Design of Experiments to determine the influence of test procedure on Experimental Modal Analysis

Abstract

This study conducts a fractional factorial design to evaluate the influence of test settings on the results of Experimental Modal Analysis (EMA) which is the method often used to determine natural frequencies and damping rate of structures. The structures in this work are simple beams cut from lightweight rail-way vehicle car bodies. The original specimens have very low structural damping, whereas others have been supplemented with Constrained Layer Damping (CLD) patches to significantly increase vibration energy dissipation. It will be shown that test settings do influence Experimental Modal Analysis (EMA) results in ways that are seldom intuitive or foreseeable. The statistical significance of some factors, or combinations thereof, proves that their effects are not entirely attributable to random phenomena. The results show that a statistical analysis like the one presented in this paper should be conducted prior to extracting accurate damping data from EMA tests. The technique reduces the number of tests needed to assess the influence of settings. This technique, not only highlights the statistically significant factors, but also quantifies their influence, thus helping the researcher make an informed decision regarding test settings. • The analysis of variance determines which factors do have a significant influence on EMA analysis. • Test settings depend on beam type, modal parameters, and support configuration. • Test settings have a strong influence on modal damping. • The fractional factorial design reduces the number of tests necessary. • The proposed statistical analysis should be conducted prior to extracting accurate damping data.