Behaviour modelling and vibrations analysis applied to predictive maintenance: Different approaches leading to the same conclusions

The paper introduces the maintenance functions and new services needed: (1) diagnostic aids for detection and location of failures; (2) historical analysed data to establish intervals for scheduled maintenance; and (3) analysed performance data for prediction of equipment degradation. It then focuses on predictive maintenance. Especially, we present two different approaches in the field of information processing and signal processing. These two different approaches have been experimented in two different contexts: (1) pumps, jacks and gates behavioural modelling; and (2) turning machines vibrations analysis. Our results show that these two approaches are efficient for prediction of equipment degradation