On the estimation of rotational frequency response functions

Rotational receptances can perform an important role when predicting the dynamic behaviour both in theoretical and experimental models, especially when a modification includes a rotational inertia or a rotational stiffness. Estimation of rotational frequency response functions (FRFs) from experimental data often leads to poor results, most of the times exhibiting spurious resonance peaks. In this paper, one tries to evaluate whether or not such spurious peaks are related to torsion modes which were not considered in the model. The method used herein to determine the rotational receptances is based on FRF coupling and on a structural modification, which can be accomplished by performing several measurements using rigid fixtures of appropriate geometry. Here, one discusses the influence that these fixtures might have on the results. To avoid one of the major sources of experimental error, the excitation force is applied only at the translation co-ordinate, without application of a moment excitation. Several experimental examples are presented and discussed so that the advantages and limitations of this methodology are identified within the practical application of the method.