Modal analysis of rotors with whirling motion under non-stationary conditions

In this study, the use of adaptive filters, which are known methodologies in the data processing field, for parameter identification was investigated. Recursive least squares is an adaptive filter algorithm that can be implemented, based on the second-order ordinary differential equations, in order to characterize the modal parameters. Within this objective, the aim of this study was to gain a better understanding of rotating machine dynamics. In many cases, when performing field analysis, it is common to encounter a situation where it is impossible to stop a piece of critical equipment. In this case, all measurements performed on the equipment and the neighboring structures will be corrupted by the harmonic noise generated during the operation of the machine. In this context, the proposed methodology contributes to the solution of this problem in two distinct ways. Firstly, this approach is appropriate for measurements obtained from the experimental modal analysis of systems even when they are corrupted by harmonic noise. The only information needed is the instantaneous frequency of this noise. Secondly, the methodology allows the identification of the modal parameters of the system while it is operating under non-stationary conditions, for example, during machine start-up or shut-down. The results suggest that the natural frequencies and the vibration modes are precisely estimated, even for timevarying systems. The numerical and experimental analyses provide good results, close to the natural frequencies. One of the most interesting characteristics of the method is the possibility to extract modal parameters in real-time while the system is in operation. (c) 2020 Elsevier Ltd. All rights reserved.