Uncertainty on measurement of elastomeric isolators effective properties

Abstract

Elastomeric isolators are subjected to a series of non-destructive tests with several repeated deformation cycles. For each cycle, effective properties are calculated and afterward averaged. Despite their variability, and therefore their inherent uncertainties, these properties are treated as deterministic values by seismic design procedures. In this research, these uncertainties are quantified, based on the Guide to the expression of Uncertainty in Measurement, GUM, and Monte-Carlo simulations, considering variability between repetitions and instrumentation errors. Uncertainties were calculated for a dataset of 2,498 isolators' test results, finding that the maximum relative expanded uncertainty was 12%. The GUM and Monte-Carlo methods lead to similar results, and higher-order effects in the GUM assessment were negligible. A comprehensive analysis to evaluate the influence of the directly-measured quantities in the properties uncertainties was performed. Results showed that forces and displacements measurement errors are equally relevant in stiffness uncertainties, but force measurement errors primarily control damping uncertainties.