Smart Predictive Maintenance Strategy Based on Cyber-Physical Systems for Centrifugal Pumps: A Bearing Vibration Analysis

Abstract

Early detection of faults in rotary machines, particularly in centrifugal pumps, has become essential in terms of avoiding unplanned or unnecessary maintenance and enhancing system reliability at minimized costs. This paper focuses on the predictive maintenance (PdM) for centrifugal pumps in Cyber-physical Systems (CPS) and proposes a concept to monitor the bearings in order to evaluate the pump’s health condition. CPS have the potential to provide technical systems with self-awareness and self-maintenance capabilities. The implementation of predictive analytics as part of the CPS framework enables the machinery to continuously track its own performance and predict potential failures. Among the different methods for monitoring the pumps, vibration monitoring is one of the most important methods to collect real-time data. Using this technique for PdM would enable maintenance prior to early failure. In the case of the bearing’s vibration monitoring the residual useful life can be predicted and even increased. Consequently, the probability that a breakdown happens is minimized and a smart PdM which guarantees an optimally safe system can be accomplished. Additionally, a conceptual economic analysis which compares two different maintenance strategies is presented in the last section.