Validation of Spectral Fatigue Assessment of a West-Africa FPSO Using Full-Scale Measurements

In order to ensure structural integrity and safe operations, fatigue assessment of structural details is a key aspect of design and verification procedures for FPSOs. Spectral fatigue analysis is widely used in the offshore industry to assess damage induced by wave loading. However, the actual fatigue accumulation endured by units in operations usually differs from predictions due to the assumptions made at the design stage. One of the sources of uncertainty is the representation of the encountered sea states. The objective of this paper is to use in-situ measurements as a reference for evaluating the accuracy of spectral fatigue analysis and to investigate the influence of wave energy description on the fatigue assessment. Structural health monitoring systems have been increasingly used in the last decade in the offshore industry as they constitute a valuable source of information regarding the actual operating conditions, structural response, or encountered environmental conditions. This data can be used to update fatigue assessment in order to determine the remaining service life, understand how the structure is aging, or support for decision making regarding inspections, maintenance, or lifetime extension. The work presented is based on such information gathered during a measurement campaign performed on a spread-moored FPSO in West Africa. Measured strain time histories at several locations on the hull have been used to derive the actual fatigue damage endured by the unit. These damages are compared to the ones determined from spectral fatigue analysis using stress transfer functions obtained from frequency domain hydro-structure computations. Multiple analyses have been performed to evaluate the impact of different sources of statistical wave data and wave energy descriptions on the fatigue assessment. The wave conditions used originate from wave buoy measurements and hindcast data. Overall, the good agreement between full-scale measurements and calculations confirms the suitability of spectral methods for determining fatigue damage. When incomplete information is available, which is often true in the case of wave statistics, assumptions have to be made regarding parameters such as spectrum shape or wave spreading. However, using the full description of wave energy spectra, if available, can be a way of reducing uncertainties and removing unnecessary assumptions in such analysis. The results of this work show how fatigue assessment can be improved by gaining insight into the different sources of uncertainty, notably the sea state representation. With increasing focus on digital solutions, these results show realistic potential for virtual hull monitoring solutions based on accurate numerical models and realistic representation of wave conditions.