discrete shear gap method based on the first-order shear deformation theory is employed to accurately describe the buckling behavior of laminated composite plates. The method shows significant advantages in reducing the computational resources due to its fast convergence rate. The strain gradient tensors are further smoothed from three partial triangles, which helps to avoid the shear locking phenomenon and makes the obtained results less sensitive to mesh distortion even with coarse mesh. Deterministic results are validated with available results in the scientific literature. The stochastic buckling sensitivity and probability are thoroughly examined for both cross-ply and angle-ply laminated composite plates. The accuracy of stochastic results is observed by its convergence to and distribution around the converged values of Monte Carlo simulations. Non-unique buckling modes exist with increasing probability due to instant switches among buckling modes when the degree of stochasticity increases.