An efficient time domain fatigue analysis and its comparison to spectral fatigue assessment for an offshore jacket structure

Abstract

Industry codes require inspections of critical joints within certain intervals to ensure the structural integrity of fixed offshore structures and suggest the use of spectral fatigue analyses to identify them. Spectral methods are considered less reliable but more efficient than approaches in the time domain, which are generally considered highly accurate but computationally unfeasible. This paper evaluates both the time and the frequency domain approaches using a large platform as an example. The results obtained from a, computationally demanding, full-scale time domain fatigue assessment are used as the standard to quantify the errors resulting from the assumptions and simplifications made in the spectral fatigue analysis. These results indicate, also, that the simplifications involved in spectral fatigue lead not only to the well-known inaccuracy but also to consistently lower fatigue lives. This excessive conservatism leads to unnecessary and costly inspections. This paper contributes to the literature in two important ways. First, it shows the main causes of the inaccuracies of the spectral method and quantifies them. Second, it proposes a novel approach for the time domain fatigue analysis which drastically reduces the computational burden, while maintaining a high degree of accuracy. Performing fatigue analyses in time rather than frequency domain increases the accuracy and the reliability of the results and extends the fatigue life of the joints of the structure since several conservative assumptions are eliminated.