A flutter reliability optimization approach for aerospace composite structures based on adaptive ensemble model

Uncertain factors generally exist in aeroelasticity systems, and ignoring their impacts can potentially result in unexpected flutter failures. Additionally, the computational cost of integrating flutter reliability with optimization is significant, as it requires a large number of expensive model evaluations to estimate the failure probability for each distribution parameter. In this paper, a new decoupled flutter reliability optimization method based on adaptive ensemble model is proposed, which fully leverages the advantages of each surrogate model and no additional original model evaluation is required. Firstly, flutter modelling is presented for supersonic composite plate embedded in Shape Memory Alloys (SMA). Secondly, an ensemble model is proposed to estimate the Failure Probability Function (FPF) with enhancing accuracy and efficiency by assigning specific weights to each individual model. The flutter reliability optimization is then decoupled using the FPF. Finally, a highly nonlinear function is employed to demonstrate the validity and computational efficiency of the proposed method compared to DLMCRO, DROAK and DROAPCK method. Two numerical applications including composite plate with SMA and wing model with engine considering the reliability and deterministic optimization are discussed.