DESIGN OPTIMIZATION OF 3D AUXETIC STRUCTURE FOR BLASTWORTHY STRUCTURE APPLICATIONS USING MACHINE LEARNING METHOD

• Main Author: Andika
• Format: Final Project
• Language: Indonesia
• Online Access: https://digilib.itb.ac.id/gdl/view/73812
• Institution: Institut Teknologi Bandung

<p align="justify"> The need for structural blastworthiness is still high due to the increasing number of military casualties from anti-tank mines. The data shows that there is an increase in the number of casualties due to anti-tank mines as many as 1220 fatalities in 2020. One of the proposed solutions to fulfill blastworthiness is the use of an auxetic structure that is able to absorb as much energy as possible due to the negative Poisson’s ratio characteristic effect. In this research, optimization has been carried out to obtain the most optimum 3D auxetic structure in maximizing energy absorption when subjected to compression load. Optimization in this research uses machine learning methods which consist of artificial neural networks (ANN) and non-dominated sorting genetic algorithms II (NSGA-II) methods. The collection of training data and analysis of blastworthiness characteristics was carried out using the dynamic finite element method from LS-DYNA software. The optimum 3D auxetic configuration produces specific energy absorption (SEA) of 117.96 kJ/kg with an error of 1.4% against the results from machine learning. The optimum configuration is applied to the sandwich structure in experiencing the blast load. The results show that the optimum structure subjected to 8 kg of TNT resulted in a maximum displacement and acceleration of 49.7 mm and 37,935 G, respectively. There was no damage to the occupant side plate (OSP), so the optimum structure complies with the standard of NATO STANAG 4569 level 3 for the protection of vehicles subjected to blast loads.