Abstract
Fatigue analyses of flexible pipes through the application of long-term response statistics, when considering bimodal seas, are particularly challenging due to the excessive computational costs involved, since they require, generally, to solve numerically a four-dimensional integral associated with time-domain finite element-based analyses of the flexible pipe. This paper investigates different dimension-reduction methods (DRMs) as effective approaches to compute fatigue damage of these structures. The performance of the DRMs was compared with that of the Monte Carlo simulation method (MCSM). Two case studies are presented: a 4 in., and a 7 in., flexible riser connected to a semi-submersible, and a spread-moored floating production storage and offloading unit platform, respectively. The studies demonstrated that some DRMs presented accurate results, exhibiting high efficacy with errors below 3%. Furthermore, the bivariate dimension-reduction method required only 18% of the computational cost associated with the standard Gauss–Hermite quadrature in space R4 (without any reduction), or much less when compared to the MCSM.