Non-Parametric Identification of Rotor-Bearing System through Volterra-Wiener Theories
The structure of the Volterra and Wiener series, which model the relationship between system response and input in terms of series of first and higher order convolution integrals, provide analytical platforms which can be utilized for parameter estimation. These are non-parametric forms of response representation. Non-parametric identification concerns modeling in a function space by input-output mapping, for systems where sufficient information on the mathematical structure or class is not available. Parametric identification, on the other hand, refers to systems where sufficient a-prioriinformation about the mathematical structure of the class to which the system belongs, is available. In the present study, structured Volterra and Wiener response representations are employed to develop identification and parameter estimation procedures for nonlinear rotor systems. Experimental investigations and validation of algorithms have been carried out on a laboratory test rig. Linear and nonlinear stiffness parameters are estimated and compared with approximate theoretical formulations and some previous experimental results.
KeywordsVolterra series Wiener series Nonlinear system identification Rotor-bearing system
- 1.Zhabanch, N., Jean, W.Z.: Nonlinear dynamic analysis of a rotor shaft system with viscoelastically supported bearings. ASME J. Vib. Acoust. 125, 290–298 (2004)Google Scholar
- 5.Fretchet, M.: Sur les fonctionnels continues. Annls. Scient. Ec. Norm. Sup. 27, 193–219 (1910)Google Scholar
- 6.Volterra, V.: Theory of Functionals. Blackie & Sons, Glasgow (1930)Google Scholar
- 9.Harris, T.A.: Rolling Bearing Analysis. Wiley, New York (1984)Google Scholar
- 10.Ragulskis, K.M., Jurkauskas, A.Y., Atstupenas, V.V., Vitkute, A.Y., Kulvec, A.P.: Vibration in Bearings. Mintis Publishers, Vilnyus (1974)Google Scholar