Abstract
When influence coefficient (IC) method is used in rotor balancing, it is possible that serious errors could be resulted in unbalance determinations due to the contamination of measurements and IC matrix being ill-posed. In this study, the availability of least squares algorithm (LSA) and Tikhonov regularization (TR) is studied by error estimations of unbalance determination with considering measurement errors. Also, the influence of illness of the IC matrix equation on the change of balancing accuracy relative to the measurement error will be simulated. The finite element method is used to analyze unbalance responses of a rotor-bearing system and the analysis results of both TR and LSA are compared in case studies. Furthermore, the validations of both methods are realized by balancing experiments of a rotor kit.
