Aeroelastic Analysis of Laminated Composite Plate with Material Uncertainty

Abstract

The involvement of complex manufacturing and fabrication process of composite material has been the main cause of uncertainty. Therefore, the stochastic analysis of flutter characteristics of laminated composite structures is essential. This research work is focused on two different approaches of stochastic analysis of flutter characteristics. First is the perturbation technique, which used the Hessian and gradient of quadratic response surface, optimized based on response surface methodology. The second is polynomial neural network (PNN)-based uncertainty model, which is apparently attempted first in this study. PNN is a group method of data handling–based algorithm developed to relate the input material properties and flutter response. Further, the reliability study is carried out using first-order reliability method (FORM). The outcome of PNN and perturbation technique is compared with Monte Carlo simulation (MCS) and FORM. Perturbation technique is found to be highly efficient and provide results with great accuracy compared with MCS results. Although the PNN-based model showed better computational efficiency compared with MCS, it is not found to be beneficial over perturbation technique. Also, the sensitivity analysis is carried out and the material properties, which greatly affect the flutter characteristics, are identified.