PSI - Issue 57

Yann Chevalon et al. / Procedia Structural Integrity 57 (2024) 633–641 Author name / Structural Integrity Procedia 00 (2019) 000 – 000

637

5

While the figure 6(a) provides an illustration of the work done to apprehend as-built geometry numerically after extrusion and before fatigue testing, figures 6(b) and 6(c) show how the SCF has been chosen from literature and FEA.

Fig 6 (a) Evaluation of a geometry indent with a CAD software (b) Estimation of SCF from literature (c) FEA analysis

The conversion from strain to stress is done with a linear relationship with the Young modulus, because the imposed deformations in service are very small for most of the time. It is also considered that wave loads are happening fast enough to neglect any creep or other viscoelastic effects in regards of cyclic loading. 2.4. Dynamic loading on flexibles Fatigue failures result from cyclic loading, and in the case of flexible pipes, this dynamic loading is primarily caused by waves. Being a stochastic phenomenon, waves are classified in a certain number of categories: their height, but also their number of occurrences. These parameters depend strongly on geographical location. Based on wave categorization, floater characteristics, pipe properties like weight and stiffness but also riser system configuration, a global analysis of the pipe response is done via an explicit solver, where the stiffness of all the different layers of the pipe are taken into account, in regards of bending and axial response of the flexible. The structure of a pipe is designed to resist tensile loads, and internal and external pressure. Therefore, most of the deformations, in the case of wave loading, will result in a change of curvature and in stresses of the tensile armour. For the polymer fatigue, the loads are therefore driven by the curvature ranges. The global analysis returns, for each wave class, curvature variations, their occurrences, and the location of the largest accumulated movement. When plotting the number of occurrences against the variation of curvatures, at the highest accumulated movement location, one can see that most of the curvature variations are small, while high wave events causing large curvature variations happen rarely. This is illustrated by the graph in the figure 7, where ordinate represents the number of occurrences, while the abscissa represents the curvature variation. The red dot represents the mark of 90th percentile. This means that 90% of the curvature variations are on the left of this mark.