National Technical University of Athens (NTUA)
Computational techniques for optimal design of composite materials based on multiscale modeling
Abstract
dc:descriptionThis dissertation focuses on the development of computational techniques for the optimal design of composite cellular materials and structures using topology optimization and multiscale modeling. The main objective of the dissertation is to formulate and assess optimization procedures in which the performance of a macro-scale structure is improved through the design and use of periodic microstructure. The theory of numerical homogenization is used to connect the two scales by computing the effective elastic properties of a periodically repeated unit cell and introducing them into the macro-scale. The dissertation first presents the standard topology optimization formulation for compliance minimization under a volume constraint. In this presentation, emphasis is given on the density-based SIMP approach and the BESO and level-set approaches. A simplified Method of Moving Asymptotes update scheme is introduced for the standard compliance-minimization problem, aiming at reducing the implementation complexity and computational cost while maintaining behavior comparable to the general MMA algorithm. The standard formulation is extended to homogenization-based material design, where the periodic unit cell is optimized with respect to macro-scale structural performance. This enables the design of anisotropic cellular materials that adapt to the loading and boundary conditions of a structure. Concurrent multiscale topology optimization formulations are subsequently introduced, allowing the macro and the micro-scale to be updated within the same optimization procedure. SIMP-based, BESO-based, and combined implementations are presented in detail. An additional two-phase composite extension is also examined, in which the void or weak phase of the unit cell is replaced by a secondary material and different volume-constraint strategies are compared.A stress-based multiscale topology optimization formulation is developed. This formulation focuses on the problem of TO for volume minimization subject to local stress constraints. The formulation uses the von Mises yield criterion and the Augmented Lagrangian function to enforce local constraints without relying on global stress aggregation techniques. Finally, model order reduction techniques, including Proper Orthogonal Decomposition, on-the-fly reduced basis construction, and Approximate Reanalysis, are implemented into the TO procedure to reduce the computational cost of repeated finite element analyses in topology optimization and homogenization-based material design. The results demonstrate that the proposed formulations provide a unified framework for stiffness-driven, stress-aware, and computationally efficient multiscale design of composite cellular materials and structures.
Degree
thesis:*- Grantor dc:publisher
- National Technical University of Athens (NTUA)
- Year dc:date
- 2026
Author and committee
dc:creator, dc:contributor.*- Authors dc:creator
-
- Kazakis, Georgios
- Καζάκης, Γεώργιος
Subjects
dc:subject × 16- Βελτιστοποίηση
- Βελτιστοποίηση τοπολογίας
- Ομογενοποίηση
- Μοντελοποίηση πολλαπλών κλιμάκων
- Ανάλυση κατασκευών
- Optimization
- Topology optimization
- Homogenization
- Multiscale modeling
- Structural analysis
- Επιστήμη Πολιτικού Μηχανικού
- Επιστήμες Μηχανικού και Τεχνολογία
- Δομοστατική
- Civil Engineering
- Engineering and Technology
- Structural Engineering
Rights
- Language dc:language
- eng
Identifiers
dc:identifier.*- Identifier
- 10.12681/eadd/62655