Multi-objective optimization of 3D-printed honeycomb-inspired thin-walled sandwich structures using an artificial neural network surrogate model

Description

This work presents an Abaqus-ANN-NSGA-II based framework for prediction, multi objective optimization of additively manufactured PLA, ABS and PETG honeycomb sandwich structures under quasi-static compression. 72 Abaqus simulations DOE were generated for PLA, ABS, and PETG by varying cell width, core height, and wall thickness. The simulation outputs were processed as P95 von Mises stress and maximum compression-axis deformation to form a robust dataset for surrogate modelling. A two-output artificial neural network with a 7-5-2 architecture was trained using material code, geometric parameters, and elastic properties as inputs.

Institutions

• Galgotias University

  Uttar Pradesh, Greater Noida

Categories

Licence