摘要:AbstractIn order to optimize the maintenance operations and limit the impact of material wearing on the galvanizing coating quality and defects, this paper proposes a data-driven model to estimate the remaining useful life of a roller system immersed in a galvanizing bath. The proposed model comes from the Hidden Markov Model family and is trained in non-supervised mode. The pros and drawbacks are given, in conclusion, for a possible use in a control loop.