Evaluation the Reverse Crashing Characteristics of Thin-Walled Shells that are Packed with FGM Foam and Forecast the Energy Absorption Through the Use of Artificial Neural Networks

Abstract

The present research investigates the reverse crashing behavior of thin-walled shells filled with functionally graded foam (FGF). The study employs a combination of nonlinear finite element analysis and artificial neural networks (ANN) to predict the energy absorption capacity of the shells. Three types of foam properties distributions are examined for aluminum porous core foam under quasi-static vertical loading. The impact parameters, including specific energy absorption, are evaluated by analyzing the geometrical parameters of the shell and the mechanical properties of the foam core. The results reveal that foam enhances the displacement mode of the shells, resulting in more regular folds compared to shells without foam. Additionally, the average crashing force and absorbed energy per unit mass increase with increasing foam density and shell wall thickness. The optimal geometric parameters obtained through a genetic algorithm and neural network demonstrate that the specific energy absorption capacity of the foam-reinforced shell can increase by up to 120% compared to a similar hollow specimen. Moreover, the distribution of foam properties has a significant effect on specific energy absorption, with an increase of up to 66.5%.