Off-Axis Input Characterization of Random Vibration Laboratory Data for Model Credibility

Abstract

The goal of this work is to build model credibility of a structural dynamics model by comparing simulated responses to measured responses in random vibration environments, with limited knowledge of the true test input. Oftentimes off-axis excitations can be introduced during single axis vibration testing in the laboratory due to shaker or test fixture dynamics and interface variation. Model credibility cannot be improved by comparing predicted responses to measured responses with unknown excitation profiles. In the absence of sufficient time domain response measurements, the true multi-degree-of-freedom input cannot be exactly characterized for a fair comparison between the model and experiment. Methods exist, however, to estimate multi-degree-of-freedom (MDOF) inputs required to replicate field test data in the laboratory Ross et al.: 6-DOF Shaker Test Input Derivation from Field Test. In: Proceedings of the 35th IMAC, A Conference and Exposition on Structural Dynamics, Bethel (2017). This work focuses on utilizing one of these methods to approximately characterize the off-axis excitation present during laboratory random vibration testing. The method selects a sub-set of the experimental output spectral density matrix, in combination with the system transmissibility matrix, to estimate the input spectral density matrix required to drive the selected measurement responses. Using the estimated multi-degree-of-freedom input generated from this method, the error between simulated predictions and measured responses was significantly reduced across the frequency range of interest, compared to the error computed between experimental data to simulated responses generated assuming single axis excitation.