A new approach for predicting the extreme response of structures

In this article a method to estimate the mean upcrossing intensity, μ(u), of a stochastic process is proposed. It is assumed that the stochastic process is a sum of a Gaussian process and a second-order correction term. The method is based on the two-dimensional Saddlepoint approximation. The accuracy of the method is tested on processes having analytical solutions for μ(u). Numerical examples are given where the stochastic process represents (i) the horizontal response of a floating offshore structure in a Gaussian sea, and (ii) the response of a structure subjected to a Gaussian wind velocity process. In addition, the estimates are compared to empirical upcrossing intensities of simulated responses. For case (ii), the obtained μ(u) estimates are compared to estimates obtained by numerical integration.