Optimizing stiffener configuration and buckling performance of composite panels under in-plane shear: A hybrid ANN-FEA approach

Abstract

Predicting buckling loads and optimizing stiffener configurations for composite stiffened panels present significant challenges due to nonlinear behavior and the computational demands of iterative simulations by traditional finite element analysis (FEA). This study addresses these challenges by integrating artificial neural networks (ANNs) with FEA to develop an efficient and accurate predictive framework. An in-plane shear load experiment was designed and conducted to validate the combined ANN-FEA model, which was further utilized to investigate buckling phenomena and provide initial predictions of critical buckling loads. The FEA results demonstrated that the stiffener configuration significantly affects load-carrying capacity, underscoring its critical role in structural performance. To reduce the computational intensity of FEA, ANN was trained on a subset of FEA-generated data, achieving high predictive accuracy for buckling loads with reduced modeling effort. The proposed hybrid approach successfully optimized stiffener parameters, offering a robust solution for improving the design and performance of composite stiffened panels under shear loading.