Optimized Kresling origami-inspired structures using Artificial Neural Network and Monte Carlo Method

Abstract

In this paper, fatigue and stability characteristics of an origami-inspired structure are investigated. The studied reconfigurable structure is designed by utilizing the Kresling origami pattern, which is used ubiquitously in various engineering applications, from aerospace to electromagnetics. The study is based on the results of a thorough finite element analysis to assess the stability and fatigue capacity of the structure based on its geometric characteristics, in ANSYS software. Based on the FE results, two artificial neural networks are trained to be used as surrogate models, to predict the fatigue and buckling load of the origami structure based on its geometric parameters. The studied geometric parameters include the length ratio, total height, circumscribed circle radius, thickness, story height, and crease parameters. The ANN surrogate models enable extensive parametric studies to assess the single or simultaneous effects of the parameters on these pair of capacities while bypassing the inherent computational cost of the Finite Element Analysis (FEA). In this study, the pair of developed ANN-based surrogate models enables feeding the Monte Carlo approach with 100,000 samples, for optimizing the geometric configuration of the designed origami structure to obtain its maximum stability and fatigue capacities simultaneously.