Identification of Journal Bearing Characteristics

Abstract

A new approach is proposed to the identification of the eight linearized coefficients which characterize the dynamics of a journal bearing oil-film. The use of pseudo-random binary forcing sequences is described together with cross-correlation and multiple regression analysis. The proposed technique is demonstrated by building a computer model of a rigid rotor running in plain journal bearings, based on Ocvirk’s short-bearing approximation. Multiple regression analysis is applied to the simulated data and the governing equations are reconstructed. In this way the advantages and limitations of the method are examined under controlled conditions.