Updating future reliability of nonlinear systems with low dimensional monitoring data using short-cut simulation

This paper proposes a novel stochastic simulation method of updating future reliability of nonlinear systems with high dimensional uncertainties when the monitoring data is low dimensional. The novelty of the proposed framework is to bypass the most difficult part of the problem: drawing samples of uncertain variables conditioning on the low dimensional monitoring data. This research proposes a short-cut simulation approach: instead of drawing samples of possibly high dimensional uncertain variables conditioning on the monitoring data, it is shown that the problem can be solved by drawing samples of the low dimensional monitoring data conditioning on the future failure event.