Presenter: Paolo Bethaz
Monday, March 9th, 2022 17:00
A Predictive Maintenance strategy for a complex machine requires a sophisticated and non-trivial analytical stage to provide accurate and trusted predictions. Manufacturing industries are not only interested in identifying a failure in advance, but they also want to have continuous monitoring of machinery’s conditions, estimating its remaining useful life (RUL). In this SmartTalk, we present a methodology to estimate the degradation of an industrial machinery over time by analyzing its vibrational measurements collected during manufacturing processes. In particular, assuming that we have measurements covering an entire life cycle of the machinery, the aim is to learn, through Machine Learning algorithms, its degradation trend. The learned trend will then be used to infer the remaining production life for machines that have similar characteristics and perform similar production cycles.
Preliminary experiments of this methodology have been tested in a real use-case and the results seem to be promising.
Biography: Paolo Bethaz is a second-year PhD Student at Dipartimento di Automatica e Informatica, Politecnico di Torino and member of the SmartData@Polito center. He received the masters’s degree in computer engineering from the Politecnico di Torino in 2019. His research interests are focused on the improvement of the automated data science applied in different contexts, including industry 4.0 and the IoT.