Machine learning techniques for the comprehensive analysis of the continuous casting processes: Slab defects

Continuous casting route is the most widely produced process in the world every year, accounting for around 95% of global steel production. Due to the various inevitable defects, including cracks, segregations, inclusions and breakout involved in the production process, tranditional control and monitoring approaches are pushed to their limits. Over the past decade, the increasing global competition has combined metals engineering with digital technology and artificial intelligence skills, to help the manufacturing industries to reap the benefits of “big data.” By targeting critical machine learning (ML) techniques, the casting industry leaders rapidly take advantage of the latest discoveries in manufacturing processes through digital twinning to enable defect-free casting. This study offers an analytical overview of ML methods applied to the examination of the continuous casting procedure. This paper examines the current research on ML in steel continuous casting, organizing the findings into categories, which aids in identification of typical application cases and approaches. This comprehensive analysis concludes with the elaboration of challenges, potential solutions, and a future outlook for further cutting-edge research and development to unleash the potential of future ML techniques and maintain casting industry's reputation.