Abstract

Additive manufacturing (AM), commonly known as 3D printing, offers unparalleled flexibility in producing complex geometries and customized components. However, achieving optimal quality, mechanical performance, and production efficiency remains challenging due to the numerous interdependent process parameters involved, such as layer thickness, print speed, extrusion temperature, and material feed rate. This paper presents an AI-assisted framework for process optimization in AM, leveraging machine learning (ML) and deep learning (DL) models to predict part quality and recommend parameter settings. The system integrates sensor data from in-situ monitoring, historical build logs, and material property databases to train predictive models capable of identifying defect patterns and performance deviations. Reinforcement learning agents iteratively refine process settings to maximize quality and minimize waste. Experimental validation on polymer and metal AM platforms shows reductions in surface roughness by up to 22%, tensile strength improvements of 15%, and material usage optimization of 12%. The proposed framework demonstrates the potential of AI-driven decision-making to enhance repeatability, reduce trial-and-error, and accelerate design-to-production timelines, contributing to the broader adoption of smart manufacturing in Industry 4.0.