First-passage probabilities for randomly excited mechanical systems by a selective Monte-Carlo simulation method

Abstract

In this paper, Monte-Carlo methods used for the reliability assessment of structures under stochastic excitations are further advanced, e.g., by leading the generated samples towards the low probability range which is practically not assessable by direct Monte-Carlo methods. Based on criteria denoting the realizations that lead most likely to failure, a simulation technique called the “Russian Roulette and Splitting” (RR&S) is presented and discussed briefly. In a numerical example, the RR&S procedure is compared with the direct Monte-Carlo simulation method (MCS), demonstrating comparative accuracy.