Experimental Study of Joint Linearity

Abstract

Joints in finite element models (FEMs) are often either ignored or simplified using linear assumptions. When experimental modal analysis is performed to validate the accuracy of FEMs, the experiments are usually performed at low excitation energies to minimize excitation of nonlinear properties within a structure. When finite element models are calibrated to low excitation level modal tests, they may not be suited for predicting high excitation level environments. This study was conducted to determine, for a particular set of hardware, how results from a low excitation level modal test compare to results for a high excitation level modal test. A set of simplified hardware, in which cones are connected through a threaded joint, was developed for this study and was subjected to random excitations as well as controlled-force step-sine testing to investigate the effects of joint linearity across different excitation levels. The goal of this work is to obtain a point of reference for how much error would be associated with a FEM, calibrated to low excitation level modal data, if the FEM was used to predict responses at a higher excitation level. Using this information, it can be determined if a linear FEM is appropriate for predicting responses or if nonlinearities need to be incorporated into the model.