Experimental vibration tests for structural durability assessment are one of the most important steps to estimate the fatigue life of mechanical components.
Tests setup where the components endure uni-axial random loading state may be easier to execute than bi-axial or multi-axial tests. The uniaxial stress can be achieved, for instance, with a cantilever beam mounted onto a shaker producing a bending random loading.
On the contrary, a multi-axial stress assessment may require one or more machines, specific testing layout and particular shape of specimen, that allows the activation of different loads in different direction at the same time, which may be a tedious task in practice. To mitigate these difficulties, a number of testing methodologies that differ in terms of testing machines, specimen geometry and type of excitation, were proposed in the literature.
This paper aims to present a new testing setup to apply random bending-torsion loading in vibratory test by a uni-directional exciters. The system is composed by a specimen, with rectangular section, mounted onto an electrodynamic shaker and a T-shaped mass placed on the specimen's tip. The eccentricity of the T-shaped mass induces a torsion loading, combined with bending, on the specimen in a specific range of frequency. A finite element model is first used to conduct a design phase analysis to define a suitable dimension of the T-shaped mass, and then to estimate the system dynamic response and the stress on the notched specimen section. The model is calibrated and validated through harmonic tests. From numerical and experimental tests, by means of evaluating the numerical stress components and accelerometers phases, it is shown that the specimen is subjected to a bending-torsion coupled loading if the second mode is excited, and to a pure bending loading when the first, third, and fourth modes are excited. After a numerical validation of the loading cases, fatigue tests with random loading are executed and the experimental results are compared with those obtained with numerical methods.