Abstract:

Porous metals are increasingly important in technology. Due to their tunable me- chanical properties, they are promising candidates in various emerging applications such as metallic scaffolds for load-bearing bones, lightweight structures for trans- port technologies, electrodes for electrochemical energy storage devices and more.
This project aims at developing a computational model to establish a quantitative understanding of the relations between the enormous variety of possible topologies of porous structures and their mechanical properties. The machine-learning based model will be employed to identify various prototype structures of new morpholo-gies via implementation of hierarchical screening and a material genome approach.
The optimal and/or statistically relevant structures will be 3D printed and tested mechanically in experiments, with the results contributing to both tuning and val- idation of the computational designs. This dissertation presents the first year as- sessments of the research project as well as the future goals and steps.
During the first year the main effort was made to create necessary computational tools in order to start collecting data for machine learning models training. Preliminary results of Molecular Dynamics simulation for several topologies are presented here show- ing the whole workflow from structure generation to determination of mechanical properties.