Adjustment of a Rotor Model to Achieve Agreement between Calculated and Experimental Natural Frequencies

Abstract

This paper is concerned with the problem of adjusting the mathematical model of a system such that the computed natural frequencies coincide with those measured experimentally. The particular system considered is a laboratory turbine-rotor model, modelled mathematically by 42 Timoshenko beam elements and lumped masses. Model adjustments are made by assuming, firstly, Young's modulus and the modulus of rigidity to be variable, a change from standard values representing overall stiffness deficiencies in the mathematical model. In this case, a best fit to the lowest six natural frequencies, as measured experimentally, is made. Secondly, stiffness diameters are assumed variable, thereby allowing for deficiencies in the model near discontinuous changes of section, and in this case, the lowest six natural frequencies are matched exactly, but an overall measure of the differences between the actual and the stiffness diameters is minimized. An analysis for the rates of change of natural frequency with the various stiffness properties (i.e. the sensitivities) is presented, and the results of the manipulation discussed.