Towards machine learning for error compensation in additive manufacturing

• Main Authors: Omairi, Amzar, Ismail, Zool Hilmi
• Format: Article
• Published: MDPI AG 2021
• Subjects: T Technology (General)
• Online Access: http://eprints.utm.my/id/eprint/95181/; http://dx.doi.org/10.3390/app11052375

Additive Manufacturing (AM) of three-dimensional objects is now being progressively realised with its ad-hoc approach with minimal material wastage (lean manufacturing) being one of its benefit by default. It could also be considered as an evolutional paradigm in the manufacturing industry with its long list of application as of late. Artificial Intelligence is currently finding its usefulness in predictive modelling to provide intelligent, efficient, customisable, high-quality and sustainable-oriented production process. This paper presents a comprehensive survey on commonly used predictive models based on heuristic algorithms and discusses their applications toward making AM “smart”. This paper summarises AM’s current trend, future opportunity, gaps, and requirements together with recommendations for technology and research for inter-industry collaboration, educational training and technology transfer in the AI perspective in-line with the Industry 4.0 developmental process. Moreover, machine learning algorithms are presented for detecting product defects in the cyber-physical system of additive manufacturing. Based on reviews on various appli-cations, printability with multi-indicators, reduction of design complexity threshold, acceleration of prefabrication, real-time control, enhancement of security and defect detection for customised designs are seen of as prospective opportunities for further research.