Mechanical Properties and Neural Network‐Aided Design of Curved‐Wall Rose‐Shaped Metamaterial Honeycombs

This study investigates the mechanical properties of a category of honeycomb structures characterized by rose‐shaped cells under various loading conditions. Analytical expressions for the equivalent mechanical parameters of curved‐wall rose honeycombs are derived, significantly reducing the errors associated with employing star‐shaped honeycombs in prior calculations reported in the literature. Subsequently, mechanical parameters such as Poisson's ratio and Young's modulus, computed using the derived formulas, are compared with results obtained from finite‐element models. The close agreement between these two sets of results confirms the accuracy and reliability of the proposed theoretical formulas. Furthermore, both forward and inverse neural network models of the rose‐shaped honeycombs are developed to investigate the relationship between geometric configurations and mechanical performance parameters, facilitating the programed design of rose‐shaped honeycomb structures. The validation results demonstrate that the average correlation coefficients exceed 98.48%, and prediction errors remain below 6%, thereby validating the effectiveness of the models. Finally, the study compares the performance differences of Kinesio tape (KT) with and without honeycomb structures across various arm models to assess the potential application of auxetic structures in KT. Additionally, performance comparisons among rose‐shaped, star‐shaped, and re‐entrant honeycombs are conducted. This research provides a theoretical foundation for the application of rose‐shaped honeycombs in engineering contexts.