PhD.Dissertation:Abdullah Kendibilir

PARTICLE BASED TOPOLOGY OPTIMIZATION METHODS FOR ADDITIVE MANUFACTURING TECHNOLOGIES

Abdullah Kendibilir
Manufacturing Engineering, PhD Dissertation, 2024

Thesis Jury

Assoc. Prof. Adnan Kefal (Thesis Advisor),

Prof. Oliver Weeger

Asst. Prof. M. Erden Yildizdag

Prof. Mehmet Yildiz

Assoc. Prof. Gullu Kızıltas Sendur

Date & Time: 22 July, 2024 –  16.30 PM

Place: FENS G029

Keywords : peridynamics, topology optimization, additive manufacturing, crack

modelling, fracture toughness

Abstract

Topology optimization (TO) is a robust design method that generates the best material distribution by improving an objective function (e.g., maximizing stiffness or minimizing strain energy) and satisfying design constraints (e.g., volume constraints, and minimum dimensions). Although TO is increasingly utilized in lightweight designs of various engineering structures, TO designs have usually been criticized for causing complex forms, which makes obtained geometry difficult or even impossible to fabricate by using traditional manufacturing methods such as machining, casting, etc. Additive manufacturing (AM) is an effective approach to fabricating these intricate shapes obtained from topology optimization (TO). However, AM methods, which entail multiple heating and cooling cycles, lead to the formation of concentrated local residual stresses in TO designs. Consequently, this results in diminished material properties in specific areas and increases the possibility of local failures, such as cracks/defects. At this point, a non-local reformulation of classical continuum mechanics, namely Peridynamics (PD), offers a practical way to model discontinuities by only breaking the bonds between particles. The PD method can accurately model the dynamic fracture behavior of the structures including crack branching and coalescence thanks to the utilization of the integral form of balance equations instead of partial derivatives.

Overall, the main motivation of the present thesis is to develop novel 3D topology optimization methods based on peridynamics (i.e., peridynamic topology optimization or PD-TO) to maximize the structural performance of the final designs produced by AM processes. In this context, three different density-based TO algorithms namely bi-directional evolutionary structural optimization (BESO), optimality criteria (OC), and proportional approach (PROP) are coupled with the PD method for the optimization of three-dimensional structures. The developed methods are extensively validated by performing on various benchmark problems. Afterward, the PD-TO method is utilized for the TO of large-scale engineering structures (ship cross-section design). According to the obtained results, the viability and superior capabilities of the PD-TO methods are revealed. Furthermore, this method is performed with a systematic optimization scheme for increased fracture resistance under the dynamic load by considering a classical continuum mechanics based failure criteria. The comparisons showed the increased fracture resistance on the optimized design by the proposed algorithms against the classical optimization results. Moreover, the PD-TO methods are utilized for decreasing residual stresses occurs on the structures produced by metal AM processes. Thanks to embedding hypotetical cracks in the initial design domain and topologically optimizing via PD-TO, we achieved residual stress minimized light weighted structures and experimentally validated the obtained numerical results.