Identification of Rotating Assembly Inertial and Bearing Parameters

The problem of estimating the state of a turbopump rotating assembly and identifying unknown products of inertia and unknown bearing parameters is considered. A linearized, extended Kalman filtering approach has been used and found to be successful for both state estimation and parameter identification in an inherently nonlinear problem. The input data was simulated Bently test data which was obtained from a verified 12-dimensional state-space model and was compared with real Bently test data obtained from NASA. The results prove the feasibility of using this model and this type of test data to obtain the hidden parameters of a typical turbopump rotating assembly. This method should lead to improved performance by allowing improved balancing of the rotor and giving improved knowledge of the bearing reactions.