A technique that refines uncertainty quantification of a stochastic structure's dynamical response by conditioning on noisy measurement data. It extracts the most informative key measurement condition and, via the principle of probability conservation and conditional-probability theory, derives quotient-form closed expressions, the key conditional quotients (KCQ), for the conditional mean, variance, and probability density of the response. KCQs are evaluated with a non-equal-weighted generalized quasi-Monte Carlo method, a Dirac-function smoothing technique, and an online-offline coupled computational strategy, and are shown to significantly reduce response uncertainty relative to non-conditional UQ on linear and nonlinear examples.
A method for quantifying the uncertainty of a stochastic structure's dynamical response after conditioning on noisy measurement data. Structures have random loads and parameters, so their vibration r…