EXPERIMENTAL EVALUATION OF ERROR PROPAGATION IN ROTOR-MODEL-BASED IDENTIFICATION OF FOUNDATIONS IN ROTATING MACHINERY

Modelling rotating machinery foundations is of wide practical interest since the availability of a sufficiently accurate foundation model is an invaluable asset for efficient operation and balancing. There are two common foundation modelling approaches. The first uses vibration measurements to identify equivalent foundation parameters; the other uses finite-elements, this latter approach being limited by difficulty of modelling. This paper concentrates on the parameter identification approach, and in particular, on identifying appropriate dynamic stiffness parameters for the foundation, as such identification does not require rotor removal. Relevant to this approach is the consideration of errors in the identified parameters as a result of the propagation of uncertainties in input data. For this, an identification strategy, which uses as inputs the rotor model and the rotor unbalance (used to determine transmitted forces) as well as motion measurements of the rotor and the foundation, is revisited to identify the bearing pedestal parameters of a rotor bearing foundation system with hydrodynamic bearings, taking into account propagation of uncertainties in input data. A formulation for the error propagation based on multivariate uncertainty analysis is experimentally evaluated by applying it to a simple rotor bearing pedestal system, viz. a laboratory rig consisting of a multi disk rotor supported on simple pedestals via hydrodynamic and rolling element bearings. The suggested methodology for determining the uncertainty in the estimated parameters has been formulated as an extension of the accepted (standardised) practice for determining uncertainty propagation for a one-dimensional measurand. The methodology involves basic multivariate statistics to process a repeated measurement set. The approach has been examined for two identification methods in frequency domain, taking into account experimental uncertainties.