Use of artificial intelligence in algorithm-based product design tools in powder bed fusion of metals : a literature review
avoin
Julkaisu on tekijänoikeussäännösten alainen. Teosta voi lukea ja tulostaa henkilökohtaista käyttöä varten. Käyttö kaupallisiin tarkoituksiin on kielletty.
Lataukset8
Pysyvä osoite
Verkkojulkaisu
DOI
Tiivistelmä
This thesis reviews how algorithm-based design tools can support metal additive manufacturing, with a focus on powder bed fusion (PBF). It concentrates on three tool families: implicit modeling (with field-driven workflows such as nTopology), generative design based on topology optimization, and AI-enabled methods that accelerate or extend these approaches.
The thesis proposes a practical framework that links geometric design freedom to PBF constraints. It treats support requirements, minimum feature limits, residual stresses, thermal distortion, and the role of process parameters as design inputs rather than downstream issues. The review scope emphasizes recent academic work and summarizes how current tools address common PBF design problems and where they still fall short.
From the surveyed literature, algorithm-based workflows can reduce mass and improve structural performance by creating lattice or cellular architectures that follow load paths and stress fields. AI contributes mainly through surrogate models and generative methods (for example GANs, CNN-based predictors, diffusion approaches, and reinforcement learning) that reduce iteration time in topology optimization and expand early design exploration. Research also shows increasing integration of manufacturability rules into optimization loops, as well as data-driven approaches for predicting and compensating PBF distortion using hybrid simulation and machine-learning methods.
The thesis concludes that these methods are promising for PBF part design, but adoption still depends on computational cost, validation effort, and tighter coupling between data-driven models and physics-based simulation. It also includes a qualitative review of key equations and optimization methods used in the discussed workflows.