|
Abstract
|
This article presents an innovative perspective on optimizing additive manufacturing processes by integrating advanced in-situ monitoring, control systems, and data-driven methods, with a particular focus on deep learning. The literature highlights a consensus on the need for increased accuracy, efficiency, and reliability in producing complex, customized components aiming for assured quality and defect-free manufacturing. Utilizing in-situ monitoring systems and advanced sensors including imaging, optical projection, acoustic, and integrated sensor networks is deemed essential for real-time detection of anomalies and defects. The deployment of AI, machine learning, and deep learning algorithms across major AM domains such as design evaluation, material and geometric optimization, real-time defect identification and mitigation, and big sensor data analytics plays a pivotal role in enhancing process control, automating parameter adjustment, and modeling mechanical properties. The transition from basic monitoring to adaptive closed-loop control (CLC) driven by ML, as well as the implementation of responsive and intelligent decision-making systems based on real-time data, are key steps identified in recent studies. Furthermore, the use of digital twins and virtual sensors is emphasized for elevating smart integration between physical and digital production environments. Ultimately, merging ML with advanced sensing technologies is fundamental for achieving continuous, precise, and high-quality automated AM production.
|