Identification of sixteen dynamic coefficients of two journal bearings from experimental unbalance responses

Identifying Bearing Rotor-Dynamic Coefficients Using an Extended Kalman Filter

“…The conclusion is that the EKF filter cannot estimate the bearing properties without the impact excitation present. Some parameter identification methods (Lee and Hong, 1989;Tieu and Qiu, 1994;Tiwari et al, 2002;De Santiago and San Andrés, 2007b) are capable of estimating bearing properties using imbalance excitation only, though the results are not as accurate as when impact excitation data is used. It is possible that this particular rotor-bearing problem exacerbates the inherent drawbacks with the EKF and that other Kalman filtering methods that overcome these problems, such as the Unscented Kalman Filter developed by Julier and Uhlmann (1997), may be more successful at identifying bearing parameters from imbalance response data.…”

Identification of Speed-Dependent Bearing Parameters

“…Nordmann and Schollhorn [5] identi"ed the sti!ness and damping coe$cients of journal bearings whereas Kraus et al [6] identi"ed the coe$cients for rolling element bearings by means of the impact method. Sahinkaya and Burrows [7] and Tieu and Qiu [8] estimated the linearized oil "lm parameters from the out-of-balance response where the shaft was excited by a known unbalance force (synchronous excitation). Chen and Lee [9] identi"ed rolling element dynamic characteristics in #exible rotor-bearing systems by using unbalance responses at all bearings and several shaft locations without a priori knowledge of the unbalance.…”

Evaluation of the determinant of identification equations for a linear model of a mechanical vibratory system

“…The first, the frequency-domain method, is based on Fourier transforms and measurement of the frequency characteristics of steady forced vibrations of the system being modelled [1,2]. The second, the time-domain method, which has recently seen considerable progress, relies on new high-precision measurement equipment, which is capable of measuring the parameters of motion as practically continuous functions of time [3][4][5][6]. A recent review of identification methods [7], which is far from complete, devotes much attention, apart from the mathematical aspects of the approach, to methods for exciting test vibrations.…”