Adequate oil supply is mandatory for optimal operation of hydrodynamic bearings, however many rotating systems in operation do not have sensors to individually measure oil supply flowrates and, eventually, identify failures in the bearing lubrication system. On the other hand, measuring the vibration of rotating machines is a basic procedure and present in any maintenance program for this type of equipment. In this context, this doctoral project proposes the development of a model-based method for measuring oil supply flowrates in hydrodynamic bearings using the information contained in the rotor vibration. The method consists of modeling the rotor-bearing system and adjusting the unknown parameters, in order to eliminate errors between the vibration predicted by the model and the vibration measured in the real machine. The parameters to be adjusted will be the oil flowrates in the hydrodynamic bearings and also the unbalance of the rotor, since both change the system vibration and are generally unknown in real operating conditions. For this, a thermohydrodynamic model with conservation of mass must be developed to determine the hydrodynamic forces in the bearings as a function of the oil supply flow. This model will be inserted into a discretized rotor model obtained by the finite element method. Thus, with the complete model of the rotor-bearing system, one can apply recursive filtering techniques, such as the Kalman Filter, to adjust the flow and unbalance of the system model in real time. Initially, the identification techniques will be evaluated using vibration signals simulated with noise. Then, the most efficient technique will be experimentally validated to identify the flowrates in the bearings and the rotor unbalance from vibration measurements of a real system on a test rig. Thus, this doctoral project aims to obtain a robust and efficient method for real-time identification of starvation failures in hydrodynamic bearings of rotating machines, capable of predicting early failure conditions and thus avoiding sudden stops in production lines.
News published in Agência FAPESP Newsletter about the scholarship: