This was presented at the NAFEMS World Congress 2025, held in Salzburg, Austria from May 19–22, 2025.

Abstract

Meta-Materials (MMs) pose a huge potential for providing solutions to advanced engineering problems and also opening a new playground for cutting-edge high-performance parts and systems. However, developing MMs and integrating them into everyday R&D processes is a big challenge. It requires know-how and expertise in MMs and computational physics. The design and simulation of systems consist of millions of unit cells (basic blocks of MM microstructures), making simulation times almost non-feasible. Optimization of such systems becomes impossible. We develop a methodology that addresses the challenges above and provides an R&D approach for MM design. We also present novel MMs with properties that do not exist in nature. We develop two MM categories. 'œConventional' MMs are materials with extraordinary properties. They are made from common engineering materials, and their properties are engineered by the design of their micro-structures. Mostly used today in optics, electromagnetics, and acoustics. Second category: Dynamic & Multi-Functional MMs involve creating cavities inside unit-cells, then placing liquids, powders, and particles inside them, and allowing them to move inside these cavities. This powerful concept aims at multi-disciplinary usage, and allows for design and optimization of several properties and characteristics of the material. For example: we developed materials that become stiffer at higher temperatures, materials that change their mechanical properties when exposed to an external magnetic field. These MMs are adaptable, responsive to external conditions, and dynamically controllable. MM design is based on combining computational multiphysics modeling and Machine Learning. Building the ML-based predictor starts with developing a parametric, high-fidelity computational model, including mesh refinement, sanity checks, and step-by-step validation. This ensures a reliable foundation for the next stages. Physics constraints and manufacturing requirements, such as tolerances and resolutions, ensure the models are robust and practical. Standard ML validation techniques verify accuracy and robustness with test, train, and validation sets, resulting in reliable predictions for FEM modeling. Modern systems face major challenges related to power consumption and heat generation. Common engineering materials often fail to provide adequate solutions, leading to a growing demand for improved heat exchangers, heat sinks, and short-circuiting prevention. We develop MMs to address these challenges. Examples include controlling the coefficient of thermal expansion versus mechanical stiffness, phase-changing elements with controlled conductivity versus temperature, and multi-physics (heat and flow) with triply periodic minimal surface geometry, enabling control of pressure and temperature drops. In mechanics our R&D explores energy absorption through fluid viscosity, center-of-mass dynamics, and the integration of pendulum principles into microstructural unit cells. Fluid Viscosity: We place viscous fluid within the unit cells. Upon impact, the fluid flows and resists motion due to its viscosity, enabling controlled energy dissipation. This mechanism ensures highly effective shock absorption and vibration damping while maintaining structural integrity. Recognizing the immense potential of MMs and the challenges in making them accessible to engineers, we develop a methodology that combines computational physics simulations with ML models. Using this methodology we develop novel dynamic and multi-functional MMs for heat transfer and shock absorption and vibration damping.