Eventos Anais de eventos
COBEM 2021
26th International Congress of Mechanical Engineering
DETECTION AND IDENTIFICATION OF OVALIZATION FAULT IN HYDRODYNAMIC BEARINGS
Submission Author:
Matheus Victor Inacio , SP
Co-Authors:
Matheus Victor Inacio, Katia Lucchesi Cavalca Dedini, Gregory Bregion Daniel
Presenter: Matheus Victor Inacio
doi://10.26678/ABCM.COBEM2021.COB2021-0639
Abstract
Due to the crucial role that rotating machines have performed in the productive sector, there is an increasing demand on monitoring systems able to detect and identify early faults in order to avoid the sudden shutdowns and high maintenance costs. As one of the most critical sources of faults in rotating machines, hydrodynamic bearings require special attention, since they have a great influence on the dynamic behavior of the rotor. In this context, this paper aims to develop a method for the detection and identification of ovalization fault. For this, an ovalized bearing model was implemented by Finite Volume Method and introduced in a rotating system discretized by Finite Element Method, allowing simulating the dynamic response of the rotor under different ovalization faults. Several response-parameters are investigated to the proper identification of this kind of fault, such as the forward and backward components of the rotor response, the inclination angle and maximum and minimum radii of the shaft orbit. Based on these response-parameters, the ovalization fault can be identified in the bearings using numerical methods, in which the error functions were defined as the relative difference of the response-parameters obtained in the simulated and reference responses. Two different methods are evaluated in this paper, being the two-dimensional bisection and the Newton-Raphson, respectively, a dichotomous and a gradient methods. The numerical simulations were performed considering different ovalization levels and angles, resulting in 16 cases of ovalized bearing. These 16 faulty cases were identified with both numerical methods, considering two different noise levels on the reference response, namely SNR of 12 and 18. The results obtained in the simulations were quite satisfactory, since the fault identification was successfully accomplished. Regarding the search methods, the Newton-Raphson method performed faster identification, maintaining the same order of errors obtained in the two-dimensional bisection. Therefore, the findings obtained in this paper contribute to the improvement of faults monitoring systems applied in rotating machinery.
Keywords
Rotordynamics, Fault Identification, Hydrodynamic bearing, Ovalization Fault
