Artificial Intelligence-based Optimization and Material Ageing Studies to improve geometrical and mechanical performances of parts manufactured through Additive Manufacturing processes

Additive Manufacturing has emerged as a transformative technology to produce complex geometries and customized parts across a range of sectors. Nevertheless, considerable obstacles remain in optimizing the mechanical properties and dimensional precision of manufactured parts, as well as addressing concerns regarding the sustainability of the materials used. This thesis investigates the integration of Artificial Intelligence techniques and Multi-Objective Optimization methods to enhance the performance of AM processes. In the Fused Filament Fabrication process, artificial intelligence-driven models were employed to optimize the performance indices. This approach permitted the concurrent enhancement of mechanical strength and dimensional precision in Polylactic Acid components, thereby facilitating multi-objective optimization for industries necessitating high-precision manufacturing. This approach markedly diminishes the necessity for trial-and-error procedures and augments the predictability of AM outcomes. While in Selective Laser Sintering, this thesis develops predictive models for mechanical properties derived from degraded Polyamide 12 powder. The present study investigates the effects of repeated thermal cycling on mechanical properties and crystallinity. The relationship between the degree of crystallinity and the ultimate tensile strength of the parts was established, thereby offering a practical method for predicting mechanical strength and determining the optimal reuse of aged powder, with the dual objective of ensuring both sustainability and part integrity. The research was conducted along three primary avenues: (i) the creation of performance indices that unify mechanical and dimensional properties into a single optimization value; (ii) AI-driven optimization of these indices in Fused Filament Fabrication, improving both mechanical and dimensional properties; and (iii) predictive modeling of mechanical properties from degraded Polyamide 12 powder in Selective Laser Sintering, enabling sustainable material reuse. The results illustrate that AI-driven optimization techniques yield enhanced precision in mechanical performance predictions, while the proposed recycling strategies for PA12 contribute to more sustainable AM practices. The findings provide substantial insights into the application of AI for both industrial manufacturing optimization and environmental sustainability in AM processes.