Sustainable maintenance and digital twin technology: a test case for evaluating integration potentialities

The concept of sustainable maintenance is quite recent in the industrial context, entailing a change of mindset compared to traditional models that requires integrating the three main dimensions of sustainability (i.e., economic, environmental and social) in the whole life-cycle of an asset, as well as in the design phase of the maintenance strategies and tools. Digital technologies could boost the development of sustainable maintenance models. The concept of Maintenance 4.0 refers to the integration of digital technologies to enhance reliable predictive models and support the identification of effective maintenance actions in real time. This work aims to present a prototype tool for supporting sustainable maintenance based on digital twin technologies. A test case that adopts machine learning technologies for prediction analysis integrated with assessment of safety and environmental impacts is discussed. The test case has been developed, based on a dataset of events (failures and non-conformities) collected through on-board sensors on a specific set of equipment. Using an open-source tool, three different ML algorithms have been selected and their effectiveness has been analyzed and compared through a set of KPIs, in order to identify the most reliable for predictive maintenance purposes.