Machine learning guided analysis and rapid design of a 3D-printed bio-inspired structure for energy absorption

Abstract

Mantis shrimps employ their telson, or tail plate, to mitigate the impact with hard surfaces, thanks to its unique double-sine shaped microstructures that absorb energy through deformation. Inspired by this natural impact-resistant design, similar lightweight energy absorbers have been developed for applications in transportation systems and personal protective equipment. This study presents a data-driven approach to analyze and optimize these structures subjected to crushing loads. The structure's geometry is defined by three simple parameters based on a sine wave shape function and fabricated using ABS-M30 polymer through 3D printing. Material tests and compression tests under uniaxial loading conditions are conducted to characterize the material properties and structural behavior. Finite element models are created to simulate these tests, and Machine Learning techniques are applied to study the structure's behavior. A total of 100 Design of Computer Experiments are generated by manipulating the design variables, and the Decision Tree method categorizes deformation modes. Intrinsic and response parameters are predicted as functions of the geometric parameters. Using these relationships, a multi-objective optimal design is achieved, enhancing specific energy absorption while reducing peak crush force. The Pareto Front, representing optimal designs for these objectives, is obtained through genetic algorithms. A multi-criteria decision-making algorithm factors in designer preferences to narrow down the optimal design dataset. This study highlights the potential of bio-inspired structures and design methodologies for innovative lightweight protective equipment in transportation systems and human wearables.