Juan Carlos Tarín Tomás obtained the Bachelor’s degree in Industrial Technologies at the Universitat Politècnica de València in 2017 and the Master’s degree in Industrial Engineering (specialization in Electronics) at the same institution in 2020. Since October 2020, he is a PhD candidate in the research group of electroactive materials in LaCàN, within the Department of Civil and Environmental Engineering at Universitat Politècnica de Catalunya, under the supervision of Profs. Irene Arias and Francesco Greco. Funded by a Spanish Ministry FPI predoctoral contract (PRE2019-087992, Ayudas para Contratos Predoctorales para la Formación de Doctores), his work focuses on the topology optimization of lattice metamaterials exploiting flexoelectricity as an alternative to piezoelectricity, with a particular emphasis on genetic algorithms and computational design frameworks implemented in Python and MATLAB.
This thesis aims to design and identify optimal lattice material architectures that exploit flexoelectricity as an alternative to piezoelectricity, through the development of a topology-optimization framework.
On the one hand, a geometric representation framework is developed to prevent two key issues. First, artificially thin junctions, which commonly arise in flexoelectric design by producing large strain gradients through non-manufacturable geometric features. To mitigate this effect, a binary discrete representation is adopted, enforcing a minimum thickness throughout the geometry. Second, the appearance of disconnected material regions, a typical artifact of topology optimization that becomes more complex in lattice materials, since it may occur both at the unit-cell level and at the lattice connectivity level. This issue is addressed through connectivity and path-search algorithms e.g., Dijkstra-based formulations that restore connectivity with minimal geometric modification. In addition, a hexagon-based discrete representation is introduced to improve isotropy with respect to the conventional square discretization.
On the other hand, a two-stage optimization strategy based on genetic algorithms is proposed. The first, single-objective stage is used to obtain high-quality seeds to initialize the second multi-objective stage, accelerating con-vergence towards configurations that preserve functional performance while minimizing material usage. Finally, a post-processing step is introduced to add material in a controlled manner to promote variable thickness, thereby enhancing the strain gradients that are relevant for the flexoelectric response.
Speaker:
Date:
Time:
Category: