EN

Machine learning based design of single and interpenetrating phase, vibration-isolating advanced materials

United Arab Emirates | Computer Sciences, Earth Sciences

Swiss partners

  • Laboratoire fédéral d’essai des matériaux et de recherche EMPA: Andrea Bergamini (main applicant)

Partners in the MENA region

  • New York University Abu Dhabi, Emirats arabes unis: Nikolaos Karathanasopoulos (main applicant)

Presentation of the projet

The project aims at the development of vibration isolating and attenuating advanced materials that are based on both single and interpenetrating-phase, architected topological designs. In particular, it targets the engineering and design of moderate and high-strength viscoelastic co-continuous composite materials that rely on triply periodic minimal surface (TPMS) topologies for the first time.

The project working packages explore the dynamic performance of advanced material designs with a wide combination of stiffness, density and viscosity attributes, that remains up to now utterly explored. As such, the project premises the exploration of an entirely new dynamic material performance space, with significant vibration isolation attributes over a wide range of frequencies that is utterly infeasible for single-phase and merely elastic architected solid materials. The project combines extensive numerical analysis and machine learning parts, that are supported by a considerable amount of advanced manufacturing and experimental testing tasks. For this purpose, extensive, experimentally-calibrated datasets will be created, serving as a reference for this utterly novel advanced material design dynamic performance space.

Machine learning will provide the basis for the development of data-based metamodels that can directly evaluate the dynamic performance of this rather vast space of advanced material designs, while providing advanced modeling capabilities that are infeasible with the mere use of traditional modeling techniques. In particular, the developed machine learning models will render the solution of the inverse optimization problem for dynamic material tasks feasible for the first time, through their coupling with standard optimization analysis techniques. By those means, the inverse on-request design of vibration isolating materials, more specifically, the identification of the inner interpenetrating phase composite design specifications that optimally satisfy vibration isolation targets at a frequency range of interest will become accessible to the engineering community.

The investigation of single phase and interpenetrating architected TPMSs will offer new insights in the mechanics that govern the interaction of two vastly dissimilar phases and allow to reach areas of the  stiffness-damping space currently not accessible to conventional materials, especially if high strengths requirements are taken into consideration.