PSI - Issue 57

Arne Fjeldstad et al. / Procedia Structural Integrity 57 (2024) 692–700 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

695

4

3. Fatigue analysis

3.1. Analysis based on S-N data

For design purpose it is appropriate to use conservative values for parameters required for the analyses. However, for inspection planning it is important to use relevant (expected or best estimate) values and associated uncertainties in order to predict accurate results that allow the inspections to be directed to the hot spot areas where the fatigue cracks are most likely to occur first. Thus, it is important to base planning of in-service inspection on fatigue analysis that has been performed in a consistent way. By consistent is understood that all hot spots are analyzed based on a similar methodology such that any inherent conservatism in the analysis methodology is similar for the different hot spots. Use of in-consistent assumptions in analysis may direct inspection to areas where inspection in reality is not necessary. Consequently, one might get a false impression of the reliability of the structure with respect to fatigue. If a reliable fatigue analysis has been performed for design, this analysis may also be used for planning in-service inspection. However, the focus during a design is often different from that of inspection planning where consistent calculated fatigue lives are preferred to obtain a good relative ranking of where fatigue cracks are most likely to occur. This may be exemplified by Fig. 2 where joint A could be selected for inspection as this joint show the shortest calculated fatigue life. However, this detail might have been designed in a conservative and simplified way because the calculated fatigue life was found acceptable based on use of a conservative approach. Thus, if a more refined fatigue analysis was performed, the calculated fatigue life may be moved to joint B. Thus, to learn as much as possible from an in-service inspection of a hot spot, the selection should preferably be based on fatigue analysis that is made for this purpose Reference is made to Lotsberg et al. (2016) and DNV-RP-C210 for a comprehensive description of typical differences between a fatigue analysis intended to be basis for inspection planning and a fatigue analysis used for design. DNV-RP-C210 also presents recommendations to fatigue analysis for purpose of inspection planning for jacket structures, semisubmersibles and FPSOs.

Actual fatigue life

A

B

Calculated fatigue life

Fig. 2. Example of calculated life versus actual life. Excerpt form Lotsberg et al. (2016)

3.2. Analysis based on fracture mechanics

S-N data are based on laboratory tests. Normally the cycles to failure are recorded, but not the development of the crack, i.e. crack size versus number of cycles. Fracture mechanics provides a method to describe the crack growth through the member thickness in more detail. This analysis is based on the same long-term stress range distribution as is used for the S-N analysis. Analysis results from fracture mechanics are dependent on more parameters than that of an S-N analysis. Therefore, a calibration of the fracture mechanics model is necessary to provide fatigue lives in agreement with the S-N analysis.