Investigations of a practical wind-induced fatigue calculation based on nonlinear time domain dynamic analysis and a full wind-directional scatter diagram

Abstract

By performing the nonlinear time domain dynamic analysis, and calculating the wind-induced fatigue using the procedure presented by this author, this paper investigates the influences from different drag coefficient specification systems. The study shows that the drag coefficient specified by Norsok is more conservative than that of the DnV system, mainly due to Norsok's rough categories of the drag coefficient values varying with the Reynolds number. Furthermore, the effects of time increment step length, time duration for simulation as well as the flexibility of wind strut-main flare boom connections are also investigated. By comparing the fatigue life from dynamic analysis with its static counterpart (without taking into account the inertia effects of the structural and non-structural installations), the significance of the contribution from the structure's dynamic response can be identified. It is suggested that the modelling of secondary structures such as a flare and vent line must be carefully considered, as this may significantly influence fatigue damage calculation. In addition, the effects of gravity on the structure's fatigue damage are also studied and the conclusion challenges traditional engineering practice. Finally, non-Gaussian responses are discussed through the statistical investigation of the local responses. The wind fatigue calculation procedure presented can be widely adopted for the similar study on high-rise tubular structures.