Leveraging Semantics and Machine Learning to Automate Circular Economy Operations for the Scrap Metals Industry

Circular Economy operations and supply chains are becoming increasingly intricate. From raw material extraction, manufacturers and retailers to consumers and recyclers, the benefits stemming from an organized and transparent circular economy network are numerous. Scrap materials have the potential to contribute significantly to the reduction of energy consumption and resource consumption. One is challenged to combine data from scrap metal collection and management, shredding and separation stages, data related to the suppliers and customers as well as background data as they are available from reference sources and classifications available by the metal industry. Integration of data from dispersed and heterogenous sources with in-house ERP systems, alongside a need to automate and extract knowledge and support, call upon a semantically enabled approach as the one demonstrated in the paper. The work combines knowledge from a sizeable actual plant (ANAMET), data technologies (Symbiolabs) and modelling expertise (NTUA). The approach proposes the means to enhance circular economy operations for scrap metal materials with the use of semantic technologies and machine learning techniques presenting examples and cases as they have been used in real-life.