Experimental assessment of test tailoring methods for single-axis and multi-axis accelerated tests

Current standards on accelerated dynamic tests are commonly intended to be implemented on single-axis shakers discarding, thus, the effect of the excitation components in other directions. The effect of those components, normally present in real operational environments, is only simulated performing multiple tests in succession for each orthogonal direction. Studies available in the scientific literature highlighted how this methodology not only is time consuming but also may lead to failures of the device under test not observable in real operating conditions. Starting from these considerations an experimental research has been performed using a multi-axis electrodynamic shaker available at the University of Ferrara. The experimental campaign consisted in single-axis and multi-axis tests varying the levels of excitation and the degree of correlation. The activity was carried out on notched steel specimens and permitted to investigate the limits of current accelerated test formulations when more than one source of excitation is present. For each type of test, the prediction accuracy of the so-called Inverse Power Law was investigated and inaccuracies have been spotted, in large part imputable to natural frequency shift caused by the different mechanism of damage propagation.