A review of machine learning (ML) and explainable artificial intelligence (XAI) methods in additive manufacturing (3D Printing)

Abstract

Additive Manufacturing (AM) (known as 3D printing) has modernised traditional manufacturing processes by enabling the layer-by-layer fabrication of complex geometries, along with advanced design capabilities, cost efficiency, and reduced production time. AM offers flexibility and customisation in product development, allowing for the deposition, solidification, or joining of materials based on computer-aided models. In recent years, the large-scale collection of AM-related data has facilitated the use of machine learning (ML) techniques to embed into AM processes and optimise quality. However, many advanced ML algorithms do not provide their underlying decision-making criteria, remaining as a black box. Alternatively, explainable artificial intelligence (XAI) methods have been employed to explain these black box models. Even though ML has been widely used in AM, the use of XAI in AM is still very limited. This paper provides a comprehensive review of the integration of ML and XAI (XAI for the first time) in AM processes, exploring current research progress and future prospects. The study outlines the various ML techniques and XAI applied in different domains of AM. Additionally, it examines ML and XAI applications across different AM technologies and life cycle stages, highlighting their functions in design optimisation, process monitoring, and sustainability analysis. Furthermore, the review discusses new developments, challenges, and future research directions in ML and XAI for AM, allowing for enhanced efficiency, innovation, and sustainability in AM processes.