Identification of oil starvation in hydrodynamic journal bearing using rotor vibration and Extended Kalman Filter
Oil starvation is a critical fault in hydrodynamic bearings caused by insufficient oil supply flowrate. When late detected, this fault can deteriorate the bearings' performance and damage the rotating machine. Thus, early fault identification techniques should be applied to avoid this scenario and allow effective maintenance. However, the literature about the identification of oil starvation fault is relatively recent and very scarce. Among the few studies available, oil starvation has been identified in early stages using a purely deterministic technique, which can be jeopardized by model inaccuracy. Alternatively, this paper proposes a novel method for the identification of oil starvation faults in hydrodynamic bearings using the rotor vibrational responses and the Extended Kalman Filter (EKF), which is a stochastic-determinist state estimator that can deal with modelling and measurement errors. Computational simulations performed in this paper show that the proposed method is suitable to successfully identify the oil starvation faults in real-time even with noisy vibrational responses, thus representing a promising tool for the monitoring of rotating machines.