Multiaxial fatigue

Abstract

The Allseas Group S.A. has made a name for itself in the field of offshore pipelaying and subsea construction by staying innovative. The company has six specialized vessels operating worldwide and the biggest vessel in the world is almost ready for operations. One of the company vessels is the Solitaire and lays pipe using the S-lay installation method. This method allows fast installation together with the applicability to lay pipe in a wide range of water depths. A stinger, a space frame structure consisting of welded circular hollow sections, is designed to prevent excessive bending of the pipeline while laying. This structure is attached to the vessel and after the pipe joints are welded on board they leave the vessel horizontally guided by the stinger. A known failure mode of the stinger structure is fatigue. Fatigue can be characterized as a mechanism whereby cracks initiate and grow in a material due to fluctuating stresses. For the stinger these fluctuating stresses originate from environmental loads, mainly waves. Although Allseas has already developed an inhouse fatigue assessment procedure for the stinger structures to calculate fatigue life, there is an interest in gaining a more fundamental understanding of multiaxial fatigue phenomena with respect to stinger joints. This leads to the following research goal: “The goal of this research is to gain a more fundamental understanding of the multiaxial stress field for a stinger structure and in the end improve the fatigue assessment.” Allseas has performed real time strain measurements on a tubular joint of the Solitaire stinger and already used the data for validating their fatigue assessment (Korakidou, 2013) and evaluating the Stress Concentration Factors (Zhao, 2010). In this study the nominal strain data of two braces were decomposed into three load components: axial forces, in-plane bending moments and out-of-plane bending moments. Their time histories are used as load input for a 4-noded shell finite element model of the stinger joint under investigation. The change of principal stress direction during loading around the braces-to-chord connections was researched, for this is an indication of the stress state multiaxiality variation which could lead to multiaxial fatigue. The saddle and in-between locations showed variations of 20 degrees or more, which is seen as a threshold, for two boundary condition sets analysed: all fixed ends and a combination of pin-rollers. Furthermore, the chord saddle location showed an 18 percent higher Hot Spot Stress for axial loading when principal stresses were used for extrapolation compared to primary stresses perpendicular to the weld toe. The hypothesis that this difference is due to a deviation in principal stress directions between extrapolation locations has been investigated. In order to achieve this, 13 simple T/Y-joint configurations with varying joint parameters were analysed in Femap through its application programming interface. The focus was only on the chord saddle location with axial loading and the chord ends were fixed. Although it was found that joint parameters β, τ and θ have an effect on the difference of principal stress directions between extrapolation locations A and B, no noticeable difference was found between the Stress Concentration Factors of the two extrapolation methods, leading to the conclusion that other factors cause the difference in Hot Spot Stress. Lastly the ratio of Stress Concentration Factor(SCF) to Strain Concentration Factor(SNCF) has been investigated for the same 13 T/Y-joint configurations. This ratio showed to be dependent on the individual joint parameters with values varying from 1.17 to 1.30. Based on the results it can be concluded that the in-between locations around the welded connection showed to be most prone to multiaxial fatigue based on the change in principal stress directions during loading. The difference in Hot Spot Stress between extrapolating principal stresses and primary stresses perpendicular to the weld toe cannot be explained by a difference in principal stress directions, so further research should be done in order to find the actual cause. Nevertheless, based on the results it can be considered conservative to use the principal stresses. Lastly, when converting strains to stresses by using a standard value of 1.20 for the SCF/SNCF-ratio this can lead to a too conservative fatigue life estimation in case the real ratio is lower and vice versa.