On the validity of planar, thick curved beam models derived with respect to centroidal and neutral axes

Most dynamic analyses of planar curved beams found in the literature are carried out based on a curved beam model which assumes that the neutral axis coincides with the centroidal axis of the curved beam. This assumption leads to governing equations of motion which are relatively simple with analysis results that have acceptable accuracy for shallow curved beams. However, when a curved beam is not shallow and/or its cross section is not doubly symmetric, the offset distance between the neutral and centroidal axes may be large enough to influence the in-plane dynamics of the curved beam even for small motion. In this paper, the validity of this underlying assumption for modeling a linear curved beam is examined. To this end, two sets of equations of motion governing the in-plane dynamics of a planar curved beam are derived, in a consistent manner for comparison, based on the linear strain-displacement relations and Hamilton’s principle. The first set of equations is derived from the displacement components measured with reference to the neutral axis of the curved beam while the second set is derived with respect to the centroidal axis of the cross section. The curved beam is considered extensional and the effects of rotary inertia and radial shear deformation are included. In addition to the curvature parameter that characterizes the wave motion for both curved beam models, an eccentricity parameter is introduced in the first model to account for the offset between the neutral and centroidal axes. The dynamic behavior predicted by each curved beam model is compared in terms of the dispersion relations, frequency spectra, cutoff frequencies, natural frequencies and modeshapes, and frequency responses. In order to ensure that the comparison is accurate, the wave propagation technique is applied to obtain exact wave solutions. It is shown that, when the curvature parameter is not small, the underlying assumption has a substantial impact on the accuracy of the linear dynamic analysis of a curved beam.