Energy Absorption Properties of Curved Wall Honeycombs Based on Neural Networks

Abstract

Honeycomb structures are used widely nowadays and honeycombs with negative Poisson’s ratio has attracted widespread attentions. The compression tests of 3D printed concave hexagonal honeycomb model is compared with the results of the finite element model, which confirms the effectiveness of the finite element models. It is known that the curved wall honeycomb can effectively alleviate the stress concentration compared with straight-walled honeycombs. The curved walls in sinusoidal shape replace the straight walls which are mainly carried in the concave hexagonal honeycomb cells in this paper. Python is used to generate a large number of finite element models and then establish a dataset corresponding to the parameters and mechanical properties of the curved wall honeycombs. The neural network is proposed to predict energy absorption properties of the honeycombs. The sensitivity analysis of the parameters is carried out to provide guidance for the design of curved wall honeycomb structures. The specific absorption energy is optimized, and the energy absorption capacity is evaluated by using the neural network. The results show that the total energy absorption of the concave straight wall honeycomb is higher, but the energy absorption efficiency of the concave curved wall honeycomb is higher.