Structural Design Analysis of Substrate with Honeycomb Core Under Normal Pressure, Using RSM and ANN

Abstract

Listen

Translate article

This study presents a comprehensive structural performance analysis of a honeycomb-core substrate under normal pressure, highlighting the superior predictive accuracy of Artificial Neural Networks (ANNs) over Response Surface Methodology (RSM). The analysis focused on critical design parameters, such as material selection, coverage rate, and wall thickness, which significantly influence the substrate’s maximum deformation, elastic stress, and mass. The ANN model, trained on these parameters, optimized the design to achieve a cell size of 60 mm, a wall thickness of 12.5753 mm, a coverage rate of 64.38%, and selected aluminum as the material. This optimization resulted in a substrate with a maximum deformation of 7.21 × 10³ mm, an elastic stress of 1.9465 MPa, and a mass of 54.949 kg. The RSM-ANN method surpasses RSM in both optimization and accuracy, enhancing the understanding of how honeycomb design affects substrate properties.